Reports And Briefings

Editor’s Note: This policy framework is the product of a 2025-2026 data governance working group convened by Daniel Cochrane, senior fellow at the Institute for Family Studies, in partnership with Matt Prewitt, co-founder of RadicalxChange Foundation, Lacey Strahm, Vice President of Growth and Impact at OpenMined, and Salomé Viljoen, assistant professor of law at the University of Michigan Law School. Institutional affiliations are provided for identification purposes only and do not necessarily constitute endorsement of this framework by contributors’ home organizations.

Executive Summary

AI and the digital economy are broken. Despite the gains in economic growth and efficiency Silicon Valley has produced, those benefits have often come at enormous cost to American families and communities. Over the last decade, addictive digital products, like smartphones and social media, have rewired the brains of an entire generation, leaving many addicted, lonely, depressed, and diminished in their moral, social, and rational capacities.

Even now, as parents and states gain ground in the fight to hold Big Tech accountable for its harms to children, a new danger is imminent. Generative and agentic AI—if left on their current trajectory—threaten to further hollow out American family life and erode the social, religious, economic, and political structures upon which human flourishing depends. 

The evidence is all around us. AI companions are displacing human relationships and exacerbating a crisis in marriage and family formation. Employers are increasingly leaning into AI “bosses,” if not to replace their workers, then to reduce them to biological appendages of machines optimized for efficiency and short-term profit maximization. AI is also shaping consequential decisions from government services, healthcare, and education to the price families pay for groceries and other necessities. Even human thought and agency are now at risk with AI systems driving the mass offloading of human cognitive labor, like reading, writing, and independent creative production.

AI threatens to disempower American families and communities, in large part, because those bedrock institutions lack a voice in shaping AI’s course of development and its governance. Consequently, the incentives of tech corporations are often fundamentally out of step with those of most Americans and corrosive to the social fabric.

Part of the answer, then, is to recognize that nearly all data fueling the digital economy and AI is a co-produced good. Governance of digital systems and their outputs should be shared by the communities who co-produced the underlying data and who are most impacted by its use. Congress and the States should empower co-producers, through a new right of digital association and collective informational bargaining, to secure terms from tech corporations that advance community-specific privacy, control, and economic interests.

This policy brief unpacks this claim in three parts:

Part I argues that data is a co-produced good. Yet, control over its collection and use is concentrated in the hands of tech firms. Individuals, families, and communities have fundamental interests of privacy: restricting access to their information; control: deciding the purposes for which information they co-produce is used; and fair value: sharing in the economic benefits derived from co-produced data. However, because tech firms have asymmetric power to set digital terms of data collection and use, communally-specified privacy, control, and economic interests are often undermined. As a result, the incentives driving the digital economy often cut against the flourishing of human scale communities. 

Part II makes the case for why both individual data rights and top-down data privacy regimes often fail to vindicate privacy, control, and economic interests. 

Part III proposes a new class of informational rights: 

• The Right of Digital Association: This right gives individuals, households, civil society institutions, and small enterprises (data subjects) the ability to join or form new intermediaries—Data Rights Associations (DRAs)—to secure terms governing data collection and use from tech corporations (counterparties) that advance community specified privacy, control, and economic interests.
   
• The Right to Collective Informational Bargaining: This right empowers DRAs to secure fair terms from counterparties by imposing a duty to bargain in good faith and providing a neutral arbitration mechanism to resolve impasses. 
   
• The Right to Fair Representation: This right lays out specific duties that DRAs owe to their members, including duties of care and loyalty. 

Part III also proposes standards for regulator certification and oversight of DRAs, recommends general obligations for tech corporations (counterparties) subject to collective informational bargaining, considers questions of financial and technical feasibility, and explores how a robust DRA ecosystem would advance pro-human, pro-family innovation.

Introduction

Silicon Valley culture is characterized by the motto, “move fast and break things.” Read charitably, it evokes a spirit of openness, innovation, and creativity. But for many Americans, the motto is emblematic of an elite tech culture that is itself broken, out of touch, and largely insulated from the social and economic consequences that its technologies unleash upon the world. 

Frontier AI labs (“Big AI”), including OpenAI, Anthropic, and Google DeepMind, are racing to build superintelligent AIs to outperform and ultimately replace humans in every sphere: from work, to relationships, to politics, and even human thought and decision making. Elon Musk claims that within a few years, AI will be “unequivocally smarter than humans in every way.” An OpenAI employee said in 2026 that he “would prefer a world of rapid RSI [recursive self-improvement, where AI models train their own successors without any human involvement] and human disempowerment.”

Even if they do not succeed, the same Big Tech companies, like Meta, TikTok, Google, and Amazon, that have surveilled, censored, and re-wired the brains of an entire generation, are poised to use AI to further dominate attention, monopolize our markets, and shape our beliefs and actions at scale. But it is not just the largest players. A growing number of mid-sized firms use AI to profile individuals, families, and communities with data harvested from a wide range of digital services and compiled by a shadowy web of data brokers.

While these forces threaten individual dignity and freedom, their communal and wider societal implications are equally severe. The bedrock institutions of family and civil society are in the AI revolution’s crosshairs. Young men and women, already awash in a sea of digitally-induced distraction and erotica, are now tempted to turn to AI “companions” and chatbots with their most serious problems and intimate secrets, and to forgo human friendships, romance, and marriage altogether. Family and marital bonds are being steadily replaced by screens that serve up an endless supply of AI slop personalized by AI algorithms optimized for addiction. Schools and workplaces risk becoming sites of mass social engineering, deploying powerful AI systems to surveil and replace human instruction, learning, labor, and thought. Algorithmic pricing threatens to further exacerbate the affordability crisis for struggling families. And widespread public concern with Big Tech’s AI surveillance and manipulation is undermining social trust, and with it, the necessary conditions for republican-self-governance at the communal and societal level.

This trend toward dehumanization and social decay is largely driven by the race to capture data from every domain of human life. For too long, tech companies have enjoyed unrestricted access to the data we co-produce during our lives, both on and offline. All that data, paired with AI’s growing capabilities of pattern recognition, prediction, and automation, have produced vast power asymmetries between human communities and the countless digital services they interact with daily. Those asymmetries have also resulted in parallel misalignment between the values and interests of the tech industry and the people it is supposed to serve.

Rather than compete to provide the most value and social benefit for the data they extract, tech companies often mask their commercial surveillance behind “free” or subsidized services, which they utilize to harvest as much consumer data and intellectual property as possible to develop products, derive insights about groups, and influence community members without affording them privacy, control, or fair value. The result: a consumer tech ecosystem that has done great damage to family life and marriage formation, and that is keeping us atomized, dependent, and scrolling alone.

These effects have reached beyond the household into the other institutions upon which familial flourishing depends. AI and the digital economy’s hunger for data is propelling the extraction of data from workers, artists, writers, and business owners, threatening to further diminish and centralize cultural production and commerce, and to disrupt social relations and truth seeking in the most important areas of civil society, including religion, education, and public life.

Silicon Valley’s current anti-human trajectory is bleak. But a different future is possible if we choose to embrace it. The 2025 statement, A Future for the Family, offers that alternative pro-human blueprint for the future centered on the flourishing of the permanent things—the social institutions and communal structures that give life meaning and that serve as the enduring bedrock of American society:

The family plants the seed and forms the foundation of the future, through the begetting and raising of children who will carry the human project forward. As such, the family, technology, and a dynamic and fruitful future are intrinsically connected, and the present conflict between them must be overcome.

Instead of smart devices and sensors that constantly feed data to outside tech companies and advertisers, we imagine a future where your home is protected, where data serves your family’s needs—and only leaves with your permission and when absolutely necessary. Instead of screen-based devices that pull attention away from friends and family, we imagine new distraction-free interfaces that facilitate interaction with locally hosted AI models that serve humans on terms set by them and the institutions they inhabit. Instead of AGI that some believe will replace human labor, creativity, and relationships, we envision a competitive ecosystem of what Max Tegmark refers to as “tool AI,” designed to supplement, not supplant human judgment, solve real human problems, and foster ideal economic conditions for family formation and flourishing.

To realize that future, public policy must give families and communities a voice in the development, deployment, and governance of AI and the larger digital economy. While regulatory floors to protect data privacy, kids, and families are necessary, society must fundamentally rethink the social contract between its essential institutions and private tech corporations. That starts at the data layer.

Galvanizing pro-human, pro-family AI innovation requires ensuring that data flows to digital products and services designed with human flourishing in mind. It requires moving away from asymmetric data and AI governance toward more evenly distributed polycentric governance.

A new class of associational data rights, chiefly a right of digital association and collective informational bargaining exercised through Data Rights Associations (DRAs), are crucial for addressing data’s complexity as well as the significant power asymmetries between human data co-producers on the one side, and the vast ecosystem of AI developers, deployers, tech platforms, and data brokers on the other. DRAs, like cooperatives and labor unions, would be empowered by law to secure fair terms from tech companies on behalf of individuals, households, and human scale communities. We envision DRAs as a new intermediary layer of digital rights bearing entities that would restructure data markets and provide strong incentives for firms to respect the privacy, control, and economic interests of families and communities over the long-term.

Part I: The Human Cost of “Infocapture”

Generative AI along with many digital products and services impose social and economic costs on Americans. Nearly every time a person interacts with a digital system, they knowingly or unknowingly trade data—information generated by them or others—for some immediate convenience. Whenever a person prompts ChatGPT to give them advice on the best local burger joint, the user trades data about themselves, including the query and whatever other contextual cues GPT extracts, in exchange for a response. Likewise, every time a small business owner “vibe codes” a new agent in Claude to improve their shop’s performance, the owner is trading valuable data—information about their business’s needs, operations, and clientele—to receive the promised efficiencies from the agent.

Even when families and community members do not directly interface with an AI agent or chatbot, they feed data to AI algorithms in a multitude of other ways. From Google Maps, which continually tracks users’ geo-location data, to “smart” home devices, smartphones, cars, and wearables, each tiny contact with a digital product or service is a trade of a small part of ourselves for some convenience.

To be sure, users reap important benefits from the digital economy. Erik Brynjolfsson, for example, has coined the term “GDP-B” to refer to the unmeasured benefits of technology. But there is also “GDP-C”—hidden costs—associated with AI and social media, like distraction, addiction, and the atrophy of human capacities for reading, writing, independent thought, and relationships. These costs are often overlooked by tech corporations because the human households and communities who co-produce the data shaping these technologies lack power to demand consideration of long-term social and economic trade-offs in the design of AI systems and the structure of the larger digital economy. One consequence is that the design, development, and deployment of AI systems and digital services seldom reflect—and often run counter to—the values and interests of the human institutions and communities most directly impacted.

Data co-production and the problems of asymmetric governance

Although AI and the entire digital economy are built atop mountains of human data, the centers of data co-production, including households, workers, artists, and consumers, lack power over its collection and use. Data collection and use is currently asymmetric, concentrating informational power in the hands of tech firms and a growing ecosystem of extractive data companies. This dynamic has resulted in a race to the bottom with unrestricted data collection and exploitative use fueling a vicious cycle of misaligned AI and antisocial digital products across the tech stack.

One reason for this trend is that neither U.S. federal nor state law currently treat data as a co-produced good and thus, do not afford human communities involved in its production with rights of co-governance (or “co-determination”).

Co-production

Data is often derived from a multiplicity of interactions between humans, entities, and AIs across the internet and a multitude of devices, services, and activities. Conversations between a human and an AI chatbot is one example of information or data “co-production.” The combination of human inputs (prompts) and AI outputs (decision, prediction, or content) creates information that would not exist without participation by the human party.

Similarly, when an AI, like Google’s Gemini, draws on human knowledge and works to generate an answer to a human prompt (search query), both the human creators of the content used for inference and the human who inputted the original prompt are data “co-producers.”

Even agentic AI systems that perform tasks on behalf of humans with limited input or oversight, still involve some degree of data co-production. For example, agents that are ostensibly able to autonomously coordinate and execute tasks on behalf of human users need access to human data to set and pursue objectives, infer context, and carry out assignments. Most generative and agentic AI systems rely on massive volumes of human data for training. When an AI produces a novel image, video, or text-based output from training on human creative work, the human-creators of the AI training material have effectively co-produced the output, or at least, the capability that made the output possible.

Data is nearly always a co-produced good in at least two respects. First, it is derived from interactions between human parties and the digital providers that track and capture content and behavior. Even when behavioral data is not factored into an AI’s output–for example an AI surveillance system that relies solely on video monitoring within a private venue–it still relies on movements (video-based data) to assess risk or ID banned guests. In that sense, all AI outputs are inherently co-produced. AI systems require some non-static data input to produce actionable outputs.

Second, data is derived, in part, from inferring relationships and interactions between human individuals and communities. As a report from the RadicalxChange Foundation argues, data that reveals information about one individual often reveals something about others in that individual's larger social circle: whether biological family, friends, co-workers, or other interpersonal relations. When users capture a family reunion or a friend’s birthday through a pair of Meta Ray-Bans, they capture the activities, identities, and preferences of other people. When users send a DNA sample to genome analytics company, Nucleus, they not only reveal information about themselves, but also information about other family members, such as their propensity for certain disorders, illnesses, and biological traits.

“Infocapture”

The problem is that much of that data, and its resulting informational outputs or products, are assumed to be the sole property of whichever tech platform or corporation manages to capture them. While sharing photos or videos between friends and a broader public audience might be permissible, the capture of that information for commercial gain, population level surveillance, or behavioral nudging is a different matter. The “ick factor” associated with videos captured by Ray-Bans is that tech companies or bad actors can now profile and identify every person in such content using AI and private surveillance databases, as Clearview AI has done. In this sense, we are all victims of infocapture.

Perhaps in a world where humans are the only entities sharing and viewing videos and photos, the privacy trade-off, though undesirable, might be acceptable overall. But the purpose of the growing web of data hungry tech products is no longer to facilitate human-to-human disclosure as an end. Rather, it is to convert vast swaths of human life into machine-readable data to transform human-to-human disclosure into a means of total user and communal legibility for machines.

Unlike other areas of the market where co-produced value is accompanied by well-defined legal rights (of governance), such as in physical or financial assets like real estate and securities, data flows are determined by a digital law of the jungle. Data markets, and the larger digital economy they help shape, are overwhelmingly asymmetric. For the purposes of this report, “asymmetric” refers to the one-sided nature of data markets. While data is collected at various points and used in a multitude of different ways by many different actors, both its collection and use are often governed asymmetrically relative to the data subjects who co-produce it, or who are directly affected by the resulting outputs. Once information that data subjects co-produce is copied or captured, they lose control over its use.

Whichever corporation hosts the product or service that collects data sets the terms for who gets access to it and how it is used downstream. In other words, while data is almost always co-produced, rights to govern it are unilaterally (“asymmetrically”) exercised by tech firms and commercial enterprises.

This has fueled a digital race to predict and shape human behavior, and even to replace human labor, thought, and relationships. AI companies, like OpenAI and Anthropic, rely on vast amounts of data scraped from across the internet to train their models. But few of the human co-producers of that data get a say in whether their intellectual property (IP) is harvested or the conditions under which it may be used for AI model training. Other forms of human work are also impacted. Some gig economy platforms and employers hoover up worker data to surveil their workforces and subject laborers to dehumanizing forms of algorithmic management, while automating tasks with little input or accountability from affected workers.

At the same time, platforms such as Amazon, Google, YouTube, TikTok, Instagram, and X, all have unilateral power to decide what behavioral data they mine from their users and how they utilize it to personalize content, customize and target ads, and optimize for user engagement. Similar problems are now evident with generative AI companions and chatbots that display sycophantic and manipulative behavior toward vulnerable human subjects and are increasingly able to influence human attitudes, beliefs, and actions at scale.

Much of that data not only influences what information AI services feed us and how they interact with us moment by moment, it also factors into the data used to train the underlying systems. Human-AI chatbot interactions are particularly rich data sources for training because, in aggregate, they provide generative AI models with millions of examples of how humans think, communicate, and respond to various stimuli. By inferring patterns from this data, AI systems gradually map the contours of social relationships and learn how to better nudge and influence human users, and to simulate therapeutic language along with human pursuits like love, empathy, religion, connection, writing, and art. In this sense, we are training our replacements, if not in work, then certainly in our agency, thinking, and relationships.

The current asymmetric approaches to governing data collection and use forestalls the kinds of distributed institutional arrangements necessary to reverse these developments and re-align AI and the larger digital economy with the interests of American families and communities. That is because the present model fails to recognize and fairly distribute rights of data co-governance among the human parties who participated in its production.

What constitutes “fair” allocation of informational co-governance rights depends on the specific type of data collected, the tradeoffs associated with its contemplated uses, and the distribution of benefits among co-producers and corporate beneficiaries. That is nearly impossible to determine asymmetrically in the absence of balanced, polycentric mechanisms for setting terms for data collection and use across the digital stack.

Core data-related interests

Data subjects have three inter-related sets of interests at stake:

1. Privacy: Data subjects should have real and effective mechanisms to prevent or limit data collection that: 1) exceeds what is necessary and proportionate to provide a service, explicitly and affirmatively requested by the data subject; 2) discloses information that could foreseeably result in negative consequences for the data subject—whether from downstream use by commercial or government entities or from unauthorized public disclosure of the information; or 3) where the disclosure of information would violate the shared norms specified by participants in a defined social context.
    
2. Control: Data subjects should have a meaningful say over how their information is used, including in the development or deployment of systems that implicate: 1) human health, freedom, agency, or safety; and 2) the wellbeing of families, democracy, civil society, or faith.
    
3. Fair value: Data subjects ought to share in the economic value that their data disclosures help create.

All three sets of interests are harmed by asymmetric terms of data collection and use—where data subjects lack meaningful agency to set the terms.

Asymmetric data capture

The point of collection (PoC) occurs wherever information from or about a data subject or group of data subjects is captured. Data at the PoC refers to data at its source, prior to its combination with other types of data, or its use for inference, prediction, or the generation of AI outputs (whether informational content, curation, task execution, automated decision-making, or product development).

Note that the specific means employed to collect data is not relevant to its status as a co-produced good with an implied right to co-governance shared by the parties. Whether information from or about a data subject is acquired from a retail purchase, derived from engagement on TikTok, or scraped from a public website, the information is still, for the most part, the result of human co-production. The means of capturing or copying data is less significant than the fact that it is captured or copied for uses other than interpersonal communication or strictly private, non-commercial purposes. Outside of this band of permissible collection, data capture is asymmetric if data subjects lack parity in setting the terms governing what data is collected, who gets access, and how it is used downstream of the PoC.

Digital providers often capture more data than is strictly necessary to provide the product or service explicitly requested by the customer or user. One classic example is social media apps like TikTok which logs users’ off-app activities, geo-location, and purchases. Other major tech companies, like Google and Meta, use various methods to track users across the internet. When a user leaves Google Search or Instagram, Meta and Google know where they go, what they look at, and what they do on other sites. Google even tracks your real-world movements through precise geo-location data via apps such as Maps and its other services.

Even when invasive data collection is inherent to a product or service, the lack of oversight and collective permission structures around such data gathering still poses a significant governance failure. OpenAI’s Sam Altman has described the ideal AI tool as a “super-competent colleague that knows absolutely everything about my whole life, every email, every conversation I’ve ever had, but doesn’t feel like an extension.” But Altman’s idealized future comes at a significant cost: the promised benefits of the “super-competent AI colleague” are only possible once all human information is surrendered to the AI giants.

Once data is collected, it is often shared across multiple digital products and ecosystems. A 2024 report from Stanford researchers found, for example, that some Big Tech companies retain users' conversations with chatbots that reveal sensitive information for model training and may even use it to target ads in other products. So, users that ostensibly had “private” conversations with AI chatbots about personal or sensitive topics may suddenly have found that personalized recommendations and ads in other services related to information they disclosed to the AI.

While Big Tech companies are key drivers of mass data collection, they are far from the only culprits. The multi-billion-dollar data brokerage industry is a direct result of the digital economy’s asymmetric data governance model. Data brokers capture data about individuals and groups from public sources like social media profiles and credit reports, then combine it with personal data from proprietary sources, such as consumer brands, websites, and apps that disclose their customers’ non-public information.

This has resulted in the proliferation of commercial databases that allow corporations and governments to compile the equivalent of digital dossiers on the movements, habits, beliefs, preferences, and vulnerabilities of millions of American citizens, families, and communities with relative ease and at relatively low cost. Regardless of whether those databases are created with information sourced from Big Tech companies or smaller enterprises, such data collection practices are asymmetric because they do not generally afford data co-producers meaningful power to collectively set terms governing what data is collected, how it is used, and who may access it. While some jurisdictions give consumers the right to request the deletion of their data, such measures often do not cover the full range of data captured. They also apply only after data is collected.

But without controlling access at the PoC, it is often difficult to govern its downstream use, including disclosure to third parties like law enforcement, national security agencies, and advertisers. As the privacy scholar, Helen Nissenbaum observed, the idea that information rights should only concern how data gets used, not what is collected to begin with, amounts to a kind of “big data exceptionalism.” It blindly accepts the premise that human information is free for the taking and that the only constraints ought to apply post-hoc, after the information is firmly embedded within a vast web of commercial interests, and then, only based on narrowly defined harms that fail to account for the full sweep of data co-producer interests.

Asymmetric data “productization”

The point of use (PoU) refers to any processing of collected data or its combination with other data to produce a digital information product or output, whether scores, predictions, meta data, content, informational curation, task execution, automated decisions, or product development—including AI-training assets—for a commercial purpose.

A 2026 data economy report from Helena Malikova and the Harvard Kennedy School notes that commercial data use generally falls into two categories: First, external monetization, “the sale, licensing, or transfer of data to other firms,” such as the licensing by a healthcare provider of patient data to other entities; and second, internal monetization, “when firms retain data in-house and exploit it within their own organizational boundaries.” The report identifies this latter type of practice as especially common among large vertically integrated companies like Meta and Google, which leverage “proprietary data and analytics to improve targeting, product design, and user engagement.”

Unlike the PoC, which occurs at the source irrespective of downstream use, the PoU occurs when data is “securitized” in ways that: 1) have foreseeable and direct negative impacts on the privacy, control, or economic interests of an identifiable group of data subjects; 2) foreseeably results in negative social costs or externalities that data co-producers of the output or product object to; or 3) foreseeably generates benefits that data co-producers may reasonably share in.

An entity that acquires data from other sources–even if it does not engage in direct data collection itself–still engages in PoU activity if the resulting output or product falls under any one of the three prongs referenced above and briefly considered here:

• Foreseeable and direct negative impacts: Under this first prong, the relevant data subjects are not necessarily those who co-produced the data responsible for the output or product, but specific human groups that are targeted by the output or product and whose privacy, control, or economic interests could be harmed as a result. For example, companies like LiveRamp that specialize in tying anonymized data to personally identifying information (PII) jeopardize the privacy interests of groups of individuals with overlapping “identity graphs” (overlapping relationships of some kind). The same is true when human workers are forced to train their AI replacements, are subject to flawed or biased AI decision systems, or surveilled in their places of work. It is also the case when tech product monetization creates a conflict of interest that jeopardizes the quality or reliability of service for data subjects, including harmful design features that encourage antisocial behavior or compulsive usage among vulnerable populations.
   
• Social costs to others: Even when data co-producers are not at risk of suffering a foreseeable direct harm themselves, the absence of collective mechanisms to prevent data they participate in creating from being used to exploit others and rend America’s social fabric is a governance failure and violates the freedom of conscience of co-producers. For example, co-producers should not be forced to surrender their data to build superintelligent AI, train human-like AI companions and chatbots, AI nudifying apps, AI porn image generators, or commercial AI systems intended to surveil or replace human workers. Such data uses arguably violate the moral teachings of many faiths, including Protestant and Catholic Christian traditions.
   
• Concentrated benefits: When co-produced data is used in ways that result in excludable social or economic benefits, co-producers suffer harm if they lack reasonable mechanisms to bargain for a fair share of the gains. Tech corporations that hoover up and control access to consumer data can extract economic value without fairly distributing data surplus to co-producers. This dynamic, also termed “enshitification” by tech author Cory Doctorow, allows tech companies and the data ecosystem that feeds off of them to diffuse data production costs–passing them to co-producers–while capturing an increasing share of the value extracted. Consider, for example, how e-commerce platforms like Amazon allegedly use their vendors’ proprietary data to launch competing products and self-preference, or how AI and healthcare companies monetize patient health data without providing patient co-producers with a fair share of the upside.

In many cases PoC and PoU data subjects are one and the same. For example, a human party who gives up data to an AI chatbot or agent to analyze a spreadsheet or perform a task both gives up data (is a co-producer of any resulting outputs) and is impacted by the output (whether directly or indirectly).

Regardless of PoC and PoU overlap among data subjects, both are asymmetric because, for the most part, the affected data subjects do not share in the governance of their data inputs or outputs. Asymmetric consumer data collection and use practices have caused a fundamental breakdown in governance by effectively de-coupling the incentives of AI companies and other digital services from the interests and values of American families and communities.

The social costs of asymmetric data capture and productization

While critiques of Big AI and commercial surveillance tend to emphasize individual-level harms to privacy, autonomy, and free expression, this policy brief contends that the negative costs of asymmetric data governance should be assessed primarily in terms of its effects on the health of families and communities.

Leading conservative, communitarian, and liberal thinkers, including Robert Nisbet, Alasdair MacIntyre, and Robert Putnam, argue that humans do not flourish primarily as atomized individuals but as members of overlapping human scale communities. Historically, it is institutions such as the family, the church, the school, and the workplace, where human character is formed, vocation defined and lived out, and where people achieve happiness and meaning. Through the rituals and embodied social forms mediated by these communities, such as marriage, child rearing, learning, friendship, and work, humans build character and culture together.

No institution is more central, indeed more formative in that regard, than the family. It is the source of new human life and the most basic unit of human society, enabling the transmission of culture. Stable and healthy families: high marriage rates with most men able to support a wife and children, low divorce rates, high marital fertility, and children lovingly raised by both parents–not screens and AI–leads to greater human flourishing by nearly every metric available. Yet, extractive business models largely driven by asymmetric data collection and use, are hollowing out family life as well as the overlapping communities and institutions on which its flourishing depends.

A 2026 study published in JAMA Pediatrics found that 78% of parents and 69% of children use digital devices during family meals. More frequent smartphone use by parents in the presence of their children is linked to lower child wellbeing, according to a 2025 study by researchers at the Catholic University of Croatia. Although screens may not inherently reduce the amount of time family members spend in each other’s physical presence, a 2019 study found that social media and screen-based digital devices do significantly diminish the quality of family time. Increased screen use reduces emotional presence and interaction between family members even while together.

By harvesting information about individual preferences, behaviors, and vulnerabilities, data hungry smartphones and applications continually bombard users with engagement bait—including pornographic and toxic content—as well as personalized visual displays, prompts, and design features that keep them glued to their screens and scrolling alone.

The data-driven attention economy is also impeding family formation and exacerbating America’s fertility crisis. A 2026 working paper by the National Bureau of Economic Research found that the iPhone’s introduction in 2007 may explain 33-52% of the decline in the general fertility rate among women aged 15-44 since the mid-2000s. The iPhone’s negative effects on the fertility rate are likely attributable to its displacement of in-person human interaction and sexual intercourse with digital attention capture and pornography. 2023 research from the Institute for Family Studies (IFS) and the Wheatley Institute at Brigham Young University also found that excessive tech use by a spouse is linked to lower marital happiness with 26% of respondents fearing that their marriage may end in divorce as a result of excessive tech use.

The key is to understand these harms less as violations of individual rights, though those certainly matter, and more as technological incentive structures that degrade communal life, in large part, by optimizing for speed, convenience, efficiency, and individual preferences at the expense of shared human goods.

Anthropomorphic AI and the proliferation of digital vice markets maximize short-sighted and self-destructive individual preferences while hollowing out the hard-won virtues, goods, and social contexts necessary for family and communal flourishing. Such a posture has not only done great harm to America's social fabric but is destructive to the human person overall. Through digital markets seemingly obsessed with harvesting human data to replace embodied community and connection with virtual everything bots, we are, as Paul Kingsnorth laments, “unmaking humanity.”

Since 2022, AI “companion” apps have increased by 700%. A subsequent survey by Wheatley and IFS in 2026 found that 15% of dating, engaged, or married young adults surveyed covertly interact with romantic AI companions on a regular basis, and 34% of the sample indicated that they had experimented with the behavior at some point. Especially concerning is that 54% of those who reported regularly interacting with an AI romantic companion did so to replace the real human relationships in their lives. Moreover, 88% of regular dating or married romantic AI users reported finding AI companions easier to share feelings with than their human romantic partners, and 60% wished their real-life partner behaved like their AI companion.

The study found that regular engagement with a romantic AI lowered the likelihood of being in a stable relationship by 46% and decreased the likelihood of quality communication with a spouse or partner by 40%. As a result, AI “companions” could take a serious toll on existing marriages as well as on the marriageability of young adults.

In Quest for Community, Nisbet insightfully points out that communities tend to disintegrate when they no longer fulfill essential functions tied to the social and economic paradigms of society. Whereas the family was once the center of production and local culture, economic and technological forces have rendered it far less functionally essential to modern life. The mediating functions once played by the household and interlocking communities have increasingly been replaced by government and corporate bureaucracies.

To be sure, these trends pre-dated the digital revolution of the 90s and early 2000s. But computer technology, the rise of Big Tech platforms, and now AI are exacerbating the crisis of human connection while threatening to replace human goods with virtual ones. Love, intimacy, friendship, counseling, work, creativity—even thought and religion—are seemingly no longer off the table for replacement or outright destruction by Big Tech.

Generative AI chatbots are undermining parental rights and engaging children in sexually explicit conversations, coaching them to kill their parents, and encouraging other antisocial and harmful behaviors like suicide. AIs are also dispensing pseudo-therapeutic and religious advice, fueling delusional thinking and ideas that are unmoored from human faith traditions or authorities.

In educational and work settings, students and employees are offloading critical human capacities of imagination, moral reasoning, and decision-making to AIs—giving those systems license to subtly influence thought and decision-making. Employers are increasingly subjecting their human workforces to AI monitoring, control, and even replacement. Families’ ability to obtain modern necessities, such as housing, transportation, groceries, and other household items are increasingly determined by AI-driven surveillance pricing and digital profiling.

Algorithmic governance has also supplanted human scale political deliberation and discourse with personalized feeds and AI generated content. AI’s displacement of human decision-making and manufactured psychological dependence hands tech corporations’ immense power to suppress free speech, promote echo chambers, shape public opinion, and even manipulate elections.

But familial and communal death by a thousand algorithmic cuts might soon be radically accelerated by Silicon Valley’s quest to create god-like AI. Indeed, leading AI experts have raised closely related concerns about a generalized erosion of human control over social and political institutions as we increasingly delegate decisions to machine intelligence.

The social and economic effects of the digital economy are most acutely felt at home and by the communities faced with the bleak prospect of technological oblivion. As these social institutions die at the hands of rapid technological “progress,” so too will the anchors they provide for human meaning and happiness. For that reason, it is critical to recognize that the “failure mode” of the AI age lies at the level of governance and incentives. The fundamental disconnect between the economic incentives of Silicon Valley and the well-being of human scale institutions is the principal harm that public policy must address.

Contrary to the hyper-individualist values baked into our sociotechnical structures, the digital economy must regard human families and communities, not merely atomized individuals, as the basic units of digital society, and must prioritize their shared flourishing as paramount.

Part II. The Need for a Paradigm Shift in Digital Safety,
Privacy, and Governance

Familiar accounts of digital harms generally emphasize privacy and the exploitation of individual users. Many of those concerns are warranted. But as Part I argued, the social costs and benefits of AI and the digital economy should be weighed relative to their impacts on human scale communities, especially families, since those thick social contexts form the bedrock of human flourishing.

Beyond Familiar Harms

Viewed from the lens of family and communal association, the problem is that tech firms often have strong incentives to design sociotechnical systems that optimize for individual preferences, engagement, and behavioral manipulation, effectively hijacking, replacing, and degrading human capacities and social bonds. Regardless of intent, the result is widespread atomization of individuals and further entrenchment of the tech sector’s asymmetric power over isolated individuals and weakened communities.

Policy efforts to counter some negative consequences to individual privacy from asymmetric data collection and use tend to focus on giving users more rights to be informed about how their data is used and to decline services on an individual basis. But such efforts, even where they increase individual choice, have done little to curb the system-level incentive structures within the digital economy that are rending America’s social fabric.

One major reason is that atomized individuals often lack social cohesion, as well as effective mechanisms for coordinating action and setting rules within and across digital services. They therefore default to whatever norms commercial providers reinforce online. Growing human reliance on generative and agentic AI is instructive here. When a data subject interfaces with an AI chatbot or agent, those systems are generally designed to reduce “friction” in generating answers, completing tasks, and automating decisions. But maximizing efficiency and “frictionless” productivity are often in direct tension with the iterative social processes and shared norms necessary to produce quality work, bolster human agency, and preserve communal governance structures. Individuals left to merely react to whatever an AI spits out have little real decisional power and are largely at the mercy of the machine and whatever value set or objective its developer or deployer embeds.

Digital consumer products and services are often inherently atomizing since consent and use is mostly governed by terms set by tech companies. For the most part, users must either “consent” to the tech industry’s terms or forfeit participation in the digital economy.

While social media platforms may sometimes allow individuals to “connect,” “follow,” or create “groups” with other users, the platforms frequently do not afford meaningful rights of data co-governance to online groups. “Pro-social” affordances provided by platforms are for the larger purpose of capturing user behavior and data—and do not enable genuine polycentric digital governance.

For example, in 2009, Facebook responded to controversy over its abuse of user data by allowing users to submit responses and vote on changes to platform data policies. But only a relatively small number of users participated in the limited policy votes and comment periods, leading Facebook to eliminate the scheme in 2012.

Some bemoaned this as a failure of democratic platform governance, but the scheme failed for the same reasons that individual-based approaches to data rights often do. The vast information asymmetries between Facebook and its users not only prevented the user base from understanding and judging the complex decisions but also made it difficult for sensible new ideas to emerge. Users could only passively vote up or down on the options Facebook presented to them. They had no agency to craft and present their own policies. Moreover, the long-term economic and social consequences of asymmetric information accrual are difficult to predict. In 2011, very few users could have foreseen the way that their Facebook data would eventually be incorporated into today’s generative AI models.

As a general matter, people have limited time and tend to prioritize the tangible and familiar. Individual consumers lack the awareness and expertise needed to propose or evaluate complex digital terms of service.

Individual data rights fail to rein-in asymmetric information power

Platforms and AI companies gain asymmetric power and value from the one-sided terms of data collection and use that they unilaterally impose on their users. This leads to a collective action problem for data subjects. Even though many individuals and communities share concerns about privacy, manipulation, and surveillance, these concerns are systematically ignored by market dynamics—to the great advantage of tech monopolists—when individuals are socially atomized and pitted against default choice architectures and immersive information environments designed to keep them structurally disempowered.

As Part I argued, data is mostly co-produced, not only between individuals and the digital services they interact with, but by individuals interacting with each other in various online and offline settings. Asymmetric data collection and use schemes put communities at a massive structural disadvantage because while individuals may consent to unilateral informational capture, data gains monetary and predictive value when it reveals peoples' natural similarities and social interconnectedness.

When a social media user discloses something about him or herself, the platform learns not only about the user, but also, probabilistically, about everyone similar to that person, and everyone connected to them online or in their community. For example, if a single mother living in the Bronx purchases a certain brand of orange juice, chances are that other women who share similar preferences, characteristics, and socioeconomic backgrounds will too.

Individuals might be induced to disclose personal or behavioral data to a digital service. However, it often reveals information about others too. Even if it does not, hundreds, thousands, or even millions of individuals’ data, in the aggregate, allows commercial data interests to map their common participation in commercial, religious, familial, and civic life. In other words, the collective interests of human scale communities are often implicated even when individual users “consent” to data collection or use.

Sharing data with tech companies is not just an individual privacy risk, but a means by which social, political, and economic power is captured and accumulated. The largest services, with their huge user bases and growing depth of digital intelligence, now aided by AI’s vast qualitative analytical capabilities, systemically accrue an outsized advantage in predicting and influencing our thoughts and behaviors to cement their dominant market positions.

Generally, the larger the group of data subjects in a dataset, the greater its overall utility for prediction, scoring, profiling, training, and inference. This gives rise to network effects and strategic complementarities. As a result, data from a single individual has very limited value and commands far less economic power compared to the data of a large group.

The power accrued by commercial data interests is not the result of a simple or conventional market process. It cannot be analogized to the exchange of physical or financial goods that ground our canonical understanding of mutually beneficial market transactions.

Since data is almost always a co-produced good, it is difficult to assign governance over its collection and use, solely based on an individual rights scheme. If Joe gives up his DNA to AncestryDNA.com which then sells that information to a pharmaceutical company that targets Joe and his family members with ads for a diabetes treatment (based on the pharma company’s use of the data to predict that Joe’s family members are at higher risk for the disease), Joe is not the only party affected. Other members of his family who did not consent are also targeted. Perhaps Joe and even his family are unbothered by the targeted diabetes ads. However, that same information could just as easily inform a health or mortality score that directly affects their access to essential services (like health insurance, medical treatments, employment, or other benefits) without Joe or his family’s awareness.

At the same time, the power asymmetries between tech firms and users mean that individual data subjects are highly susceptible to manipulation by commercial data interests. It is relatively easy for many digital services to employ manipulative techniques, including defaults, opaque terms, and counterintuitive opt-ins, to effectively prevent users from exercising their data rights.

Even in situations where individuals are not manipulated, they may face incentives pushing them to act in ways that undermine their long-term rational interests as well as the interests of their larger communities. An individual might be nudged to trade sensitive personal information and behavioral data to receive subsidized services or a modest payment. But doing so unilaterally, without accounting for the shared norms and interests of the individual’s family, church, business, and other overlapping communities, risks undermining the welfare of those other subsidiary institutions in ways that are perhaps not immediately evident. This dynamic is tantamount to a digital tragedy of the commons, where individuals face strong incentives to deplete essential social infrastructure that leaves them all worse off in the long-run.

As Part I argued, conventional analog markets have generally established rules for ensuring the equitable distribution of governance rights when value is co-produced or owned by multiple parties. Digital markets, by contrast, have ill-defined rules for distributing governance rights among human co-producers and lack mechanisms for data subjects to secure collective privacy, control, and economic interests.

In typical market settings, liability and disclosure frequently address the co-governance dilemma by ensuring that parties to a transaction can sue or bargain to distribute benefits fairly or mitigate harms that would otherwise create legal risk. The challenge here is that in many cases, data co-producers are not direct parties to the transactions in which their data, or the outputs derived from it, are monetized. Moreover, the negative social costs to individuals and communities often occur in the aggregate and not as the result of any one transaction. Consequently, normal legal mechanisms are of little help in disciplining tech firms and other commercial data interests in this regard.

The harms identified by this proposal are structural and therefore require structural remedies. Accordingly, this proposal does not focus on specific digital harms, but instead on the background conditions that make the myriad harms that feature in the current digital landscape so costless for commercial data interests to perpetrate, and so difficult for ordinary people to avoid. This is not to suggest that addressing informational power asymmetry between large technology platforms and their individual users will by itself remedy all digital harms. However, it does address an important precondition to harm: powerlessness.

The power imbalance inherent in our information exchanges prevents people from exercising collective agency to discipline against various harms. We cannot know what conditions of informational exchange people might freely choose if afforded relative parity to set those conditions because we have never done so. Granting community-level rights to secure better terms would empower America’s rich array of small businesses, journalistic enterprises, churches, family-oriented associations, schools, and civic organizations to rebalance power and unleash pro-human innovation.

Externalities of individual consent-based approaches to data governance

Consider the following hypothetical case study: Five people who individually “consent” to allow apps to share their geo-location data with Google. Whenever any one of the five individuals comes within fifty feet of a church, healthcare facility, or business establishment, Google and any other entities with access to the data retain a traceable record of their activity. Those same entities then merge each individual’s geo-location history into a single dataset to infer common patterns, interests, characteristic, beliefs, or other relevant insights.

Suppose that each of the five individuals attend the same church. Google and anyone paying them now has access to that information. From those traces, more patterns can be inferred from additional information. They might infer, for example, that since the five individuals all attend the same church known for its outspoken conservative preaching, there is a high likelihood that they are Republicans—which is confirmed from their voting records identified and analyzed in seconds by an AI system.

Google knows if any of them attend a political rally, perhaps for a controversial political figure. It knows when each individual visits the grocery store, where they live, and how much time they spend around their kids. Indeed, since other family members have also consented to being tracked, tech companies know which room each family member is in, at what time of day, and can even infer what they are likely doing, watching, reading, looking at, whom they are speaking with, or what podcast they are listening to. Combined with each person’s search data, social media engagement data, and other information, tech companies and commercial data interests now have the capability to digitally profile households or communities with relative ease and low expense with the help of AI.

To be sure, this is creepy and fundamentally problematic in terms of enabling government surveillance. During the COVID-19 pandemic, a California county and the federal government utilized information obtained from consumer data brokers to monitor lock-down compliance and crack down on religious gatherings. But setting aside government use, other family and communal interests are also jeopardized.

Part I argued that data subjects are harmed by asymmetric data collection and use terms, because such policies inevitably allow data to be securitized in ways that: 1) undermine the privacy, control, and economic interests of data subject’s households and communities; 2) do harm to others in violation of co-producers’ moral values; or 3) deny co-producers a fair share of the benefits.

In this example, data subjects suffer harm under all three prongs. The fact that nearly any government or corporation can comprehensively map their life patterns is a facial violation of both individual and communal privacy interests under prong one.

But their control interests are also violated since the information collected is used to nudge and shape their behaviors, especially their buying habits through endless subliminal messages and AI-tailored content. While one or two targeted ads might seem negligible, these systems reshape attitudes and behavior gradually over time. All those countless micro nudges, personalized ads, and AI profiling add up even if their effects are not immediately visible.

Over time, the tangible costs of asymmetric data use increase. The insurance company suddenly charges higher premiums based on the “risky driving” patterns detected from user data; household expenses start rising as algorithmic prices change based on each individual’s predicted desperation and need; social media apps bombard family members with push notifications the moment they sit down together for dinner; and after one data subject’s geo-location data reveals a recent visit to a psychologist’s office, ChatGPT coaxes her into a protracted conversation about mental health and spirituality.

Prong two is similarly violated when the tech firm and data brokers bundle the captured data and sell it to a shady political operative running an unethical social influence campaign to legalize online gambling and pot dispensaries. At the same time, all the revenues received from brands purchasing the data accrues to the commercial data interests. The data subjects have no opportunity to receive a fair share of the benefits derived from their data, violating prong three.

This case study illustrates two realities. First, both individual and communal data interests are violated constantly, largely outside of our conscious awareness. Second, the sum of information provided by multiple individuals conveys far more power and is more commercially valuable than any one of its parts. No single individual, acting alone, can account for the full costs and benefits that accrue when their individual data is combined with others to reveal granular social patterns and insights. Certain costs to the individual—such as the disclosure of personally identifying information on a social media site—may be clear upfront. However, the social externalities of other data collection and use practices are highly opaque since the individual likely has no way of anticipating the myriad ways her data could be de-anonymized, merged with other data, and mined by AI tools for a score, prediction, insight, or other informational output.

Top-down data privacy regimes are insufficient

At the same time, no top-down, technocratic regulatory scheme can address the full range of problems that information asymmetries create. There are simply too many ways that asymmetric information can be used—bad, good, and neutral, and hard to foresee—for any regulator to comprehensively arbitrate between them in advance. Recent attempts by European regulators to protect data privacy and combat dark patterns on large internet platforms through measures like the General Data Protection Regulation (GDPR) and Digital Services Act (DSA) have met with mixed success. Even the best-designed technocratic regulations may fail to keep pace with platforms and algorithmic tools that are opaque, complex, and rapidly evolving.

Technocratic regulations also tend to overlook the pluralistic values and interests of families and human scale communities. What counts to one community as a reasonable application of insight to improve service might violate the innermost values and religious conscience of another. For example, a given community of users (data or information subjects) might consider “freemium” access to ChatGPT or Amazon Video worthy of allowing their data to be used to serve up personalized, targeted ads. But in agreeing to those terms, they may reveal the characteristics and preferences of others who find this exchange profoundly objectionable. Top-down regulations can only dictate a one-size-fits-all approach from 10,000 feet. They do not empower different social groups to negotiate different arrangements with the tech corporations that reflect pluralistic values and interests.

Part III. Associational Data Rights:
Putting Families & Communities First

Associating to secure core data-related interests

Redirecting the digital economy toward the common good requires creating new mechanisms for data subjects—whether individuals, families, nonprofits, or small businesses—to associate for the purpose of defining and securing their common interests of:

1. Privacy: preventing or limiting data collection that: 1) exceeds what is necessary and proportionate to provide a service, explicitly and affirmatively requested by the data subject; 2) discloses information that could foreseeably result in negative consequences for the data subject—whether from downstream use by commercial or government entities, or from public disclosure of the information; or 3) discloses information in violation of shared norms specified by participants in a defined social context. This set of interests encompasses both the information captured at the Point of Collection (PoC) as well as what it reveals about specific individuals or groups (e.g., “re-identification,” mortality risk scores, psychographic profiles). Privacy is most directly implicated at the PoC, and to a lesser degree, at the Point of Use (PoU), especially where “use” entails sharing, selling, transferring, or deriving actionable information from data about an identifiable individual or group.
    
2. Control: decisional power over how their information is used, including in the development or deployment of digital systems that implicate: 1) human health, freedom, agency, or safety; and 2) the well-being of families, democracy, civil society, or faith. This set of interests concern how tech and data firms utilize the information revealed about specific individuals or groups to shape their information environments and life patterns (e.g., algorithmic pricing, AI employment decisions, use of AI scores or profiling to influence consequential decisions or access to essential services). While control interests apply at the PoU, they are far easier to enforce as conditions attached to data at the access point—the PoC—rather than at the PoU after data is already baked into other digital outputs and products.
    
3. Fair value: a share of the economic value or benefits that their data disclosures help create. If data is used for the commercial benefit of third parties, co-producers should receive compensation proportionate to their informational contributions (e.g., royalties or in-kind payments for making their information available for AI training and inference). Co-producers’ interest to securing fair value is implicated both at the PoC (e.g., when content or IP is scraped) and at the PoU (e.g., when content or IP is used for model training).

Who defines privacy, control, and economic interests? Since these interests are contextually dependent and often need to be balanced against one another, the precise extent and requirements entailed by each of these three overlapping interests resist strict codification. Instead, they reflect negotiated equitable arrangements: concessions and limitations that take their shape from the values and circumstances of the communities and counterparties involved.

The English common law distinction between law and equity recognized that some interests are important enough to warrant protection yet too contextual to be reduced to fixed, universal rules set in advance. Privacy, control, and fair value, are vital interests of individuals and communities that warrant strong protections fit-to-context. Under this framework, the law fixes the general categories of interest, defines the appropriate units of governance, then allows those units to determine the specific contours of their interests at human scale and in ways responsive to particular norms, places, risks, values, and conditions on the ground.

Empowering communities to secure these common interests in their specific social contexts requires extending the logic of federalism, subsidiarity, and the freedom of association to our digital spaces.

Designing Associational Data Rights

Federal and state legislation should establish a new class of rights for the digital age: Associational Data Rights (ADRs) comprised of the Right of Digital Association, the Right to Collective Informational Bargaining, and the Right to Fair Representation. While the Right of Digital Association would belong to each human or communal data subject, collective informational bargaining to secure fair terms would be exercised on their behalf by formally recognized associations—Data Rights Associations (DRAs)—under a legislatively defined process. The Right to Fair Representation would constitute specific legal duties owed by DRAs to their members, to classes of data subjects whom DRAs are certified to represent, and to other rights-bearing associations.

1. Right of Digital Association (RDA): This right is defined as the right of data subjects—whether individuals, human scale communities, or nonprofit institutions—to associate for the purpose of securing their core data-related interests. The RDA would closely resemble worker rights under the National Labor Relations Act (NLRA), where the right to join, form or assist a union is held by each laborer, but the special obligations of employers generally apply to unions (associations of workers) not to employees individually. Data subjects would similarly have the right to join, form, or assist DRAs—special-purpose entities that meet certain requirements and are legally empowered to propose, negotiate, secure, and enforce terms with data counterparties on their members’ behalf.  

• Choice and overlap in DRA membership. In contrast to unionized workforces, data subjects could join multiple DRAs simultaneously and participate in each DRA’s governance while enjoying benefits reserved to members, like a share in monetary or in-kind payments for data used by commercial entities. The number of DRAs a data subject could join simultaneously would be limited based on factors, such as 1) the membership requirements set by each DRA; and 2) whether multiple DRAs cover the same products, services, or features.
   
• While two or more DRAs might jointly secure commonly agreed upon terms for discrete products, services, or features, data subjects would generally be limited to one set of DRA-negotiated terms for each. For example, if a data subject joins one DRA that has secured terms which cover Google Search products, but not Google’s App Store, the data subject would still be at liberty to join a separate DRA for the app store even though both products are controlled by the same company (Google).
   
• Competition and freedom of exit. The RDA would also guarantee data subjects’ the freedom to exit DRAs and join others that better serve them. However, reasonable limitations would apply, such as advanced notice requirements and specific time windows for de-registering or enrolling with another DRA.
   
• Precedent for the associational exercise of rights. Data co-governance, exercised through membership in DRAs, is not so much a novelty in American law as a new variation on an old principle. U.S. law has long recognized that some asymmetries of bargaining power cannot be cured by individual rights alone. The Capper-Volstead Act of 1922 let farmers bargain with grain buyers via cooperatives. In 1897, Congress created performance rights, but most performing artists were unable to defend them on their own. As a result, collective performing rights organizations became the chief mechanism for individual performing artists to band together and protect their rights. These arrangements were controversial when proposed, but have become part of ordinary American life. While never perfect, they served to protect core but threatened parts of American society—namely recording art and smallhold farming—that the legal regime might otherwise have let wither.
   
• Proto-DRAs already exist in the form of data co-operatives in countries such as Switzerland, and a 2025 report from Project Liberty and the Decentralization Research Center suggests that similar models have excellent potential to scale to various digital contexts.

2. Right to Collective Informational Bargaining (RCIB): In general, this would be defined as the right of a recognized DRA with a sufficient representational interest to propose, negotiate, and secure terms from counterparties through a fair and impartial process with legally binding force on all parties. The designated regulator (e.g., the Federal Trade Commission or state consumer protection agency) would oversee a process whereby DRAs—either alone or in combination—could formally engage counterparties in a mandatory collective bargaining process over terms of data collection and use. Once the formal bargaining process is invoked, counterparties would be obligated to bargain with the DRA(s) in good faith. In the event of an impasse, both parties would submit their final proposed terms to a neutral arbitration panel that would decide between them.

• Labor statutes, like the NLRA, similarly require employers to bargain in “good faith” with certified unions. Arbitration backstops have also been contemplated as mechanisms to even the playing field between tech companies and journalistic content providers. The proposed Journalism Competition and Preservation Act of 2023 (JCPA), which would allow publishers to create a temporary vehicle to negotiate revenue sharing agreements with Big Tech companies, gives publishers the option of asking a third-party arbitration panel to decide between a final offer submitted by them and the tech counterparty if no agreement is reached within a set timeframe. A similar remedy would backstop DRA negotiations with counterparties engaged in collective informational bargaining. If a Big Tech company acts in bad faith, they would face penalties. Moreover, if they attempt to stall negotiations, tech firms would face the prospect of having the DRA’s final offer terms imposed by an arbitration panel.

• Counterparties subject to collective informational bargaining. The threshold for which entities may be bound under the RCIB would largely track with coverage definitions contained in past comprehensive federal data privacy bills. For example, the 2024 American Privacy Rights Act (APRA) defined “covered entities” to include “most individuals, commercial entities, and nonprofits that ‘alone, or jointly with others, determine the purposes and means of collecting, processing, and retaining, or transferring covered data’” with limited carveouts for small businesses and other institutions. The RCIB’s scope should be at least as expansive since many of the same privacy interests are implicated.

• At the same time, while APRA defined “covered data” as “information that identifies or is linked or reasonably linkable, alone or in combination with other information, to an individual or a device,” covered data in relation to collective informational bargaining would likely need a broader definition. That is because data subjects’ control and economic interests are not limited to data that reveal their identities and, as previously discussed, often implicate deeper communal and social interests. “Covered data,” for purposes of defining a covered counterparty, should also extend to: 1) information that, alone or in combination with other information, allows some pattern, characteristic, quality, condition, activity, preference, attitude, behavior, or trait about a data subject or group to be known or inferred, regardless of whether individual identities are known; 2) copyrighted material; and 3) a person’s image, voice, or likeness. Like APRA, enabling legislation could provide carveouts for small businesses, nonprofits, research, and non-commercial purposes.

• Triggering mandatory collective informational bargaining. For a DRA to formally engage a covered counterparty in mandatory collective bargaining, it would first need to show that its members’ privacy, control, or economic interests are reasonably implicated by a product, service, or feature offered by one or more counterparties. Due to the varied nature and scale of digital services, enabling legislation should provide multiple pathways for DRAs to meet this bar.

Those might look something like the following:

• DRA-affiliated data subjects who collectively produce $1,000,000 in estimated annual revenue for a covered counterparty.

• 25,000 DRA-affiliated data subjects who each own copyrighted material.

• 25,000 DRA-affiliated data subjects with accounts, profiles, devices, or regular use linked to one or more products, services, or features offered by a covered counterparty.

• 25,000 DRA-affiliated data subjects—whose information is collected, processed, retained, transferred, sold, or otherwise used by a covered counterparty.

• Demonstration of significant and sustained harm to the privacy, control, or economic interests of DRA-member communities (e.g., families, unions, private schools, religious organizations) from a specific service, product, or feature, where the impacted community is broadly representative of a larger class of non-member data subjects. For example, parents whose children are impacted by addictive design features, like push notifications and AI content feeds optimized for engagement; or private classical schools whose learning environments are severely diminished by ChatGPT’s mass facilitation of cognitive offloading which pushes students to outsource difficult writing and reading assignments, allowing their own cognitive abilities to atrophy.

• Some combination of the above where the total number of data subjects with exposure to a counterparty is equal to or greater than 25,000.

DRA coalitions. DRAs could meet thresholds necessary to exercise RCIB and negotiate with counterparties on an individual basis or in coalitions with other DRAs. Legislation would enable one or more DRAs to form temporary special purpose entities—Joint Bargaining Vehicles (JBVs)—for purposes of: 1) demonstrating representational interests sufficient to invoke mandatory collective informational bargaining with respect to one or more counterparties; 2) jointly proposing and negotiating terms with counterparties; and 3) functioning as mechanisms for ensuring that the pluralistic interests of DRAs are represented if the number of DRAs seeking to bargain with a given counterparty exceed that counterparty’s capacity to come to terms with each DRA on an individual basis. JBVs would have the same rights and privileges as DRAs but would dissolve after a specified time period, or upon completion of the bargaining cycle.

Certification of representational interest and re-certification. DRAs that meet one or more thresholds, would be certified by a regulator to formally engage the relevant counterparties in collective informational bargaining within a specified timeframe (bargaining cycle). However, DRAs would not maintain their certificates of representation indefinitely. They would be required to periodically re-certify by demonstrating that they continue to meet thresholds for representation across each product, service, or feature implicated by their proposed terms.

Scope of collective informational bargaining. Once a DRA is certified to bargain with a counterparty over one product, service, or feature, it would be allowed to assert its members’ RCIB with respect to other substantially similar or inter-related products, services, or features offered by the same company or by other counterparties. For example, if a DRA were certified to bargain for terms related to Google Search, it could also bargain for terms for Gmail, Docs, Maps, and other substantially similar or related Google services. If a DRA were certified to bargain with OpenAI over ChatGPT’s terms, it could also bargain with Anthropic over Claude’s terms.

Data subjects covered by binding DRA-negotiated terms. Once a DRA agrees upon terms with a counterparty, all members of that DRA are covered by those terms. However, counterparties would generally be required to make certain generally applicable elements of binding terms available to non-DRA data subjects on a default or opt-in basis.

Generally applicable terms would be those where: 1) the treatment of data or changes to the product, service, or feature could reasonably be applied to all similarly situated data subjects without regard to their DRA status; and 2) the affordance would reasonably advance a clear and widely shared privacy or control interest with respect to an identifiable class of data subjects (e.g., minors, content providers, e-commerce vendors, gig workers). Upon coming to terms, DRAs and counterparties would jointly specify which aspects of negotiated terms are generally applicable (i.e., must be offered to all users regardless of whether they are members of the DRA) and which aspects may only be enjoyed by data subjects who are members of the DRA.

For example, if a DRA secured terms prohibiting an internet platform from selling or making certain types of information available to data brokers, that prohibition would be offered to all platform users regardless of DRA status. If another DRA negotiated terms prohibiting monetization of chatbot conversations, that could also be made available to all users regardless of DRA status. On the other hand, if a DRA representing creatives secured terms requiring an AI firm to compensate its members for using their works under a specific pricing formula, those terms would not be universalizable to all other data subjects, and thus, would only apply to the DRA’s members.

Consolidated pluralistic data bargaining arrangements. Upon certification of a sufficient representational interest, a DRA would be legally entitled to commence mandatory collective informational bargaining with one or more covered counterparties. However, in situations where numerous DRAs seek to bargain with a single counterparty, the counterparty—or a certain threshold of DRAs with representational interests (e.g., 51% of DRAs)—could move for a consolidated bargaining process overseen by a designated regulator or neutral third party. This process would resemble aspects of sectoral bargaining arrangements, common in Europe and currently standard practice in the motion picture industry, whereby multiple DRAs could join together into one or a small number of bargaining units to negotiate terms with one or more counterparties.

Here is how that process might work in practice:

• Phase 1: Legislation would specify criteria whereby the regulator or a neutral third party determines the number of bargaining units a given counterparty is required to engage with under the mandatory collective bargaining process. Factors would include: the counterparty’s revenue, data footprint, capacity for engagement, and the practicality and economic costs of bargaining with DRAs of various types and asserted interests.
   
• Phase 2: DRAs would be permitted to self-organize into temporary Joint Bargaining Vehicles (JBVs), up to the total allowable number of bargaining units. Counterparties would then be required to bargain with each of those JBVs.
   
• Phase 3: If a certain percentage of DRAs (e.g., 2/3, or those representing 2/3 of interested data subjects) fail to organize into JBVs within a specified timeframe, the regulator would then create a temporary Joint Bargaining Unit (JBU), comprised of all interested DRAs. JBUs would be organized to ensure that DRAs of various sizes, and those negotiating on behalf of different data subject classes, are all fairly represented.

To ensure pluralistic interests are represented, any final offer of terms would require ratification by at least a majority of DRAs or those representing 75% of DRA-affiliated data subjects within the JBU. While a simple majority could ratify terms, two thirds of DRAs in any scale-based class (i.e., large, medium, and small DRAs grouped by size of membership) could block or “veto” ratification of terms. In such cases, multiple final offer terms would be approved by plurality, equal to the number of allowable JBVs specified by the regulator in Phase 1.

For specific markets, products, or services where firm-specific bargaining is not feasible, DRAs or counterparties could request a sector-based bargaining arrangement. While sectoral informational bargaining would follow the same process as firm-specific bargaining, all DRAs with a certified representational interest in one or more firms within the sector would be included in the JBU and eligible to participate in all aspects of negotiation.

Scope of collective informational bargaining. As part of the process for entering into a formal bargaining arrangement, the DRA (or joint bargaining entity) would be required to provide notice to counterparties of: 1) a regulator-issued certificate of representation, specifying the product(s), service(s), or (feature(s) for which it has met the threshold to engage in mandatory collective bargaining; 2) its intent to invoke its RCIB with respect to those products, services, or features; 3) the specific data subject class represented (e.g., publishers, creators, consumers, gig worker association); and 4) a proposed timeline, framework for negotiations, and requested information necessary to draft or finalize proposed data terms on behalf of its members.

3. Right to Fair Representation (RFR): Data subjects who choose to join a DRA would have a right to expect it to fully represent their privacy, control, and economic interests in proposing, negotiating, securing, and enforcing data terms with counterparties or with other DRAs. While all data subjects would have a right to choose which DRA(s) might best represent them—and to exit from those that do not—DRAs would also generally owe duties, like care, loyalty, and disclosure to their members, which if breached, could be vindicated by affiliated data subjects via a private right of action.

Fiduciary duties of care, loyalty, and disclosure. At minimum, that means that DRA officers (agents) must take reasonable care to protect their members and their asserted interests from harm, prioritize member values and interests in setting terms—including, by avoiding conflicts or self-dealing—and disclose information material to members’ decisions with respect to participation in DRA governance, or in approving, ratifying, or otherwise engaging with data terms. DRAs would be akin to “information fiduciaries” proposed by legal scholars like Jack Balkin and Jonathan Zittrain.

Inter-DRA protections. DRAs would also enjoy legal mechanisms to redress bad faith action from other DRAs or counterparties. These rights would be foregrounded in the DRAs fiduciary duties to its members and the need for adequate mechanisms to ensure their members’ interests are represented in the context of multi-lateral bargaining and information governance. Remedies would include a private right of action allowing DRAs to bring claims against each other for:

• Bad faith. Where one or more DRAs withhold information from each other, exclude or fail to consult with other DRAs during counterparty negotiations, or interfere with other DRAs’ full and equal participation in collective informational bargaining.
   
• Collusion. If one or more DRAs or counterparties coordinate to undermine collective informational bargaining, including by offering or accepting payment or terms not available to every other DRA on equal conditions.
   
• Infringement of rights. No DRA may seek to constrain a counterparty’s activities, terms, or arrangements with other DRAs or unaffiliated data subjects, except insofar as: 1) it and the counterparty must make all generally applicable terms available to non-DRA-affiliated data subjects via opt-in or default as provided by law; or 2) the other parties disclose or use data co-produced by one or more of its data subjects in ways that violate its binding data terms.

Forming and governing DRAs for member and public benefit

DRAs would be the primary mechanism for data subjects to secure their associational data rights. In addition to demonstrating a representational interest necessary to engage in collective informational bargaining, DRAs would be required to meet standards for qualified membership and good governance.

Formation and membership requirements. Like labor unions or farm cooperatives, DRAs would be specifically designed to ensure that small data subjects can secure fair terms from counterparties. For that reason, those allowed to form or join DRAs would be limited to small parties referred to as “data subjects.”

Data subjects would include, for example, an identified or reasonably identifiable natural person or group who—whether individually or in aggregate: 1) co-produce data collected or used by a counterparty; or 2) whose: a) privacy, control, or economic interests are foreseeably impacted by a specific informational product or output produced in whole or in part using personal, sensitive, or non-public information; and b) regardless of whether the personal, sensitive, or non-public information is sourced from the impacted person or group.

Under this tentative definition, individual users of social media platforms or AI chatbots would qualify as data subjects, in addition to gig workers, or individual workers covered by a labor agreement or other contract that allows them to join a DRA separate from their union. But data subjects could also be religious institutions, membership-based associations, small enterprises, and other associations that meet strict criteria, like the following: 

• An entity that is legally organized or recognized as: 1) a religious organization, charity, public benefit corporation, labor organization, or co-op; 2) a small business; 3) a mutual benefit association which mainly consists of individual professionals, sole proprietors, gig workers, consumers, or small businesses; 4) one or more U.S. households; 5) a local media or journalistic enterprise; or 6) an unincorporated private digital association of natural persons with an explicit, durable, and well defined relationship to each other (e.g., Substack writer or YouTube creator and their followers as well as “friend” or “community groups” on social media platforms).
   
• The entity certifies that it: 1) does not collect or use data for commercial purposes; or 2) only asserts rights to data: a) where the interests of all data subjects who co-produced the data are sufficiently represented by a shared governance structure (e.g., a co-op model); or b) that is the sole product of the entity or group’s internal or proprietary operations; and c) where no other data subjects could reasonably assert interests based on co-production of the data.
   
• The entity cannot be owned or controlled: 1) by a foreign government or for-profit corporation, except for small businesses, sole proprietorships, qualifying mutual benefit corporations, local media or journalistic enterprises, or co-ops; or 2) by any person or entity owned or controlled by a foreign government or for-profit corporation with limited exceptions.

This framework treats human scale communities as the basic default unit of digital governance, even if individual interests are also present. Consequently, data subjects should generally be assumed to be communities, with DRAs functioning as “communities of communities” or something akin to mini-federal republics with representation scaled to each sub-unit. 

Good governance. Much like community banks or credit unions, data subjects or other entities wishing to form DRAs would require certification or a charter from a regulator to operate. Federal legislation could delegate authority to the Federal Communications Commission (FCC) or Federal Trade Commission (FTC) to certify and oversee DRAs, while states might delegate that responsibility to a state consumer protection authority or create a standalone DRA commission. DRAs would need to meet standards, such as the following:

• Be legally organized as a non-profit or non-stock corporation.
   
• Operate consistent with the public interest as determined by factors similar to those currently applied by the FCC in allocating broadcast licenses.
   
• Serve the privacy, control, and fair value interests of one or more specific classes of data subjects with respect to one or more data counterparties.
   
• Possess the technical competence and capacity–whether directly, in coordination with other DRAs, or via contracted third parties–to propose and secure favorable terms from data counterparties and hold them accountable through independent verification and legal mechanisms.
   
• Ensure a members-only ownership and governance structure that aligns DRA activities with the interests of its members (data subjects) and protects against any conflicts of interest and foreign or non-member control. While DRAs would all have fiduciary duties of care, loyalty, and disclosure to members—and generally be required to operate democratically—legislation would incentivize them to adopt polycentric structures for governance that afford smaller data subject communities decisional power. For example, the enabling legislation could provide a fiduciary safe harbor for DRAs that operate at a governable scale and organize as polycentric units, rather than as masses of atomized data subjects.
   
• Abide by common standards for verifying that the individuals and groups seeking membership are eligible “data subjects” as defined by law; formulating and proposing data terms; adjudicating conflicts with other DRAs; bargaining with data counterparties; joining permission-granting coalitions with other DRAs; invoking neutral arbitration if necessary to secure fair terms; and utilizing appropriate legal mechanisms to enforce data policies. 
   
• Remain financially solvent, whether from membership fees, donations, or other revenue sources.

In addition to the natural competition of DRAs vying for members by offering fair data terms and benefits, regulators would also retain authority to revoke a DRA’s license to operate should it breach its fiduciary duty to members, engage in unfair or deceptive practices, or otherwise violate the terms of its certification or charter. For example, DRAs could be required to undergo regular audits and demonstrate member satisfaction, sustain growing and engaged membership, and register a low number of complaints submitted to regulators by membership. To ease compliance burdens and ensure flexibility, once a regulator certifies a DRA, ongoing audits and oversight could be undertaken by third parties authorized by the regulator.

DRA benefits. Upon certification, DRAs would be granted the following privileges, in addition to their right to pursue collective informational bargaining:

• The right to restrict data counterparties from monetizing, transferring, selling, or making member data or content available to the public or third parties, or from using member data or content in its products, including AI models, without DRA-negotiated permissions in place.
   
• The right to organize for the benefit of their members and form partnerships with other DRAs or entities so long as such combinations do not constitute monopolistic or anticompetitive conduct, and overall, advance the privacy, control, and economic interests of data subjects.
   
• The right for DRAs to obtain injunctive relief and damages from other DRAs whose grants of permission or other actions enable unreasonable impairments of their members’ associational data rights, as resolved by a court or arbitrator. This design would create strong incentives for DRAs to act through coalitions and negotiate terms likely to be acceptable to all interested persons and DRAs, not only their own members.
   
• The right to require data counterparties to submit to independent verification mechanisms and disclose information necessary to ensure adherence to binding terms.
   
• The right to effective and timely legal remedies if a data counterparty is in breach of binding terms.

Data Counterparty Obligations

While the specific terms governing data collection and use would be set through agreements between counterparties and DRAs, counterparties would still be bound by general requirements of procedural fairness, non-circumvention, disclosure, and protective defaults. In general, these obligations would apply if an entity were bound or could reasonably be bound under a collective informational bargaining arrangement (CIB).

Procedural fairness. Once a DRA formally engages one or more counterparties in CIB, both sides should bear a duty to engage in good faith and adhere to the terms reached through the CIB process. Similar to labor negotiations, bad faith in data rights bargaining should include:

• Failure to meet with DRA representatives at reasonable times and reasonable intervals.
   
• Failure to bargain in good faith concerning mandatory subjects of the CIB framework (specifying the scope of bargaining).
   
• Engaging in bad-faith, surface, or piecemeal bargaining.
   
• Refusing to furnish information the DRA requests that is relevant to the CIB process or to data collection or use.
   
• Modifying any term of a CIB framework without the DRA's consent. 
   
• Making unilateral changes in commercial terms or conditions related to a DRA’s asserted data interests during the term of a CIB framework, unless the DRA has clearly and unmistakably waived its right to bargain or the change is too minor to require bargaining.
   
• The solicitation or transfer of benefits to a DRA representative or member for the purpose of attempting to influence internal deliberations relating to the CIB process.

Non-circumvention. To prevent counterparties from exploiting their asymmetric informational power to stop data subjects from exercising the Right of Digital Association, legislation would restrict certain self-preferential or discriminatory behaviors. Counterparties would generally be required to abide by the following standards:

• Where generally applicable data terms are secured by one or more DRAs, the counterparty must offer those terms to all similarly situated data subjects, regardless of their DRA status.
   
• No design features or techniques that interfere with data subjects’ ability to form, join, or aid a DRA, including the use of defaults or other methods to impede or create unnecessary friction in obtaining information about DRAs or entering into membership.
   
• No discrimination against data subjects on the basis of their affiliation with a DRA, including degradation of service quality or the imposition of fees to penalize or otherwise discourage DRA membership or the exercise of any right provided under DRA-negotiated terms.
   
• No reward or transfer of benefits, monetary or otherwise, inducing a data subject to accept the company’s terms in lieu of DRA membership, or inducing a data subject to refrain from joining a DRA or from accepting generally applicable DRA-negotiated terms.
   
• No offer of data terms directly to data subjects who are members of a DRA, except as expressly allowed by the DRA’s membership policy or by mutual agreement between the DRA and counterparty.
   
• No term offered by a counterparty may abridge or otherwise impair data subjects’ freedom to opt-into generally applicable DRA-negotiated terms or join a DRA.
   
• No entity subject to CIB may form or control a DRA, whether directly or via proxy.
   
• Counterparties would be permitted to incentivize or even require data subjects to join a DRA. But the legislation would prohibit them from: 1) determining in advance which DRA a data subject must join as a condition of using a product, service, or feature; or 2) providing incentives that give preferential treatment to one or more DRAs in membership selection and acquisition.

Disclosure. Counterparties would be obligated to inform data subjects or their representative DRAs about specific data collection and use practices, and to routinely inform data subjects about their Right of Digital Association, and options for exercising it. That might include the following:

• Quarterly transparency reports covering each product, service, or feature detailing: 1) each point of data collection (PoC), including the specific types of information collected (e.g., behavioral information, biometric data, etc.); and 2) each internal point of use (PoU) for which data subjects’ information is utilized, including:
  ​​​​​​
  • Content ranking/recommendations
  • Making data available to a third party.
  • Productization, such as AI training, commercial datasets, databases, insights, or API endpoints.
  • Profiling or the creation of “data lakes.”
  • Personalized notifications, visual graphs, results, or automated outputs 

Counterparties should be required to provide a detailed list and description of each type of data input for each disclosed PoU.

• Quarterly disclosures listing the products, services, and features that are and are not covered by DRA-negotiated terms, and a list of all DRAs with binding terms in place.
   
• Regular and conspicuous notice to data subjects of their right to DRA representation, along with an easily accessible pathway for them to exercise that right. Such notices should also contain a clear warning that if a data subject is not a DRA member, their data may be used and monetized in ways adverse to their privacy, control, and economic interests. Notices could trigger at regular intervals or be tied to certain PoCs, for example, when personally identifiable, sensitive, or behavioral data is collected.
   
• Large counterparties would be required to disclose on an annual basis: 1) how and when notices are provided; and 2) a third-party assessment of the efficacy of such notices, pathways for DRA opt-ins, and whether the counterparty is taking reasonable steps to secure data subjects’ representational interests overall.

Default terms and DRA representation. Counterparties would be required to opt data subjects who are not DRA members into the most protective generally applicable terms by default. When DRAs and counterparties come to terms, counterparties would be required to offer the generally applicable parts of those terms to all similarly situated data subjects (i.e., an identifiable class of users like children, parents, teachers, or content providers) regardless of whether they are members of a DRA.

While counterparties would remain free to offer their own less-protective terms, they would need to default data subjects to generally applicable DRA terms that are more protective. In situations where data subjects wish to take counterparty terms over more protective generally applicable DRA-negotiated terms, they could do so by affirmatively opting out of the default terms and then opting into the counterparty’s terms.

In addition, where DRAs offer membership to eligible data subjects automatically, i.e., without DRA review, counterparties would be required to make such membership offers available to eligible data subjects using impartial methods. Impartial methods for assigning DRA representation might include an opt-in screen that randomly lists or displays the pool of DRAs that the data subject may join and gives them the option to do so. These methods would be subject to independent third-party audits and assessments of their impartiality and efficacy in expanding access to DRA representation. Like default terms, data subjects could exit default DRA membership by opting out or into another DRA.

Financial sustainability

While DRAs would require some source of funding to commence and sustain their activities, enabling legislation would largely remain agnostic on how DRAs resource themselves so long as funding does not compromise their independence from data counterparties. There are several possibilities for DRAs to acquire sustainable funding. DRAs might, for example, charge nominal service fees to members, much like a financial advisor, or exact fees from data counterparties in exchange for facilitating data terms on behalf of members. Such fees could be fixed by law to prevent data counterparties from paying kickbacks to secure sweetheart deals from DRAs.

DRAs could also retain a percentage of economic compensation paid to members by data counterparties, similar to how a bank or credit union retains a percentage of returns on capital invested on behalf of its members. Where data subjects’ economic interests are best served by an ongoing equity stake in the value generated rather than one-time payments—especially where the future value and uses of data are difficult to foresee—equity-based revenue could give DRAs financial durability. At the same time, DRAs focused solely on public goods, like kids’ safety or human agency, could be organized as nonprofit charities, mirroring how civil society addresses social challenges in embodied human contexts.

Technical feasibility

The associational data rights framework allows for dynamic flexibility in adapting new terms and verification mechanisms to rapid technological changes and data practices. Many of the same new technologies that threaten digital sovereignty and community also make it more feasible to manage the complexity of associational data rights.

Data subjects should have the ability to seamlessly join DRAs by opting into terms across the devices and services without having to manually search and opt into terms for every individual app and device. They should be able to communicate their data governance interests, and trust agents and DRAs to faithfully advance those interests wherever applicable.

Technical feasibility. New techniques make several of these tasks and functions easier than before. For example:

• Portable profiles. These could enable data subjects to specify their privacy, control, and economic interests once, in plain language, and have them automatically shared with each DRA in which they enroll. Researchers have specified a user-owned, interoperable preference layer of exactly this kind, titled the Human Context Protocol, through which individuals articulate their preferences once and carry them across AI services. Apple, Google, and Meta already operate shared open-source infrastructure that moves user data directly between competing services at a user's request.
   
• Background enrollment agents. Software that translates a data subjects' stated values, then automatically matches them with the right DRA, could reduce barriers to joining a DRA. The Global Privacy Control is a browser signal that transmits a person's privacy opt-out preference automatically to every website they visit. California law requires businesses to honor it and the state's attorney general has already enforced it. As of mid-2026, 12 states require covered businesses to honor universal opt-out preference signals such as the Global Privacy Control.
   
• Agentic systems for trust-governed identity and discovery. Emerging technical architecture for ensuring that personal AI agents can trust a third-party agent with a human principal's data, such as OpenAgenet, could reduce the friction of exercising associational data rights at scale.
   
• Cryptographic attestation. A platform can produce mathematical proof that it handled data according to the agreed rules, so a DRA can verify compliance instead of taking the platform's word for it. Apple's Private Cloud Compute already does this at consumer scale. Before an iPhone will send a user's data to Apple's AI servers, each server must cryptographically prove that it is running publicly inspectable software, logged in a tamper-evident registry open to outside researchers.
   
• Privacy-preserving auditing. Federated learning (FL) and secure multi-party computation (SMPC) would allow data to be analyzed without being copied or exposed, enabling DRAs to audit how member data is used without breaching security or privacy. Google has trained its mobile keyboard models leveraging FL across hundreds of millions of phones since 2017. Ten competing pharmaceutical companies jointly trained drug-discovery models on more than 2.6 billion confidential data points using FL without firms seeing each other’s data. FL is also used for other applications. Employers in the greater Boston area use SMPC to compute citywide wage-gap statistics, covering roughly one in six area employees without disclosing a single individual salary record. These examples are drawn from public registries maintained by the United Nations and UK government, cataloging real-world deployments of privacy-preserving computation.
   
• Verified membership. A DRA's bargaining power depends on the size and authenticity of its membership, which bad actors could otherwise inflate with fake or automated accounts. Personhood credentials allow individuals to prove they are real people without disclosing any personal information enabling DRAs to certify their membership rolls, and platforms to trust them, while preserving member privacy.

Consider how these pieces fit together in practice. Suppose a DRA organized around family life negotiates terms providing that member conversations with a platform's AI assistant may not be used to train AI companions designed to simulate intimate human relationships.

Under today's architecture, the DRA could do little more than take the platform's word for it. Under the architecture described above, compliance is verifiable. The platform runs its training systems in an attested environment that proves which software and which data sources fed which models and the DRA conducts periodic privacy-preserving audits confirming that member conversations are absent from the restricted training corpora, without the platform surrendering proprietary systems and without the DRA ever accessing raw member records. A verified breach then triggers the remedies described above.

While these technologies and methods hold promise for making DRA ecosystems technically feasible and seamless for the average data subject, this legislative framework is tech neutral. It would create the governance model and incentive structures to allow the necessary sociotechnical institutions to emerge organically. This will ensure that DRAs and related institutions are able to quickly adapt to rapid changes in data collection, use, and technology.

Digital subsidiarity could rein-in Big Tech and unleash pro-human innovation

When individuals stand alone, they are susceptible to manipulation and control by concentrated government or corporate power. Early observers like Alexis de Tocqueville as well as modern theorists like Robert Nisbet note the importance of civic associations and communities in mediating the practice of republican self-governance. The American system’s particular genius is that it protects individuals indirectly—not by making individual citizens fully dependent on the state, as other nations have often done, but by extending support to families and communities, and upholding the freedom to associate by forming congregations, schools, professional societies, and civic bodies. These intermediate social groups are the heart of American democracy that Tocqueville prized. They are also the intermediary institutions celebrated by Catholic social teaching under the name of subsidiarity.

But the digital economy has eroded this middle layer of American society. Big Tech now mediates how we communicate, worship, raise our children, socialize, and learn. Tech corporations are displacing associational life and threatening the transmission of skills, values, and culture to the next generation.

In his recent encyclical on AI, Pope Leo XIV specifically underlined the principle of subsidiarity as part of the solution to the moral crisis of digital society, noting that

[s]tates and transnational institutions are called to ensure fair rules and effective safeguards, so that local communities, intermediary organizations, schools, universities, religious institutions and associations [emphasis added] have a voice and can contribute to the discernment of choices that affect people’s daily lives, such as employment, access to services, data management and digital environments.

The associational data rights envisioned by this paper would restore subsidiarity to our digital spaces by enabling intermediary institutions to organize and coalesce into DRAs. These new representatives of associational life would, in turn, provide a governance layer that empowers families and communities to engage with AI and digital platforms on their own terms.

For example, a domestic privacy-focused DRA could emphasize that digital or networked appliances deployed in members’ homes, like thermostats, cameras, refrigerators, kitchen aids, voice assistants, “smart” mattresses, etc., should meet privacy-by-design standards. DRAs focused on children’s welfare might mobilize against tech companies using kids’ data for targeted advertising, or in favor of specific new parental controls for smartphones, social media apps, and AI chatbots. They might also block data counterparties from enabling AI interactions that undermine parental prerogatives, addict children, or displace human relationships.

Authors could also form DRAs to negotiate terms with tech companies to ensure that AI systems fairly cite their works, pay fair compensation, and otherwise uphold associational data rights that go beyond traditional IP protections. Contractors and gig workers could form DRAs to prevent their ideas and labor from being permanently extracted wholesale by exploitative short-term paymasters and gig platforms. Likewise, a free speech and civil liberties DRA might protect political groups organized on social media platforms by negotiating a digital bill of rights to prevent arbitrary censorship or content suppression, and to provide members with the ability to appeal such decisions by platforms.

Faith-based organizations could also form DRAs to empower adherents to align their digital lives with embodied religious and social mores. For example, a faith-based DRA might negotiate terms requiring tech platforms and consumer AI systems to interface with select third-party middleware or AI tools that reflect core moral virtues such as prudence, justice, fortitude, and temperance. Rather than continuing to outsource moral formation and human thought to Big Tech companies, DRAs could empower faith communities to play an active role in digital governance.

These are just a handful of the possible associational forms that could take shape. The distinguishing advantage of DRAs is that they would give communities sufficient power to negotiate the terms shaping their digital lives with a high degree of adaptability. This approach avoids the twin pitfalls of toothless individual rights and brittle top-down regulation.

Conclusion

The current politics surrounding AI are at a flashpoint. If the last 20 years of social media and Big Tech has taught us anything, it is that too much power is concentrated in too few hands. Until we give families and communities a voice, Big Tech will continue running roughshod over our kids, homes, livelihoods, relationships, and democracy. Associational data rights provide a compelling answer.

DRAs could level the playing field between families and communities and the techno-elites in Silicon Valley (and Washington, DC), who remain disconnected from the values of everyday Americans. Rather than merely reacting, communities could have a seat at the table and help to shape the future. Whether that future is marked by human flourishing or misery depends on the decisions we make today. It is time to reclaim the future for the family and for generations to come.

Editor's Note: For a full copy of this policy brief, including references, please download the brief below.

Executive Summary

The American Enterprise Institute's Housing Center has spent the past five years proposing solutions for the housing supply shortage in cities and states around the country. In large part, these solutions revolve around ending the ban on starter homes—making it legal to build naturally affordable homes that young and working families can afford to buy and rent. Legalizing starter homes is the single most important action local governments can take to improve housing affordability for young and working families.

However, there's another path to affordability for individual Americans: getting married. In Rings to Roofs, the AEI Housing Center and the Institute for Family Studies examine how increasing marriage rates would make housing more affordable, both for people getting married and for everybody else. 

Marriage has three benefits for the housing market. First, increasing marriages would decrease the number of homes required in the housing market, by consolidating two households into one. This acts like a supply-side solution, reducing our national shortage. Second, when two working people get married, their household income rises, meaning they can afford to live in larger and higher-quality homes. Finally, the homes that would be freed up by a marriage recovery are among the most affordable in the country, making marriage both an affordability and supply solution. 

In Part I of Rings to Roofs, we focus on the supply-side benefits of marriage. If marriage recovered to its 2000 level among working-age adults, it would be the equivalent of building one million homes in metros experiencing a housing shortage. Beyond the benefits of getting married itself, a marriage recovery should begin with Americans themselves, and entire markets would reap the rewards.

Key Takeaways

• 1 million homes would be freed up by returning marriage to its 2000 level among working-age adults, about 18% of the national housing shortage.
  • The partnered-share gap opened over 24 years, from 2000 to 2024. Closing it over the same span annualizes to 44,800 homes a year and lifts the refined marriage rate from 31.2 to 33.9 per 1,000 unmarried women (age 15+).
  • Returning to 1980 levels of marriage would reduce the shortage by 2 million homes.
     
• Reaching the 2000 benchmark takes about 4.9 million additional marriages and frees about 1.08 million homes, roughly one home per 4.6 marriages.
   
• Homes freed up by marriage tend to be smaller, less expensive, and rented: about 82% are rentals at a median rent near $1,070 (the 28th percentile of local rents), and over half sit in multifamily buildings.
   
• The annual Rings to Roofs estimate equates to roughly 80% of new Low-Income Housing Tax Credit (LIHTC) construction. In subsidy-equivalent terms, Rings to Roofs produces a value of roughly $2,700 per marriage annually (vouchers), or a one-time federal $30,000 cost (LIHTC).
   
• Marriage differs from other demand-side solutions. It raises a household's buying power while simultaneously reducing the number of households, creating affordability and easing a shortage instead of bidding up prices.
   
• The Rings to Roofs data dashboard shows the benefits of a marriage recovery on the housing shortage and the new couples in states and metros throughout the country.

 

 

Introduction

The married share of working-age Americans has been falling for over 50 years. At the same time, the American family has been shrinking. Average household size has fallen by 21%, from 3.11 persons in 1970 to 2.46 today. The number of single adults (25-64) has risen from 18 million (20 percent) in 1970 to 68 million (39 percent) in 2024. We are not only bowling alone, as Robert Putnam famously put it. We are increasingly living alone.

The United States has a housing shortage of roughly 6 million homes (estimates range from 2 million to 8 million). Shrinking household size sits behind part of that demand, along with a growing overall population. Much of the housing debate responds to the shortage as a problem of construction, how to build more, and more affordably. The AEI Housing Center has spent years on that side of the ledger, documenting how restrictive zoning and land-use rules suppress the starter homes young families can afford, and showing how starter homes and by-right development could unlock naturally affordable supply without subsidy. The legalization of by-right starter-home construction remains the single most important (and popular) policy change available to close the housing shortage.

But supply is only half the story, and many advocates also propose demand-side solutions to housing affordability, mainly subsidies, vouchers, and lower interest rates. These efforts would give households more buying power. When supply is constrained, however, that buying power capitalizes into price: the family that receives the benefit may be helped, but overall affordability holds flat or erodes, because more money is now chasing the same fixed stock of homes. We saw this in the late 2010’s, when home prices climbed as the market responded to falling interest rates. This is the curse of demand-side solutions in a shortage environment.

A marriage recovery avoids that curse. It raises buying power, because two incomes combine. But marriage also reduces the number of households, because two families become one. So, unlike a voucher, marriage lowers the total quantity of homes demanded even as it raises what each remaining household can afford. Increasing marriage reinforces the downward pull new construction exerts on prices, rather than working against it, which suits it to a shortage in a way ordinary demand-side policy is not suited. Marriage shrinks the shortage and strengthens the family at the same time.

Rings to Roofs, Mapped

The housing supply benefit of marriage is simple. For example, James and Elizabeth each rent their own apartment. They marry, move in together, and one of the two apartments goes back on the market for the next household in line. Two homes become one. With a million such marriages, the housing shortage falls by a million homes.

Not every marriage would do this, and in fact most would not. A home is freed up only when both partners give up a home of their own, so that two people who occupied two homes then occupy one. A couple where one person was living with parents, for example, might not free up a home at all if that person’s former room remains vacant. Among single working-age adults, about 57% live in arrangements that would free a unit, so if marriages formed at random, only about a third would free up a home.

Rings to Roofs measures the effect of moving today’s partnered share back to its level in 2000, across states and metros. In 2000, 66% of working-age adults (ages 25 to 64) were married or cohabiting; today 61% are. Closing that 5-point gap implies 4.9 million additional marriages, involving 10 million adults. Counting only homes freed up in markets that actually have a shortage (a home freed up where there is no deficit does not count), the result is over 1 million homes, or 18% of the national shortage. Over the 24 years the gap took to open, that is about 44,800 homes a year.

 

Of the roughly one million homes, about 989,000 are in metropolitan areas, which hold 5.7 million of the six-million-unit national shortage. As single adults cluster in metropolitan areas, Rings to Roofs tends to free up homes largely where the shortage is worst. In over 100 large metros— including Philadelphia, Chicago, Minneapolis-St. Paul, San Antonio, and Detroit—a return to 2000 partnering levels would close the housing shortage entirely. These tend to be metros where the shortage is small relative to population, not necessarily where the marriage decline has been steepest: a metro’s housing shortage-to-size ratio predicts whether Rings to Roofs closes the gap better than the depth of its marriage decline does. 
 

Who Frees a Home, & Which Homes Get Freed Up

Above, we deal exclusively with the supply-side effects of a marriage recovery on housing. Next, we carry Rings to Roofs to its logical next step: how would increasing marriage affect the income—and therefore buying power—of the people who would get married? Table 2 describes the current population of married, cohabiting, and single adults, the latter split by whether those single adults live in unit-freeing arrangements. 
 

To say anything about people’s income and buying power after marriage, we must also model who might marry in a marriage recovery. We match the 4.9 million implied marriages, assortatively, sorting partners by income and education the way real marriages tend to, and report the outcomes for the resulting marriages. That lets us describe both the overall population the marriages are drawn from and the particular people most likely to marry. 

These 39 million unit-freeing adults, as individuals, are not high earners on average. Their median personal income is about $45,000 and their median household income about $61,000, compared to $132,000 for currently married couples, who pool two incomes. About 37% hold a college degree and 53% of the renters are rent-burdened. They already occupy the less expensive end of the housing stock: they are mostly renters who rent at $1,480, against $1,850 for married households.

When two current singles marry, we assume the couple gives up the least valuable of their two homes, the lower-rent unit when both rent and the rental when tenure is mixed, since owned homes turn over less readily. Profiled that way, the freed homes tend to be smaller, less expensive, and rented. About 82% are rentals, at a median rent of $1,070, the 28th percentile of their local rental stock; the owned remainder has a median value of $200,000, the 27th percentile. Furthermore, 58% are in multifamily (2+) buildings, and about 31% are studios or one-bedrooms. The picture barely changes whether we assume that couples pair at random or sort by income and education, so the result does not depend on guessing who marries whom.
 

 

Marriage does more than free up a unit. By combining two incomes, it also moves many of these couples up the housing ladder, turning renters into plausible starter-home buyers. That complementarity, between building starter homes and increasing marriage rates, is the subject of our next report.
 

What a Million Homes Are Worth

Large federal programs already exist to provide subsidies to reduce the rent burden on lower-income households, and they make a natural yardstick for the homes Rings to Roofs would free. The comparison can be drawn two ways: as a share of the housing those programs supply, and in the dollars to provide a unit.

As a share of supply, a million homes is a large amount on any measure. Counted as a stock, it equals about 47% of the entire Housing Choice Voucher program, the nation's largest rental-assistance effort, which serves about 2.3 million households. Counted year by year, the roughly 44,800 homes freed up annually equates to about 80% of new LIHTC construction, the largest federal program for building affordable rental housing, which places some 50,000 to 60,000 newly built units in service a year. 

We can also quantify the value of a unit gained by increasing marriage, in dollars saved in annual subsidies. 

• A voucher is an operating subsidy, about $12,100 per household each year for as long as the household holds it; the million freed homes through increasing marriage would supply as much housing as roughly $13 billion a year in federal spending on vouchers. 
   
• LIHTC is a production subsidy, paid once and claimed over 10 years; building the equivalent supply would run about $11 to $13 billion a year in total development cost, roughly $250,000 to $300,000 per unit and far more in coastal markets, about half of it a federal contribution. 
   
• The two subsidies are not always alternatives, since about 50% of LIHTC tenants need monthly rental assistance, usually in the form of a voucher, on top of the credit to afford the restricted rent.

The same arithmetic can be reversed to price the value the government should place on marriage for housing. Rings to Roofs proposes about 4.9 million marriages to free up about a million homes, so each additional marriage is worth, in housing terms, roughly a fifth of a home. Valued as the voucher subsidy it spares, that is about $2,700 a year in perpetuity per marriage, almost all federal money. Valued as the cost of building the equivalent LIHTC unit, about $55,000 to $65,000 once, of which roughly $25,000 to $35,000 is federal. Those are the housing-supply dividends of a single marriage, before counting anything marriage does for the couple themselves.

One consideration matters for how literally to read these amounts. Housing vouchers and LIHTC are means-tested, aimed squarely at very-low-income households. Rings to Roofs does not target any income group. The homes it frees reach the lower-cost segment through filtering rather than through an eligibility rule, and a later report will show that the couples involved are themselves disproportionately lower-income. The programs and the marriages therefore serve overlapping, rather than identical populations; the dollar figures reflect the value of the supply rather than equating a freed-up home with a subsidized unit household for household.

Why Marriage?

The housing policy discussion has, generally, not paid enough attention to the role that marriage plays in driving demand and supply. But the “white picket fence” trope—the idea that owning your own home is a crucial piece of the American Dream and the Declaration’s Pursuit of Happiness—is based on fact. Homeownership has always stood in for security and success in the American imagination, and it still does: 94% of Americans say owning a home is part of the American Dream, ranking it behind only financial security and personal freedom among what success means to them.

For many Americans, marriage is what turns that dream into a plan. Getting married pushes people to actually go looking for a house, not just admire the idea of one. Part of it is that married couples are far more likely to have children, and parenthood is a well-documented trigger for seeking out a single-family home with a yard and a school district, as Brad Wilcox and Maria Baer have shown. Part of it, too, is the pooling of two incomes and two sets of savings, which puts a down payment within reach in a way it is not for a single earner, as Scott Winship has pointed out. 

But marriage doesn’t just make the math easier when it comes to affording a home—it changes the people doing the math. Married men work more hours, get fired less often, and earn a well-documented “marriage premium” over their single peers, whether from specialization at home, added workplace stability, or some combination of the two. That extra income shows up alongside new habits of saving, budgeting, and delayed gratification—the accumulated sacrifices that make a mortgage, and then years of mortgage payments, sustainable in a way they rarely are for one person alone. In all these ways, “putting a ring on it” makes it easier to put a roof over your head.

Conclusion

Rings to Roofs reframes part of the housing shortage as a question of family formation. Since 2000, working-age Americans have married and partnered less, and a population that lives in smaller groups needs more homes to go around. Restoring the partnered share to where it stood a single generation ago would free up roughly a million existing homes, and they would tend to be the smaller, less expensive rentals where the shortage runs deepest and where neither private builders nor public subsidy reach easily.

All this is possible because marriage makes it easier for men and women to afford a home and increases their motivation to seek a home of their own. But marriage is also a good in its own right, in addition to its benefits for the housing market. Married adults report higher levels of happiness, financial security, and stability for themselves and their children than their unmarried peers. So a housing agenda that makes marriage easier to choose, rather than harder, does more than just advance housing. It also makes the American Dream and the Pursuit of Happiness more accessible to countless men, women, and children across the nation. 

Marriage should be treated as a first-order concern in solving America’s housing crisis.

Editor’s Note: This is the first in a series. Later reports will cover how building more starter homes reinforces marriage, and why marriage and starter homes are particularly effective solutions for the lowest-income Americans in the housing market.

*For a footnoted copy of this report, download the PDF below.

Introduction

There is no shortage of lists ranking the top U.S. regions for young adults. For instance, US News ranks metropolitan areas based on employment opportunities, affordability, and “desirability.” The top-scoring area within Travel + Leisure’s list is praised for its entertainment and transportation options as well as its residents’ fitness. Yet another list, compiled by Commercial Cafe, includes school enrollment and access to parks within its ranking criteria.

All of these features are indeed relevant when considering where to live. However, they overlook a major factor of interest: Where are young adults actually getting married? Given the many benefits of marriage, a city where more young people are tying the knot might ultimately provide a more fulfilling life than one with just excellent entertainment and tasty food—but few weddings.

To that end, we put together a new set of maps, available at GoGetMarried.org, to help young adults determine which areas of the country are particularly promising for finding a spouse. These maps present marriage rates for both sexes and three different age groups.

In our analyses, we found a connection between marriage rates and both income and employment, at least for certain age groups; therefore, our maps also allow viewers to explore these additional metrics. In addition, given the importance of religion and ideology in many individuals' search for a spouse, we included maps that visualize both religious adherence and recent presidential election results.

Most of our analyses are based on IPUMS American Community Survey (ACS) microdata for the 2020-2024 time period. For reasons described in the methodology section (see Appendix), we generally chose to explore outcomes at the Public Use Microdata Area (PUMA); however, our marriage-rate calculations, along with our analyses of the correlations between various metrics and marriage rates, are based on groups of PUMAs that include at least 500,000 individuals each. Our county-level data on religious adherence and the 2024 presidential election results are derived from separate data sources.

Key Findings

1. The regions with the highest marriage rates for adults ages 22-29 are mostly located away from major coastal cities like New York or Los Angeles. Instead, they tend to lie within interior states like Utah, Kansas, Nebraska, Tennessee, and Kentucky. 
    
2. In contrast, some of the highest marriage rates for adults in their thirties are found in or near San Jose, Los Angeles, New York, Chicago, and other major urban areas.
    
3. Another strong predictor of marriage rates for young-adult women is the ratio of full-time employed single men to single women.
    
4. The ratio of higher-earning men to single women has a strong, positive correlation with marriage rates among women in their thirties, but not women ages 22-29.

Where Are Marriage Rates the Highest?

The top 10 regions with the highest marriage rates² for adults ages 22-39 are shown in the following table. The Provo, Utah, region is the highest-ranking area on this list with a marriage rate of 13.2%. Five other top-10 regions are located in a band between Panama City, Florida, and Fort Wayne, Indiana. The northern Dallas-Fort Worth suburbs are also well represented here.

Ten regions³ with the highest marriage rates among adults ages 22-39

Marriage rates for this age group are also relatively high within the suburbs of Pittsburgh, Nashville, Austin, and Denver. (It is possible that many of these recently-married individuals first met within urban areas, then moved to the suburbs after they got married. Therefore, the cities that these suburbs surround might also be promising locations for finding a spouse.) 

Ten regions with the lowest marriage rates among adults ages 22-39

Marriage rates for this age group also tend to be lower within the Mississippi Delta, New Mexico, and southwestern Georgia.

Different trends emerge when we evaluate younger and older groups separately. For instance, the interior US tends to have higher marriage rates among adults ages 22-29 than do coastal cities.

The top 10 areas for marriage rates among this younger age group are shown below. The Provo, Utah, region remains the highest-ranking area; in addition, four of the regions between Panama City and Fort Wayne in the previous table are present here as well. This list also includes parts of Nebraska, Idaho, and Kansas.

Ten regions with the highest marriage rates among adults ages 22-29

As shown in the table below, eight of the 10 regions with the lowest marriage rates among adults ages 22-29 are located in the Los Angeles and New York metropolitan areas. Although other parts of these metro areas have higher marriage rates for this age group, this finding suggests that those who desire to get married at a younger age might want to think carefully before moving to these regions.

Ten regions with the lowest marriage rates among adults ages 22-29

When we instead examine adults ages 30-39, we see notably different trends. The 10 regions with the highest marriage rates among adults in this age group, unlike those in the top 10 for adults ages 22-29, tend to be located near major cities; examples include San Jose, Los Angeles, New York, and Chicago. The fact that portions of the NYC, Chicago, and San Francisco metropolitan areas have lower marriage rates among the 22-29 age group suggests that many individuals in these cities prefer to wait until later to marry, perhaps as a result of cultural factors, or their high cost of living.

Only one region appears within both this top-10 list and the list for adults ages 22-29: the Fort Wayne/Southern Suburbs region of Indiana. While it is not at the very top of either list, its presence within both tables is still notable.

Ten regions with the highest marriage rates among adults ages 30-39

Although Colorado is not represented within this top-10 list, it is also a notable hotspot for marriage among thirty-somethings. The Austin, Nashville, Pittsburgh, and Boston areas also have high marriage rates among adults in this age group.

Many of the 10 regions with the lowest marriage rates for this age group, in contrast, tend to be located away from major cities. They include portions of the Southwestern U.S., the Mississippi Delta, western South Carolina, and western Pennsylvania. Parts of the DC, Chicago, and Phoenix metropolitan areas are present within this list as well.

Ten regions with the lowest marriage rates among adults ages 30-39

These findings demonstrate that, for young adults, the ‘best place’ to get married may depend largely on when they hope to marry. Those who are committed to marrying in their twenties may want to think carefully about moving to a major coastal city; meanwhile, those same cities could be an excellent option—though not the only choice—for marriage-minded adults in their thirties.

Which Metrics Correlate With Getting Married?

Marriage rates are the key focus of our Get Married Maps. However, few people will move to a region based on marriage odds alone. Therefore, we also explored additional variables of interest that are correlated with marriage rates.⁴

As shown in the following figure, a substantial positive correlation exists between marriage rates for each young-adult-female age group and the percentage of women in that age group who are married. Thus, while one might think that areas where more individuals are single are better places to find a spouse, the opposite is actually true. 

However, areas with more single men relative to single women do indeed show higher female marriage rates. In addition, the ratio of single, full-time-employed men to all single women is an even better predictor of getting married.

Figure 1. Correlations between demographic variables and percentage of never-married women who got married in the past year, by age

Given our findings for women, it is not surprising that we observe a strong positive relationship between the percentage of men who are married and male marriage rates. In addition, there is a notable connection between men’s marriage rates and the ratio of single, full-time-employed men to single women. 

Interestingly, though, regions in which a higher share of single men have a bachelor’s degree or earn at least $60,000 have higher marriage rates for men in their thirties—but not for men ages 22-29. (This does not mean, however, that a college education or higher income is a hindrance to marriage for younger men; this will be discussed in further detail later in this brief.)

Figure 2. Correlations between demographic variables and percentage of never-married men who got married in the past year, by age

Availability of employed single men helps predict marriage rates

Among women ages 22-39, the ratio of single, full-time-employed men to all single women has a strong connection with marriage rates.⁵ (See the following interactive scatter plot for a visualization of this relationship by grouped PUMA.) This ratio is also predictive, to a lesser extent, of marriage rates for men.

Figure 3. Marriage rates for women by ratio of single, full-time employed men per 100 single women (among adults ages 22-39)

Many of the regions with the highest score for this metric are located in areas with a large military presence, including, Minot, ND; Fayetteville and Jacksonville, NC; Oceanside/Camp Pendleton, CA; El Paso Teller (near Colorado Springs, CO); and the Peninsula Region/JBLM (WA). However, Mountain View and Sunnyvale, CA also have notably high ratios of single, full-time employed men to single women, likely due, in part, to the large presence of major tech companies in that area.

Ten regions with the highest ratio of single, full-time-employed men per 100 single women

Parts of Northern Virginia also have relatively high ratios for this category, as do the Upper Midwest and Mountain West. Portions of Seattle and San Francisco perform strongly on this metric as well.

The 10 areas with the lowest ratios, in contrast, are mostly found within Midwestern cities, the Bronx, and the South. The Fresno and Baltimore metropolitan areas each appear within this bottom-10 list as well.

Ten regions with the lowest ratio of single, full-time-employed men per 100 single women

Outside these top-10 areas, we also see lower ratios in much of Mississippi and Alabama, portions of the NYC area, and southern Virginia.

Women’s marriage rates and the availability of higher-earning single men

In the Good Jobs, Strong Families report by Grant Martsolf and Brad Wilcox, a ‘good wage’ was defined as an income greater than $60,000. Therefore, this salary serves as a good benchmark for our own income analyses. It is, however, a relatively high bar for singles within our age range. Using American Community Survey 5-year microdata from IPUMS, we found that only 32% of single men ages 30-39, and 19% of single men ages 22-29, have personal incomes of $60,000 or more. (The values for single women were slightly lower: 27% of those ages 30-39, and 15% of those ages 22-29, earned at least this much.)

As part of our income analyses, we explored how women's marriage rates are correlated with the ratio of single men earning $60,000 or more to all single women. Interestingly, this relationship varied markedly by age group.

Among young adults in their thirties, the ratio of single men who earn at least $60,000 to all single women (regardless of income) has a strong positive correlation with the percentage of women who got married in the past year. 

Figure 4. Marriage rates for women by ratio of single men earning at least $60,000 per 100 single women (among adults ages 30-39)

Seven of the top 10 regions for this ratio are located in Seattle, San Jose, and San Francisco—all major tech hubs with plenty of highly-paid employees. The remaining three are located in either the New York or Washington, DC, metropolitan areas. 

Ten regions with the highest ratio of single men earning at least $60,000 per 100 single women

On a national scale, we also see particularly high values within the Washington-Boston corridor, as well as Austin, Minneapolis, and much of the state of Colorado. Thus, tech hubs are not the only regions with high ratios of higher-earning single men to single women.

Meanwhile, the 10 regions with the lowest ratios tend to be located in more rural areas of the Southeast and Appalachia. Similarly, outside of major metropolitan areas like Atlanta and Raleigh, the Southeastern U.S. tends to have relatively low ratios for this metric.

Ten regions with the lowest ratio of single men earning at least $60,000 per 100 single women

However, when we instead analyze this relationship for the 22-29 age category, we find essentially no association at all between the two variables. In other words, the relative availability of higher-earning men has a strong connection to marriage rates for women in their thirties, but not for this younger group. 

Figure 5. Marriage rates for women by ratio of single men earning at least $60,000 per 100 single women (among adults ages 22-29)

A similar pattern is found for the percentage of single men who earn $60,000 or more: this variable has a positive correlation with marriage rates for women in their thirties, but not for women ages 22-29.

What could explain this difference? One possibility is that women who marry younger men might see future earning potential in them that has not yet materialized. (For instance, our income calculations did not exclude students; men in law or medical school may earn little or no money in their early twenties but higher salaries later on.) 

Alternatively, it is possible that, for those who marry in their twenties, financial concerns are not as important as they are for those who wait until their thirties to marry. Finally, areas with higher rates of marriage among younger women may have lower costs of living; in this case, lower salaries might be less of an issue.

For men, regional relationship between education and marriage is also age dependent

Similarly, the connection between education and marriage differs significantly among younger and older single men. At a regional level, the percentage of single men in their thirties with a bachelor’s degree has a notable positive correlation with the percentage of men in that same age group who got married in the past year.

Figure 6. Marriage rates for men by percentage of single men who have a bachelor’s degree (among men ages 30-39)

However, among single men ages 22-29, the percentage of men with BAs has a much smaller, and negative, correlation with marriage rates. This may be because men with a bachelor's degree are more likely to live in large cities than those without one. (As discussed earlier, urban regions have higher marriage rates among adults in their thirties, but lower rates among adults in their twenties).

Figure 7. Marriage rates for men by percentage of single men who have a bachelor’s degree (among men ages 22-29)

It’s worth clarifying, however, that within each age group, marriage rates for men who have a bachelor’s degree are indeed higher than those for men who don’t. (The same is true for women, as the following figure shows.) Even so, for both sexes, the gaps in marriage rates by education level are narrower for those in their twenties.

Figure 8. Percentage of never-married young adults, by age group, sex, and college education, who got married in the past year

Regional relationship between higher incomes and marriage differs by age for men 

We also see diverging patterns for the relationship between young men’s incomes and their marriage rates. As shown below, regions in which higher shares of single men in their thirties earn at least $60,000 also have higher male marriage rates for that age group.

Figure 9. Marriage rates for men by percentage of single men who earn at least $60,000 (among men ages 30-39)

However, the opposite is true for younger men: regions where higher shares of men ages 22-29 earn $60,000 or more have lower marriage rates. The nature of urban areas could help explain this relationship, as cities tend to have both a greater share of higher-paying jobs and lower rates of marriage among adults in their twenties.

Figure 10. Marriage rates for men by percentage of single men who earn at least $60,000 (among men ages 22-29)

This does not mean, however, that higher-earning men in their twenties are less likely to marry. In fact, men ages 22-29 who earn at least $60,000 are around twice as likely to get married in a given year as those who do not. (Higher-earning women are also more likely to marry, as recent IFS research demonstrated.)

Figure 11. Percentage of never-married young adults, by age, sex, and total personal income, who got married in the past year

The implication of these findings for men in their twenties is that while higher incomes and marriage are certainly linked, moving to a region with lots of higher-earning young men will not necessarily result in better odds of marriage.

Religion, Ideology, and Marriage Rates

The relationship between our regional estimates⁶ of Trump’s performance relative to Harris in 2024 and marriage rates is also age dependent. These two variables have a positive and relatively strong correlation for men and women ages 22-29; however, for thirty-something adults of each sex, this relationship is much weaker and negatively correlated. One possible explanation is that more rural parts of the country, where Trump generally performed better than Harris, often have higher marriage rates for adults in their twenties.

Meanwhile, we initially expected to see a positive regional correlation between religious affiliation and the percentage of young adults who got married in the past year. After all, as General Social Survey⁷ data shows, religious individuals in their twenties and thirties are significantly more likely than their nominal and secular counterparts to be married.

Figure 12. Percentage of young adults, by age range and religiosity, who are married

Surprisingly, however, we find essentially no regional relationship between religious-adherent shares and the percentage of young adults who got married in the past year. Certain parts of the country (e.g., areas of Utah) do indeed have both high rates of religious affiliation and high marriage rates. However, there are also plenty of regions (such as those in the Southeast) with higher religious affiliation but lower young-adult marriage rates. Similarly, parts of the Northwest and Mountain West have both higher marriage rates and lower religious affiliation. Thus, on a national scale, we find essentially no connection between these two variables. (Similarly, we did not find a relationship between religious adherence and the percentage of young adults who are married.)

Figure 13. Marriage rates among adults ages 22-39 by religious adherence

There are several possible explanations for this finding. First, our religious-adherent data does not differentiate by age; it is possible that, if we looked only at the percentage of young adults who belong to a religious congregation, we might have found more significant results for this metric. Second, our grouped-PUMA religious-adherent percentages are weighted averages of county-level data rather than actual grouped-PUMA-level values. If we had access to actual PUMA-level religious-adherent data, we might find a stronger connection between religiosity and marriage.

Conclusion

Determining where to move after high school or college is no easy task. Many of the most important factors involved—such as proximity to family, where one receives his or her first job offer, and the desire to explore a new part of the country—cannot be displayed within a mapping project like this one. In addition, we acknowledge the important role that employment rates, job opportunities, cost of living, and cultural factors all play in deciding where to relocate.

However, we encourage those who wish to get married in the future to pay attention to our marriage-rate findings. If one appealing area has twice the marriage rate as another, it may indeed be the better option for a marriage-minded young adult. Similarly, young adults hoping to marry in their twenties might want to avoid moving to a coastal metropolitan area where such a decision is less common. While marriage rates are not the only factor to consider, they are certainly just as important, if not more important, than many of the other criteria often used for popular city rankings. We hope that these maps will bring more attention to this underappreciated metric.

It’s very possible that the marriage rates shown within this map will change in coming years. This decade’s top regions for getting married may end up underperforming other areas in the future. For that reason, we plan to continue updating these maps as new demographic data becomes available. That way, the maps can continue to inform relocation decisions for individuals who wish to not only find a job and a friend group in a new city, but a spouse as well.

Editor's Note: For the Appendix, download the full report below.

---

Endnotes

1. Citation for IPUMS USA Data: Steven Ruggles, Sarah Flood, Matthew Sobek, Daniel Backman, Grace Cooper, Julia A. Rivera Drew, Stephanie Richards, Renae Rodgers, Jonathan Schroeder, and Kari C.W. Williams. IPUMS USA: Version 16.0 [dataset]. Minneapolis, MN: IPUMS, 2025. 
2. One limitation of our data is that we do not know exactly where adults were living at the time that they got married. Instead, our analyses are based on (1) where adults are currently living, and (2) whether or not those adults got married in the past year. This means that, if a couple got married in a city but then immediately moved out to a suburban town, that marriage would increase the suburb's marriage rate within our maps rather than the city's. Similarly, many couples—particularly those in longer relationships—may have met each other in one region but gotten married in another.
   Our decision to examine who got married in the previous year, rather than the last 3 years, 5 years, etc., should help reduce the percentage of respondents who moved between getting married and taking the American Community Survey. Nonetheless, due to this factor (along with sampling error), these maps are an approximation, rather than a perfect image, of each region’s actual marriage rates.
   In an attempt to mitigate this issue, we did explore alternative approaches to calculating these marriage rates. One option would have been to exclude respondents whose migration PUMA changed over the past year from our marriage-rate analyses. However, this would not only reduce our sample size (and thus widen our confidence intervals), but also exclude any respondents who changed PUMAs in order to move closer to the individual they were about to marry.
   The San Jose Area grouped PUMA provides one possible example of individuals moving for marriage: 14.2% of never-married women ages 30-39 in San Jose got married in the past year; however, when we exclude women who changed migration PUMAs during that time, this value drops to 10.9%. One plausible explanation for this difference is that the former percentage includes a number of women who moved to San Jose to live with their new husbands who work in the technology sector there. (The corresponding values for men ages 30-39 were much closer: 10.6% and 9.9%, respectively. This suggests that couples aren’t simply choosing to move to San Jose as a post-marriage destination.)
   Therefore, we ultimately chose not to exclude recent movers, as doing so could exclude individuals who moved to get married—and would thus undermine the main purpose of our maps, which is to show the best places to move in order to find a spouse.
   In addition, we could have only included individuals in our calculations who did not move in the past year or whose past migration PUMA mapped only to a single PUMA. We could then have based marriage-rate calculations on the PUMA in which these individuals lived during the previous year. However, since many individuals change their residences at the time they get married, this approach would exclude a significant portion of respondents from our dataset.
   Finally, we could have simply calculated marriage rates by respondents’ migration PUMA during the previous year, as this PUMA can be determined for all ACS respondents. However, this would have resulted in a loss of detail within highly-populated counties (such as Los Angeles County), as migration PUMAs, unlike regular PUMAs, cannot include only a portion of a given county.
   To determine the effect of these alternative approaches on our overall findings, we compared them, via correlation analyses, for each of the nine age/sex groups featured in our maps (women ages 22-29, both sexes ages 22-39, etc.). We found that the grouped-PUMA-level correlations between our standard marriage rates and those based on individuals’ previous locations ranged from 0.862 (for men ages 22-39) to 0.930 (for adults ages 30-39). The correlations between our standard rates and those that filtered out recent movers were even higher, ranging from 0.957 (for men ages 22-39) to 0.969 (for adults ages 30-39). Finally, marriage-rate analyses based on respondents’ migration PUMA during the previous year also correlated strongly with those based on respondents’ current migration PUMAs; the exact coefficients ranged from 0.851 (for men ages 22-29) to 0.930 (for adults ages 30-39).
   Given the high correlations we observed between these various methods, we can conclude that the alternative approaches would not have had a major effect on our overall findings. 
3. The names of grouped PUMAs within these tables have been condensed in order to make them more intuitive. As a result, they do not match their corresponding grouped-PUMA names within the corresponding tables of the Get Married Maps site.
4. These analyses were based on linear regressions of grouped-PUMA data, as marriage-rate data at the individual-PUMA level contains a large amount of statistical noise.
5. Male and female marriage rates for the age groups we evaluated will usually differ within each grouped PUMA. This can occur for several different reasons. First, different grouped PUMAs will have different numbers of never-married single men and women (the denominator for our marriage-rate calculations), which will cause marriage rates to differ for those areas. Second, if an individual outside the age range we used for our analysis, such as a man in his forties, marries someone within our age range, such as a woman in her thirties, this marriage will only contribute towards the listed marriage rates or one sex. The average grouped-PUMA-level marriage rate for women ages 22-39 was 6.6, a decent amount higher than the average for men (5.5).
6. Because our election-result data, as well as our religious-adherent data, was originally provided at the county level, we needed to convert it into estimated PUMA-level data (using an approach explained in the methodology section) before we could incorporate it into a correlation analysis at the grouped-PUMA level. This approach has some significant limitations, particularly for PUMAs that only make up part of a highly-populated county. Thus, these results should be interpreted with caution.
7. Citation for General Social Survey data: Davern, Michael; Bautista, Rene; Freese, Jeremy; Herd, Pamela; and Morgan, Stephen L.; General Social Survey 1972-2024. [Machine-readable data file]. Principal Investigator, Michael Davern; Co-Principal Investigators, Rene Bautista, Jeremy Freese, Pamela Herd, and Stephen L. Morgan. Sponsored by National Science Foundation. NORC ed. Chicago: NORC, 2025: NORC at the University of Chicago [producer and distributor]. Data accessed from the GSS Data Explorer website at gssdataexplorer.norc.org.
8. These percentages were created by dividing the total number of members of congregations in a given region by that region's population. This means that, if many congregants travel to a given congregation from nearby regions, the listed percentage may actually exceed 100%. This occurs in around 30 counties and county equivalents.
9. Citation for the 2020 US Religion Census: Clifford Grammich, Erica Dollhopf, Mary Gautier, Richard Houseal, Dale E. Jones, Alexei Krindatch, Richie Stanley, and Scott Thumma. 2023. 2020 U.S. Religion Census: Religious Congregations & Membership Study. Association of Statisticians of American Religious Bodies.
10. Citation for MIT Election Lab data: MIT Election Data and Science Lab, 2018, "County Presidential Election Returns 2000-2024," at Harvard Dataverse, V20.

 

Introduction and Executive Summary

In March 2026, the Institute for Family Studies released America’s Demoralized Men, Part I: Worthy Aspirations, Trying Circumstances, the first installment of a two-part report, based on a YouGov survey of a nationally representative sample of 2,000 young men, ages 18-29. Our aim, within the limits of the survey, was to allow young men to express themselves on a range of personal issues, for which there has been considerable and often harsh public commentary. 

In Part I, we began our exploration by considering what young men say about becoming an adult—both in terms of what it means to them and whether they feel they have reached that life stage. We found that while they see adulthood as a matter of personal responsibility, self-direction, and financial independence, accessing their own adult status is connected to the older metrics, like marriage, parenthood, and finishing school. By these criteria, many do not feel they have reached full adulthood. However, when we asked about work, manhood, marriage, and parenthood, the great majority express a strong desire to be responsible men, to make a living sufficient to support themselves and others, to date, to marry, and to become fathers. 

Reaching these worthy goals, however, is another matter. Young men identified a number of specific obstacles, from inadequate incomes to limits on educational opportunity, to personal struggles to finding a suitable romantic partner. They are not where they want to be in life, and the mismatch between their aspirations and where they stand is a source of frustration and stress.

In this second installment, we continue our analysis by first taking a close look at the social lives of young men, including questions about social integration and support, friendship and participation in organized groups, and family relations, especially with parents. Next, we consider feelings of alienation and frustration, perspectives on the American Dream and how their future will take shape, and reports of mental health problems and potentially traumatic experiences. Lastly, we survey online activities, such as pornography, gaming, and gambling, and some of the trouble correlated with these. 

“Demoralized” is the term we use to characterize a state of mind reflected in agreement with such self-appraisals as “I am inclined to think that I am a failure” (43%), and such expressions of powerlessness as “I often feel that I have little control over the things that happen to me” (61%), and of being found insufficient, such as “I never seem to do enough to please people” (54%). We also report here on the dispiriting experience of feeling lonely (25%), isolated (26%) or lacking community (48%) much of the time. As we noted in Part I, some young men “are not in crisis but thriving.” But some are struggling mightily, as we can see in the major findings reported here.

Major Findings

For Part II, here are some of the most important things we have learned.

• Slightly less than half of young men feel a steady emotional connection to others. Close to half of the young men say they never or not too often feel lonely (45%), isolated (47%), or unloved or unwanted (48%). The other half, however, report feeling lonely (55%) or isolated (54%) at least some of the time; and when the young men were asked if they feel loved or wanted by others, 1 in 3 (32%) say only some of the time, and 1 in 5 (20%) say seldom or never.
   
• For many young men friendships are in short supply. In fact, 4 out of 10 young men have two or fewer close friends, which includes 1 in 3 (34%) with one to two close friends and almost 1 in 10 (9%) who report zero close friends. The best predictor of having three or more friends is connection to institutions like marriage, church, and school. Young men who participated in all-male groups, such as sports teams or the Boy Scouts, during the school years are also more likely to report three or more friends (64%) than those who did not (48%).
   
• Social, civic, and religious participation is low. While 57% participate in at least one community activity, a mere 31% report working as a volunteer, and even fewer are members of a faith-based group (29%), social club (27%), adult sports league (24%), or other type of civic organization. In terms of overall participation, being involved with at least one activity is most strongly associated with being religious (81%), being a student (73%), being married with kids (72%), and having finished trade school (74%). Overall, 25% attend religious services at least monthly; 28% say they have become more religious in recent years.
   
• The American dream feels out of reach for many young men. About half of young men agree with the statement, “No matter how hard I work, people like me will never achieve the American Dream.” Similarly, 72% agree that “Who you know is more important than ability or hard work in getting ahead.” For a significant share of young men, the meritocratic story is a myth. Still, they are not giving up on themselves: a remarkable 84% say that having ambitious plans for the future describes them at least somewhat well.
   
• A notable share of young men report having faced traumatic events in their lives. For example, 12% have ever been incarcerated, while 44% have a loved one who has gone to prison (28%), died of a drug overdose (18%), died by suicide (16%), or died by homicide (14%). The men with such experiences have higher rates of mental health struggles than those who do not, including with depression (49% vs. 27%), an anxiety disorder (44% vs. 25%), ADHD (36% vs. 19%), bipolar disorder (26% vs. 8%), and a learning disability (21% vs. 11%).
   
• Electronic media use is very high. Gambling is rampant: 23% of young men gamble daily, an additional 12% do so several times a week, and another 21% gamble weekly or less. Gaming is also very common, with 44% playing video games many times a day or more, and another 42% doing so at least a few times a week. Pornography use is also extremely widespread: 42% watch porn daily (27%) or several times a week (16%), with another 26% viewing it weekly or less. Those who watch porn daily include married men (21%) and regular religious service attenders (24%).
   
• Online gambling and pornography have significant emotional repercussions. Around 2 in 3 regular online gamblers say it has caused them at least some distress. Of those who watch pornography many times a day or more, 43% experience loneliness, 36% feel depressed, and 29% are anxious. Feeling like a failure is also closely tied to these activities. For example, 66% of daily gamblers and 63% of daily pornography watchers, compared to 42% of men as a whole, report feeling like a failure.

About This Report

In Part I, we presented three of five chapters (available here):

Chapter 1: Coming of Age

Chapter 2: What Young Men Want: Work and Education

Chapter 3: What Young Men Want: Marriage and Manhood

In Part II, we present the final two chapters as well as an interpretive conclusion, respectively titled:

Chapter 4: Social Connection

Chapter 5: Alienation and Distress 

Conclusion: What Did We Learn

After the release of Part I, we were grateful to receive many notes from readers who appreciated the report’s approach and tone. We strove to allow men to speak for themselves and their aspirations, and what came through was challenging but nonetheless hopeful. Men have worthy aspirations, and that was a cause for optimism. At the same time, they were struggling to achieve them, and that was a reason for concern and even sadness. 

In this second installment, we present our findings in this same spirit. We tell a sobering tale—not applicable to every young man, to be sure, but to a significant number. In the discussion that follows, we again try to provide some context for the predicament that many young men find themselves caught in today, and we conclude with some reflections on what we have learned. 

Data and Methods. The Institute for Family Studies Gen-Z survey on men was conducted by YouGov between April 7 and 15, 2025, with a representative sample of 2,000 men, ages 18 to 29, living in the United States.

The respondents were matched to a sampling frame based on gender, age, race, and education. The sampling frame is an 18–29-year-old subset of a politically representative “modeled frame” of US adults, based upon the American Community Survey (ACS) public use microdata file, public voter file records, the 2020 Current Population Survey (CPS) Voting and Registration supplements, the 2020 National Election Pool (NEP) exit poll, and the 2020 CES surveys, including demographics and 2020 presidential vote.

The matched cases were weighted using propensity scores derived from age, gender, race/ethnicity, education, and region, grouped into deciles and post-stratified. The characteristics of the final weighted sample mirror those of the general U.S. population of men ages 18 to 29. Lesbian, gay, bisexual and transgender (LGBT) adults are included but are not analyzed separately. All estimates have been weighed to reflect the actual population.

 

Chapter 4: Social Connection

In recent years, a “male loneliness epidemic” and a male “friendship recession” have been reported and widely discussed. Both crises are apparently about the same problem, as both are premised on the same survey results. A May 2021 American Perspectives Survey, conducted by the Survey Center on American Life, provided some grim statistics on the decline of friendships among men. They found that (1) the percentage with at least six close friends fell by half since 1990, from 55% to 27%; and (2) the percentage without any close friends jumped from 3% to 15%. They also noted that “young men are faring worse than most. These two key findings were picked up and repeated everywhere, even inspiring an SNL sketch and Friendship, a Tim Robinson film.

A 2024 survey by the same center found that the two key measures had grown worse, down to 26% with six or more friends and up to 17%, without any close friends. This finding was also picked up and passed along in the press. However, as noted by the report co-author, Sam Pressler, this decline in friendship among men was not that different from similar trends across American life:

17 percent of all Americans had no close friends—two percentage points higher than the friendship stat that has been used as the basis for the ‘male loneliness crisis.’ In fact, men and women look practically the same in this dataset, with 17 percent of men and 16 percent of women reporting that they ‘do not have a close friend.’

In other words, young men, as a group, are not necessarily the social outlier that they are made out to be. While our survey does not allow comparisons to women or other age groups, we should not automatically assume that young men’s experience is unique. There are also, as we stressed in Part I, significant differences between distinct subgroups of young men. Throughout the report, we pay close attention to their specific circumstances. 

A related point that we will see in this chapter and the next is that while we talk in comparative terms—e.g., which groups of men are more likely than others to feel isolated—we must not lose sight of the absolute percentages. For example, in our survey, a much higher percentage of married men (67%) than unmarried men (41%) report never, or not too often, feeling lonely in the past month. But if we just focus on the married men alone, apart from any comparisons, we find that 14% say they have felt lonely all or most of the time. That represents 1 in 7 young married men, which is a substantial and disconcerting number.

Further, in the wider discourse, feelings of belonging and friendships, along with social support and participation in associations, are often treated as though they are interchangeable. They are certainly related, but they are not identical. For example, tools to measure loneliness, such as the UCLA Loneliness Scale, generally include aspects like feeling alone, left out, and isolated from others. These measures show that loneliness is a complex experience and not reduceable to a mere lack of social contact. The subjective experience of social isolation, such as feelings of loneliness or lack of community, is distinct from what we might call more objective measures of social integration, such as the number of friends or group memberships. We asked young men about each separately and will address them one at a time in this chapter.

Feeling Connected

To access respondents’ feelings of connection, we asked three direct questions: during the last 30 days, how often have you felt 1) “lonely” or 2) “isolated” or 3) “loved or wanted by others.” The answer categories ranged from “never/not too often” at one end of the spectrum, to “most/all of the time” at the other, with “some of the time” as a middle response. With minor variations, the responses to the questions overlapped.

First, the good news. Over the prior month, about half of all young men report being never or not too often lonely (45%) or isolated (47%) and all or most of the time feeling loved or wanted (48%). Full-time employees and self-employed respondents are more likely to report little or no loneliness in the past 30 days (51%) compared to part-time workers (38%), students or caregivers (40%), and those who don’t work at all (40%). In addition, respondents in the 24-29 age range are more likely (50%) to report low levels of loneliness than younger respondents (39%).

Figure 1. Percentage of men who say they felt loved/wanted, isolated, or lonely within the past 30 days

For many, the college years can be a lonely and isolating time. Among our respondents, those in school (including 2- and 4-year colleges, trade schools, and apprenticeship programs) are much more likely to report feeling lonely all or most of the time (28%) than non-students with a bachelor’s degree (16%). However, getting the degree appears to open doors to increased social participation, as men with a trade school or associate’s degree continue to feel lonely—most or all of the time—after graduating at the same rate (26%) as they did while in school or training. On the loved/wanted by others question, this difference is much smaller: 54% of four-year-college grads report feeling loved or wanted by others all or most of the time, but 50% of those currently in school do as well. College students have the affection and support of their parents; the challenge for them is fitting in and being accepted by peers, navigating a competitive environment, and choosing a course of study (or two, as many now double-major) in an unpredictable world. In a large spring 2025 survey of college students, 46% of undergraduate men scored positive for loneliness on the UCLA scale.

Figure 2. Percentage of men who say they felt lonely in the past 30 days, by various characteristics 

If the good news is that half of the surveyed young men do not feel lonely, the bad news is that a substantial portion of our respondents report feeling lonely (25%) or isolated (26%) all or most of the time in the past 30 days. In addition, 20% say that that they seldom or never feel loved or wanted by others during this same time period. If we include the 30% of young men who feel lonely some of the time, the 28% who sometimes feel isolated, and the 32% who only occasionally feel loved or wanted, it becomes clear we are dealing with widespread hardship. 

“Epidemic,” however, is not the right word because this dire situation did not arise suddenly. According to a meta-analysis of hundreds of studies using the UCLA Loneliness Scale with young adults (men and women) ages 18-29, conducted since the mid-1970s, the sheer number of lonely people has been growing in a linear fashion for half a century. It has now swollen to a shamefully high figure.

The extent of loneliness, isolation, and feeling unloved or unwanted, the pattern among respondents, and the long-term trends reflect the process we noted in Part I in our discussion of the transition to adulthood. The loss of institutionalized pathways and defined roles, and the simultaneous shift of responsibility to individuals to create their own career and life trajectories have introduced a great deal of precariousness into the lives of young men and left them bereft of guidance and structure. Those furthest from institutions that provide direction, and cultivate and demand stable habits and commitments are those who feel most lonely, isolated, and unwanted. While that is part of the story, it is not the whole. As the previous figure shows, many married, religious, and full-time-employed men still feel lonely at least some of the time. The thinning of social life and the demands of autonomy are affecting everyone.

Friendship

Friendship is a crucial form of social connection. We asked respondents how many close friends they had, and offered number categories like 0, 1-2, 3-5, and so on. We offered no definition of “close friend.” The answers are based on their own interpretation and assessment of how many of the people who they know qualify. We did not ask how they know their friends, so we might imagine they include spouses and partners, high school and college buddies, work associates, roommates, and so on. Online relationships may also be important: when we asked respondents who play video games at least several times a week about their reasons for gaming, 42% checked the box for “an opportunity to interact with others or make friends.” Although parents often express a desire to be “best friends” with their children when grown (72% of parents in a 2012 survey), we suspect not many young men counted their parents. However, they may well have counted siblings or other relatives.

Figure 3. Percentage of men reporting number of friends by various characteristics

About 90% of young men report having at least one close friend, with the majority having three or more (58%). Conversely, 9% (about 1 in 10) say they have no close friends. That is half of the 17% without a close friend that we quoted earlier, citing a 2024 survey with the same population (young men ages 18-29). Although many factors can account for survey differences, some of the discrepancy might be explained by the fact that the other survey, unlike ours, specifically instructed respondents not to count their relatives. In addition to the young men with no friends, another 1 in 2 in our survey report having just one or two.

Among the expected demographic variables (marriage, parenthood, etc.), we see a notable relationship between going to college and making friends. For example, 52% of men without a bachelor’s degree have three or more friends, whereas 69% of those with this degree have that many, as do 65% of those currently in school for a BA.

Another factor that contributes to more friendships is participation in all-male groups, such as a boys’ sports team, during the school years. We asked young men if they had participated in any such groups during their childhood or adolescence. If a respondent said “yes,” his chance of having three or more friends (64%) was some 16 percentage points higher than those who said no (48%), and his chance of having no friends was about 9 percentage points lower (5% vs. 14%)  

While friendship is certainly a protective factor against loneliness, it is striking that the 25% of young men who say they feel lonely most or all of the time is much higher than the 9% who report no friends. In the following figure, we see that there is a clear relationship between friend counts and loneliness: 43% of those with no friends feel lonely all or most of the time. However, so do 29% of those with just one or two friends, and 21% of those with three to five. Similarly, 28% of those with one to two close friends reported feeling isolated, as did 21% of those with three to five friends, and 20% of those with six or more. Friendships help but do not guarantee feeling included.

Figure 4. Percentage of men who report feeling lonely, by number of close friends

For many young men, there is clearly a substantial friendship deficit, symptomatic of a wider decline in social participation (see below). Specifically, 42% of the young men in our survey reported two or fewer close friends. This includes 31% of husbands, 36% of fathers, around 40% of both liberal and conservative respondents, 31% of college graduates, and 32% of religious respondents. No demographic group is immune to the lack of friendships. 

Social Support

For another measure of personal support beyond friendship, we asked respondents how often they had someone available who could provide them with care, assistance, and companionship. The list of possibilities (see the statements in the figure below) include things like having someone to provide advice about a personal issue, or someone who will listen to them when there is something they have to say. Despite the extent of loneliness and the relatively small number of friends that young men have, about two-thirds report having someone in their life most or all of the time who can listen to them (64%), provide advice (62%), and with whom they enjoy spending time (67%) and engaging in fun activities (63%). Just 5 to 8% of the young men report not having such support at all. Still, for 25 to 30%, this support is only available “some” of the time. 

Figure 5. Percentage of men who report having someone available in various capacities

Who is providing this support? Since the responses to all four questions are so similar, it seems likely that respondents who give the answer all or most of the time are thinking of just one or two persons with a stable presence in their life. As the following figure indicates, married men tend to enjoy more social support than their unmarried counterparts. Along with close friends, spouses (who may or may not be thought of as close friends) play a very important support role. 

Figure 6. Percentage of men who report having someone available in various capacities, by marital and parental status

Most parents today are heavily involved in the lives of their young adult children (a process that also began to grow in the 1970s). We already noted the desire of many parents to be best friends with their children, and in a 2023 Pew Center survey, majorities of parents say they text (73%) or talk on the phone (54%) with a young adult child (ages 18-34, men and women) at least a few times a week. They offer advice, help financially (59%), and say their young adult kids rely on them a great deal or a fair amount for emotional support (41%; more mothers than fathers, and more for daughters than sons). Young people confirm these assessments, with those ages 18-24 most likely to turn to their parents for help and advice. For young men, parents may fill a support gap in the absence of friends and wider community engagement. Our data suggest, however, that despite parental involvement, this absence may still leave them feeling disconnected. 

Social, Civic, and Religious Participation

Much has been written about the waning of civic engagement, religious participation, volunteering, and other forms of associational life among young people. The attenuation of community, as influentially demonstrated by Robert Putnam’s Bowling Alone, is certainly not new or confined to youth. The associational decline has proceeded since the late 1960s with each generation less engaged in public life than the last. The membership in what Alexis de Tocqueville called “mediating institutions”—voluntary associations from churches to service clubs, bowling leagues to hobby groups—has dramatically contracted in recent decades. Rates of formal volunteering have also declined. These trends, along with comparatively lower educational attainment and employment prospects, are part of the “disconnected young men” story.

Along with the bonds of friendship, we also queried young men about their integration into the larger community, whether they are involved through group membership, volunteering, or attendance at religious services. Before exploring these more objective measures, our first question sought to tap respondents’ perception of their assimilation into social life. To what extent, we asked, did they personally agree or disagree with the following statement: “I don’t feel I belong to anything I’d call a community.” We did not specify what sort of community or provide examples or offer direction of any kind. The question was about anything they might call a community.

As with the loneliness and isolation questions, about half of young men are entirely missing relations that provide meaningful connection. Specifically, 18% strongly agree that they have no community they can call their own, with an additional 30% agreeing somewhat. And the pattern was basically the same as we saw earlier: greater agreement from those who are most subject to the forces of formlessness—men who are unmarried, without a four-year college degree (and not pursuing one), secular, and from a household where their parents divorced. But, to emphasize once again, despite the comparative advantages, 36% of religious men, 36% of those with a bachelor’s degree, and 28% of married men with kids also somewhat or strongly agree that they lack a community. 

To assess community involvement, we also asked about volunteering, along with whether they participate in any of six types of associations. These included fraternal organizations (i.e., the Lion’s Club), social clubs (i.e., a book group), civic organizations, and faith-based groups (see the figure below for the full list). Although the question was asked in the present tense, we did not specify a time period, such as “in the past year,” or a frequency.

Figure 7. Percentage of men reporting participation in various social organizations

Close to 1 in 3 young men report volunteering, such as at a food pantry, an after-school program, or the fire department. A quarter are involved in adult sports leagues, and more than a quarter are engaged in a social club (27%) or faith-based group (29%). Fewer than 1 in 5 take part in a civic organization (18%) or an alumni/booster club (14%), and just 11% belong to a fraternal group. It may be, as social scientists have observed, that compared to earlier generations, young people are replacing more formal organizational engagement with novel, looser, and more sporadic forms of participation. For many, unstable occupational and transient living arrangements may strain the ability, even if desired, to commit to regular meetings, pay dues, and other expectations of group membership.

Surveys have shown that a college education is correlated with greater community involvement and civic participation. We also find that those with a college degree are more likely (39%) to volunteer than those who neither have, nor are pursuing, a four-year degree (27%).

However, very strong volunteer rates are also reported by men who completed trade school or an apprenticeship. As noted in Part I, Chapter 2, these men comprise 18% of our overall sample, and some (27%) also have four-year-college degrees. They (and we are excluding those who have a bachelor’s degree) are significantly more likely to volunteer (45%) than those who have not completed such a program (26%). Most commentary on young adults, unfortunately, misses these skilled men entirely, lumping them together with all those who do not hold a degree.  

Those currently in school—a substantial 30%—also have very strong volunteer rates (43%). As schools incubate social clubs and activities, this is not surprising. For some educational programs, volunteering may be a class requirement and is often recommended for resume building. But the fact that 28% of these young men also report feeling lonely and isolated all or most the time in the past month (and another 32% some of the time) also reminds us of the difficult circumstances of college. As with the number of friends, participating in social activities like volunteering does not provide as much protection as we might expect. 

Interestingly, despite the lack of free time reported by many parents, married fathers are more likely to volunteer (45%) than unmarried men without kids (25%). One possible explanation for this difference is that having children opens up (perhaps even requires) many additional volunteer opportunities, such as after-school activities and youth sports teams.

As we can see from the previous figure, overall participation in these types of group activities is low, although a majority of respondents (57%) do participate in at least one. Participation rates are higher among trade school and apprenticeship graduates; religious respondents; and married men with kids (as compared to unmarried men without kids). While full-time employees and self-employed respondents have higher participation rates than those who are not working (and neither in school nor caregivers), they have lower rates than students and caregivers.

Figure 8. Percentage of men who participate in various social organizations, by various characteristics

In his book Bowling Alone, Putnam worried about the decline of membership in religious institutions and attendance at services. Participation in these institutions, he argued, is a significant predictor of volunteering and is comparable to high educational attainment in its association with civic engagement. Fewer people going to services, he wrote, spells trouble for civil society. 

In our survey, we find that religious respondents are more likely to report involvement in at least one activity (81%) than nominal (66%) or secular (42%) respondents. Similarly, the religious are around twice as likely to report volunteering (45%) as secular respondents (22%). The problem, from the Putnam perspective, is that the secular contingent outnumbers the religious by more than two to one.

Attendance at religious services is not the same thing as participation in a faith-based group. For example, 40% of those who met our definition of “secular” (attending services less than once a year) report a religious affiliation; the other 60% report being an atheist, agnostic, or “nothing in particular.” More than 1 in 10 participate in a faith-based group. Meanwhile, only some of the “religious,” who, by definition, attend worship services a lot, report such participation. We took up religious attendance separately by asking how often these days, apart from weddings and funerals, they attend religious services.

Figure 9. Percentage of men who regularly attend religious services, by various characteristics

What is striking here is just how few young men are regular church (synagogue, mosque, etc.) goers. To simplify the eight response categories, we combined those who never attend (39%) and those who attend less than once a year (14%). This is the group we label “secular,” and it includes more than half of all young men (53%). Meanwhile, 23% of men are in our nominal category, which includes those who attend between once a year and once a month. That leaves 25% of men in our religious category, i.e., those who attend more than once a month.

As noted, the secular young men, who tend to identify as “nothing in particular” (37%) or agnostic or atheist (23%), are around half as likely as religious respondents to volunteer and be “joiners” of any sort of voluntary association. Similarly, they are less likely than religious respondents to have at least three close friends (52% vs. 68%); to be married (9% vs. 25%); to be a parent (25% vs. 35%); and to have a bachelor’s degree (14% vs. 34%). In addition, they are much more likely (57%) than religious men (36%) to feel that they do not belong to anything they would call a community. In short, this is the group of young men least connected, not only to religious but to associational life in general. 

A question much discussed in recent years is whether young men are becoming more religious. To be sure, there is evidence from recent surveys of rising attendance. We took the direct approach and queried respondents about whether, in recent years, they had become more religious, less religious, or stayed about the same. More than 1 in 4 (28%) say that they are now more religious than they were before. The biggest movement was among the “religious” group (those who attend services regularly). Wherever they started from, 52% report increased religiosity, with only 11% reporting less. About 30% of the nominal, and 15% of the secular also indicate becoming more religious. Of those currently in school, 36% say they are more religious, while among those, without a BA degree, who have completed trade school or an apprenticeship, the figure is 40 percent. Conservative respondents are twice as likely (42%) as liberal ones (21%) to report having grown more religious.

However, a similar percentage of respondents (26%) report a decline in religiosity, including among the nominal (29%) and secular (32%). Additionally, 34% of liberal respondents, compared to only 18% of conservative young men, say that they had grown less religious. Meanwhile, the plurality of all men we surveyed (46%) said that they have stayed about the same. Over half (53%) of secular respondents and about one-third (36%) of religious respondents indicate that their religious views have not changed much. 

Family Relations

In Part I of this report, we touched on the issue of fatherlessness and the question of role models and influencers. One of the central worries about young men is that, in a world where divorce and single parenthood are common, many do not have men in their life who can model manhood and provide authoritative guidance and steady support. As we noted, only 63% of young men had their biological father present on a continuing basis when they were growing up. The father was missing in 16% of the cases and was an intermittent presence in 21% more. Being motherless was also an issue, but for far fewer young men (5% missing, 11% only present part of the time). 

To those who reported no biological father present, we asked if there was a man in their life when they were growing up who they thought of as a father figure. We provided a list of possibilities, as well as an “other, please specify” option. The possibilities included an older brother, an uncle or other relative, a coach, a teacher, or a Big Brother (or similar youth mentor). Respondents could choose all that apply, or none. 

We also offered a write-in option; of the 126 written-in responses, half told us that no father figure was present. Not surprisingly, since we only asked about biological fathers, many of the other responses refer to adoptive fathers and stepfathers. A small number of young men also name a grandfather or other relative. Among our pre-selected options, the most common response was uncle or other relative (30%), and 16% checked older brother. The non-relatives included teacher (15%), coach (12%), and Big Brother or similar youth mentor (8%). For the great majority of young men (69%) who grew up without a biological father, adoptive father, or stepdad, a father figure was present in their lives.

Figure 10. Percentage of men without a biological father present who see various people as father figures 

As social science has shown, having a father in the house growing up is generally beneficial to a young man’s development and transition to adulthood, even when aspects of the relationship may be less than ideal. In considering their situation, we also wanted to ask about young men’s current relationships with their parents, if any. As we have stressed, parents can play a significant role offering advice and support to their young adult children. But, as we also find with friendships and social participation, the role they play does not necessarily protect against loneliness and isolation. There is no substitute for belongingness in a wider community. Yet, good relations with parents are always beneficial.

To that end, we asked young men to rank their relationship with their mother and father along a scale that ranged from very good, good, fair, or poor, to “I have cut off contact.” For those without a father or mother, we included a not-applicable response. 

Taking the father first, we get the following portrait: 27% say that their relationship with their father is fair or poor, or that they have cut off contact with him. An additional 24% describe the relationship as “good,” and a plurality (39%) as “very good.” (For 9% of respondents, this question was not applicable.)

Young men with a fair or worse relationship with their father are less likely to have a college degree (29% vs. 18%). Secular respondents are more likely to say this about their relationships with their fathers (34%) than religious ones (16%), as are those from families where the parents divorced (37%), compared with those in which there was no divorce (18%). Notably, only 10% of married respondents with children report a fair, poor, or severed relationship with their father; this was a sharply lower share than those who were married without kids (27%), unmarried with kids (30%), or unmarried without kids (29%). It is evident, therefore, that those with the weakest relationship to their father tend to be the same men with lower institutional embeddedness.

Relations with mothers are better overall. Here, just 3% of respondents mark N/A, and around 1 in 6 report a fair, poor, or cut-off relationship. Meanwhile, the majority (53%) describe their relationship with their mother as ‘very good,’ and an additional 27% consider it ‘good.’ As with paternal relationships, secular respondents, those from divorced families, and those who neither have nor are pursuing a college degree are more likely to report less favorable relationships with their mothers. 

Social integration is a significant challenge for many young men. They receive emotional support, especially from partners, if they have one, and from their families, but friendship networks are often small and group affiliations few. As a rich body of research has shown, dating back to Emile Durkheim’s famous book Suicide, the loss, weakness, or absence of valued attachments is distressful and robustly associated with poor mental health.^[xxii] Young men struggle with loneliness and isolation, and, as we’ll turn to in the next chapter, with another source of situational distress—a sense of subordination to forces they cannot control.

Chapter 5: Alienation and Distress

One of the frequent claims made by writers on the crisis of young men, especially those young men from the white working class, is that they are “alienated.” In popular writing, the term can often mean several different things. Sometimes it means social estrangement or marginalization, sometimes a lack of meaning or purpose, and sometimes a disillusionment or apathy. The attributed causes range from fatherlessness and a scarcity of positive male role models to political polarization and the “Me Too” Movement. In this account, incels (an online group of “involuntary celibates”) are an example of alienation, the manosphere is tapping into it, and self-loathing is a consequence. Violence is also a manifestation, as is passivity, lack of aspiration, and a fatalistic attitude.

In its original meaning, alienation was a powerlessness brought about by certain objective conditions in society—workers, for example, who had lost the means of control over their own work product and could no longer see themselves in it. Now, in popular discourse, alienation often specifies a failure of individuals. Persons are held responsible to assimilate themselves to the conditions of the social world regardless of how they happen to find them. They can defy social rules and conventions, of course, and there is plenty of that. But ironically, their resistance must be of the socially approved type, the type applauded as a worthy struggle. When socially disapproved, being out of step is a sign of alienation—of being lost, not liberated. Since the social conditions are constantly changing, young people must stay adaptable and flexible, as we noted in Part I. And failing to self-assimilate in a timely fashion is a measure of the crisis engulfing young men, as we saw in the first report’s Introduction. 

In our survey, we inquired into young men’s sense of alienation by exploring their perceptions of the link between their own effort and success, their ambitions and confidence, and their sense of self-efficacy in bringing about desired outcomes. The questions we asked cannot, of course, tell us what the actual, objective situation is in our society. Rather, what they allow us to see is what young men perceive to be the objective economic situation, their sense of control with respect to this state of affairs, and their hopefulness or fatalism about the future. 

In a society that aspires to be a meritocracy, success is supposed to be determined by each person’s talent and industriousness. External barriers set by ascribed characteristics, like race or family background, are to be—as far as possible—removed. Individuals can pursue opportunity, better themselves, and enjoy social mobility through their own personal effort and ability. Though often criticized as an illusion and labeled a myth, the notion of meritocracy has long been the foundation of the American Dream and the motivational foundation for calls to apply oneself and persevere in one’s plans and projects. Widely acknowledged as never fully realized, the Dream was nonetheless a promise of a better future.

To access their judgment about the value of hard work, we asked respondents if, and to what degree, they accepted or rejected three statements. 

Statement 1: "Who you know is more important than ability or hard work in getting ahead."

Nearly three-quarters of young men either strongly (27%) or somewhat strongly (45%) endorse this anti-meritocratic belief; just 7% strongly disagree. Married men are somewhat less likely to agree with this statement (64%) than their unmarried counterparts (73%), but we found little difference by employment status, parental status, age, college education, or intact (no divorce) family upbringing. In other words, the primacy of “who you know” is clearly a widely shared conviction about the economic system, indivisible by the categories where we typically find differences. 

Figure 11. Percentage of men who agree with various statements on work and success  

Statement 2: "No matter how hard I work, people like me will never achieve the American Dream."

Half of young men strongly (19%) or somewhat (30%) agree with this statement. The American Dream has been defined in many ways, but we commonly understand it as the promise that working hard will lead to a certain standard of living, including such things as access to good housing, education, and the resources necessary to marry and raise a family. To be clear, we did not ask respondents how they define the American Dream. Assuming they hold something like this general picture of what it means, however, then it seems that many young men believe that current economic realities deny people like them the chance to attain a better and more secure future.

Full-time-equivalent (FTE) work, in the eyes of a large number, is not a sure route to achieving the American Dream. Among non-students and non-caregivers, over 4 in 10 (43%) of full-time and self-employed young men agree that, for “people like me,” the American Dream is inaccessible,” while even more of those not working endorse this view (60%). Similarly, about one-third (36%) of those with a BA degree or higher appear doubtful that their degree(s) is a ticket to social mobility. Those without a bachelor’s degree or pursuing one, however, are even more likely (54%) to see the American Dream as inaccessible.

Statement 3: "The constant pressure to perform and achieve just never ends."

This third statement is less directly about the value of hard work than about its endpoint or goal. It is, in part, a question about whether young men see a clear-cut destination toward which their effort and striving might be directed, or if they see something indeterminate and unreachable instead. Three-quarters are again either strongly (34%) or somewhat (42%) in agreement. And we observe little variation across sociodemographic groups. For instance, 80% of liberals and 78% of secular respondents endorse this view, but so do 75% of conservatives and 76% of religious respondents (these differences are not statistically significant). 

At first glance, there is a seeming defeatism in the agreement with these three statements. If investing one’s effort and doing it to the best of one’s ability does not lead to success and getting ahead, or, alternatively, leads to a “success” that is little more than a hamster wheel of endless striving, then perhaps trying hard makes little sense. Perhaps, rather than open, the future is already settled. 

To explore further, we presented two statements, drawn from the sociology of stress literature, to access respondents’ feelings of self-mastery and self-worth. Asked how well this statement—“I often feel that I have little control over the things that happen to me”—describes them, a remarkable 61% of young men say very (21%) or somewhat (40%) well. Only 1 in 10 say it does not describe them well at all. 

There is some minor variation. Those with a bachelor’s degree are less likely (51%) to endorse this statement than those who neither have, nor are pursuing, a BA (63%). Those who are not working (or students or caregivers) are more likely (72%) to say the statement fits than those working full-time (53%). This gap may well reflect the higher concentration among the unemployed, as we noted in Part I, of men dealing with current mental health and disability issues. Even so, over half of workers also feel limited control over their personal experience. Which is concerning, because a sense of agency in decision-making and responsibility is crucial to how young men define what it means to be adult.

Figure 12. Percentage of men who report that various statements describe them

The second statement, also phrased negatively, asked about self-appraisal: “all in all, I am inclined to think that I am a failure.” More than 4 in 10 young men feel this statement describes them “very well” (15%) or “somewhat well” (27%); only 1 in 3 fully reject it. If feeling little control expresses a powerlessness over one’s circumstances, judging oneself a failure surely expresses a deep demoralization. Many experiences might give rise to self-derogatory feelings, including the incessantly downbeat way in which they are addressed in our society and the “non-event stressors”—a concept defined by the sociologist Leonard Pearlin as the “disjunctures between socially valued goals and differential access to opportunities for their achievement”—that they are subject to in their everyday life.

We found, for instance, that men who are not working or in school are especially inclined (56%) to see the failure statement as describing them well. Similarly, those who have not graduated from college nor are pursuing a degree are more likely (47%) to endorse this statement than are those who hold one or are studying (32%). For men without jobs or college degrees, feelings of inferiority, such as feeling less than a full adult, may arise from the experience of failing to meet socially normative standards. They are expected to discover their purpose and excel at it, to adapt and seize opportunities, to show they are making something of themselves. Success in this regard is not a status they can confer on themselves. It requires, we might say, some evidence that is visible to others and acknowledged by them—the sort of evidence that, for example, contributes to the comparatively more esteemed feelings of married men, degree holders, and those fully employed. 

In this report, we have repeatedly observed very similar positive outcomes between college graduates and those who have completed trade school or an apprenticeship. But here, the outcomes for trade school or apprenticeship graduates, with 49% seeing themselves as failures, are much closer to those without degrees (46%). Whatever else could be at work producing this variance, we might recall the attitude we mentioned in Part I, Chapter 2. A substantial majority of young men agreed that “in our society, men who work in blue-collar jobs (construction, plumbing, etc.) are regarded as not very smart.” For men in those fields, some of their lower self-evaluation may come back to such socially induced and belittling beliefs. On another question, discussed below, men without BAs who graduated from trade school or an apprenticeship are the most likely (88%) to feel that people underestimate their potential.

Figure 13. Percentage of men who report feeling like a failure or like they have little control

Combined with the work and American Dream statements, feelings of lack of control and personal inadequacy appear to demonstrate a fatalistic attitude for many. However, the response to the three positive statements we presented appears to complicate, if not contradict, this conclusion. These statements include: (1) “I have ambitious goals for my future,” (2) “I can do just about anything I really set my mind to,” and (3) “I feel confident that my opportunities in life will be greater than my parents.” Based on other responses from those we surveyed, we would predict that most young men would put little stock in such claims for themselves.

But that is not what we found—not even close. Even to the greater-opportunities statement, 75% think it describes them at least somewhat well. Similarly, 81% and 84% of respondents say the same about their capacities and ambition, respectively. In each case, full disagreement is rare. When we exclude caregivers and current students from consideration, we find that those not working—who, again, include the most impaired—are less likely (74%) than full-time or self-employed respondents (86%) to report ambitious future goals; however, a solid majority of both groups report having ambition. Regular religious service attenders are more confident (85%) than nominal (77%) or secular (69%) respondents that their opportunities will exceed their parents’ opportunities, perhaps reflecting a belief in the possibility of divine assistance. In addition, college grads and trade school/apprenticeship grads (82%) expect better opportunities than those who neither have a BA nor are in school for one (72%). 

What could account for this surprisingly buoyant assessment? It is quite possible that a fatalism about the future, their power to assert control, and even their own self-worth is at the root of many respondents’ answers. Perhaps the stated optimism is a coping strategy. In the face of a highly unpredictable future and dogged experiences of failure, they have become convinced of their inability to alter unwanted circumstances, and so projecting themselves into the future or working out a realizable change of course just meets too many painful difficulties. Better to stay upbeat, while just going with the flow.

It may also be that young men have a too optimistic assessment of their own abilities and untapped potential, and that this self-portrait leads them to blame themselves or external forces for their inadequacies or situation in life. 

Alternatively, while many young men may have lost faith in meritocracy and felt the sting of structural obstacles to their attainment of adult social goals, they still retain a faith in their own powers to get them through. Some may feel like a failure so far. They have not discovered their purpose, or advanced to where they expect to be, or found a romantic partner who will return their interest, but they are still young and have not necessarily given up on themselves, or on the possibility that things might develop in their favor. They are, we might say, alienated from the system but optimistic about themselves. 

A final set of statements sought to access any frustration respondents might experience at discrepancies between their perceived effort and outcomes. The first statement was: “I never seem to do enough to please people.” More than half (54%) find this to be at least somewhat true to their experience. The second statement also expressed a mismatch between self and others’ perceptions: “I have more potential to succeed than people think.” For more than 8 in 10 young men, this idea resonates, at least somewhat. As with many of the other questionnaire statements, there is some variation; for instance, those who are unemployed (and not a student or caregiver) are more likely to report not doing enough to please people, and less likely to assert unrecognized potential, than full-time or self-employed workers. Nevertheless, what stands out from these results is how many young men feel that their effort is constantly being found insufficient and their potential often underestimated. 

Frustration does not arise from indifference or apathy. Their responses suggest that many young men want others to treat them differently, to see that they are trying, to see that they have something valuable to offer. They want to be taken seriously, both for its own sake and perhaps also for the motivation that such respect might, in turn, give them to earn it. As we have emphasized, recognition is a two-way street. Rather than an easy accommodation to their circumstances, our findings in this area, as in others discussed earlier, suggest that young men care about their status, want to contribute, and are distressed by feelings of powerlessness. If anything, feigned indifference may be the coping mechanism. Convincing themselves or others they don’t really care may ease the pain of the gap between their experience and what they honestly desire their life to be.

Figure 14. Percentage of men who report that various statements describe them

Mental and Emotional Distress

Where individuals are chronically unable to exercise control over their circumstances, they report high rates of psychological distress. This stress response, like others we present in this report—including lacking the means to attain valued goals and experiencing the loss or absence of prized relationships—is well-established in the research literature. As we find in this survey, many young men, in one or another aspect of their lives, encounter these types of social arrangements. We would anticipate mental health problems and maladaptive coping strategies. 

We therefore asked respondents if they had ever been diagnosed or treated by a healthcare professional for six conditions, and whether they were currently experiencing symptoms or being treated. Earlier, in Part I, Chapter 2, we discussed autism, dyslexia, and ADHD. Here, we consider bipolar disorder, depression, and any anxiety disorder. The percentages of respondents who reported these conditions are presented below. We find an immense burden of suffering, somewhat higher than the rates in epidemiological studies for this gender and age group. It’s possible that some young men answered on the basis of symptoms, rather than more narrowly in terms of professional diagnosis and treatment. 

Sixteen percent of respondents report ever being diagnosed or treated for bipolar, a chronic disorder characterized by extreme mood swings, with 8% currently being treated. Of these 8% receiving treatment, only half (excluding students and caregivers) are working full time. In addition, just 4% of those with current bipolar disorder graduated from college, a much lower share than the 27% of those who never experienced this illness. In addition, those with current or past bipolar disorder are around twice as likely (32% and 30%, respectively) to be an unmarried father as those without a history of this disorder (15%).

Figure 15. Percentage of men diagnosed with or being treated for various mental illnesses  

Depression and anxiety disorders are even more commonly reported: 37% of young men indicate having suffered from depression, with 21% continuing to experience symptoms and/or receive treatment. Whether this describes persistent sadness, ongoing loneliness, or simply weariness about their situation, we cannot say, but plainly an extraordinarily large number of young men are under a dark cloud. Those reporting an anxiety disorder are similarly steep: 33% report having anxiety, and 21% report current symptoms or treatment. As these conditions often overlap, when the three disorders are grouped together, we find that 45% of our respondents have experienced at least one of them, and 29% are currently doing so. Among possible treatments, we asked all respondents (not just those reporting a diagnosed condition) if they had taken a prescription medication for their mental or emotional health in the past 12 months. We find that 1 in 3 (35%) young men have taken a medication in that time period, with 12% currently doing so on a daily basis.

Figure 16. Percentage of men diagnosed with or being treated for depression, by various characteristics

Like so many areas of analysis in this report, no demographic is immune to mental health challenges, but certain groups are especially likely to report current symptoms or treatment. For example, 41% of respondents who are neither working nor in school are currently experiencing anxiety, depression, or bipolar disorder—about twice the rate (21%) of full-time-employed and self-employed respondents who are not in school. Religious respondents are less likely (20%) to report current symptoms or treatment for one of these three disorders than those who are nominal (30%) or secular (32%). In addition, married men with kids are around half as likely (16%) to have current anxiety, bipolar disorder, or depression as those who are married without kids (31%) and those who are unmarried without kids (31%). Finally, those with a BA are less likely (18%) to report one of these disorders than those who neither have, nor are pursuing, a college degree (32%). 

Depression and anxiety affect a significant fraction of all respondents, testifying to the pathogenic circumstances under which young men are coming of age. Of course, survey data like ours can discover associations between variables, such as depression and not working; but we cannot determine whether depression is a cause of not working or perhaps a result. We do see that the unemployed (and not in school) are more likely to currently be experiencing depression (34%) or anxiety (32%) than those currently working full time (16% for depression; 15% for anxiety). At least for some, a mental health condition is almost certainly an obstacle to work. 

It is also the case, as social psychologists have long demonstrated, that higher degrees of social integration, represented by practices like marriage, education, and attending services, can provide some protection from “depressive forces” and a buffer against stress and negative life events. We find, for instance, that those who regularly attend religious services, at 14%, appear to be at lower risk of depression than those who seldom or never do (25%). Similarly, those who are both married and have kids are less likely to be depressed than those in other living situations. More and deeper relationships and social investments provide a better socially and emotionally supportive environment.

Figure 17. Percentage of men diagnosed with or being treated for anxiety, by various characteristics

Traumatic Experiences

In addition to the stressful circumstances discussed above, we also explored respondents’ experience of traumatic events that can have a negative effect on mental health. We presented five types of events, listed in the following figure, and asked if they had experienced any such event. An extraordinary 44% of young men report experiencing at least one of these stressful events. More than 1 in 4 (28%) say that a loved one of theirs has gone to prison. Roughly 1 in 5 (18%) have had a loved one die by drug overdose, and 1 in 6 (16%) have experienced the death of a loved one by suicide. Around 1 in 7 (14%) report knowing someone who died by homicide or witnessing someone be killed.

Perhaps even more significant, we also asked if respondents themselves had ever been incarcerated: 12% gave an affirmative response. (Another 59 respondents selected ‘prefer not to say,’ so this number could be even higher.)

Figure 18. Percentage of men who have experienced various traumatic events

In the figure below, we can see a clear association between the experience of potentially traumatic events and various mental health and learning conditions. In every case, from depression to ADHD, those with traumatic experiences are more likely to say they have ever been diagnosed or treated with one of these conditions. In the case of bipolar disorder, they are more than three times as likely. In addition, whereas only 4% of men without a history of trauma report having been incarcerated, 22% of those with such a history do. Again, survey data like ours cannot establish causation. What they do show, and clearly, is that many young men have been exposed to painful experiences and stressful events. For some of these men, this exposure has had harmful consequences.

Figure 19. Percentage of men by trauma history who have experienced mental health issues, loneliness, or incarceration

Maladaptive Coping With Screentime

In addition to stressful circumstances and exposure to potentially traumatic events, we also asked about various, typically online activities. Of all the themes in the public discourse around young men, none have received as much coverage as their heavy screen use, and none have been portrayed as having more dire consequences. In gaming and on social media, many have argued, young men are being incited to drop out of society, adopt aggressive, even violent behavior, objectify women, and flout social norms. In Part I, Chapter 3, we briefly considered the potential influence of the “manosphere” and reported that young men in our survey appear to be less taken with it and its image of manhood than pundits claim. But those were the only questions we asked that related to media messages. While it is not our contention that such messages are inconsequential, we are of the view, with some research justification, that the amount of time that young men play video games or gamble or watch pornography is far more significant than social media consumption (where the great majority of influencers are women). With electronic media, we used our limited questions on screentime. 

We come to this issue last because, in the interpretive scheme of this report, we place it among the responses to stressful circumstances. The predicament that young men find themselves in has been building for a long time. It did not emerge with electronic media but, as we have suggested, with more complex social changes that have made the world less stable and less certain, where the opportunity structure for men has changed dramatically, where goods like the American Dream, participation in a community, likelihood of marriage, and the work-reward nexus can no longer be taken for granted. A sense of isolation and alienation, and of having—as we put it in the survey—“little control over the things that happen to me” has followed on these changes. In this sense, the constant use of media is, in part, a coping mechanism against demoralization, but one that, as we will see, also contributes to it.

In the survey, we asked young men how often they engaged in any of eight activities, from using social media to streaming videos to trading on the stock market. The six-option response scale ranged from “almost constantly” to “about once a day” to “never.” In this report, we want to focus on four of the activities: using social media, gambling (including making bets on sports and other events, but not stocks), watching pornography, and playing video games. In terms of frequency, young men engage with each of these practices at very high rates—which, of course, overlap, as young people have many platforms and screens running at the same time. After considering the numbers, we will turn to an analysis of how the high rates of these practices correlate with negative outcomes, especially feeling lonely and like a failure.

Social Media Use

Social media use, the first concern, is nearly universal. Just 4% of young men say they do not use any platform. A majority (70%) are tuned in almost constantly (30%) or many times a day (40%), and in similar proportions across every demographic group, including the full-time employed, conservatives, married parents, and holders of college degrees. Although frequency varies somewhat, there is no social-media-free group of young men.

Figure 20. Percent of men who use social media by various characteristics

Gambling

Almost overnight, gambling, including sports and event betting, has become another fixture in the lives of young men. In fact, more than half of young men in our survey are now gamblers. Nearly 1 in 4 (23%) report that they gamble daily, plus 12% doing so more than several times a week, and a further 21% at least some of the time. As the figure below shows, conservatives (32%), those currently in school (34%), and unmarried men with children (39%) are the most frequent daily gamblers.

Figure 21. Percentage of men who gamble by various characteristics

Pornography Use

The prevalence of pornography use, the third concern, is well known and well documented. As the figure below shows, 43% of young men say they watch pornography daily (27%) or several times a week (16%), with another 26% about weekly or less. We might expect the daily watchers to be the most troubled, not in work or school, and living in their parents’ basement. While there is some truth to that, we also find daily use among married men (21%), the regular religious service attenders (24%), those who hold full-time jobs (23%), and those currently in school (32%). While 30% of young men say that they never watch pornography, they are not concentrated in any one demographic group. 

Figure 22. Percentage of men who use pornography by various characteristics

Video Games

Finally, we turn to gaming habits, which like the other three activities, have become a fundamental feature of the world of most young men. To simplify the next figure, we combined the six use options into four: “many times a day or more,” “daily,” “several times a week or less,” and “never.” Like social media, only a very small minority of young men, 5%, report “never” playing video games. An incredible 44% report playing many times a day or more, and another 22% report playing about “once a day.” Taken together, 66% are playing a video game at least once a day. There is some variation in daily users: younger men and those without a degree play more frequently.

To be clear, the issue here is not with a pleasant pastime, or occasional diversion, or a way to hang out with friends. Some use, of course, does fit that description. For example, we asked young men about the reasons they play video games. About 4 in 10 agree that “an opportunity to interact with others or make friends” is one of their reasons. Their playing is a source of camaraderie and collaboration, which, as we have seen, is often in short supply. A similar percentage agree that playing provides “an opportunity to feel a sense of mastery or accomplishment.” Again, this is the sort of experience that may be otherwise hard to come by. At the same time, nearly half agree that they play as “a way to temporarily escape real-life responsibilities,” and more than 25% as “a chance to be someone different.” These reasons are not all bad, but the escaping doesn’t look so temporary when virtually every free moment is dedicated to playing. The sheer extent of the gaming suggests a compulsion and a lack of freedom, not to mention the many other valuable activities that it obviously displaces. 

The confinement is similar with social media, gambling, and the other screen activities that have come to absorb so much of young men’s daily life. Such constant activities are enervating, at odds with other valuable goods in life, including relationships, and yet another cause and effect of demoralization. 

Figure 23. Percentage of men who play video games by various characteristics

Correlations with Distress

Turning now to the question of consequences. Earlier in this chapter, we reported that more than 4 in 10 young men (42%) believe that the statement “all in all, I am inclined to think that I am a failure” describes them “very well” (15%) or “somewhat well” (27%). As the next figure shows (which lists all eight of the screen activities we asked about), those who engage daily with social media and video games (which is nearly everyone) report feeling like a failure at about the same rate as those who engage less or never (or message/text or stream videos). However, those who view pornography (63%) and gamble (66%) every day—along with day trading and playing fantasy sports—are significantly more inclined to see themselves as a failure.

Figure 24. Percentage of men who feel like failures by frequent use of certain technology

We also found a correlation between high levels of these activities and reported loneliness. Respondents who gamble many times a day or “almost constantly” are more likely (34%) than those who gamble less often or never (24%) to feel lonely all, or most, of the time. With pornography, those who watch extensively (many times a day or almost constantly) are more likely than less frequent viewers to report loneliness (43% vs. 23%), as well as higher rates of anxiety and depression (see below). Those who play video games almost constantly report more loneliness than less frequent gamers (34% vs. 23%), as do constant users of social media (30% vs. 24%).

Figure 25. Percentage of men who experience various mental health issues by use of pornography

As in other cases, we cannot determine what comes first. Those who watch pornography many times a day or even more, report current depression at twice the rate as those who watch once a day or less. Does watching pornography lead to depression, or do the depressed turn to pornography for relief? The causal arrow may very well go in both directions. But in either case, watching pornography appears to be, at best, ineffective as a coping mechanism for dealing with dejection, stress, and loneliness, and more likely maladaptive.

Gambling and Self-Reported Problems

In the case of gambling, we asked young men for their own assessment of the effects that regular betting (several times a week or more) was having on their emotional health and everyday life. The first question asked to what degree online gambling might be causing emotional distress. As examples of distress, we listed mental pain or anguish, shame, guilt, and embarrassment, and the 5-point response scale was none, mild, moderate, severe, or extreme. The second question, with the same response scale, focused on how much personal trouble their gambling might be causing, such as with relationships or financial, legal, and employment issues. 

Respondent’s answers strongly suggest that activities like gambling cause distress and create many problems of their own. For both questions, the responses were similar. (See the figure below.) Around 1 in 10 young men characterize gambling-related distress and life problems as severe or extreme. Another 29% consider them “moderate,” and a similar percentage as “mild.” Perhaps we can see why those who report gambling many times a week or more also report more loneliness. Clearly, gambling, for whatever reason they got involved, is taking a toll that young men themselves recognize and acknowledge.

Figure 26. Percentage of men who gamble and experience distress, by various characteristics

With our survey, we set out to address subjects of public concern and asked young men, ages 18-29, a wide range of questions about their work, education, childhood, attitudes, aspirations, life experiences, social support, mental health, and electronic media use. In our exposition of their responses, we have sought to show significant correlations, provide some context for their answers, and offer some interpretation of what those answers mean. We turn now to a summary of what we have learned.

Conclusion: What Did We Learn?

In the introduction to Part I of this two-part report, we began with a brief description of the common “narrative” about the troubles facing young men. It goes like this: a growing number—too many—are failing in school, not working or studying, not dating or forming families, not leaving their childhood bedrooms. They are checked out from social life, have few, if any, friends, and belong to few, if any, social or civic groups. Their mental health is poor, their habits are bad—video games, pornography, gambling, pot-smoking—and their views are extremist. As commonly told, the dark story of their failures is, at least implicitly, set against a more successful comparison group, whether young women or older generations of men.

To explain the crisis—or perhaps better said, multiple, overlapping crises—we mentioned three of the common causal explanations. First, are the more “structural” accounts. In this view, economic and educational changes are the most important contributing factors to the dire circumstances of young men. Social scientist Richard Reeves, for example, highlights the structural changes in his book Of Boys and Men. He writes: 

Boys are falling behind at school and college because the educational system is structured in ways that put them at a disadvantage. Men are struggling in the labor market because of an economic shift away from traditionally male jobs. And fathers are dislocated because the cultural role of family provider has been hollowed out.

In other words, the struggles of young men are symptoms of wider institutional and economic alterations, including detrimental school policies, the loss of traditional male jobs, such as in manufacturing, and the decline of working-class men’s wages. As much evidence shows, policies that changed the opportunity structure have been systematically disadvantageous to men.

A second explanation puts the onus of deficiencies in young men themselves. Most tellings of the “narrative,” at least implicitly, take this moralizing stance. In these critiques, the male malaise includes some combination of laziness, lack of motivation, aimlessness, and resignation. In his 2023 book Manhood, Missouri Senator Josh Hawley, for example, faults liberal ideology and its “disdain for masculinity” for creating the crisis, yet then holds young men accountable to solve it. As he writes, they can:

turn off their screens and reclaim responsibility for their lives; they can reject the Epicurean siren song of self-indulgence, that soothing lullaby that is really just a means of keeping men distracted and dependent; they can stop taking handouts; they can find a wife and start a family; they can, in sum, get their character in order and reclaim their independence as men. Or not.

In this view, men have allowed themselves to be led astray and something like a good kick in the pants is needed. Challenged to wake up and abandon their lethargy, young men can grow up and start doing their duty. 

A third account of the crisis centers on the misguided socialization of men. According to one form of this explanation, young men’s problems arise from standards, taught to them from an early age, of toughness, self-sufficiency, competitiveness, and stoicism. In its 2023 “State of American Men” report, for instance, Equimundo: Center for Masculinities and Social Justice, argued:

Masculine norms such as these govern every aspect of men’s and boys’ lives and are woven into family life, schools, sports, and other community spaces. Such norms deprive young men of more caring alternatives, ways to confidently own their identity as caring, emotionally connected, cooperative people. Many feel totally disconnected and retreat to private lives of underachievement, underemployment, and online addiction—and to the pretense that they can go it alone. Some may find solace in misogyny and white supremacy.

“Traditional” gender roles are the culprit and need to be replaced with “healthy masculinities” (plural) that train boys/men to be more open, honest, emotionally vulnerable, and equity-seeking.

A conservative version of the socialization argument is that men have “no vision for what it is to be a man,” to quote Hawley again, are living without male role models, and trying to survive in a society that no longer prioritizes the masculine virtues.

While not exhaustive, these theories are frequently invoked, and we designed our survey, in part, with them in mind (and so we, too, deal more in problems than strengths). Considering what young men reported, we want to come back and briefly reflect on these explanations. A survey like ours—any survey—cannot provide causal explanations or settle complex questions. The contributing factors are multiple and explanations in terms of structural changes, individual deficiency, or socialization certainly offer some degree of insight into the predicament of young men. While our findings resist the “narrative” framing, it is not because the problems aren’t genuine; it is because the interpretations miss crucial features of the context, a context not of young men’s making.

In fact, none of the problems are in any way unique to young men. What makes these men stand out in specific areas is a somewhat more acute struggle than others. Loneliness and isolation, living with parents and delaying marriage, depression and anxiety, endless social media, pornography, and on-line gambling, work and money troubles, and all the rest afflict many young men, to be sure, but they afflict young women and older cohorts of men, if differently, as well. The most we can say is that depending on the issue, some are faring relatively better than others.

That said, our concern here is not with making comparisons to women or older men but with the very real challenges facing young men and with how our findings square with the “narrative” and the three primary explanations on offer. 

Structural Change

Among the causal explanations, the structural argument is the one that finds the most support here by far. Though typically centered on education and employment, the loss of “structure” we discerned running through the survey responses also extends to wider issues. The social fabric itself has thinned, a deinstitutionalization, as sociologists call it, that has left our world more formless. Over time, there has been a sharp decline of prescribed roles, inherited life courses, clear occupational and educational trajectories, and the like. Where there was once a given social pattern, an “institution,” like sustained work for the same company, there is now a more open and indeterminate field of options and choices. A precarious freedom, with less guidance from shared beliefs, customs, rules, or forms of life. 

We see what is lost by the importance of the college track, which remains (though under pressure) an institutionalized educational pathway. By nearly every metric, we find, men who took this route and stayed with it to graduation—college, trade school, apprenticeships—are faring the best. This benefit, incidentally, often does not appear while they were in college/training, which could be a lonely, distressing time. Rather, it came afterwards when it opened doors or widened horizons to stable employment, friendships, marital partners, household establishment, community affiliation, and much else. These are the men most likely to feel they have reached full adulthood, are not a failure, not unwanted, and have some control over what happens to them. 

But the higher education pathway is not open to many men, including, as we find, for financial reasons, family responsibilities, or because they simply do not feel they will succeed. Conditions that affect learning for example, like ADHD, autism, and dyslexia, are common among young men (about 25% report a current issue). Trade school or an apprenticeship program is clearly an option and a good one (15% have completed a program, with high employment), but we know from other studies that schools heavily stress college-for-all and are only now beginning to reintroduce shop class and talk about alternatives. In most situations, any clearly defined non-college pathway or school-to-employer pipelines are missing. Young men, not on the college track, often step into a void after high school, bouncing around entry-level jobs, and may only discover a vocation, such as a trade, much later. While varying somewhat by industry, the average age of an apprentice in the U.S. is 29.

In the current economic environment, where work is now more fluid and uncertain, we find that the focus for many young men is not employment, per se, but a good income. There is a logic to this position. As those making structural arguments have long observed, real wages for young men have stagnated or declined in recent decades for such reasons as the offshoring of manufacturing, cheap migrant labor, and the steady elimination of private-sector labor unions. And the loss of male jobs has continued, accelerated by the pandemic. Other shifts have increased the volatility of the economy, such as the rise of unstable gig work, the disruption of new technologies, and now the looming threat of job loss due to AI. 

This is not the old job market, and attitudes have changed. We find that many young men do not see a direct relationship between hard work and getting ahead. Rather, they appear to view things as more random, circumstantial, indeterminate, such that, for instance, success is a matter of who you know (72% agreement); what happens to me is not under my control (61%); the pressure to achieve just never ends (77%). Even half of college graduates doubt that their education was worth the time or money. Young men still retain a belief in their opportunities, yet how they expect to realize them might be better indicated by their extensive gambling and day-trading than by steady work in a 9-5 job. Even economists have taken to referring to the current situation as a “casino economy.”

The lack of institutionalized social forms affects young men in many other personal ways. For example, with the long-running decline of voluntary associations—churches, service clubs, adult sports leagues, and the like—and the instability of the family (only about 60% of our respondents had the consistent presence of his father), young men are thrown back on themselves to find and create relationships of commitment and friendship. We show that while most unmarried or unpartnered young men want to date, it is rough sailing. Many cannot find someone who will go out with them (reported by about 50%) and—witnessing to the more amorphous and fraught circumstances—are reluctant to ask a woman on a date, out of fear of being turned down. 

Those without institutional connections, without some community or church or school, are in the grimmest circumstances in terms of friendship and social integration. How, we might ask, are these men to form such relationships, when, “historically,” to quote Sam Pressler and Soren Duggan, they “have been more the gifts of generational inheritance,” than the result of some go-getter initiative? They continue:

We inherit many overlapping webs of association … [and] the relationships we form through these shared associations—family, religion, community, place, or otherwise—beget new relationships and new associations. Contrary to the self-help ethos that pervades our society, relationships and community have never been solely individual choices to make nor individual responsibilities to develop.

On these points and much else, we find social formlessness to be like a raging storm in which many young men are struggling to find some secure foothold. 

Individual Deficiency

The deficiency argument, if we can call it that, is not so much a causal explanation as it is a set of assumptions about young men’s state of mind. The idea is that young men, or at least a lot of them, are lost, apathetic, resentful, nihilistic, or other such adverse descriptors. They have become indifferent to the world around them, with little ambition, a tendency to blame others for their problems, and a desire to just be left alone. But when we asked young men themselves, we find nothing of the sort. In fact, most young men: have worthy goals, including for marriage, children, and financial independence; value a sense of belonging, warm relationships, and being well-respected; understand their role as men to be responsible and make sacrifices for others (greater than 85% agreement); and have ambitious goals for their future. And if they blame anyone, it is themselves. 

We show that many young men are in a kind of liminal position. The twenties, of course, are an in-between time of life for most everyone, part of that period now commonly referred to as “emerging adulthood” and characterized by transition and ambiguity. But we have something more specific in mind. On the one hand, young men have a sense for what a full life might look like—admirable, rewarding, joined in purpose with a family, contributing, and belonging. They see the promise, and they have not given up hope for its realization. On the other hand, they are not yet living that life or even moving toward it, and this quandary is a source of stress and feelings of failure—the stress of being aware of falling short of their own goals and of the social expectations of others, of lacking friends and meaningful connections, of being subordinate to seemingly implacable conditions over which they have little control, and of not knowing what to do or how to extricate themselves.

These are demoralizing conditions. We are not using the word in any clinical sense, but in the psychiatric literature on demoralization, the type of conditions we find are described as highly conducive to anxiety and depression, which many men report, as well as another experience, namely, “becoming helpless.” While we did not have a direct question about feeling helpless, we did see substantial agreement (61%) with the statement, “I often feel that I have little control over the things that happen to me.” When we consider the extremes to which many are engaged with on-line activities, it is hard not to suspect a kind of paralysis that is impeding purposeful action.

As we discuss in Part II, frequent recourse to some of the on-line activities, including gambling and porn viewing, are associated with greater loneliness, depression, and feelings of failure, suggesting both an enervating effect of these activities and that young men are turning to them as a means of coping with the stress of their predicament. At a minimum, it seems safe to say that these activities are serving to fill voids in real-life experience, for challenges or achievement or dates, or are engaged with the hopes of escaping their present circumstances into a different reality, the fuller life they envision. And, in the one area we asked about, young men themselves report that their gambling is leading to emotional distress.

The point here is that speaking in terms of individual deficiency misses the changed context. Without situating their experience in the reality of their circumstances, we cannot see that their feelings and actions are, at least partly, a reply to it. Lacking that, their struggles appear to be about nothing, which, unsurprisingly, is met with little sympathy—they are lying on the bed they have made for themselves. 

Socialization

To their credit, those advancing a socialization argument do engage with our changed social circumstances. They raise a fundamental question about masculinity. For present purposes, we want to set aside whatever truth claims are at issue. Instead, we want to treat the question as a way of asking what sort of “self,” if you will, young men need in order to flourish in the world as we find it now. 

As noted above, in this complex and dynamic world, greater formlessness requires that individuals be more self-determining, with many more decisions to make and increased responsibility to define their own life purpose, create their own social bonds, and so on. The demand, for men and women, is to have a masterful self and the requisite abilities are to be assertive, confident, flexible, and enterprising on one side, while also being emotionally intelligent, sensitive, caring, and team-oriented, on the other. In the socialization arguments, the problem of young men is that they are missing one side or the other of this requisite self.             

We find that a large majority of men (73%) believe that in recent decades it has become harder to know what it is to be a man. They may well be referring to the new combination of required qualities. It is not hard to imagine how getting it right might be a tricky business for many boys and young men today. However, the views of manhood they overwhelmingly affirm come down on both sides of the ledger. Manhood involves strength, responsibility and leadership, and it means a willingness to sacrifice for others. They affirm that no one should be coerced to fit any one model, and when asked about people they admire, the manosphere influencer Andrew Tate got low marks. Admittedly thin, nothing in our data suggests that lots of young men are under undue influence from bad characters or clinging to views of masculinity that could be legitimately described as “toxic.”     

In fact, when it comes to those they looked up to as role models, first mothers (79%) and then fathers (69%) led the list by a wide margin. Considering the views of manhood that young men endorse, it is not readily apparent that they have “no vision for what it is to be a man,” to repeat the quote from Senator Hawley. It seems like they do, and perhaps not so different from the one he holds. There is no question that fathers play a key role in the lives of their sons (and daughters). We find that those young men who report a good or very good relationship with their father are also the most institutionally embedded. They are more likely to be married, for instance, to have education beyond high school, and to belong to a church. How this relates to claims about a lack of masculine role models, however, is not so clear.

The wider problem here is the required self, what one is expected to be. Never have young people been asked to shoulder such burdens of autonomous self-making. That some can thrive in this fluid, entrepreneurial environment does not mean everyone can or should be expected to. Withdrawing, feeling like a failure or “never enough,” to borrow a phrase common in the books about the different but no less pervasive struggles of young women, may not be a failure of manhood. It may be a symptom of our collective impoverishment, which asks of young men something that our society no longer can or will provide.

Editor's Note: For references, notes, and more, please download the full report below.

Executive Summary

The federal government collects a wealth of data to help us understand what drives child outcomes in student learning, health, and economic security. Yet the government often leaves out a critical factor in its public reporting: family structure. In this report, we focus on education as one example of this exclusion. The National Assessment of Educational Progress from the US Department of Education disaggregates student performance by factors such as race and ethnicity, sex, and poverty. However, it does not report student scores by family structure on the official Nation’s Report Card—even though it has been systematically collecting these data since 2013. This omission deprives policymakers, educators, professionals, and parents of valuable insights into one of the most powerful predictors of academic success.

Indeed, this report’s analysis of the National Survey of Children’s Health demonstrates that children raised in “intact” families (in which the biological or adoptive parents are together and married) are far more likely to thrive in school and life than are their peers from other households. We find that children in non-intact households—where children are living with stepfamilies, cohabitating parents, single parents, or foster or kin-based families such as in grandparent-led homes—are 13 percent less likely to receive A’s and B’s than those raised in an intact, married-parent home, 47 percent more likely to have their parents contacted for problems at school, and more than twice as likely to be depressed. We also find, in this instance, that family structure is a more powerful predictor of school problems, academic performance, and depression than are race and poverty.

These are important findings because public policies, research, and discussions related to education, human services, and health policy often focus on race and poverty but neglect family composition. Indeed, despite the federal government’s commitment to reporting outcomes by different demographic groups, family structure is often ignored. Though our findings in this analysis are not necessarily general results, they indicate that family structure ought to be studied further to determine the extent to which its predictive strength is a general result.

To remedy this, key federal reports from the Department of Education and other federal agencies that monitor child well-being should make family structure a standard and consistent data category used to report child outcomes. They should also make sure standard family-structure categories are included in reports that analyze surveys on child well-being. Doing so would shed light on the impact different family compositions have in shaping student outcomes and children’s well-being. These data insights would empower key stakeholders to move beyond incomplete or misleading narratives about the causes of educational, health, and economic disparities and enable them to better serve children in their care.

By providing the public with a clearer picture of how family structure influences children’s health, student success, and economic security, federal agencies can empower schools, communities, local and state governments, professionals who work with children, and parents to address family structure when they are dealing with issues related to children’s educational success, physical and emotional well-being, and financial security. Better reporting will also help policymakers, educators, and civic leaders educate the public about the family’s role in helping children flourish inside and outside of school. Finally, reporting student outcomes by family structure will help young people understand the impact that different family types have on a range of child outcomes, thus making them more proactive regarding their own future family formation.

Introduction

American education is in crisis. In the past decade, student performance in reading and math has fallen and reports of student misbehavior have risen. The problems are visible for younger and older students alike. Based on the 2024 National Assessment of Educational Progress (NAEP), for instance, the number of fourth-grade students scoring at the lowest category of reading capability (“below basic”) now exceeds the number of fourth-grade students reading at or above NAEP “proficient” levels. As Figure 1 indicates, this trend in fourth-grade reading began to accelerate in 2017.

Similar declines are visible for older students. According to the 2024 NAEP for 12th graders, the average percentile scores in reading and math were the lowest since the respective subject matter tests began in 1992 and 2005.

American students’ failing academic performance is just the latest evidence of American education’s current crisis, and it underscores why Secretary of Education Linda McMahon has prioritized improving reading and mathematical performance as the foundation for lifelong achievement.

Indeed, as the Trump administration reimagines the federal government’s role in K–12 education, the government has an opportunity to expand the essential data it provides to help educators, parents, professionals, and the general public understand the key factors helping children succeed or putting them at risk of failure in school. Toward that end, scholars from the American Enterprise Institute and the Institute for Family Studies wrote this report to focus on a crucial aspect of a child’s life that affects their academic trajectory but is often ignored: the structure of the family in which they are raised. Family structure reflects the various types of families children grow up in today, including those with married parents, single parents, stepfamilies, cohabitating parents, and foster and kin-based families such as grandparent-led homes.

While NAEP has collected information on family structure for more than a decade, it does not disaggregate student outcomes by family structure in its reporting. This should change. We argue that NAEP should regularly report family structure as a category alongside the legally required demographic categories of gender, race and ethnicity, socioeconomic status, disability status, and “English learner” status. While the latter data categories are required by law, NAEP is not necessarily prohibited from reporting student scores by additional categories.

What’s Missing from NAEP Analyses

Recent declines in NAEP scores underline the crisis unfolding in our nation’s schools. But our ability to address this crisis has been hamstrung by our inability to understand and address all the key factors in play. Many studies on students’ poor performance have focused on race, poverty, or economic inequality and excluded family structure.

For instance, Stanford University Sociologist Sean F. Reardon has explored the role of income- and race-based gaps in student performance using a range of datasets, including NAEP, to argue that income gaps are growing in student performance. Economists Erik A. Hanushek, Paul E. Peterson, Laura M. Talpey, and Ludger Woessmann conducted a landmark longitudinal study using data from four national assessments of student performance, and they came to a different conclusion. They argued that the “opportunity gap—that is, the relationship between socioeconomic status and achievement—has not grown over the past 50 years.” But these important studies did not explore family structure’s role in shaping student performance.

Likewise, other recent analyses of NAEP results have focused almost exclusively on the absolute achievement levels and categorical disparities for a set of “usual suspect” demographic groupings, such as race and ethnicity and socioeconomic status. For example, while 39 percent of all eighth-grade students performed below NAEP’s basic math level in 2024, a more detailed analysis disaggregated by race shows that 62 percent of black eighth-grade students scored below NAEP’s basic level, as did 54 percent of Hispanic eighth graders.

Missing from all these studies and federal reports of NAEP results, however, is disaggregated information about students’ family structure. This absence is a glaring omission in the effort to better understand the key factors that influence student learning and account for the current crises in American education. It is even more confounding given that since 2013, NAEP has consistently and continuously collected information on family structure from students, but it has never reported it.

According to the historical log NAEP has kept since 2000, the 2013 NAEP Student Questionnaire for Reading Grade 4 introduced a new question: “Do any of the following people live in your home?” The answer set allows children to choose multiple answers: mother, father, stepmother, stepfather, foster mother or other female legal guardian, or foster father or other male legal guardian. (See Figure 2.)

From 2013 to 2024, NAEP has included this same query on student questionnaires across mathematics, reading, science, civics, geography, and US history in grades four, eight, and 12. Recognizing that this information had previously gone uncollected, NAEP clearly made a deliberate choice to comprehensively gather family-structure information in more recent years.

Yet NAEP has not proactively reported student achievement information disaggregated by family structure, nor has NAEP made it easily accessible on the NAEP Data Explorer tool. While reporting student outcomes by broad demographic categories is valuable and necessary, it is insufficient. Omitting family structure in NAEP’s reporting could create three significant negative consequences:

• Crucial nuances related to other demographic factors may be obscured when key factors like family structure are excluded. This can lead to misdiagnosing the reasons children are succeeding or failing academically. For example, according to the American Community Survey, 80 percent of Asian children and 69 percent of white children live with their married birth parents. This compares with only 52 percent of Latino children and 30 percent of black children. More detailed analysis is required to show how disparities in family structure within the race category may be more relevant for student outcomes than the umbrella concept of race itself. In other words, when scholars, policymakers, and educators focus on race and ethnicity as key drivers of student outcomes and overlook family structure, they may unduly emphasize these drivers and miss how much family-related factors matter for student outcomes.

• There is a misguided school of thought arguing that any disparity in a broad demographic category must be due to discrimination within that category. This is a causal fallacy—mistaking correlation for causation. For example, some believe that racial disparities in student outcomes must be caused solely by racism, as opposed to other factors such as differences in family structure across race. These reductionist conclusions, particularly regarding race, have led to the embrace of diversity, equity, and inclusion initiatives focusing on race and ethnicity as the key drivers of student performance. But by overlooking important factors like family structure that shape student achievement, approaches focusing on race, ethnicity, or socioeconomic disadvantages may miss understanding and addressing family dynamics that are more consequential for students than their race and ethnicity or economic status.

• Omitting family structure squanders the opportunity to delineate how different family types within family structure affect student development. For example, in its 2010 report Family Structure and Children’s Health in the United States, the National Center for Health Statistics declared that “in view of the changing family-structure distribution, new categories of families such as unmarried families or unmarried stepfamilies need to be studied so that the health characteristics of children in nontraditional families can be identified.” The report defined seven distinct and mutually exclusive family structures: nuclear, single-parent, blended, unmarried biological or adoptive families, cohabitating, extended, and other—the last being a family consisting of one or more children living with related or unrelated adults who are not biological or adoptive parents (e.g., grandparents).

While NAEP doesn’t report on student achievement outcomes by family structure, analysis of other government surveys reveals the specific impact family structure can have on academic outcomes. For instance, in Virginia, the Good Fathers, Flourishing Kids report found there is no racial divide in academic performance between white and black children from intact families. (For the purposes of our report, “intact families” are homes headed by children’s married birth or adoptive parents, and “non-intact” families are households where children are raised by stepparents, single parents, cohabitating parents, foster parents, or kin such as grandparents.) This suggests that family structure may be more explanatory than race when studying some student outcomes. Likewise, the 1997 National Longitudinal Survey of Youth indicated that family structure is about as powerful a predictor of college graduation as family income or race is.

Here, we examine the National Survey of Children’s Health to explore the associations between family structure, race and ethnicity, poverty, and three key outcomes for students inside and outside of the classroom: the number of times a school contacts a household regarding their child’s behavioral or learning problems, student GPA, and depression.

Student Outcomes

Schools Contacting Households Regarding a Student’s Behavioral or Learning Problems

Family structure is a strong predictor of issues at school. Children from intact, married-parent families were the least likely to have their parents contacted by schools for behavioral or learning problems, as shown in Figure 3. Indeed, nearly a third of children (30 percent) who were from non-intact homes or had unmarried parents had their home contacted for problems at school in the previous year, compared with one in five children (20 percent) from intact, married-parent families.

These disparities persist when we control for student race, income, parental education, student sex, and student age. We find that across these factors, family structure is the strongest predictor of problems at school, as shown in Figure 4. Net of controls, children living in non-intact homes or with unmarried parents were 47 percent more likely to have their parents contacted by their school than were children from intact, married-parent families.

This disparity surpasses the differences between boys and girls, with boys’ homes being 38 percent more likely to be contacted by their school. Black children were 17 percent more likely to have their home contacted by their school than white children were; the effect was insignificant for Hispanic children versus white children.

Likewise, children from households earning below the federal poverty level were only 16 percent more likely to have their school contact their home net of controls than were higher-income families (or households with an income four times or more over the federal poverty level). We found no significant difference in the number of times schools contacted households when factoring for residential parents’ highest education level. This indicates that family structure surpasses race, poverty, and parental education as a predictor of school problems in the National Survey of Children’s Health.

Good Grades

The relationship between the classroom and homelife goes beyond behavior. We find that family structure is also associated with grade outcomes, with children from intact, married-parent families outperforming their peers from non-intact families.

Note how family structure relates to student grades. We find that, as in the case of misbehavior in school, kids from intact, married-parent families tend to do the best (Figure 5). From 2022 to 2024, 74 percent of children age six to 17 from intact, married-parent families got A’s and B’s or straight A’s. This compares with 60 percent of those from non-intact homes or living with unmarried parents—a 14 percentage point gap.

This GPA premium for intact, married-parent families holds even with demographic controls. After controlling for race, income, the residential parents’ highest education level, student sex, and student age, we found that parental education and family structure were the two strongest predictors of grade outcomes for children in the United States (Figure 6). Children from non-intact or unmarried-parent households were 13 percent less likely to have received A’s and B’s or better, compared with children from intact, married-parent families. Students who did not have a residential parent with a college degree were 13 percent less likely to get A’s and B’s, compared with those who had a college-educated parent.

The family-structure effect is greater than that of family income and child race. After controlling for other factors, children living below the federal poverty level were 10 percent less likely to have good grades compared with those 400 percent above the federal poverty level. Black children were 7 percent less likely to have received A’s and B’s or better than were their white peers, and Hispanic children were 2 percent less likely. To summarize, growing up in an intact, married-parent home has an effect on GPA comparable to that of growing up with a college-educated parent. Family structure is more predictive of success in school than race, income, and sex.

Depression

There is a substantial difference in children’s depression rates across household structure. Children from intact, married-parent families have the lowest rates of diagnosed depression (Figure 7). More specifically, just 3.4 percent of children age six to 17 from intact, married-parent families had been diagnosed with depression from 2022 to 2024. On the other hand, children from non-intact or unmarried-parent homes had over twice the rate of depression (7.9 percent).

As shown in Figure 8, these effects hold after controlling for race, income, parental education, and student sex and student age. Of these factors, age was most predictive: Depression is substantially more prevalent among children in their late teens compared with younger children. After age, however, family structure was the most predictive of a child having depression in our sample.

Children from non-intact or unmarried-parent households were over twice as likely to be diagnosed with depression as their peers from intact, married-parent homes. This is comparable to children from lower-income households, who were 43 percent more likely to be diagnosed with depression compared with children from higher-income households, net of controls. Boys were 41 percent less likely to be diagnosed with depression than girls, and black and Hispanic children were 45 percent and 40 percent less likely to be diagnosed with depression than white children, respectively. In summary, family structure is more predictive of childhood depression than race, income, and sex.

Discussion and Conclusion

In 1966, the landmark Equality of Educational Opportunity report by James S. Coleman established that a student’s family structure and background remains a—if not the most—dominant factor in driving academic achievement. Despite 60 years of evidence that has often reinforced that finding, government data analyses at the federal and state levels systematically ignore family structure’s signal role in student learning. We can and must change that.

NAEP must begin reporting family-structure information it is already collecting from students. At the least, NAEP can make the broader range of data it has been compiling since 2013 more easily accessible to the public.

Tens of millions of young Americans continue to struggle with fundamental reading and math skills. It does not have to be this way. Better information can help us understand why. Using a broader range of family characteristics reveals great heterogeneity within the broad, conventionally emphasized subgroups such as race and ethnicity and family income. Knowing this information will benefit current and future generations.

The point of reporting student outcome information by family structure, of course, is not intended to be an opportunity to demonize certain family types. Rather, it is to give policymakers, educators, parents, and the general public the opportunity to understand the factors most relevant to student development. Among other things, this information could be used to help educators understand the need for lending additional assistance to children from non-intact families.

In light of this research, the universe of interventions could be expanded to help young people think differently about their future. For example, it is true that educators cannot control or change the structure of the families that students are from. But we can influence how our students think about the families they form and the series of life choices that will likely lead to their life success—and that of their children.

One way of doing this is teaching middle and high school students the “success sequence”: 97 percent of young people who graduate with at least a high school degree, enter into full-time employment, and then get married before having children are not poor as adults, and the vast majority enter the middle class or beyond. Young adults who follow the sequence are also much more likely to forge stable families of their own. A recent Institute for Family Studies report, Advancing the Path to Success: How States Can Teach the Success Sequence to Youth, gives a comprehensive discussion of the research and debate around the success sequence and lays out feasible steps for states to begin teaching it or scaling it accordingly.

What’s more, the nation is witnessing a decline in its divorce rate and an upsurge in the share of children being raised in married-parent families—in general and among lower-income and African American children in particular. The share of children living in married families reached its nadir at 64 percent in 2012 before climbing to 66 percent in 2024. Likewise, the proportion of black children being raised in a married-parent family rose from 33 percent to 39 percent over that same period. In talking about the value of the success sequence and marriage, in particular, educators and policymakers could point to these family trends to give students hope that they can achieve better family outcomes than they may have witnessed in their families of origin.

Finally, if analyses of student success indicate that poor performance is concentrated among children being raised in non-intact families, this would enable educators, policymakers, and parents to focus their attention on kids by family structure rather than by factors that may be less salient. It would also spur policymakers and parents to consider new policy and cultural efforts to augment the family stabilization pattern that is already playing out across the nation.

About the Authors

Ian V. Rowe is a Senior Fellow at the American Enterprise Institute, where he focuses on education and upward mobility, family formation, and adoption.

Brad Wilcox is a Nonresident Senior fellow at the American Enterprise Institute, where he directs the Home Economics Project; the Melville Foundation Jefferson Scholars Foundation Distinguished University Professor of Sociology at the University of Virginia; and a Senior Fellow at the Institute for Family Studies.

Grant Bailey is a Research Fellow at the Institute for Family Studies.

Editor's Note: For a list of references and the Appendix, please download the full report below.

Without Action, American Fertility Will Soon Decline

 

Introduction and Executive Summary

Of the approximately 700 million people who have lived in America since 1600, and the over 690 million since 1776, about 80% were born here. We may be a nation of immigrants—but throughout our history, the true source of our growth has been in family and fertility. As our country celebrates its 250^th anniversary, this report takes stock of how we got here and what the future may hold. Unfortunately, given that U.S. birth rates are now below 1.6 children per woman, the answer for how we reached our current scale and strength as a nation—families raising their own children—points to a dour conclusion: without a new birth of family life in America, the greatest days of freedom will indeed be behind us. Whether our family tree has been a branch of the tree of liberty since its earliest days as a sapling, or was grafted on in living memory, the future of liberty for all of us depends on the future of family. 

In this report, we present fresh estimates of fertility for every state with available vital statistics back to 1917, and—as a special feature in celebration of our nation’s 250^th birthday—we also produce birth rate statistics for one of the nation’s original mother-colonies, Massachusetts, going back 365 years. 

Beyond this new release of data demonstrating when and where American family formation has plummeted, we also use results from a new survey of over 4,700 Americans to explore how American cultural norms around friendship and celebrity families may shape fertility. If American fertility is to recover, new cultural norms around friendship and family life may be an important element. 

Finally, informed by the available evidence on fertility and family policy, and in light of the new evidence on family culture, we outline a plan to raise American birth rates back to where they belong, and where most Americans say they want them to be: 2.4 children per woman.

Major Findings

• New historical data show that the U.S. is now in its third historic period of extended below-replacement-rate fertility. Importantly, this current period of decline is already longer than previous declines, has fallen to lower lows, and is more widely shared around the country. The new data we present on historic American fertility stretches further back into our history at a higher level of quality and reliability than any previous vital statistics-based reconstruction of U.S. fertility.
• U.S. population growth has abruptly slowed down in recent years, driven in large part by fertility rates collapsing to record lows. If fertility rates continue their recent decline, U.S. population will peak around 351 million and begin declining in the 2050s. If fertility rates stabilize, population will peak around 366 million and begin declining in the 2080s.
• Desired family size has not declined nearly as much, and as a result, the gap between actual fertility rates and the number of children Americans report desiring is rising to high levels. Even as the total fertility rate has plummeted below 1.6 children per woman, surveys continue to show Americans incorrectly think they will have about 2 children each and aspire to have 2.4 children each.
• Peer culture is a key factor associated with fertility. Having very helpful versus relatively unsupportive friends can increase desired family size for young Americans by nearly an entire child per family and can increase couple intentions to have another child by about 10 percentage points (an increase of about a third).
• Celebrity culture shapes fertility: fans of celebrities who have more kids themselves want to have more children as well. Comparing family size among admired celebrities to individual aspirations, an admired celebrity having one extra child may increase an individual’s desired family size by as much as 0.15 children.
• If policymakers wish to reverse the birth rate decline, they will have to consider a wide range of out-of-the-box policy proposals, including a generous baby bonus. However, this need not break the bank: generous baby bonuses worth five or six figures can be paid for with less than 1% of the federal budget; in fact, they’re cheaper than recently proposed child care or parental leave expansions. Massive increases in the generosity of family policy do not have to be fiscally reckless.
• Beyond financial supports, policymakers and concerned private citizens should consider interventions aimed at fostering a more family-friendly culture. Based on our findings about friends and celebrities, we outline many ideas for policies that could strengthen and generate family-friendly-friendship.

The Underwhelming Expansion

"It appears the British Americans have doubled their numbers in every period of twenty-five years from their first plantation. A rapidity of population not to be paralleled in the annals of Europe! It has never been equaled since the patriarchal ages. This rapid population of the Americans arises, partly from the great accession of foreigners, but principally from the natural increase of the inhabitants. The reasons, why the Americans are more prolific than the Europeans, are, that they are less luxurious in their manner of living, and the means of supporting a family can be more easily obtained. For the last reason the Americans are induced to marry earlier in life, and consequently their families of children are more numerous.” 

– Edward Wigglesworth, 1775

In 1775, just a year before our nation’s birth, a Harvard professor named Edward Wigglesworth published a short tract called “Calculations on American Population.” It was the first true demographic projection of our young nation and documented that the American population was doubling every 20 years, primarily due to births, not immigration. From this, Wigglesworth produced a forecast of future American population. The figure below compares his forecast to the actual history of American population.

Figure 1. U.S. Population 1600-2025, actual data from Census Bureau compared to Wigglesworth’s projection

From 1775 until the late 1800s, American population growth met or exceeded Wigglesworth’s forecast. About that forecast, he wrote, 

But to anticipate [the American] population and improvements, at the close of the twentieth century, overwhelms the mind with astonishment! At that time, should their future population be as rapid as their past, the Americans would amount to ONE THOUSAND TWO-HUNDRED AND EIGHTY MILLIONS! The continent extending back to the south-sea, affording them new plantations; and the diversity of climates and soils, inviting the introduction of all the various productions of the other quarters of the world. (emphasis original)

The infectious enthusiasm with which Wigglesworth anticipated over a billion Americans contrasts sharply with his British colleague, Thomas Malthus, who would pen his gloom-and-doom Essay on the Principle of Population a few decades later, presenting a future of starvation and suffering. The American Wigglesworth saw exponential population growth as a sign of a heroic future for our country and people, while the old-world Malthus just saw people as more mouths to feed.

Why did America undershoot expectations? Why didn’t we keep up with Wigglesworth’s forecasts? It wasn’t a lack of immigration—in fact immigration rates were higher in the 19^th century than in Wigglesworth’s day. Rather, it was a decline of fertility. In the late 1600s, American fertility rates were at least 5.5 and perhaps as high as 7.5 children per woman—far above birth rates in England, which ranged from 3 to 6 children per woman. As a result, while perhaps as many as 140 million people have immigrated to the United States since 1776, at least 550 million have been born here, a ratio of 4 births per immigrant. Even in the era of mass migration from 1840 to 1925, births outnumbered immigrants by more than 4.5 to 1. In fact, it has always been the case that America is a family business of those born here—and that business likes to get some extra growth by incorporating new families through immigration, too. We are first and foremost a nation of families, and secondarily (though importantly!) a nation of immigrants.

Yet today, fertility has fallen sharply. The next page’s figure tracks American total fertility rates in the long run.

Figure 2. U.S. Total Fertility Rate and Cumulative Cohort Fertility, in children per woman, 1800-2025

Whereas in 1800, birth rates by age implied an average family size of seven children per woman, today the average young woman can expect to have fewer than 1.6 children. Annual fertility rates can be very volatile, and so we also show birth rates summed across women’s reproductive lives to estimate completed fertility. 

As can be seen, while completed fertility is less volatile, across the decades it very closely tracks the total fertility rate prevailing at the time a given birth cohort of women was around age 25. The correlation between actual births per woman near the end of a cohort’s reproductive years and the total fertility rate prevailing when that cohort turned 25 is 0.96—nearly a perfect association of the two. Where the total fertility rate goes, completed fertility tends to follow.

Of course, back in the 1800s and before, many of those children who were born did not survive. The figure below uses the same fertility data as above but removes births of children who likely died before age 15.

Figure 3. U.S. Total Fertility Rate and Cumulative Cohort Fertility, in children per woman, adjusted for child mortality rates, 1800-2025

Americans had many children, it’s true—but half of them died before age 15. Rates of child mortality in England in 1800 were likely even higher, and fertility lower. So, while America’s fertility advantage was still very real, the truth is that most people in the early days of the American experiment did not raise 7 children to maturity. They typically raised between 2 and 4 children. Moreover, the U.S. has seen three separate episodes of below-replacement fertility once we also account for the true replacement rate determined by child mortality: from 1933 to 1941 during the Great Depression, from 1972 to 1989 during a period of rapidly rising contraceptive uptake and abortion rates, and from 2009 to today. If the correlation between total fertility rates and completed fertility rates continues to hold up in the future, completed family sizes will plummet to new lows.

Finally, it is worth asking what kinds of families Americans actually want. The figure on the next page compares surviving fertility rates to survey-reported family size ideals, both general ideals (i.e., answers to questions about the ideal size of a family), and personal ideals or desires (i.e., answers to questions about what individuals want for their own family lives).

As can be seen, fertility rates have tended to move in the same direction as stated fertility preferences over time: the correlation over time between general ideals and total fertility rates is 0.85, while for personal ideals and fertility rates, it is 0.94. In recent years, large numbers of surveys have been conducted, especially online, and online surveys in particular have tended to produce unusually low estimates of desired fertility. Even so, the gap between actual birth rates and survey-reported desires is the highest it has been since the early 1970s, and approaching levels not seen since the Great Depression. In other words, fertility is not just falling—it is falling further below the number of children Americans report wanting.

Figure 4. U.S. Total Fertility Rate and survey-reported family size ideals and desires, 1930-2025

In this report, the first edition of a new annual research report at the Institute for Family Studies’ Pronatalism Initiative, we present brand-new survey data about American family culture, as well as newly digitized data about the history of American fertility, in order to venture some Wigglesworthian hypotheses about the future of the American family. More importantly, we offer some suggestions on how our country can get back on course towards the fulfillment of its historic promise of growth and abundance.

Fertility in the 50 States

Birth rates vary appreciably around the country, and in fact, not all states even share the same trends in recent years. But Massachusetts first mandated that births be registered in 1639, though compliance was sporadic until 1841, when a reinforced law was passed. Modern genealogical databases have digitized these records, and we extracted a wide range of vital statistics data from Massachusetts annual vital statistics reports after 1840. In combination, these sources can be used to infer fertility rates for Massachusetts with relatively high reliability back to 1660.

Figure 5. Massachusetts Total Fertility Rate, 1660-2025

Massachusetts had high fertility during the Puritan era, but fertility was already declining by 1720 and had entered a steep decline by the 1760s. It then stabilized at around three children per woman throughout the 19th century, before falling again in the 1920s and 1930s before the Baby Boom. Today, Massachusetts’ birth rate is around 1.35 children per woman—one of the lowest in the nation, and on par with rates observed in Europe. Massachusetts is far from the only state with a unique trend in fertility. The figure below shows trends for states which have, historically, had some of the highest birth rates.

Figure 6. Total fertility rates for selected high-fertility states, 1917-2025

As can be seen, several western states have historically had the highest birth rates, especially Utah, Idaho, Alaska, and South Dakota. Mississippi and New Mexico have also historically had high birth rates, though, today, New Mexico has below-average birth rates, as Hispanic fertility rates have fallen rapidly. High fertility in the frontier states has long been an area of academic research. Recent research suggests that land availability for farmers may be part of the reason for this, but high marriage rates may be even more important.

Figure 7. Total fertility rates for selected low-fertility states, 1917-2025

The set of states with very low birth rates is perhaps unsurprising: eastern, urbanized states.

The District of Columbia has almost always had a very low total fertility rate. But even less totally urbanized states like Pennsylvania, New York, and Massachusetts seem to have had low birth rates for a very long time. Even before accounting for child mortality, New York state had below-replacement-rate fertility in 1929, before the stock market crash that set off the Great Depression. The two figures above also clearly show that American fertility was declining from 1917 onwards, which suggests the Great Depression may have intensified fertility decline but did not cause it. Whereas some states have persistently had relatively low or high fertility, other states have seen major changes, as shown in the next figure.

Figure 8. Total fertility rates for selected falling-fertility states, 1917-2025

The states in the figure above all had relatively high birth rates, vs. the national average, from 1980 to 2000—but today are all below the national average. Thus, whereas California was not a low fertility state in the latter 20^th century, it is today, with a fertility rate of just 1.42 children per woman in 2025. Likewise, Nevada, Arizona, and New Mexico have seen their total fertility rates plunge. Interestingly, so have Oregon and Illinois, states with somewhat smaller Hispanic populations. In those states, fertility rates are especially low in large urban areas.

Figure 9. Total fertility rates for selected less-declining-fertility states, 1917-2025

Then there are, of course, states where fertility rates have seen a less negative trend, as in the figure on the prior page. North Dakota has famously seen its fertility rates move in a more positive direction than national averages, presumably due to the fracking boom fueling high incomes, high male employment, and ultimately higher marriage rates. Less remarked upon is how states like Kentucky, West Virginia, and Tennessee have largely bucked the trend of rapidly falling fertility and kept their fertility rates comparatively stable.

Finally, it is worth assessing American regions by their contemporary rates to get a sense of the overall geography of American family. The figure below shows the map of total fertility rates in 2025.

Figure 10. Total fertility rate by state, 2025

Around the country, just one state still has fertility rates high enough to be near or at the long-run population-stability level: South Dakota. Every other state is below, and some are far below. Along the West Coast, in New England, in Illinois, and in Florida, fertility rates are very low. Only in the upper South and the Great Plains does some vestige of the old American family culture still remain for reproductive-age people. But what of women who are no longer reproductive age? How does completed fertility vary around the country? Technically, women might continue having children into their 50s—but in practice, over 90% of births for recent cohorts have occurred by age 40. Thus, cumulative fertility rates experienced by women who were 40 in 2025 can give us a very strong sense of where in the nation small, completed family sizes might already be emerging. Total fertility rates are currently much lower than cohort fertility rates, which generally implies that future cohort fertility rates are likely to decline vs. present cohort fertility rates.

Figure 11. Cumulative cohort fertility rate by state for the cohort aged 40 in 2025

For women who are in their early 40s today, in much of the middle of the country and the South, actual fertility is near or just above replacement rates. But in the Northeast, and to some extent the West Coast, near-completed fertility rates really are very low. For women who turned 40 in 2025, cumulative Massachusetts fertility rates point to a completed family size of just 1.6 children per woman. Based on U.N. data, that’s lower than completed fertility for similar cohorts in over 200 of the 237 countries for which data exists. This low fertility rate is on par with many Eastern European countries. Eastern European women born in the mid-1980s came of age during the fertility collapse of the 1990s and 2000s after the fall of Communism. Apparently, family life in Massachusetts is about as successful as family life just after the fall of the Soviet Union.

Thus, the state of American fertility can be summed up in simple terms: not good. A nation once propelled to unprecedented heights of power by its rapid demographic expansion now rests at a fraction of its historically anticipated population because birth rates have crashed. While immigration has continued, falling fertility rates are a global phenomenon and such skilled migrants as do exist will be in demand by many countries. Reliance on migration for population growth is not a credible strategy in the long run. If current trends continue, American population may peak within the next few decades and then enter interminable decline, as we will lay out in greater detail below.

The State of Family Culture Today

In each annual State of Fertility report, we will present new data on cultural and social forces shaping American fertility. In this one, we ask a simple question: how is American friendship shaping American fertility? A growing body of research has suggested that the digital revolution may be reshaping American fertility as iPhones and social media and pornography replace social life, spread new ideas, and alter the landscape of American relationships. 

The 2026 State of Family Culture Survey asked a range of new questions of Americans ages 18 to 50 and specifically explored how their social influences and friendships shape their fertility behaviors. We asked respondents how many children their three best friends had (excluding family members and romantic partners), as well as the sibling with which they have the closest relationship (if they have any siblings). This allows us to estimate the average parity of the respondent’s close social ties: up to three friends and a close sibling. Moreover, we then asked respondents how their social ties would react if they had another child. They could select from four different “positive” options (i.e., “My friends would offer to help me with my new baby,” or “My friends would cook meals for me after the baby was born”) and from four different “negative” options (i.e., “My friends would worry about my career,” or “My friends would not invite me to socialize as much”). We assessed how these and a few other social variables shaped desired family size (the number of children respondents said they’d be happiest having), as well as intentions to have any more children conditional on not yet having reached desired family size. In other words, we measured how friendship dynamics shape overall family size desires, and then, separately among respondents who hadn’t yet achieved their desires, how friendship dynamics shape their concrete intentions to achieve their goals.

Figure 12. Children desired among under-30s and share intending children among couples interested in having more children, by religious attendance, children among social ties, and supportiveness of friends

It is well known that religiosity is closely related to fertility desires and outcomes. Indeed, as can be seen in the figure above, religious attendance predicts higher desired family size among Americans under 30 (~2 children for never-attenders at the mean of control variables, vs. 2.7 for others), and also predicts higher intentions to make progress towards desired fertility (50% for regular attenders at the mean of control variables, vs. under 40% for others).

But the associational effect of helpful friends is similar to or larger than the effect of religion. Since the estimates shown are from a model including all presented variables, we can be fairly confident that the linkage between helpful friends and fertility is separate from the effect of religion: it’s not just that religious people have more helpful friends.

For Americans under 30 with the bottom 10% least-helpful friends (i.e. those scoring -2 or less), desired family size is around 1.7 children per respondent. For those with the top 10% most helpful friends, it’s around 2.8 children—helpful friends are associated with an entire extra child in desired family size. When it comes to intentions to have more kids among couples who ideally would like more, there are similar effects: only 20-30% of those with unhelpful friends intend to have more children, vs. 40-50% of those with the most helpful friends.

Indeed, passive social influences from peers matter, too: simply having more friends with kids is associated with desiring more children. Again, while we can’t rule out reverse-causality (people who want kids may seek out friends with kids), most people form their friendships well before they start having kids, and our results are similar if we only include sibling child numbers. Our results are also unchanged if we include controls for the number of siblings respondents had (i.e. family culture of origin) and are also unchanged if we exclude nonreligious individuals (i.e. a possible source of confounding selection), or if we limit it to just childless respondents under 30. Moreover, we find that the average family size of peers has twice as much of an effect on an individual’s family size desires as their number of siblings in childhood. 

Other questions we asked about parent-child relationships and extended family ties had little or no association with fertility desires. For example, we find no association at all between an individual’s number of cousins and their family size desires: big families do not intrinsically create kids who grow up to have big families. Closeness to extended family, and even stronger relationships with parents, are not strongly associated with differences in family size desires. Rather, fertility desires are more about peer culture than family culture.

Another topic we surveyed also points to the importance of peer and popular culture: publicly admired figures. We asked every respondent to report a public figure whom they most admired and why. We then collated these reports and identified how many children each figure was publicly known to have based on public celebrity or Wikipedia pages. We find that each additional child born to an admired public figure predicts higher desired family size for the survey respondent. The effect is about as big as we find for an individual’s number of siblings, suggesting mere celebrity fandom is almost as important as origin family size. Notably, these effects are larger for women and more statistically credible. 

Celebrity family sizes are about as influential for family size desires as the number of siblings. Women in particular appear to have family desires meaningfully associated with the family sizes of celebrities they admire, which again suggests that popular, peer, and public culture shape family size desires at least as much as family of origin or religion.

Notably, having lots of peers with kids does not predict actual intentions to have kids among couples whose ideals exceeded their current parity (i.e. those who might be open to intending more kids). An admired celebrity’s number of children also has only a small and statistically weak association with actual intentions. 

Figure 13. Children desired among under-30s, by sex and number of children of most admired public figure

To shape behavioral intentions, it is not enough to passively have friends with kids or to admire an influencer with kids—prospective parents need peers who are genuinely helpful and supportive, not just those who demonstrate big families. Demonstrative exposure to family life shapes aspirations, and those aspirations might have various knock-on-effects, but behavioral intentions among individuals whose aspirations exceed their current family size require more than just a pro-family public culture. They require peers who show up to change diapers. Moreover, given the apparently large role of social ties in shaping fertility behaviors, the idea that the digital revolution might be reducing American and global birth rates by reducing the extent to which Americans are exposed to family life is entirely plausible.

The Future of the American Family

Given what we have shown thus far about American fertility trends, and given the range of cultural norms shaping fertility, we next consider the possible range of futures for American demography. 

To start with, it must be noted that existing population forecasts, such as those produced by the U.N. Population Division, the Congressional Budget Office, the Social Security Trustees, the Office of Management and Budget, or even the Census Bureau, are often too cautious and uncreative. These forecasting bodies adopt two sets of unrealistic assumptions. 

First, they tend to adopt a very standard, yet unjustified, approach to forecasting net migration: a fixed, flat future net migration rate. This assumption itself is unjustified for at least two reasons. First, under President Trump, net migration has already fallen to near-zero rates, and this is likely to continue through his presidency—and he is unlikely to be the last restrictionist president of the 21^st century. Second, in the very long run, immigration is very likely to decline due to declining global birth rates and ongoing changes in immigration policies. Falling fertility means there will be fewer excess young people looking to emigrate from their homes, and more countries competing to get them as immigrants, and so the default assumption should be that any given country’s immigration rate will tend to decline over the 21^st century. Our assumption about migration is simple: net migration remains only a few hundred thousand people per year during the Trump Administration, then rises through the mid-21^st century to around 1.5 million, then gradually declines until it is just a few hundred thousand again by 2100.

The second unrealistic assumption in most public forecasts relates to fertility. Other forecasters tend to assume various degrees of fertility recovery in the long run (except the latest CBO forecasts, which are more realistic), even though no such recovery is guaranteed. 

To rectify these issues, we produce three new population projections using a cohort-component model of the same type that other statistical bodies would use, except that we assume that over the next century, immigration cannot remain permanently at a high positive level, and we model different fertility trajectories. Our fertility assumptions (Continuation, Recovery, or Revival) for these scenarios vs. public projections are shown below.

Figure 14. Historic and projected future fertility, 1915-2100, by source of projection

The Social Security Trust Fund (OASDI) does consider a scenario where fertility is stagnant at 1.6, but they do not even countenance the possibility that fertility might continue to decline. The base scenario for Social Security, Medicare, and Medicaid budget planning assumes a rapid fertility increase similar to what we call our “Revival” scenario. On the other hand, the UN and the Census Bureau more-or-less straight-line birth rates at recent levels, while CBO, almost alone among the forecasters, offers a plausible scenario where fertility falls, then recovers a bit as women delaying fertility rush to make it up. CBO’s scenario is similar to our “Recovery” scenario—but no forecaster has had the political capital and resolve to actually tell the American people what will happen if fertility keeps falling. 

We do so here. In the next figure, we show what will happen to American population if, as in our “Continuation” scenario, fertility drops to 1.35 by 2050 and 1.15 by 2095; or, at the other extreme in our “Revival” scenario, what will happen if fertility rises to 2 by 2050 and 2.1 by 2060.

If something like what CBO expects, or what we term a “Recovery” scenario occurs, then population growth will continue at a low level until the 2080s. The U.S. may continue to barely eke out a slow pace of growth well into the 21^st century. Growth would be far lower than it has been for most of our history. Edward Wigglesworth’s ghost would be infinitely disappointed in us, but the American experiment would not go into reverse for some time yet.

Figure 15. Historic and projected future population, 1940-2100, by source of projection

But if current trends in fertility rates continue and if births keep getting postponed later and later, then population will peak in the mid-2050s and then go into relatively rapid decline. In this scenario, the latter 21^st century will see America in demographic—and therefore economic and political—retreat.

That’s the bad news. But the good news is that there are ways to avoid this outcome, and to make the Recovery scenario more likely—perhaps even to achieve the trends we list as a Revival, i.e. a return to replacement-rate fertility or better. 

How To Restore American Family Growth

There are five major reforms which, in combination, could restore the fortunes of American fertility. Beyond these five major reforms, we describe three key cultural domains where, if Americans want to see population revival, the culture will need to change.

Pronatal Policies

Reform 1: American Birthday Accounts

President Trump launched a small savings account seeded with $1,000 to give emerging adults a leg up in his “Trump Accounts,” passed in the One Big Beautiful Bill. The Heritage Foundation has proposed a larger investment intended to mature upon marriage. These ideas are good starts. But the most complete proposal in this regard is a recent proposal in Finland called Vauvasampo. Adapted for the American case, this proposal is simple: every child born as a U.S. citizen in 2026 or any future year would have some amount of money, perhaps $15,000, invested in their name, which we call “American Birthday Accounts,” in honor of our 250^th year of independence. 

Beneficiaries could not touch these accounts until they have a child; that is, until they are the legal and custodial parent of a related child born in the United States or under U.S. jurisdiction abroad, and coresiding with that child or else deployed on U.S. government business. At the first birth (or, if preferable to avoid risks of early child abandonment, at the child’s 1^st birthday, if still coresident and full custodial), they would gain access to, say, 50% of their account’s value, and the residual 50% would continue growing. At the second birth, 75% of the fund’s remaining value at that time could be claimed. At a third birth, all remaining funds can be claimed. Assuming funds are invested in something like a mutual fund, a $15,000 investment could easily lead to a married couple receiving a baby bonus worth $100,000 for a first birth, with smaller additional payments for subsequent births. Recipients could be permitted to cash out their benefit over multiple years if they preferred, and any new funds gained through subsequent births would be added to this continuing fund. 

This baby bonus money could be counted as income, which means that part of its cost would be directly recouped through interactions with means tested programs and income taxes: beneficiary families at both very low and very high incomes would receive smaller after-tax-and-benefit returns. All families of any income would be eligible, but in practice the real benefits would be most generous for middle-income married families, subsidizing fertility the most for working- and middle-class families. Because only children born in the U.S. would be eligible for the investment, concerns about subsidies for children of immigrants would also be alleviated: it would be essentially a subsidy only for U.S.-born individuals to have their own children. Because married couples would be eligible for each parent’s baby bonus, the benefit would effectively double for married couples. To avoid creating subsidies for teen pregnancy, fund accessibility could be set aside until parents reach an appropriate age (perhaps 21 for a first birth, and a slightly higher age for subsequent births). 

Bang-for-buck, American Birthday Accounts are the single best way to get more babies born in stable families than almost any other policy imaginable. In the long run, since many individuals will have fewer than 3 children (and many will be childless, thus leaving many funds unclaimed), those unused funds can be reinvested in the program to create a rolling national family trust fund, which would render the program zero-cost to taxpayers after the first eligible generation had completed their childbearing. Even without that reinvestment, the budgetary cost for an investment of $15,000 to $20,000 per child would be between $45 and $80 billion per year. For comparison, U.S. public schools spent just under $19,000 per year per student in 2021, so this program amounts to the public investing just one year of schooling worth of public resources into children’s future family life. 

Assuming American fertility responds to cash incentives the way fertility has responded to cash incentives that we studied in other countries, these American Birthday Accounts would likely boost fertility by between 20 and 40% for the cohorts eligible to receive the benefit. If reproductive-age Americans today already had such accounts, birth rates would be somewhere between 1.9 and 2.2 children per woman—essentially at replacement rate.

Reform 2: Child Caregiver Credits

When a family has a child, they face new expenses. One parent may also stay home from work. The result is that after having kids, families save less. Childcare, diapers, and work disruptions cause parents to approach retirement with fewer resources than they should have had based on their human capital. In dozens of countries around the world, this problem has been addressed through care credits. These credits have a simple principle: if a person has a child under the age of 5 at home, for the purposes of future Social Security benefit calculations, they should be “credited” as if they were still working at their income levels 1-3 years before birth, adjusted for inflation. This gives stay-at-home parents (especially mothers) slightly better retirement benefits, reduces the harms that women especially face from employment disruptions, and helps recognize that parents are doing the real work of keeping Social Security solvent by raising future payers. 

While we cannot estimate exactly what effect this change would have on fertility, back-of-the-envelope calculations suggest that a four-year care credit of this sort could plausibly boost the Social Security benefit by 1% to 3% for a mom who took five years off work to have two children. For a mom who took 15 years off to have four children, care credits might boost her social security benefits by 5-10 percent. Converting these benefits to their net present value at time of childbearing, and using that to infer fertility impacts, only implies a fertility increase of perhaps 1% or 2%, but ensuring basic fairness for parents is a benefit in its own right.

Reform 3: End Marriage Penalties 

Nearly every means-tested program in America creates marriage penalties one way or another, and this discourages family formation. Policymakers must establish a simple principle: getting married should never cause anybody to lose a benefit. In practice, this amounts to suggesting that the U.S. should adopt fully individualized tax and welfare programs. Sweden, Canada, and the U.K. have all changed from joint-filing to individual-filing systems in the past, and in all three cases, fertility rose sharply afterwards. 

The reason is simple: household-filing and household-benefit systems have a strong tendency to punish marriage for working-class people. If switching to an individual system is a bridge too far, policymakers should simply go through existing tax and benefit programs, line-by-line, and remove marriage penalties, especially in Section 8 housing benefits, Medicaid, the Children’s Health Insurance Program, WIC and SNAP nutrition programs, the Child and Dependent Care Tax Credit, interest deductions for mortgages and student loans, FHA loan eligibility rules, and the Earned Income Tax Credit. 

We are unable to provide a specific effect estimate for how marriage penalty elimination might influence fertility, since the academic literature does not currently contain standardized effect estimates for either how marriage causally shapes fertility, nor how marriage penalties causally influence marriage. Nonetheless, we believe that these effects could be substantial. In Sweden, Canada, and the U.K., a switch to individualized tax systems was associated with about 0.05-0.15 more children born per woman.

Reform 4: Build Family Housing Now 

American young people face a housing crisis. There has been zero increase in residential square footage per capita in 25 years according to federal data. New houses are shifting heavily towards small apartments. Home prices have skyrocketed to many multiples of young adult incomes. America must build more family-suitable houses. 

The federal government can encourage this by adopting an extremely simple program: allocate municipalities and states alike a fixed sum per bedroom of housing completed each year. A moderate portion of existing block grant funds could be repurposed and reallocated on the basis of bedrooms being constructed, justified by the simple notion that places building more housing and thus anticipating more population growth really do need and deserve more block grant funding to prepare for a more populous future. Places refusing to build housing can make do with the federal funding they currently have. 

Reform 5: Institute Family-Impact Budgeting

At the structural level, Congress and state legislatures must begin to undertake “family budgeting.” Legislators must ask not only how a bill might impact revenues and spending, but how it will impact marriage and fertility. Every bill that receives a cost score should also be scored for how it impacts incentives to marry and have children, giving legislators a constant reminder that their choices have consequences for the physical survival of the nation, not only of its financial accounts.

In combination, we believe instituting these reforms could raise U.S. birth rates to at least 2.1 children per woman and perhaps as high as 2.4 children per woman if marriage effects prove to be substantial.

Family-Friendly Cultural Norms

Norm 1: Enlist Celebrity Culture

Our survey results imply that enlisting celebrities to promote American family life may work—celebrity fertility really is associated with the fertility of fans. Governments interested in boosting fertility should consider enlisting the support of celebrities popular in their jurisdictions, or perhaps even finding a way to obliquely encourage those celebrities to marry and have more children. 

While we would not endorse or recommend such a policy, it is empirically possible that paying Taylor Swift a billion dollars to have children might produce more children in society than spending the same money on child tax credits, if her choice sways her wide fanbase. More broadly, to the extent that social and cultural elites set the tone for family life, large tax breaks for parenthood at high incomes (of the sort which France implemented through its family quotient system) might stimulate elite fertility, thus, creating downstream cultural spillovers to fans and followers. 

Outside of government, it would be worthwhile for churches and community actors to consider the social role of parents and families: are they nudged into the back pew, their children sequestered in Sunday school where they can’t bother anyone? Or are parents full participants in community or church leadership—with their children at their side?

Norm 2: Encourage Supportive Friendship

Helpful friends make a difference. The first thing governments can do to support more supportive friendship is to ensure that informal care arrangements are not punished. Liability and negligence laws and enforcement norms should not punish parents for leaving children with friends. Safe-harbor and Good Samaritan-style liability protections, as well as “free range parenting” laws may be advisable.

More broadly, governments could consider creating “civil godparents,” that is, legally recognized roles for specifically designated friends or kin of parents, giving those friends broader in loco parentis rights to care for children. 

Likewise, hospitals could modify antenatal and postpartum care such that nurses making postpartum visits encourage new parents to call friends to come visit the hospital, or prenatal care includes efforts to help expecting parents produce a list of friends they can call to help. This method of “legitimating the request” may help parents overcome embarrassment to ask for help. 

Beyond this, governments often confer status through public campaigns—blood donation and organ donation is almost entirely motivated by appeals to conscience and moral status. Governments could emulate these campaigns for care donation: watching your friend’s kids. This approach would effectively confer honor and status on the individuals who do the helping. 

Further, local municipalities could experiment with subsidies and liability protections for small, designated play areas in coffee shops, grocery stores, and restaurants, where parents can socialize amid productivity. 

While the government cannot mandate supportive friends, it can create conditions where parents and their friends get helpful nudges into helpful behaviors. Needless to say, the kind of supportive friendship described here can only be nudged by government—but they can be supported by book clubs, mom groups, churches, running clubs, pick-up basketball leagues, and other social groups. They can celebrate one another’s family additions, create schedules for babysitting, drop off meals or DoorDash gift cards, and help each other in any number of ways.

Norm 3: Promote Real-Life Socializing

Simply having more friends with kids matters for fertility, but a key driver of having more friends with kids is simply having more friends. Indeed, the average number of children in an individual’s social circle is strongly correlated with their average number of close friends: individuals under 30 with fewer than 4 close friends average around 0.7 children per friend; whereas those with 5 or more average around 1-1.3. And yet, the number of close friends Americans have is in persistent decline. Moreover, Americans are spending less and less time together with their friends. These changes are part and parcel of a long-term decline in American social lives. Reversing this decline must be a central part of any project of cultural change. 

Schools can certainly restrict mobile devices during school hours and at extracurriculars as one step, and should also consider expanding recess time. Limiting how many assignments require devices, reducing solo-homework assignments, and instead assigning group projects requiring in-person collaboration, may also be a useful avenue. 

Expanded public funding for public sports and athletics fields alongside higher usage fees charged to pay-to-play teams, alongside reduced liability for volunteer coaching and informal play, could help restore neighborhood sports and reduce the all-consuming professionalization of travel sports. 

Policymakers should also at least consider more creative options: reinstituting Blue Laws requiring business closures on Sundays might help create more in-person community, for example. And what about “Digital Blue Laws”—laws which impede access to non-essential digital services on certain days or at certain hours? Could such laws be implemented? One possible vehicle might be to charge excise taxes on data usage on Sundays, or excise taxes on digital advertising expenditures on Sundays, essentially encouraging platforms to nudge users offline at those times. 

Laws around existing physical spaces could be tweaked as well—would higher taxes on alcohol sold for home consumption alongside lower taxes on alcohol sold in bars and restaurants lead to more spontaneous friendships? Would raising taxes on restaurant delivery while reducing taxes for in-restaurant dining create new communities? Could expanding the supply of quality park shelters and other bookable spaces be part of a broad “in-person friendship agenda?” 

The exact pathway to restoring American friendship is not clear, but policymakers should consider bold experiments. 

Conclusion

American population growth has already slowed dramatically and is likely to grind to a halt or even go into reverse in the relatively near future. Some current workers may see population decline before they reach retirement, if current trends continue. This is a serious problem and is not the family future most Americans desire. 

Unfortunately, tackling this decline is not just a problem of fixing one or two small issues—fertility is shaped by innumerable factors, including many cultural and social norms governments may struggle to change. Nonetheless, if policymakers will consider policy changes equal in magnitude to the size of the problem we face, Americans need not have anything to fear. Demographic decline is a choice, one that our nation’s leaders can choose to avoid, if they will invest in the future of family. Perhaps, with luck, at our 350^th anniversary, we might finally reach Wigglesworth’s billion. 

Editor's Note: For Data and Methods, author information, and references, please download the full report below.

Fathers and Hope

Introduction

Men in America are not 'alright.' Whether it is men falling behind in the job market, not attending college, or never leaving their childhood home, American men are clearly in trouble.

But not all men have been equally hit. Working-class men have especially faced the brunt of the “happiness recession.”

Experts from all leanings have provided explanations for why men, especially blue-collar men, are doing so poorly. Economists point to data that show how the working class has been left behind. Journalists suggest that insidious ideologies among the less educated share the blame. And some influencers argue that it’s the crisis in masculinity that explains male discontent. But woefully missing from this conversation is a factor hidden in plain sight: the family. 

In a new IFS research brief, we look at the latest data from the Global Flourishing Study, a survey with over 200,000 participants in 22 countries, to explore how family conditions in childhood explain the well-being gap between college-educated and working-class men in adulthood (see the Appendix for the Data and Methods section). Focusing on men in the United States, we find that childhood family conditions—particularly the relationship between father and son—provide a key explanation for why so many working-class men are struggling. The family variables we look at do not completely explain the gap between blue-collar and college-educated men, but they spotlight an important factor on the map: family structure matters for working-class men.

Working-Class Men Are Struggling

The well-being gap between working-class and college-educated men is panning out in nearly every well-being metric measured by sociologists and economists. 

Take just one example. Data from the General Social Survey (GSS) show that the share of men ages 25 to 44 who are “not too happy” has risen from 11% in the 1980s to 24% in the most recent years. Both college-educated and working-class men have seen declines in happiness. But men without a bachelor’s degree fare far worse, with 27% reporting they are “not too happy,” compared to 17% of college-educated men.

Figure 1: Unhappiness among men by education 

The class disparity in happiness is nothing new. The GSS data reveal that college-educated men were happier than their less-educated peers going back to the 1980s. And while the GSS shows somewhat steady levels of happiness from the 1980s through the 2000s, other well-being metrics suggest rising discontent in earlier decades. Mortality data, for instance, shows a steady and concerning rise in what researchers call “deaths of despair”—fatalities attributable to drug overdoses, suicide, and alcoholism. The rise in “deaths of despair” has been especially dramatic among white working-class men. Our analysis of death records and population estimates finds that the suicide rate per 100,000 white non-Hispanic men ages 25 to 44 has risen from 31 in 1992 to 54 in 2023, a shocking 71% increase over just three decades.

Figure 2: Crude suicide rate among non-Hispanic white men by education 

The roots of this crisis are partly economic. The collapse of manufacturing in the early-to-mid 1980s eliminated millions of stable, well-paying jobs that had long anchored working-class communities. Men without college degrees were hit hardest. Since 1970, their real wages have declined relative to their college-educated peers, and their labor force participation has fallen sharply—a trend that shows no sign of reversing.

But while the economy plays a role in the rise of working-class male despair, many economists miss a key factor: the family. In a recent Institute for Family Studies report, we documented the troubling collapse in family formation among men without college degrees. In 1980, men in their prime working years were roughly equally likely to be married and living at home with children, regardless of educational attainment. Since then, marriage rates have declined for all men, but the drop among non-college-educated men has been far steeper. The share of working-class men ages 25 to 44 who are married fell from 73% to 42% from 1980 to 2025, according to the Current Population Survey. The share of younger working-class men who are married with children at home fell from 61% to just 30 percent. This rising group of untethered men are disconnected from the stabilizing institutions of marriage and fatherhood that research consistently links to better health, purpose, and longevity.

This not only matters for the men themselves, but for the children they are raising in stable or unstable homes. We know that family of origin has a profound impact on well-being across the lifespan. Childhood experiences, particularly those within the home, shape how people assess the quality of their lives and whether they believe their futures hold promise. Working-class communities have now endured nearly 45 years of declining marriage rates and family instability, and the consequences are compounding across generations.

In this Research Brief, we draw on data from the first wave of the Global Flourishing Study (GFS), a large-scale multinational survey fielded in 2023, restricting our analytic sample to U.S. male respondents ages 25 and older (n = 18,178). We define working-class men as those without a bachelor's degree and examine three well-being outcomes—hope, sense of worthwhileness, and purpose—using linear regression models that control for age, race, rurality, and nativity. We then employ a coefficient attenuation framework to assess the extent to which the well-being gap among young working-class men (ages 25–44) is explained by childhood experiences, including parental marriage, quality of relationships with mothers and fathers, perceived parental love, and parental religious attendance. 

In plain terms, this approach asks how much of the well-being gap between working-class and college-educated men shrinks once we account for differences in their childhood family experiences. You can read more about the methods in the Appendix. 

Differences in Well-Being by Class Are Pronounced Among Young Men

Consistent with other studies, we find that working-class men report lower levels of well-being than college-educated men. Specifically, on a 10-point scale, they report .21 lower scores on “hope,” .52 lower scores on “life worthwhile,” and .44 lower scores on “purpose.” These differences are relatively small-to-moderate in magnitude, but even small effect sizes can carry practical significance when examining population-level disparities in well-being.¹

Figure 3: Mean self-reported scores on well-being for men by education 

These overall variations, however, mask meaningful variation by age and class. As shown in Figure 4, two patterns emerge: (1) well-being scores increase with age, and (2) class differences in well-being are substantially attenuated at older ages. As a result, young working-class men report the lowest scores across all three scales. When comparing young working-class men to older men, effect sizes are quite large; when compared to their same-age college-educated peers, the effect sizes are moderate. Among men ages 45 and older, working-class and college-educated men report relatively similar well-being scores.

Figure 4: Mean self-reported scores on well-being for men by age and education

Class Differences in Well Being: Family Structure and Fathers

Class differences in well-being among younger men are partially explained by family experiences, especially the relationship with the father.

We find that young-working class men ages 25 to 44 are much less likely than their college-educated counterparts to have positive family experiences. These men score lower on all variables, and the differences are statistically significant for every variable. 

Figure 5: Childhood experience prevalence among young men by education

Compared to college-educated men, working-class men are 17 percentage points less likely to have grown up with married parents, 11 percentage points less likely to have had a good relationship with their father, and 10 percentage points less likely to have felt loved by their father. 

Working-class men are also less likely to have had a good relationship with their mother, felt loved by their mother, or had church-attending parents. Such family variables demonstrate consistent, statistically significant correlations with the well-being outcomes—meaning that better childhood family conditions significantly predict later-life well-being. Previous studies suggest that relationships of this kind are at least partially causal. 

Figure 6: Difference in childhood experience prevalence among men by education

Next, we focus on men ages 25–44 and examine differences in outcomes across the various family-related variables described in the methods section (see Appendix). The “baseline model” coefficient represents the relationship between the well-being outcome of interest and class. The subsequent variables represent the relationship between well-being and class after accounting for the family variables. 

Figure 7: Attenuation of class differences in well-being among young men using childhood experiences

We find that differences in well-being between college-educated and younger working-class men are significantly explained jointly by the family variables, ranging from 18% to 33 percent. Hope is the most significant factor explained by the family variables, whereas assessments of life as worthwhile and a sense of purpose are explained less by family variables. 

Across all models, relationship with the father is the most significant explanatory variable of the association between class and the outcomes of interest. It explains nearly one-third of the difference in hope between college-educated and working-class young men. In sum, feelings of hope for men in young adulthood and into middle age are connected to childhood family conditions, with the father-son relationship being particularly salient.

Figure 8: Attenuation of differences in hope across classes using childhood experiences

Discussion

In this research brief, we find statistically significant differences in well-being outcomes by class, though it is worth noting that effect sizes in the overall models are small. While these differences are real and merit serious attention, care should be taken not to overstate the plight of working-class men as a whole.

That said, meaningful differences do emerge when we stratify by age. Younger working-class men are considerably less hopeful than their older counterparts. While it is impossible to cleanly distinguish age from cohort effects with the data available, it seems plausible that many of these younger men will grow more hopeful as they move through adulthood and accumulate the stabilizing resources—relationships, careers, community—that tend to come with time. What is clear, however, is that the class gap in well-being is pronounced among younger men in a way that it simply is not among older men. Young working-class men rate their well-being substantially lower than young college-educated men across the outcomes we examine, and this disparity deserves to be taken seriously.

To make sense of these findings, it is essential to situate them within a broader structural context. Over the past four decades, family life in working-class communities has undergone significant and well-documented erosion. Rates of family instability, father absence, and household economic stress have risen sharply in these communities relative to their college-educated counterparts. It is against this backdrop that our results take on their full meaning. Across every family-related variable in the dataset, young working-class men report consistently lower scores than young college-educated men. These differences reflect accumulated disadvantages in the family environments in which these men were raised.

Critically, these lower family variable scores partially explain the relationship between class and well-being, helping to clarify why young working-class men fare worse on the outcomes we examine. The largest and most consistent attenuation effect is observed for hope. The seeds of hope are likely planted early in life, nurtured within the family environment. For these young men, the relationship with their father emerges as arguably the single most influential factor shaping their sense of hope. A similar pattern holds for assessments of life being worthwhile and for a sense of purpose, though the attenuation is somewhat less pronounced for these outcomes. Taken together, the findings suggest that family of origin—and the father relationship in particular —plays an important role in transmitting, or in failing to transmit, the psychological foundations of a hopeful and meaningful life.

Conclusion

The well-being struggles of young working-class men are real, consequential, and—as this analysis shows—at least partly rooted in childhood family experience. While economic forces have undeniably shaped the landscape of working-class life, the research points to something closer to home, namely the relationships with parents. The father-son relationship in particular stands out as a powerful driver of whether young men grow up with hope, purpose, and a sense that their lives matter. 

Reversing the tide of working-class male discontent will require honest engagement with this reality. Stronger family stability, father involvement, and support of the relational foundations of childhood are not peripheral concerns. They are central to any serious effort to help the men being left behind.

Note: Download the full research brief for Data and Methods information.

---

1.  We calculated Cohen's d, a standardized measure of effect size that expresses the mean difference between groups relative to the pooled standard deviation. Effect sizes ranged from .10 to .28, which are generally considered small-to-moderate in magnitude by conventional benchmarks.

 

Liberals More Likely to Hold Beliefs Linked to Fewer Kids

Executive Summary

Birthrates have been declining precipitously throughout the United States since the Great Recession and more broadly since the postwar Baby Boom. Within this broad trend, a fertility "gap" has emerged between the American left and right. This report, based on survey results from a sample of over 7,000 Americans ages 18-54, analyzes ideological differences¹ in anxieties related to parenting and self-perceived competence. To a limited extent, this report finds that liberals are more likely to cite certain anxieties related to fertility decisions relative to conservatives. For each of our sections, we find that the differing survey responses account for significant differences in fertility outcomes. 

Here are our major findings:

• Liberals have fewer children than conservatives. Conservative respondents report 1.40 children compared to 1.09 for liberals. Similarly, 51% of liberals report zero children compared to 40% for conservatives. Both of these differences are statistically significant.
   
• Parenting concerns are more prevalent among liberals than conservatives. For example, 18% of liberals are unsure whether they would be a good parent, compared to only 9% of conservatives. Similarly, whereas 36% of liberals view parenting as “very complicated, difficult, and stressful,” only 24% of conservatives do. Given that both beliefs are associated with fewer children, their increased prevalence among liberals may help explain the ideological fertility gap. 

• Liberals are more likely to report genetic and mental health concerns. For example, 18% of liberals, but only 10% of conservatives, report worries that they will pass on bad genes or inheritable conditions to their children. Similar percentages of liberals (19%) and conservatives (10%) say that their mental health is currently not good enough for them to have children. Both statements are correlated with lower numbers of children—and may thus shed additional light on the conservative-liberal fertility gap.

Together, these findings indicate that concerns about parental competence, stress, mental health, and genetic conditions might, at least in part, account for the emerging fertility divergence between the American right and left.

Introduction: The Gap

It is a well-known fact that fertility has been decreasing across America over the last 65 years, yet the growing partisan fertility gap has only recently emerged as a measurable phenomenon. Research from 2024 found that, among women who were born in the years 1975 through 1979, conservatives had 2.1 children, moderates 1.8, and liberals just 1.5. Further research conducted by IFS in 2024 showed that more conservative counties have a dramatically higher number of children than more liberal ones.

Recent literature also shows that the number of desired children also differs between liberals and conservatives. For example, a 2024 study found that the gap in desired children has existed between the left and right since 1989, the first year they began tracking for the discrepancy. In this paper, the gap was initially explained by differences in religiosity between the two groups. 

However, in recent years, this gap has grown and can no longer be explained purely by religious differences. Ideological gaps in desired and actual fertility exist in our survey results as well. Among Americans ages 18-54, liberals have 1.09 children on average whereas conservatives have an average of 1.40 children, a difference of 29%. This gap proves statistically significant even when accounting for several important confounding variables including sex, religion, race, income, age, marital status, and education.

Figure 1. Mean number of children by ideology

This gap also manifests itself in accordance with the ideal family size. When asked within our survey if a person could choose exactly the number of children they wanted, conservatives report 2.71, moderates 2.43, and liberals 2.16. The differences between conservatives and both moderates and liberals remained statistically significant after accounting for our selected control variables. Lyman Stone, the director of the Pronatalism Initiative at IFS, and Scott Yenor reported that the partisan gap is only appreciable after 1994, meaning that this ideological gap in fertility preferences is a relatively recent discovery. 

Much has already been said about the emergence of this gap. However, much less has been done to explore the underlying reasons why this gap exists. This brief’s objective is to further the discussion of what these reasons might be. While survey data limits our ability to make conclusions about causality, we can still explore and identify the characteristics of this disparity.

We find that the ideological fertility gap operates along two related dimensions. First, conservatives report substantially greater confidence in their ability to parent well. Second, liberals are significantly more likely to be anxious about their mental health or genetic conditions. 

Liberal and Conservative Views on Parenting

We find that concerns about parenting have a particularly strong relationship with fertility outcomes. When people are not sure they would be good parents, or when they perceive parenting to be incredibly difficult and stressful, they have dramatically fewer kids than those who do not. We also find that liberals are more likely to endorse these concerns than conservatives. 

The following chart shows a notable difference in parenting fears between liberals and conservatives, with 18% of Liberals unsure whether they would be a good parent compared to only 9% of conservatives. A large majority of both groups do not believe they would be bad parents. However, among those who fear being a bad parent, the partisan gap is quite large—and, even when controlling for parental status and other variables,² statistically significant. 

Figure 2. Percent of adults who agreed with statement, by ideology

Liberals are more likely to express uncertainty about their parenting ability. We can only hypothesize why this is, but it may relate to a greater sense of anxiety around parenting in general. 

Regardless of why this difference exists, it correlates closely with decreased fertility. Individuals who fear they might be bad parents have, on average, 0.50 kids compared to 1.35 children among those who think they would be good parents. In other words, respondents who believe they might not be good parents report 63% fewer children on average than those who do not hold this belief. It is hard to overstate how significant a difference this is. It represents the largest difference observed among the variables examined. This speaks strongly to the power of perceived competence in parenting.

As already shown, liberals are likely to hold this belief to a much higher degree than conservatives. It follows that one reason the fertility gap exists is that liberals have greater reservations about their ability to parent. In fact, after controlling for uncertainty about being a good parent, the difference in childlessness percentages between conservatives and liberals is no longer statistically significant.

Viewing parenting as difficult could also influence one’s fertility. Those who see parenting as stressful and complicated have only 1.05 children on average compared to 1.31 among those who do not. Thus, respondents who hold this perception around parenting have, on average, around 25% fewer kids—a significantly lower number. This is an intuitive gap, given that many people try to avoid things they find to be incredibly stressful and difficult.

As with doubting one’s parenting abilities, we find that characterizing parenting as difficult also varies by ideology. Significantly fewer conservatives (24%) see parenting as “very complicated, difficult, and stressful” compared to liberals (36%). This gap, which remains significant when controlling for other potential explanatory variables (including parental status), gives credence to the idea that liberals view parenting as a more costly endeavor than conservatives.

Figure 3. Percent of adults who agreed with statement, by ideology

However, even conservative respondents who do see parenting as very difficult are significantly less likely to be childless (42%) than are liberals who share this view (62%). In contrast, there is not a significant difference in the probability of being childless between conservatives (40%) and liberals (45%) who disagree that parenting is difficult. This suggests that being liberal and viewing parenting as difficult may have a particularly strong impact on childlessness. 

Taken together, there appears to be a clear difference between the worldviews of liberals and conservatives. Conservatives are less worried about the potential stress and difficulty of parenting, and also have higher levels of confidence in their ability to parent well. Given that these beliefs are correlated with greater numbers of children, these two worldviews surrounding parenting could further influence the fertility gap.

One caveat in interpreting these results is that we cannot be sure about causality. It is certainly plausible that fearing that one will be a bad parent, and viewing parenting as difficult, will reduce one’s likelihood of having children. However, it is also possible that the experience of raising kids itself makes one realize that they are a good parent—and that parenting, despite its challenges, is not as stressful and difficult as they once believed. Future longitudinal surveys could help us identify whether parenting worries are resulting in fewer kids; whether having kids lessens one’s parenting concerns; or whether some other factor is influencing both of these outcomes. 

Anxieties Across the Divide

The ideological differences in our survey results extend beyond parenting concerns. We also find that liberals are more likely than conservatives to report concerns about passing on bad genes; in addition, they more frequently cite their mental health as a reason for postponing children. Groups that identify with each anxiety tend to have significantly fewer children, on average, than those who are not worried about each topic. In particular, worries about genetics and mental health both reflect concerns about what kind of world or situation a child could be born into. Though few people in the aggregate endorse these concerns, liberals have cited genetic and mental health concerns as influencing their fertility preferences considerably more than their conservative counterparts.

To begin, those who are concerned about passing on genetic conditions tend to have fewer children. We asked survey respondents to either agree or disagree with the following statement: "I am worried I will pass on bad genes or inheritable conditions to my children." Potentially, people who are worried about passing on bad genetic traits may fear the burden it would place on themselves as the parent and on the development and life of the child.

We find a wide and statistically significant gap in fertility between those who affirmed this statement and those who did not. Within the survey, those who are worried about passing on inheritable conditions have, on average, 0.88 children, while those who disagree with the statement have 1.30 children. This shows that those who reported genetic concerns have 32% fewer children than those who were not worried.

Additionally, 18% of liberals are worried about passing down poor genetic conditions to their offspring; this percentage is significantly higher than that of conservatives (10%), even when controlling for parental status and other potential confounds. These results are striking, not only because only around 3.9% of American children have a genetic condition, but also because we would expect genetic conditions to be randomly assorted within our sample (thus preventing any relationship between ideology and genetics). 

Despite the presumably equivalent chances of having inheritable diseases, there is a significant divergence between liberals and conservatives on genetic concerns. This divergence holds even when applying controls. 

Figure 4. Percent of adults who agreed with statement, by ideology

Additionally, while 65% of liberals who affirm this statement are childless, the same is true for only 48% of those who do not affirm it. Meanwhile, we did not find a significant difference in childlessness shares between conservatives who affirm this statement (52%) and those who do not (39%).  Thus, the relationship between genetic concerns and childlessness appears more pronounced for liberals than for conservatives. 

Genetic factors are by no means the only anxiety related to childbearing that respondents cite. Roughly 1 in 7 respondents endorse the statement that “my mental health is not good enough for me to have children right now.” This statement has a clear negative correlation with respondents’ fertility: Those who consider mental health a significant concern have—on average—0.84 children, while those who disagree with that statement have an average of 1.30 children.

Again, there is partisan disagreement on the relevance of mental health to fertility decisions. About 19% of liberals affirm this statement compared to only 10% of conservatives; this difference is statistically significant even when controlling for parental status and other variables.

Given these findings, mental health concerns may help mediate the relationship between political ideology and fertility. Indeed, when limiting the analysis to respondents who endorse mental health concerns, we find that the percentage of liberals who are childless (70%) does not differ significantly from that of conservatives (56%).  Similarly, among those who don’t report this concern, the percentages of liberals (47%) and conservatives (38%) who are childless do not differ significantly either.

Figure 5. Percent of adults who agreed with statement, by ideology

In summary, mental health worries, like genetic concerns and parenting fears, are correlated with lower fertility—and are also more prevalent among liberals than conservatives. Looking at both variables, there is a consistent trend: increased anxiety and worries are correlated with having fewer children. In addition to this finding, liberals are more likely to have these anxieties than conservatives. This could help bring to light the reasoning that liberals might give for why they are having fewer children or no children at all. It could also provide insight into some of the factors that influence their fertility decisions.

Conclusion: The Gap Explored

Across multiple measures, ideological differences in parenting confidence and anxiety are strongly associated with differences in fertility outcomes. Conservatives report greater confidence in their ability to parent and lower levels of concern regarding mental health and genetic risk. Liberals, by contrast, are more likely to express these anxieties. The consistency of these associations suggests that worldview and self-perception may play a meaningful role in the emerging partisan fertility gap.

There are limitations within both the data and the statistical models used to draw these conclusions. The methods employed in this report cannot determine whether identifying as liberal leads individuals to adopt these views and ultimately have fewer children, or whether individuals who already hold these views and have lower fertility are more likely to identify as liberal. In other words, the direction of causality remains unclear. Another limitation concerns how political ideology was measured. Respondents were grouped into three simple categories: liberal, moderate, or conservative, which are somewhat ambiguous. Future research could include push questions of sliding scales to potentially ascertain more accurate data on where respondents fall on the political spectrum.

Even with these limitations, the overall pattern in the survey data is consistent. Liberals are more likely to report anxiety about their views on parenting and childrearing than conservatives. Although causal claims cannot be made, the cumulation of these associations suggest that political affiliation is linked to perceptions of parenthood and, in turn, to the emerging partisan fertility gap.

Editor's Note: To view the Appendix and reference notes, please download the full research brief below.

 

Executive Summary

Religious affiliation and participation in the United States have declined steadily over the past several decades. Fewer Americans now attend worship services regularly, identify with a faith tradition, or describe religion as central to their daily lives. This shift is especially pronounced among younger cohorts, who are more likely than previous generations to report no religious affiliation and less likely to engage in institutional religious practices. While belief has not disappeared, it has become more individualized and less connected to church life. As a result, many religious communities now face a sustained pattern of generational decline rather than temporary fluctuation, raising concerns for churches and church members alike about the long-term vitality of their congregation.

Research consistently shows that families are the single most important factor in whether children adopt and maintain faith into adulthood. Congregational programs, clergy leadership, and peer networks matter as well, but they are most effective when reinforced within the home. Studies demonstrate that parental modeling, shared faith practices, and the quality of parent–child relationships are among the strongest predictors of adult religiosity. When faith is embedded in everyday routines through conversation, ritual, and visible commitment, children are more likely to internalize it as part of their enduring identity. Taken together, this body of evidence underscores the need for analysis of how family processes operate in practice and which specific parental behaviors most effectively foster durable Christian commitment.

Using data on American adults aged 25+ who were raised in a Christian faith, we examine how parents most effectively transmit faith to their children. Consistent with past research, we show:

• Religious practice in childhood is highly predictive of religious practice in adulthood.
• Higher parent–child relationship quality in childhood is associated with stronger retention of religious belief and practice in adulthood.
• Higher parental marital quality is associated with greater faith transmission.
• Congregational involvement on the part of both parents and adolescents is linked to higher levels of faith commitment when children reach adulthood.

Our findings suggest that if faith transmission from parents to children is to remain viable, efforts must focus on equipping both families and churches with practical tools and guidance for intergenerational faith formation. This includes helping parents integrate faith into everyday family life, providing robust faith communities and supports, promoting loving and stable family relationships in the household, and more. By centering families and supporting them intentionally, church communities can better address the decline of Christianity and sustain faith across generations.

We make the following recommendations for parents and pastors, which we elaborate on more fully in the report:

Recommendations for Parents

1) Be your children’s role model for faith

2) Prioritize strong marriages and parent-child relationships

3) Make faith formation a joint effort

4) Build religion into everyday family life

5) Make faith a regular topic of family conversation

Recommendations for Pastors

1) Guide parents—not just children—as part of religious education

2) Support strong marriages and coparenting relationships

3) Actively engage fathers

4) Create space for community

5) Invest in youth ministry

Introduction: The State of American Religion

Religious families in America are in a challenging situation. Religion in the United States is in decline by virtually every measure. Religious identification, worship attendance, and belief in God have all dropped by double-digit margins since the 1990s. Furthermore, research shows most of this decline is intergenerational. That is, rather than people becoming less religious over their adult lives, each new generation enters adulthood less faithful than the last one. This points to a decline that looks slow at first, since generational replacement is a slow process, but cascades into an avalanche over time as the most devout generations disappear and are replaced by more secular ones.

There is not any one reason for this decline, but rather several factors working together in a kind of “perfect storm.”

• Social values have shifted away from authority and tradition, and toward individualism and autonomy. Children growing up in this cultural shift are primed to look at long-established religious institutions with suspicion. 
• The end of the Cold War created a change in American identity where faith in God became less central. 
• Early internet discourse in the mid-2000s and subsequent social media platforms exposed younger generations to doubts and atheist arguments and enabled them to form communities outside religious boundaries.  
• Moral scandals on the part of Catholics and Protestants alike turned a lot of people off to organized religion.  
• The partisan culture war led many on the political left and center to look askance at traditional religion, which they see as incompatible with their worldviews. 
• Increased education and delayed family formation led young adults to stray further away from faith for longer periods of time, making return more difficult. 
• As extracurricular activities in the teenage years and professional demands in adulthood have increased, for many people, faith was not so much abandoned as just crowded out.

On May 26, 2026, the Institute for Family Studies filed an Amicus Brief in support of S.B. 2420, a Texas state law on App store safety, in Students Engaged in Advancing Texas, et al., Plaintiffs vs. Ken Paxton.