Minwoo Hyun

How do financial frictions shape the labor market costs of environmental regulation? I study the 2004–05 Clean Air Act nonattainment designations for fine particulate matter and ozone, which tightened permitting and abatement requirements for plants in newly designated counties. I link pre-designation (2002) firm balance sheets from the Census Bureau's Quarterly Financial Report to establishment employment in the Longitudinal Business Database and to worker earnings histories in the LEHD. Designation reduces employment at the average treated firm by about 18 percent by 2009. The loss is concentrated among highly leveraged firms, reaching about 48 percent for a firm one standard deviation above mean leverage, with no difference before designation. Balance-sheet evidence points to the mechanism: highly leveraged treated firms shift assets toward fixed capital and earn less on them, without borrowing more. Workers bear the cost with a lag, as nonemployment among incumbent workers peaks in 2012, three years after firm employment bottoms out. Older and high-earning workers, who have fewer outside options, absorb the shock mainly through lower earnings. Black workers suffer the largest losses on both margins: the steepest earnings decline and the sharpest rise in nonemployment of any group. This suggests that the labor market incidence of environmental regulation depends on firms' capacity to finance compliance.

How do local economic shocks affect mortality when people can leave? Existing work links local downturns to mortality through income and pollution but says little about migration. We study the 1970s coal boom and 1980s bust across U.S. counties, combining 1940 coal employment shares with national coal prices. The boom raised mortality from procyclical causes such as cirrhosis and motor vehicle accidents. The bust instead displaced people: exposed counties lost population through out-migration, births fell, and all-cause mortality was roughly unchanged. Among those who stayed, however, the bust raised suicide at ages 20-44 and pneumonia mortality at ages 65-84, with no comparable effects during the boom. Out-migration thus shields aggregate mortality but not everyone: the costs of local decline fall on the young adults who remain and on the elderly who are less able to move. Accounting for who leaves and who stays helps reconcile procyclical mortality in national recessions with the deaths of despair documented in declining places.

Many environmental regulations price pollution only when emissions cross a threshold within a short averaging window. Because firms control the mean of their emissions but not the shocks around it, the same threshold prices each source at a different point of its emission distribution. We study the resulting "variance tax" using half-hourly monitoring data from Korean smokestacks and a tightening of nitrogen oxide limits. Tightened stacks cut mean emissions by 22 percent, while their absolute variability fell far less, so relative variability rose; a two-component shock model estimated only on pre-period data predicts these responses with no free parameters. Abatement follows the change in the 30-minute excess fee, not the semiannual basic fee that made up most of the pre-period marginal charge, and stacks that faced no new excess price did not respond. Treating the tightening as a shift in each stack's statutory fee schedule, we recover stack-level marginal abatement costs: about $100 per short ton for binding stacks, the same order of magnitude as U.S. NOx allowance prices. The fee explains about one third of the additional abatement by binding stacks; the remainder is consistent with the threat of administrative sanctions. Obtaining the same reduction with a uniform price would roughly halve abatement costs: the threshold design costs 99 percent more. Longer averaging windows equalize prices and reduce this excess cost, but they deliver less than half the abatement and leave short-run peaks unpriced.