The proposed method evaluating population risks from chemical exposure is based on estimation of the individual variabilities in the internal concentration of the chemical of concern and estimation of the magnitude of toxic effect caused at a given internal concentration. We assume a log-normal distribution for individual variability and propose loss of life expectancy (LLE) as a measure of the magnitude of toxic effects. We applied it to the evaluation of the governmental action of prohibiting the use of chlordane as a termiticide in Japan. Because the method is applicable to both carcinogenic and noncarcinogenic chemicals, the cancer risk due to chlordane and the noncancer risk due to chlorpyrifos, a major substitute, were evaluated and compared. The estimated values of risk represented in terms of LLE are 0.10 days for residents of untreated houses, 1.9 days for residents of treated houses, and 4.4 days for termite control workers when chlordane is used, and 0, 2.8, and 31 days when chlorpyrifos is used, respectively. From the result of the case study, we can see that the prohibition of chlordane is not necessarily effective in reducing risk.