Why Are We All Cowards?

The Rising Premium of Life, Or: How We Learned to Start Worrying and Fear Everything

How much is your life worth? Not just in the abstract: I mean literally, what dollar value would you assign?

If you're American, the Environmental Protection Agency (EPA) says $11.4 million¹. That's the ratio they use when deciding whether it’s worth enacting regulations that would in aggregate save one life. In 1980, you were worth $3.3 million (inflation-adjusted). Your great-grandfather in 1940? Maybe $500,000.

But here's the strange part: we're not getting 20x more life than our great-grandparents. We get maybe 10-20 extra years. So why has the price tag on those years exploded? And why is this happening everywhere? From Thailand to Switzerland, the value we place on staying alive is skyrocketing far faster than wealth alone can explain.

Nobody talks about it, but over the course of the last century, something has dramatically changed in how our species thinks about life and death.

In this article, I’ll argue that “our premium for life” – that is, our willingness to keep living, and our willingness to pay to avoid incremental chances of dying – has gone up over time. I trace three main factors that might result in this change: wealth effects, safety-risk feedback loops, and secularization, and argue that this shift has had profound implications for how we organize society.

Intro

“YOLO, you only live once (Ooh)

The battle cry of a generation

This life is a precious gift

So don't get too crazy, it's not worth the risk” YOLO by the Lonely Island

We live in the safest era in human history, yet we're more terrified of death than ever before.

The numbers tell a stark story. Americans now value a statistical life at $11.4 million, likely over an order of magnitude higher than a century ago in real terms. We shut down the global economy to prevent COVID deaths, sacrificing tens of trillions in wealth, as well as substantial social costs. Parents who once roamed miles as children now won't let their kids leave the front yard. Young people take fewer risks than any generation in history. Less sex, less drinking, less driving, less... everything.

This isn't just about having more money to spend on safety, though that's part of it. Something deeper has shifted in how our species relates to mortality. Each generation raised in greater safety than the last develops a lower tolerance for risk. Secular worldviews make this life feel like all we have. Smaller families concentrate our emotional investments. Safety culture in organizations creates competitive arms races toward zero risk.

The result? A civilization increasingly organized around the principle that death is the ultimate evil, to be avoided at any cost. But there's a potential problem: our preference for life can come at the sacrifice of other social values. The same impulses that make us safer might also make us stagnant.

This essay traces how we got here, why it's accelerating, and what it means for humanity's future. Because if we're going to spend the next century eliminating every last source of mortality risk, we'd better understand what we're trading away in the process.

Pro-life evidence

Value of a (Statistical) Life

(Source)

Suppose that you’re a government official, trying to figure out whether to enact regulations that can potentially save lives, but have a cost in economic activity. How would you balance the two? Slogans like “we must be willing to save lives at any cost” or “the government should just step out of the way” or “how could you possibly balance incommensurate moral goods?” don’t cut it; in the real world, difficult decisions must be made. Inaction, too, has a cost.

While different governments around the world make these choices differently, the US government makes this decision in a surprisingly egalitarian and decentralized way: economists study how individuals value and prioritize safety measures, and aggregate them, and then many government agencies take those inferred numbers as a given to guide collective decisions.

In particular, the Value of Statistical Life (VSL) measures how much people are willing to pay for small reductions in mortality risk. Economists calculate it by observing real-world trade-offs, like how much extra salary workers demand for riskier jobs, or how much extra people pay for safer cars. For example, if people will pay $100 for a safety feature that reduces death risk by 1 in 100,000, that implies they value a statistical life at $10 million ($100 × 100,000). Importantly, this isn't the value of saving a specific person's life: it's what we collectively reveal about mortality risk through thousands of small decisions. Government agencies use these estimates to decide whether safety regulations are worth their cost: if a regulation costs $50 million but saves 10 statistical lives valued at $11.4 million each, it passes the cost-benefit test.

Let’s look a graph of VSL² over time:

As you can see, the inferred VSL (accounting for inflation) has gone up dramatically over time. Further, the difference can’t just be accounted for by increasing wealth, as you see the same pattern after adjusting for real income:

Note that VSL is evidence for an increased premium of life in two different ways: evidentially: as this is how (some agencies in) our government believes individuals value risk reduction, and causally: as this is how our government makes life-and-death decisions on how to set regulations.

I don’t want to overstate the case: The evidence from VSL, while suggestive and strong, is frustratingly limited and by itself insufficient. Most of these government figures are made based on a relatively small number of studies, different methodologies would result in different figures, and they have not conducted recent studies. That said, I speculate more recent studies would have shown an acceleration in VSL over time for private decisions, and the current official figures are slightly too low.

Fortunately, we don’t need to speculate much further, as we have other lines of evidence for the increased premium of life thesis:

Healthcare Spending

Source: https://www.kff.org/health-costs/issue-brief/snapshots-health-care-spending-in-the-united-states-selected-oecd-countries/

In this survey of OECD (developed) countries, we see a continuous, large, and monotonic increase of healthcare spending over time across all countries. As with VSL, this growth is clearly larger than just changes in GDP or average income.

People often decry increases in healthcare spending. Common reasons cited for increases in healthcare spending include healthcare-specific reasons like an aging population, technological advancements and expensive R&D for new drugs, and administrative complexity and regulatory bloat. They also include reasons that are specific to a specific country’s healthcare: for example, the US’s American Medical Association artificially limits the supply of doctors and thus increases wage premiums for doctors.

However, I suspect healthcare-specific and country-specific explanations are only part of the explanation, and miss the forest for the trees. Instead, much of the increase in healthcare costs across countries can be readily explained by a broad, cross-cultural, secular change in people’s increased premium of life.

Evidence from developing countries: Mixed but indicative

Is this just a rich-world problem for very wealthy and perhaps WEIRD countries? After consulting the literature, my tentative answer: Probably not.

There are some tentative stylized facts I gained while researching for this article:

• Income elasticity for life is broadly above 1. That is, people on average are willing to spend proportionally greater fractions of their income on buying life-years as their incomes go up.

• Income elasticity for life is probably higher for poorer countries than rich countries

  • In other words, going from $2500/year to $5000/year will see proportionally greater changes in willingness-to-pay for health than going from $25000/year to $50000/year

  • For example, in this study of willingness to pay to increase road safety, the authors model elasticity as ~2.5 for low-middle and middle-income countries (though after reading the study, I find their methodology quite weak)

• Western academics are often surprised by the value people in developing countries place on reducing risks, relative to income.

  • For example, in a Thailand study, an estimate of VSL of poor villagers was about $470,000 ($250K in 2000 dollars), low by Western standards but high compared to what past foregone earnings models would suggest (at the time Thai GDP per capita was $2000 in today’s dollars).

  • This is weakly indicative that VSL has gone up over time, relative to Western countries at the same development level

That said, however, the evidence is sufficiently limited and mixed that it’s hard to draw any firm conclusions. For example, below³ is a plot of all the different developing-world studies on VSL, normalized against average incomes:

Source: https://forum.effectivealtruism.org/posts/uTSxcDzPLiifryqH6/technical-updates-to-our-global-health-and-wellbeing-cause#Aggregating_these_considerations

As you can see, the results are all over the place. One can read the tea leaves however they want, and honestly the data is broadly consistent with your story regardless.

Covid Lockdowns

March 2020 gave us an especially dramatic worldwide natural experiment in life valuation. Faced with a disease that threatened primarily the elderly and infirm, the world chose economic shutdown over mortality. Schools, businesses, social events, and much of civil society chose to shut down, voluntarily and otherwise. The revealed preference was staggering:

• Estimates of economic costs range from ~20 trillion on the low end, to ~80 trillion on the high end

• While separating out causality is hard, in my estimation more than half the economic costs (excluding deaths) likely comes from our responses to the pandemic (including individual behavior and government actions), rather than the disease itself.

• Lives potentially saved: ~10-50 million (estimates vary wildly)

• Implied economic cost per life saved: ~$400,000 to ~$8 million

While much of the debate for the “covid lockdowns” comes from arguments for individual freedom, autonomy and voluntary economic activity on one end, and collective benefits of health, lives saved and collective action on the other, I think the framing of individual behavior vs government mandates is largely misguided.

At least in developed countries, traffic camera data and smartphone location data broadly shows most of the social distancing and isolation measures came 0.5-2 weeks before official lockdowns were announced or came into effect. The interplay between civil society and government mandates were mostly about the tail wagging the dog, rather than the other way around.

Further, public support for lockdowns remained high even as costs mounted. Polls showed 70%+ approval for restrictions in most developed countries through 2020, with the stricter countries broadly more positive about their governments’ responses. We knew the economic costs and chose life preservation anyway.

Compare this to historical pandemics. During the 1918 flu (which had an unvaccinated infection-fatality rate ~5-8 times that of COVID, and was unusually lethal among adolescents and young men), quarantines and business closures were narrowly targeted and limited in scope, often restricted to schools⁴ and heavily crowded businesses like movie theaters and dance halls. The longest among them lasted about 15 weeks. St. Louis's aggressive closures, which included smaller businesses and churches, were considered radical overreach. A century later, Sweden's relatively light touch was the outlier.

Our response to the COVID-19 pandemic was also instructive as it was not just a tradeoff between economic values and years of life. Instead, it pitted our values of life against other sacred values – our children’s education, our maintenance and growth of social relationships, our freedom of movement, and our desires for individual rights, privacy, and autonomy. With a few exceptions, life won.

Individual Behaviors

On X, I recently saw this map go viral:

Source: https://www.npr.org/sections/krulwich/2012/10/01/162079442/do-you-know-where-your-children-are-is-that-always-a-good-thing

In it, we see the ranges of 4 generations of 8-year olds (from Ed to his great-grandfather George in 1919). As NPR puts it:

The Thomas family has been living in Sheffield, a town toward the north of England, for at least four generations. When great-grandpa George Thomas turned 8 in 1919, he was allowed to walk six miles — by himself — to go fishing. But each generation after has been given less and less room to roam.

In 1950, when Jack, the grandfather, turned 8, he was allowed to go just a mile on his own to visit the woods.

In 1979, when Vicky, the mom, turned 8, she was allowed to ride her bike around the immediate neighborhood, walk by herself to school, and could visit a swimming pool on her own. Her zone of play was a half-mile wide.

And then we have the current generation, Ed.

[...]In 2007, Vicky said her son, then 8, was "driven the few minutes to school, is taken by car to a safe place to ride his bike and can roam no more than 300 yards from home." Basically, he stays on the block.

In fact, she says, he prefers the family yard to the street outside. "He doesn't tend to go out because the other children don't," she said.

That was 2007. Now, 18 years later, post-pandemic and with ubiquitous smartphones and indoor entertainment, I’m not sure how many 8-year olds are even allowed to venture outside their own yards.

This is hardly specific to parenting. Across nearly every domain, individual risk-taking is converging toward "minimal acceptable."

Transportation: Motorcycle ridership in the US peaked in the 1970s and has declined 50% since. Teen driving licenses have plummeted: in 1983, 46% of 16-year-olds had licenses; today it's 25%. Not because cars are less available, but because both teens and parents see teen driving as unnecessarily risky.

Recreation: Extreme sports participation has plateaued despite better safety equipment and practices. Meanwhile, safer activities like yoga have exploded (up 500% since 2000).

Social risks: Young people have less sex, drink less, fight less, and commit fewer crimes than any generation in recorded history. The "wild youth" stereotype has been completely inverted: today's young are radically risk-averse compared to their parents.

Across the board, we see a wholesale shift in what constitutes acceptable risk. Activities that were considered normal parts of growing up: unsupervised play, teenage employment in manual labor, high-school boxing, and hitchhiking have now been reclassified as almost unconscionably dangerous.

The Overarching Pattern

There are multiple disparate strands of evidence that the premium of life has clearly gone up. While any given line of evidence can be disputed or explained away through some other mechanism, the overarching pattern is clear: we care more about living longer than we ever have in the past, and take heroic efforts to avoid death.

Why caused this change? And what does it mean for the future? We turn to our next section:

Why might people value life more?

Money

An obvious explanation here is money: We’re richer, so we can afford to care more about living. Hall and Jones (2007)⁵ presents a simple formal model of this. In essence, more money means more capacity to buy life. Further, once basic needs are met, your other options for spending (compared to healthcare) are more limited. Thus, if your other options for spending money on material goods do not “spark joy”, healthcare is a better option in contrast. In economics jargon, healthcare is a “superior good.”

For more, see Hall and Jones (2007).

On Money and Happiness

Finally, being richer might mean your life is better, and thus more worth protecting. Up until now, I have elided the question of whether richer people are actually happier. My thesis does not necessarily rely upon this assumption (People might fight harder to preserve life for reasons other than greater subjective well-being). However, I think the evidence is pretty overwhelming:

Source: https://www.reddit.com/r/dataisbeautiful/comments/mczgk9/does_a_rich_country_mean_a_happy_country/

When self-reported happiness from World Happiness Surveys is plotted against log (income), we see a clear and positive trend. A fairly parsimonious explanation is that richer people are in fact happier, and this gives them prima facie greater reason to want to live longer.

That said, I believe money alone is not a sufficient explanation for our greater value of life. In particular, it does not explain why people are increasingly willing to not just pay more money to extend their lives, but also trade non-monetary goods to live longer, as well as take less risks with their lives across the board.

Safety-Risk Aversion Feedback Loops

“A bat is just a rat with wings” – The Riddler, apocryphally

Greater risk aversion isn’t just a cause of having a safer environment, but also partially a result. To understand why, we naturally turn to a closely-related question: Why do bats live longer than mice?

Source :https://apomorphic.com/2019/09/09/why-we-age-1-intro

Bats and mice are evolutionarily and morphologically quite similar. The main difference, of course, is that bats could fly. And that, arguably, made all the difference.

See, one of the more interesting stylized facts in animal surveys on aging is that an animal’s intrinsic mortality rate – how likely it is to die without external pressures like starvation and predation – is closely linked to its extrinsic mortality rate. An animal’s lifespan in captivity is closely linked with its prevalence of predation in the wild.

Source: https://apomorphic.com/2020/01/12/why-we-age-2-nonadaptive

In addition to bats, birds on average also live much longer than most mammals. Fliers in general live longer than landlocked animals, since they can escape more easily, controlling for body weight (Being bigger of course also makes it harder to eat you).

In addition to flight and bodyweight, you are also more likely to have a longer maximum lifespan if you’re tree-borne, burrowing, armored, spiky, poisonous, or social – all factors that reduce your extrinsic mortality rate.

Why? There are many theories for this, but according to Bradshaw’s blogpost above (and I agree), the most plausible comes from natural selection pressures: mutations that decrease fitness later in life are more punishing if you have a longer natural lifespan (lower extrinsic mortality rate). In contrast, if predators have a 99% chance of eating you before the ripe old age of 2, any mutation that reduces fitness by the age of 10 is ~ irrelevant.

For more, see Will Bradshaw’s sequence on the theoretical biology of aging: Why We Age, Part 1; Evolution is Sampling Error; Why We Age, Part 2: Non-adaptive theories.

Why is this relevant to the human value of life? There are two different angles:

The Rational Actor Model

As baseline mortality drops, the expected value of safety investments increases. If your baseline rate of death is 10%/year, then your expected lifespan is 10 years, so getting rid of something that has a 10% chance of killing you buys you one year of life (on average). In contrast, if your baseline rate of death is 5%/year, then getting rid of something that has a 10% chance of killing you buys you two years of life.

The longer your baseline mortality, the more valuable each additional improvement in safety might be.

The Psychological Angle

Rational calculations aside, it just seems really intuitive to me that safer environments mean you care more about safety. If your children have a 30% chance of dying from smallpox before the age of five you probably aren’t going to be spending that much effort reducing their chances of drowning from 0.05% to 0.02% in the local pool – you just have much bigger problems to worry about. But when you have a child mortality rate like Norway’s (0.19%), drownings might suddenly become the biggest problem in your child’s life, and fretting parents might naturally worry more about these small (in absolute terms, on a historical scale) risks.

There are also conditioning considerations. When organisms evolve in high-mortality environments, they develop psychological patterns suited to that reality: risk-taking, fast reproduction, lower long-term investment. But what happens when the environment suddenly becomes safe?

One hypothesis: each generation raised in greater safety than the last becomes psychologically calibrated to that new baseline. They don't just expect safety; they need it at a deep emotional level. What was paranoid overprotection to one generation becomes baseline common sense to the next.

To be clear, this is just one hypothesis among several, and I’m still confused about the details. Because this happened on such a quick timescale that there can’t be a clear selection mechanism like in animal evolution, I do find these psychological hypotheses to be somewhat tentative and confused. Still, intuitively it really feels like there’s something there.

(One piece of potential evidence for this hypothesis would be longitudinal data on developing countries⁶)

The Secularization Hypothesis

While discussing these models with the language model Claude Opus IV, they suggested another intriguing possibility: Secularization. In Claude’s own words,

This might be the most profound driver. For most of human history, death wasn't the end—it was a transition. Whether you believed in heaven, reincarnation, or joining your ancestors, mortality had an escape clause.

But as traditional religious belief declines, this life becomes all there is. The stakes of mortality go from high to infinite. Losing 30 years of life when you expect eternal paradise afterward is tragic. Losing 30 years when those years are all you'll ever have? That's existentially catastrophic.

The timing fits suspiciously well. The acceleration in VSL starts in the 1960s-70s, exactly when secularization took off in the developed world. Countries with the most dramatic religious decline (Scandinavia, UK) often show the most extreme safety cultures. The US, with higher religiosity, has been a relative holdout—though we're catching up fast.

This also explains the super-elasticity in developing countries. Modernization doesn't just bring wealth; it brings secular worldviews. A Thai farmer who starts earning more money and stops believing in reincarnation will value life preservation far more than income alone would predict.

I find this argument intriguing but somewhat shaky. In particular, I’m unsure how well it aligns with data from China and other broadly always-secular countries.

Smaller Families and Greater Per-Child Investment

As family sizes have gotten smaller and the world becomes increasingly specialized, each individual child represents greater familial and state investment in their education, health, and well-being. This in turn means that it becomes costlier for them to die, and their life matters more.

That said, this theory is incomplete: one thing I’m confused about is how does this explain the potential delta in selfish preferences? It’s intuitive to me when families and states care more about preserving their children, but how does this affect the view from the inside? How does this explain why children (and adults) want to live more than they ever did before, rather than rebel (as they always did) against parents and the government in less-than-safe ways?

More research and thinking needed. Comment if you have thoughts!

Alternative Theories, Countervailing Evidence, and Implications

I’ve presented a potentially very important trend, many lines of (what I think is) good evidence for that trend, and (what I think is) some good reasons for why that trend might be occurring. But is the trend real? What are some countervailing strands of evidence that might contradict this story? And what, if true, are the most important implications of this trend?

In a future post, we’ll discuss those questions, as well as topics for future research. Stay tuned!

In the meantime, I’ll encourage readers to consider these questions for themselves, and/or discuss these questions with your friends and/or comment with your objections here. You only have one life to live, so why not spend it on manufacturing internet beefs on Substack?

1. EPA currently uses $7.4 million in 2006 dollars, which this handy inflation calculator tells me is about $11.4 million in 2025 dollars.

2. Graph made with Claude. Sources include Thaler & Rosen(1976), Viscusi and Aldi (2003), and various studies by Viscusi. I spot-checked the sources and calculations but cannot guarantee a complete lack of hallucinations

3. H/T Isabel Juniewicz for the article

4. Note that school were a more natural target for closures in 1918 than 2020-2021, since Spanish flu affected adolescents much more than COVID did

5. H/T Isabel Juniewicz for the article

6. Developing countries lifespans have increased a lot in the last 50-100 years. In general, catch-up adoption of public-health related technologies like vaccines, sanitation, and antibiotics have been relatively quick, whereas economic development have been relatively slow. This means that (e.g.) countries with GDP/capita closer to the US in 1930 have had lifespans closer to US in 1970. If my safety-feedback loop story is broadly correct, you’d expect developing countries’ willingness to pay for life to be higher than the US and other developed countries’ past willingness-to-pay at the same past income level.