Abstract: Each government agency uses a uniform figure to measure the value of a statistical life. This is a serious mistake. The very theory that underlies current practice calls for far more individuation of the relevant values. According to that theory, the value of statistical lives should vary across risks. More controversially, the value of a statistical life should vary across individuals - even or especially if the result would be to produce a lower number for some people than for others. One practical implication is that a higher value should be given to programs that reduce cancer risks. Should government use a higher VSL for programs that disproportionately benefit the wealthy - and lower VSL for programs that disproportionately benefit the poor? A serious complication here is that sometimes the beneficiaries of regulation pay only a fraction or even none of its cost; when this is so, the appropriate VSL for poor people might be higher, on distributional grounds, than market evidence suggests. An understanding of this point has implications for foundational issues about government regulation, including valuation of persons in poor and wealthy nations.