The Truth about Demographic Decline

Population, with its booms and busts, is a macabre fascination in our modern, quantified world. It was Stalin who gave the famous quip about one death being a tragedy and a million a statistic; it is demography which has made of this quote an entire academic discipline. Yet, when prompted to write on “demographic decline,” as I often am, I find it perplexing: what even is demographic decline? Total population in the world as a whole or in the United States specifically is not declining, nor is it likely to decline for the next several decades. In the United States, if our population starts declining, we can, for the next 50 years or so at least, just open up to more immigration. Why bother about “decline”? 

Others interpret decline in other ways: it’s about changing population structure, an aging population, the social transformations attendant on rarer youth and more common elderhood. People with these concerns are worried about dependency ratios for pension programs, economic dynamism, or even intergenerational economic mobility. These are all fair concerns, but, truth be told, hardly demographic emergencies. Keeping transfer programs solvent requires only modest changes in taxation, and economic dynamism and mobility are influenced by demography but far more sensitive to policy choices related to education, housing, and criminal justice. It is not strictly and absolutely necessary to tackle the demographic question in order to address the problems of a lopsided age pyramid.

Some will see demographic decline in another, more perverse way: population is rising, yes, but driven by non-white races, “dysgenic” fertility, or others whom modern antinatalists don’t dare call “undesirables,” at least not in public, not yet. Demographic decline, then, becomes too broad even to discuss: it means too many old people, too many brown people, too many disabled people, or not enough people… it is all things to all people, a stick with which to beat today’s bogeyman.

I wish to suggest a very specific definition for demographic decline: it means demographic outcomes that are explicitly and emphatically undesired by the people most immediately affected, in growing prevalence. For example, people don’t generally desire premature death. Yet death at young ages is rising rapidly in America. That is demographic decline. People generally desire children, often very deeply, and we know empirically that fertility does actually rise when economic and policy support for childbearing increases, indicating not just a stated but a revealed preference. And yet, fertility is falling far below what people say they want. That is demographic decline. Most people want to get married, and most at a reasonably youthful age (not 20 perhaps, but not 37 either): and yet fewer people are getting married, and more of them are marrying later than they would have liked. That is demographic decline. 

We should not construe demographic decline as if we are omnipotent central planners, trying to argue about the ideal ratio of old to young, or manage the population to produce the right kinds of citizens; this is pure hubris. We don’t know the “right” demographic outcome, best for human flourishing. But we can make a good guess that people know for themselves their own best outcome, and when we ask them about that in surveys, we find most people are experiencing “demographic decline”: seeing young people around them suffer and die excessively from drugs, alcohol, suicide, and homicide; struggling to find a suitable and stable partner while youth remains to enjoy them fully; confronting infertility due to long delays in initiation of childbearing. These are what we should mean by “demographic decline” because these are the “tragedies” under the “statistics.” We should avoid the fallacy of aggregation, refuse to specify the “right” fertility rate for a society, and simply say, “demographic success” means nothing more and nothing less than people living healthy and long lives, shared with long and stable relationships of love yielding cared for and beloved children. Not because we (policymakers and think tankers) think those things are good for society, but because, empirically, that is what most people do in fact desire. 

Even before Covid, U.S. life expectancy was stagnant or declining thanks to “deaths of despair.” This is indisputable “decline.”

With this definition in mind, we can think about demographic decline more helpfully. First, we must address immigration, the ever-present red herring of demography. Most research suggests immigration is instrumentally useful for migrant-receiving societies (it makes them richer and more productive), and of course migration is directly beneficial to migrants themselves, dramatically increasing their options in life and their wellbeing. That’s all well and good, and these may be fine reasons to support higher levels of immigration. But neither of these beneficial features of immigration enables migrants to serve as a solution for demographic decline as I have articulated it.

Will higher immigration make some other person less likely to overdose on fentanyl? If we issue 150,000 more green cards, will it make the young men debating whether to buy a ring or a new car decide for a ring? Can we expect immigration to have any effect at all on a family deliberating whether or not to have a third child? Economists will leap to offer entertaining examples: immigrant labor may reduce the cost of household services which are used for childbearing! Faster aggregate population growth may make marriage markets more liquid! These arguments are fair but also a bit silly; they are so marginal to the core decisions being made that they cannot be seriously entertained as the main tools at society’s disposal for helping people overcome barriers to their individual freedom and flourishing. Immigration may be good or bad, but it just isn’t responsive to demographic decline, as it does not get to the basic, individual-level tragedies of varying intensity which are constitutive of that “decline.”

Discussions of demographic decline often harp on fertility, with debates about how to interpret changing birth rates, but that’s the wrong starting point. Dead people have no children; we must start our discussion with the fact of our own impending doom, not as a society, but as fleshly and mortal beings. People may debate if falling fertility attests to women’s emancipation or declining living conditions, but surely nobody argues that the explosive increase in opioid deaths, or the recent increase in homicide deaths, is progress. Even before Covid, U.S. life expectancy was stagnant or declining thanks to these “deaths of despair.” This is indisputable “decline.” There can be absolutely no debate that we have entered a period of serious demographic decline on this metric. Since the early 2000s, the odds that a 30-year-old dies in a given year have risen more than 30%. That’s decline. That’s horrible. 

I won’t brook any of this progress-and-optimism codswallop: yes, modernity has generally been quite good to humanity, but we are presently in America inhabiting one of the exception cases to this general rule. People who spout platitudes that demographic decline is overstated because actually modern life is so good must then explain why so many young Americans are opting out of modern life, and indeed life itself, in suicide. By starting our discussion at mortality, we begin with a very clear and emphatic reality: demographic decline is indisputably real. We are in a precarious moment. The putatively inevitable march of progress has begun to stall out in America. Where we go from here is anybody’s guess.

The decline of marriage in America is not a story of working class Americans being liberated from the shackles of an arcane institution, but of social and economic chaos among poorer communities in America, depriving many of the poor of fair access to a widely-valued institution.

Likewise, marriage is being increasingly delayed, or not occurring at all. There is no serious argument to be made that American society has turned against marriage in the sense of adopting some anti-marriage ideological position: for wealthy and educated Americans, marriage remains close to universal. It is only among poorer Americans, who face stiff penalties on their taxes and welfare applications if they have the misfortune of having a spouse who works, that marriage has declined so much. According to data from the Current Population Survey, the share of college-educated 30-year-old women who had never married barely rose between 1990 and 2019; for less educated women, it more than doubled, such that college-educated women were more likely to be married than non-college-educated women. At age 35, for college-educated women, there’s been virtually no change in the share of women who are married between the 1960s and today. For non-college-educated women, the share who have never married has risen six times, from 5% to 30%. It’s tempting to suggest this is just because Americans of different socioeconomic classes desire different things, but there’s little evidence for this: in both 1988 and 2012 when the General Social Survey asked if people thought that married people were happier, college-educated and non-college-educated women had basically the same patterns of beliefs about marriage and happiness. 

Rather, marriage has declined among poorer Americans for other reasons. Liberalized divorce meant marriage didn’t guarantee a coparent even as poor families lack the hard-to-divide assets that help tie richer couples together in the absence of strict anti-divorce norms. High rates of incarceration of men stripped many communities of males and scarred those men’s employment records once back in society. Means-tested programs actively punish marriage for people with working-class and lower incomes even as the tax bracket structure rewards marriage for many richer families. None of these are stories of the victory of progress. They are stories of sadness, lack, and personal tragedy: a couple that is kept in relationship limbo by pernicious EITC eligibility rules, families lacking resources to stick it out together in hard spells of life, and the blight of crime and punishment ravaging lower-income communities. The decline of marriage in America is not a story of working class Americans being liberated from the shackles of an arcane institution, but of social and economic chaos among poorer communities in America depriving many of the poor of fair access to a widely-valued institution. Thus, when we discuss marriage and demographic decline, we are discussing a real problem, a real kind of deprivation, a real cost of decline.

Finally, we come to the topic where “demographic decline” usually starts: fertility. Across the decades, hundreds of surveys have collected responses from hundreds of thousands of Americans asking about their fertility preferences. At no time since 1955 have birth rates exceeded stated fertility desires, and those stated desires have been stable around 2.2 to 2.5 children per woman. These stated desires don’t change based on how a question is worded, and they are resilient to survey priming: in an unpublished survey experiment I find that strongly priming respondents to think about either parenting difficulties or crime and public disorder has no effect on any measure of stated preferences, a finding shared in common with the small number of other fertility preference priming studies

Furthermore, in longitudinal surveys, stated fertility preferences are extremely strong predictors of actual fertility behaviors, stronger than covariates like religion, education, race, income, or any other socioeconomic or cultural variable. No research has ever identified a stronger predictor of actual fertility behavior than stated intentions and preferences. Those statements are robust to survey method, and exogenous beneficial shocks to household finances and childrearing costs lead to increases in fertility, pointing to pent-up demand for kids. The evidence that stated preferences more-or-less reasonably proxy true underlying “real” preferences (whether revealed or not) is quite strong; sufficiently strong that a considerable body of research has begun to explore genetic roots of fertility dispositions.

Fertility preferences have not declined much. Whatever role economic and technological shocks may have had, they have not led people in most countries to report desiring fewer children.

Thus, when I say that fertility rates are below stated preferences, this is not a trivial statement: this is a real loss. Yes, people might still rationally choose to have fewer children than they desire due to cost or other reasons, but they pay a cost in welfare and happiness when they foreshorten their desires. And indeed, I find in surveys I have conducted of U.S. women that correspondence between stated fertility desires and actual fertility outcomes is related to happiness: women who hit their targets (neither more nor fewer children than they desired) are happier than other women. Moreover, whether or not a woman has been diagnosed with a mental illness is a key predictor of undershooting fertility preferences: anxious and depressive mental states impose a barrier between what a person believes would make them happiest in life and their ability to take action in that direction. The problem of low fertility is not primarily its effect on population growth. The problem of low fertility is that it represents real suffering, a toll of infertility and miscarriage, delay and expense, and treatment after treatment always wondering, will I live to meet my grandchildren?

Why fertility is falling is a matter of great debate. But the fact that fertility has fallen in the last 20 years synchronously across all the industrialized countries regardless of their cultural trends, militates against attributing the change to attitudinal factors. It probably isn’t the case that all the rich countries of the world suddenly in tandem adopted different dispositions towards childbearing. It’s much more likely that they were all exposed to common economic shocks: a financial crisis which devastated the economic trajectory of a generation, for example. The U.S. Survey of Consumer Finances shows that young peoples’ households experienced no recovery whatsoever in their net worth between 2009 and 2019, despite employment rebounding. Likewise, common technological shocks, like the expansion of social media, might have a role to play, especially if theories linking social media and increased “screen time” to rising anxiety, depression, and unhappiness are true, since these kinds of negative affectual states cause lower fertility. 

But it should be noted, stated fertility preferences have not declined much. Whatever role economic and technological shocks may have had, they have not led people in most countries to report desiring fewer children. Actual fertility has fallen even as desired fertility has not in most of the high-income countries of the world. Thus, as with marriage, the likeliest story on falling fertility in the last two decades is not one of people simply freely choosing not to have so many children. Rather, fertility has most plausibly fallen because of economic “failure to launch” among young people, long delays in career stability, excessive housing costs, exploding childcare costs, rising student debts, and other adverse circumstances, not least the oppressive panopticon of social media which makes prisoners of us all. 

Low fertility, like high mortality or rising singleness, will have consequences. Many will be adverse, though some may not. It is not necessary (and possibly even counterproductive) to advance a consequentialist argument about demographic decline when the phenomena in question are in some sense intrinsically bad; that intrinsic badness should be the focus of argument for those of us worried about worsening demographic conditions. It is bad that people aren’t succeeding in living healthy and long lives as much as in the past. It is bad that people aren’t getting married as quickly as they would prefer. It is bad that fertility rates are so far below what people say they want. Consequentialist arguments about what the outcomes may be in 10 or 20 or 200 years are interesting, and for policymakers should certainly be weighed, but only after we have understood the more basic truth that the constituent parts of demographic decline are bad in and of themselves. Fertility rates far below (or far above!) desires are prima facie evidence of something rather unfortunate in a society, regardless of what the impact on economic growth may be.

What, then, is to be done? I will not offer a specific agenda, but, following the perspective here outlined, will suggest a couple of guideposts. First, any coherent demographic agenda has got to think about more than just fertility. Confronting demographic decline means dealing with drug and alcohol abuse, because drug and alcohol abuse contributes to criminality, to unemployment, to non-marriageability, to lost years of health, and ultimately to premature death. Experts in criminology will debate how to handle this issue, but note that the objective here is to reduce usage of these substances, not simply to reduce criminal usage. Legal usage of addictive substances (like prescription opioids) can as easily lead to social dysfunction, ruined lives, and ultimately to criminality as illegal usage. 

Debating whether one substance or another should be regulated, legal, or scheduled is not relevant to my concern here; my concern is with how we can get fewer people to come home drunk and beat their spouse (or kill a fellow driver on the road). My concern is with how we can get fewer people to kill themselves, but just as much how we can get fewer people to want to kill themselves. Tackling demographic decline requires us to consider the suffering entailed in “deaths of despair,” a top-tier national policy problem demanding a relatively dramatic response. Raising alcohol taxes to their 1950s levels (which would at least double the total price of alcohol) would be a helpful first step. For illegal drugs, it’s more difficult to know what to do, but that topic should be a centerpiece of our political debates.

For marriage, we must think about the life course followed by young people today. Marriage is delayed partly because young people don’t want to marry too young; there’s no plausible way marriage age for women in America is going to fall below age 25. (Today it is around 29 for women; it was 25 in the mid-2000s.) But helping young people achieve stability, hit the milestones of adulthood, and feel ready for marriage a few years earlier is eminently achievable. Universities could be encouraged to repackage 4-year degrees into 3 years, for example, or make the 4th year include graduate courses for interested students. It’s also entirely reasonable to demand that U.S. state and federal policies adopt a strict nondiscrimination principle for marriage. Policies should be designed to keep marriage penalties to a minimum: getting married should not lead a couple to pay extra taxes, or lose benefits on which they depend. 

And finally, there is fertility. Supporting marriage and tackling serious health threats would already help to boost fertility, but some additional support is likely necessary. Child allowances and family leave are the standard recipe for pronatalism, and they do tend to boost fertility. But they are limited in total effect and come at a considerable cost. Other policy approaches are needed too: housing costs can be mitigated through liberalized zoning policies, for example, which would have a considerable impact on fertility, since housing costs are a key element of the cost of raising children. School voucher programs may also help some families. In a pluralist society like the United States, one can hardly expect the U.S. government to adopt full-throated pronatalism, such as we see in Hungary, for example. But there are plenty of non-demographic reasons to support child allowances, family leave, zoning reform, or school vouchers; those of us concerned about demographic decline can contribute additional arguments and supports to these movements without reinventing the wheel.

Demographic decline as a description of inhuman aggregates is well-trodden territory, and yet on some level fruitless to discuss. The long-term, often diffuse consequences of demography are endlessly debatable, a skein we cannot readily unwind. Moreover, the debate is unnecessary: none of the forces driving demographic decline are neutral. Prime-age mortality is rising, marriage is getting later, and fertility is low mostly because of bad things happening to people. It would be a good idea for policymakers to try to address those bad things. If that happens to yield long-term social benefits, all the better, but such benefits are ancillary to the basic question of human flourishing today: will we help the addicted, the lonely, and the infertile, or will we hang them out to dry?