Picture this: you are an analyst at a federal regulatory agency, coffee going cold, staring at a spreadsheet that will determine whether a proposed pollution rule lives or dies. The science is solid. The projected deaths are real. But somewhere in the discount rate field, someone has typed 7, and you already know the rule is not going to make it.
That number, usually sitting between 2% and 7%, does more to determine which safety regulations get written into law than any lobbying campaign, any risk assessment, or any act of political will.
The Clock That Shrinks Future Lives
The logic of discounting is not, in itself, sinister. A dollar today is genuinely worth more than a dollar promised in ten years, because you can invest it now and have more than one dollar in a decade. Economists extended this logic to costs and benefits spread across time: a regulation that costs a billion dollars now but prevents a billion dollars of harm in forty years does not break even under discounting, because that future billion is worth less in present-value terms.
Apply a 5% annual discount rate, and a harm occurring thirty years from now is worth roughly 23 cents on the dollar today. Apply 7%, and it drops to about 13 cents. The arithmetic is straightforward compound discounting: divide the future value by (1 + r) raised to the power of the number of years. A statistical life saved three decades out, valued at five million dollars by the relevant agency, becomes worth somewhere between 650,000 and 1.15 million dollars in present-value terms. That gap is not a rounding error. It is the difference between a regulation passing its cost-benefit test and failing it.
This is not hypothetical. The U.S. Office of Management and Budget has historically instructed agencies to use a 7% real discount rate as a primary estimate, derived from the average pre-tax return on private capital investment. The Environmental Protection Agency has used rates ranging from 2% to 7% across different analyses. The United Kingdom's Treasury Green Book specifies a declining schedule, starting at 3.5% and falling for harms more than thirty years out. Each of those choices is a policy decision wearing the costume of a technical assumption.
A Tale of Two Pipelines
Two proposed safety rules land on an analyst's desk. The first requires chemical plants to install updated pressure-relief systems. Costs are immediate: roughly 400 million dollars in industry compliance spending over five years. Benefits are also near-term, an estimated 80 statistical lives saved over the next decade, concentrated among plant workers and nearby residents. At a 5% discount rate, the benefits comfortably clear the costs. The rule passes its analysis. It gets written.
The second rule requires a reduction in a slow-accumulating industrial pollutant that damages cardiovascular tissue over decades of low-level exposure. Compliance costs are again immediate: 600 million dollars. But the lives saved, roughly 200 of them, are distributed across a forty-year window, with most of the mortality benefit appearing after year twenty. At a 7% discount rate, those 200 lives, each nominally valued at five million dollars, shrink to a present value of perhaps 280 million dollars in aggregate. The rule fails. At a 2% rate, the same calculation yields roughly 740 million dollars in present-value benefits. The rule passes easily.
The science did not change. The number of people who will die did not change. The discount rate changed. That is it.
What People Get Wrong About This
The common objection is that discounting is just prudent economics, that we should not let speculative future harms crowd out present spending. This misunderstands the specific problem with applying financial discount rates to human mortality.
When you discount a cash flow, you are accounting for the opportunity cost of capital: money invested elsewhere genuinely grows. But a person who dies of a preventable disease thirty years from now does not become less dead because we could have invested the prevention money in the meantime. The analogy between financial returns and human welfare is borrowed, not derived. It is like pricing a house fire by the interest you could have earned on the fire extinguisher. As the economist Nicholas Stern argued in his review of climate economics, using a high discount rate to evaluate long-run mortality and welfare is not neutral technical practice. It is a substantive ethical claim: that future people's suffering counts for less simply because they are future people. That claim deserves to be argued in public, not buried in a guidance circular.
There is a second confusion worth naming. People often assume the debate is between discounting and not discounting. It is not. Even economists who think pure time preference is ethically indefensible accept some discounting to account for uncertainty, perhaps the harm will not materialize, and growth, since future generations may be richer and better equipped to cope. The fight is over magnitude and structure. A declining discount rate, which the UK and France both use in different forms, treats near-term and long-term harms differently on the sensible grounds that uncertainty compounds over time. A flat 7% rate treats them the same, which means it treats them very unequally in practice.
And one more thing people get wrong: they assume this is primarily a climate change argument. It is not only that. Discount rates shape decisions about mine tailings, about chemical contamination of groundwater, about the long-term consequences of particulate exposure in industrial corridors. Regulatory economics runs the same machinery everywhere, and it cuts the same way every time.
The Number Nobody Votes On
The choice of discount rate is typically made by budget offices or treasury departments, embedded in guidance documents that receive no legislative scrutiny. A parliament or congress can spend months debating the provisions of an environmental bill. The cost-benefit analysis that will determine whether those provisions survive regulatory review uses a discount rate set in a circular nobody read.
So ask yourself: when did you last vote on a number?
The consequence is that some regulations are, in a real sense, dead before they are drafted. An agency that knows its analysis will be reviewed against a 7% benchmark has no incentive to propose rules whose benefits are concentrated past the twenty-year mark. The technical parameter shapes the policy space before the policy conversation begins. Call it pre-emptive arithmetic.
The discount rate is not a description of how much we value future lives. It is a prescription. Governments that use high rates are not discovering that their citizens prefer the present; they are constructing that preference in the only place it ultimately matters, the formal arithmetic that decides what becomes law. The regulations that would have protected the most people never failed a public vote. They failed a spreadsheet, authored by people whose names you will never know, before the rule ever had a name. What gets built into that one field, quietly, between budget cycles, is the true statement of whose lives a government has decided to count.