The invisible market that prices human life

In February 2024, the EPA tightened the annual standard for fine particulate matter from 12 to 9 micrograms per cubic meter. The agency’s economists calculated the rule would prevent up to 4,500 premature deaths per year. They valued those deaths. They did arithmetic. The net health benefits came to $46 billion — roughly $77 returned for every dollar spent — and the rule went forward.

By January 2026, the same agency announced it would no longer calculate the health benefits of air pollution regulations at all. The value of the lives that cleaner air would save was set to zero.

Between those two moments — $46 billion and zero — is the entire architecture of how modern governments decide who is worth protecting.

The price on your head

“You can’t put a price on human life.” People say it without noticing what they’re actually proposing: that the decision not to install airbags, not to remediate a toxic site, not to limit particulate matter emissions, should be made on something other than evidence. Modern regulators have a different view, and the bureaucratic apparatus they built to implement it is one of the stranger intellectual achievements of the twentieth century.

The value of a statistical life is not a claim about what your life is worth. Thomas Schelling, who introduced the concept in his 1968 essay “The Life You Save May Be Your Own,” was at pains to distinguish two categories of deaths: identified and statistical. The identified life — the child trapped in a well, the miner caught underground — commands near-unlimited resources. We watch. We donate. We feel the cost as personal. The statistical life is the anonymous person who will not die from particulate exposure next year because an air quality rule passed. That person has no face, no family pleading on television. They are a probability, distributed across a population. And they can be valued — not as a person, but as a risk reduction.

The dominant methodology is empirical. W. Kip Viscusi of Vanderbilt University has spent decades refining the hedonic wage approach, producing more than 60 published studies on wage-risk tradeoffs. The method reads labor markets directly: workers in high-fatality occupations — logging, roofing, mining — accept wage premiums in exchange for elevated mortality risk. If a worker requires $1,400 more per year for a job carrying a 1-in-10,000 higher annual fatality risk, the implied VSL is $14 million. Scale that preference across 10,000 workers and you have one expected additional death per year, collectively accepted for $14 million in total compensation. Viscusi’s meta-analyses of US wage data cluster around a median of $7 to $8 million, with estimates ranging from $4 million to $9 million depending on methodology and dataset. The approach has a known limitation: it assumes workers understand and voluntarily accept occupational risk, which is questionable for low-wage, non-union workers who may have little real choice about where they work.

A second method asks people directly. Stated-preference surveys present respondents with a specified reduction in their annual mortality risk and ask how much they’d pay for it. If the mean willingness-to-pay for a 1-in-100,000 annual risk reduction is $140, the implied VSL is $14 million. This approach has its own problem: hypothetical bias, where stated preferences in surveys diverge from actual behavior when real money is involved. The two methods have different failure modes. They produce roughly similar numbers. That convergence is the main reason economists trust the framework.

Current US government figures: the Department of Transportation uses $14.2 million per statistical life (2025 base year). The Department of Health and Human Services uses $13.6 million as its central estimate. The EPA’s guidance, anchored at $7.4 million in 2006 dollars and updated for inflation and income growth, yields approximately $12 million for current analyses. Three agencies, one country, figures spread across a range of several million dollars for the same population. The spread isn’t noise — it reflects genuine methodological disagreement about which labor market studies to use, what income adjustments to apply, and how to handle the tails of the distribution. The empiricism is real. So is the judgment embedded inside it.

The Ford Pinto wasn't what you think

In 1973, Ford's engineers produced a cost-benefit memo titled "Fatalities Associated with Crash-Induced Fuel Leakage and Fires." The Grush/Saunby report calculated that modifying fuel systems across all US cars and light trucks would cost $137 million. The benefits — 180 burn deaths valued at $200,000 each, 180 serious burn injuries at $67,000 each, 2,100 burned vehicles at $700 each — totaled $49.5 million. Mark Dowie's 1977 Mother Jones exposé "Pinto Madness" turned this into the definitive story of corporate murder-by-spreadsheet: Ford had decided that $11 per car wasn't worth human lives.
The story that circulated omitted several things. The memo never mentioned the Pinto. It addressed rollovers across all vehicles, not the rear-end fuel tank fires that actually killed Pinto occupants. And the $200,000 per-life figure was NHTSA's number — derived from the government's own methodology, not Ford's accounting. The public instinct — that something was wrong with valuing a human death at $200,000 — was correct. The regulatory response was not to stop pricing life. It was to price it at roughly sixty times higher.

The machine that runs on death

A VSL is just a number until it enters a spreadsheet. What gives it power is the legal architecture built around it.

Ronald Reagan’s Executive Order 12291, signed in 1981, required all major federal regulations to pass a cost-benefit test — the first time this had been mandated across the executive branch. Clinton’s Executive Order 12866 in 1993 refined the framework and established the Office of Information and Regulatory Affairs as its permanent enforcer, but kept the cost-benefit requirement. Every significant health and safety regulation since has moved through this apparatus, and the benefits side of the ledger is dominated, overwhelmingly, by the dollar value assigned to prevented deaths.

The defining case is the Clean Air Act. The EPA’s 1997 retrospective study of the Act’s impacts from 1970 to 1990 found cumulative compliance costs of $523 billion in 1990 dollars, against estimated total benefits with a mean of about $22 trillion — a ratio of roughly 42 to 1. The Second Prospective Study extended the analysis through 2020: by that year, the Act was preventing over 230,000 premature deaths annually, with benefits exceeding costs by more than 30 to 1. Mortality risk reduction accounts for about 85 percent of those monetized benefits. Remove VSL from the calculation, and the Clean Air Act cannot be justified in cost-benefit terms. The regulation that has arguably saved more American lives than any other piece of domestic legislation survives its cost-benefit test only because VSL exists and is large enough to make the math work.

NHTSA’s airbag requirements, the FDA’s drug safety approvals, OSHA’s workplace exposure limits — the same VSL logic runs through all of them. VSL is not one input among several. It is typically the dominant one, the figure that determines whether a regulation is economically viable at all.

The 1982 OSHA case established the template. The Office of Management and Budget had rejected the agency’s proposed hazard communication regulation — the rule requiring employers to disclose chemical hazards to workers — as too costly. The rejection used the “human capital” approach to valuing mortality risk: foregone lifetime earnings, which generated a figure roughly one-tenth the size of VSL-based estimates. Viscusi’s wage-risk research provided the counterargument. Shifting from human capital to VSL boosted the regulation’s calculated benefits by a factor of ten, making compliance costs look trivial by comparison. The regulation was approved. VSL entered the regulatory machine not to restrain safety rules but to justify them. The irony embedded in that origin has never fully resolved: the tool that makes protection possible also sets its ceiling. No regulation whose cost per statistical life saved exceeds the VSL figure will clear the bar.

The exchange rate

The methodology travels well. The problem is where it arrives.

The standard technique for applying VSL across borders is benefit transfer, driven by a single equation: VSL(target) = VSL(base) × (Income_target / Income_base)^elasticity. The income elasticity parameter captures how steeply VSL scales with income. Within a single rich country, Viscusi’s meta-analyses produce elasticities around 0.5 to 0.6. The OECD recommends 0.8 for transfers between high-income countries and 1.0 to 1.5 for transfers to lower-income ones.

Run the formula for Bangladesh. US GDP per capita: approximately $85,000. Bangladesh: approximately $2,800. Using a US base VSL of $13 million and an income elasticity of 1.0: Bangladesh VSL ≈ $13 million × ($2,800 / $85,000) ≈ $428,000. One-thirtieth of the American figure.

Not because Bangladeshis value their lives less. The formula says nothing about that. It measures income and records it faithfully wherever income leads — across borders, and within them. India’s VSL has been estimated at $0.5 to $1.6 million using benefit-transfer and life-insurance proxy methods, per Guttikunda and Dammalapati’s 2024 analysis. Within Mexico, VSL ranges from $400,000 in Chiapas to $3.3 million in Mexico City — the within-country spread nearly matching the between-country one, because the formula follows income wherever it goes.

The regulatory consequence is arithmetically plain. A safety measure costing $5 million per death prevented clears the cost-benefit test in the United States by a factor of nearly three. The same measure, the same hazard, the same physiology — only the income-derived VSL differs — fails in Bangladesh by a factor of twelve. This isn’t the same regulation calibrated for local conditions. It’s the same arithmetic producing different verdicts about whether the same people are worth protecting.

April 24, 2013: the Rana Plaza garment factory complex in Dhaka collapsed. The death toll reached 1,134. Building inspectors had identified structural cracks the day before. Factory owners sent the workers back in. The Accord on Fire and Building Safety, negotiated afterward with international clothing brands, committed to safety upgrades across approximately 1,600 factories at roughly $500,000 per factory — around $800 million total. In a country where the implied VSL runs below $500,000, the cost-benefit arithmetic for mandated building safety inspection was always going to return a different answer than it would in Germany, where VSL derived from German wages makes the same math come out the other way. This isn’t a failure of the formula. It’s the formula working exactly as designed.

The strongest defense of locally calibrated VSL comes from Lisa Robinson, James Hammitt, and Lucy O’Keeffe, whose work in the Journal of Benefit-Cost Analysis argues that using income-adjusted figures produces better policy outcomes in lower-income countries than imposing a rich-country baseline would. An American VSL applied in Bangladesh would set the threshold so high that only the most expensive interventions would pass, crowding out cheaper life-saving ones that could easily meet a locally calibrated bar. The argument is technically coherent.

What it describes, stated plainly: a global regulatory architecture structured so that the less wealthy a population, the less its government is expected to spend to keep it alive. The methodology calls this revealed preference. The other description: a structure that converts poverty into a lower claim on protection.

The OECD’s 2012 report on mortality risk valuation found fewer than 30 studies for the 172 countries classified as low- or middle-income. Its 2025 update expanded the database to more than 4,000 estimates from 277 studies covering 49 countries. The map of VSL is a map of which lives were measured and which were inferred from someone else’s wages.

The senior death discount

In 2003, the EPA used an age-adjusted VSL for its Clear Skies Initiative, valuing the lives of Americans over 65 at 37 percent less than those of working-age adults — reducing estimated benefits by billions of dollars. AARP organized protests. Senators demanded hearings. The EPA abandoned age-adjusted VSL within months and has not returned to it.
Americans found it politically intolerable that a 70-year-old's statistical life was valued at a third less than a 40-year-old's, and generated enough political pressure to reverse the decision. The cross-country differential — a Bangladeshi garment worker's life at one-thirtieth of an American office worker's — has produced no equivalent political response. That disparity is its own kind of evidence about which deaths register as a political problem and which are accepted as an arithmetic outcome.

The number and its absence

The natural response to all of this is to reject the framework. That response is historically testable.

Before Reagan’s 1981 executive order mandated cost-benefit analysis for major regulations, safety decisions were made through political negotiation, industry pressure, and crisis response. Regulations appeared after catastrophes visible enough to generate public outrage — mine collapses, chemical spills, the Cuyahoga River burning in 1969. The chronic, statistical deaths from air pollution, from low-level carcinogen exposure, from occupational hazards that killed slowly over decades — these produced no equivalent political event, because they produced no equivalent image. No single death, no single face, no moment of concentrated horror that demanded a named response. Cass Sunstein, who ran the Office of Information and Regulatory Affairs under Obama and has developed the argument across multiple works including The Cost-Benefit Revolution (MIT Press, 2018), describes this as availability bias: humans respond to vivid, concentrated catastrophes and ignore diffuse, probabilistic ones regardless of which kills more people. VSL and cost-benefit analysis exist, in part, to correct for that bias — to force regulatory attention onto chronic, invisible, statistical deaths that the political system would otherwise ignore in favor of dramatic ones. Without the number, government spending on mortality risk reduction is allocated by media coverage and political salience, not by lives saved per dollar.

Frank Ackerman and Lisa Heinzerling’s Priceless: On Knowing the Price of Everything and the Value of Nothing (The New Press, 2004) attacks the framework from the opposite direction. Cost-benefit analysis doesn’t eliminate distortion — it substitutes its own. It privileges outcomes that can be monetized over values that resist quantification: equity, dignity, the fair distribution of risk across people who didn’t choose it, obligations to people not yet born. And the willingness-to-pay principle at VSL’s foundation systematically undervalues the interests of people with less money to express as preference. This isn’t a sentimental objection. It’s a structural one. Ackerman and Heinzerling were describing a domestic distortion. The arithmetic in the previous section is what that distortion looks like scaled to a global income distribution.

Both arguments are, stubbornly, correct. The pre-VSL regulatory world was worse — not because it lacked a principle, but because the absence of a systematic method meant that only the politically visible deaths triggered response. And the VSL-based world is structurally biased in the direction of wealth. Both statements are simultaneously accurate, and no version of cost-benefit analysis resolves the tension between them.

Then January 2026.

The EPA announced it would no longer estimate the economic value of health benefits — including lives saved — from air pollution regulations. The change appeared first in the Economic Impact Analysis for Stationary Combustion Turbines Rule, published in January 2026. The agency’s stated rationale, quoted from the document: “to rectify this error, the EPA is no longer monetizing benefits from PM2.5 and ozone but will continue to quantify the emissions until the Agency is confident enough in the modeling to properly monetize those impacts.” Industry compliance costs continued to be tallied as before. Health benefits were not. Cass Sunstein, who built the cost-benefit apparatus inside OIRA, described the move as unprecedented.

Now work through the arithmetic. The Second Prospective Study found that roughly 85 percent of the Clean Air Act’s monetized benefits derive from mortality risk reduction. Strip that valuation out and the benefit column approaches zero. With VSL = 0, every air pollution regulation dependent on health benefits fails its cost-benefit test — not because anything changed about the regulation, the hazard, or the people being protected, but because one column of the spreadsheet was set to zero. This is not a refusal to price human life. It is a pricing decision. The most consequential variant of the same choice the original methodology was making all along.

The options are not “price life” or “don’t price life.” The options are which number to use and whether to acknowledge it.

The February 2024 rule and the January 2026 decision sit eighteen months apart in the same agency’s history. By now the reader knows what is inside those numbers. The $46 billion was built from wage premiums accepted by loggers and roofers, from survey responses about marginal risk reduction, from a methodology that records what people will pay for safety and nothing else. Applied in Bangladesh, the same formula returns $428,000. Applied by the EPA in January 2026, it returns zero.

VSL has never measured what its critics and its defenders both speak as if it measures — the worth of a human life. It measures how much a regulatory apparatus will spend, expressed as the mortality risk preferences of the population whose wages were used to calibrate it. That gap — between what the number actually measures and what everyone, on all sides, argues as if it measures — is where every serious disagreement about cost-benefit analysis gets stuck and stays.

What remains when the gap is named: the number is large in America, and more Americans are alive because of it. The same number, one-thirtieth as large in Dhaka, produced a different answer about building inspections. Set to zero in Washington, it dismantled the mechanism that had been authorizing protection for four decades — not because anyone decided the deaths no longer mattered, but because the column that counted them was emptied.

$14 million. $428,000. Zero. The same formula, operating correctly on populations with different incomes.

Gen AI Haftungsausschluss

Einige Inhalte dieser Seite wurden mit Hilfe einer Generativen KI erzeugt und/oder bearbeitet.

Medien

Main entrance of U.S. EPA Headquarters; the William Jefferson Clinton Federal Building on 12th Street, N.W., Washington, D.C. The pylon and elevator entrance for the Federal Triangle Metro station is seen on the left. – Wikipedia

Wichtige Quellen und Referenzen

Thomas Schelling, “The Life You Save May Be Your Own,” in S.B. Chase Jr., ed., Problems in Public Expenditure Analysis, Washington DC: Brookings Institution, 1968, pp. 127–176.

W. Kip Viscusi, “The Value of Life,” Harvard John M. Olin Center for Law, Economics, and Business Discussion Paper No. 517, 2005; and peer-reviewed publications on wage-risk methodology, Vanderbilt University Law School.

U.S. Department of Transportation, “Departmental Guidance on Valuation of a Statistical Life in Economic Analysis,” 2025. transportation.gov.

U.S. Department of Health and Human Services, “HHS Standard Values for Regulatory Analysis, 2025,” ASPE, aspe.hhs.gov, February 2025.

U.S. Environmental Protection Agency, “Mortality Risk Valuation,” epa.gov/environmental-economics/mortality-risk-valuation. Central estimate: $7.4 million (2006 dollars), updated to year of analysis.

Executive Order 12291, Federal Register, February 17, 1981.

Executive Order 12866, Federal Register, October 4, 1993.

U.S. Environmental Protection Agency, “The Benefits and Costs of the Clean Air Act, 1970 to 1990: Retrospective Study,” EPA 410-R-97-002, October 1997.

U.S. Environmental Protection Agency, “The Benefits and Costs of the Clean Air Act from 1990 to 2020: Second Prospective Study,” EPA-410-R-11-002, March 2011. epa.gov/clean-air-act-overview/benefits-and-costs-clean-air-act-1990-2020-second-prospective-study.

U.S. Environmental Protection Agency, “Final Regulatory Impact Analysis: Reconsideration of the National Ambient Air Quality Standards for Particulate Matter,” February 2024. epa.gov/system/files/documents/2024-02/naaqs_pm_reconsideration_ria_final.pdf.

Ford Motor Company, Grush and Saunby, “Fatalities Associated with Crash-Induced Fuel Leakage and Fires,” Environmental and Safety Engineering, 1973. Primary document: autosafety.org/wp-content/uploads/import/phpq3mJ7F_FordMemo.pdf. Reproduced in litigation records and widely documented in regulatory economics literature.

Mark Dowie, “Pinto Madness,” Mother Jones, September/October 1977.

Sarath Guttikunda and Sai Krishna Dammalapati, “Value of Statistical Life (VSL) to Support Cost-Benefit Analysis of India’s Air Quality Management,” SSRN, abstract_id=5250231, October 2024.

Luis Armando Becerra-Pérez, Roberto Alonso Ramos-Alvarez, Juan J. DelaCruz, and Benjamín García-Páez, “Value per Statistical Life at the Sub-National Level as a Tool for Assessing Public Health and Environmental Problems,” Inquiry: A Journal of Medical Care Organization, Provision and Financing, SAGE, 2024. PMC11032065.

Accord on Fire and Building Safety in Bangladesh, 2013.

Lisa Robinson, James Hammitt, and Lucy O’Keeffe, “Valuing Mortality Risk Reductions in Global Benefit-Cost Analysis,” Journal of Benefit-Cost Analysis, 2019. PMC7473065.

OECD, “Mortality Risk Valuation in Environment, Health and Transport Policies,” OECD Publishing, 2012. ISBN 9789264130807.

OECD, “Mortality Risk Valuation in Policy Assessment,” OECD Publishing, October 2025. oecd.org.

U.S. Environmental Protection Agency, age-adjusted VSL, Clear Skies Initiative, 2003. Congressional hearings and AARP response documented in Congressional Record.

Cass Sunstein, The Cost-Benefit Revolution, MIT Press, 2018.

Cass Sunstein, “Will the EPA Stop Considering Lives Saved?” Substack, January 12, 2026. casssunstein.substack.com/p/will-the-epa-stop-considering-lives

Frank Ackerman and Lisa Heinzerling, Priceless: On Knowing the Price of Everything and the Value of Nothing, The New Press, 2004.

U.S. Environmental Protection Agency, “Economic Impact Analysis for Stationary Combustion Turbines Rule,” January 2026. epa.gov/system/files/documents/2026-01/combustion_turbines_eia_final_2026-01.pdf.

Bryan Hubbell and Alan Krupnick, “How the US Environmental Protection Agency Got It Wrong About Monetizing Benefits of Air Pollution Regulations,” Resources for the Future, January 2026. rff.org/publications/reports/how-the-us-environmental-protection-agency-got-it-wrong-about-monetizing-benefits-of-air-pollution-regulations/