The majority of the increase in the duration of unemployment that is caused by UI benefits actually is attributable to the liquidity effect, which he links to cash on hand at the time of job loss, rather than to the net wage effect.
One of the classic findings in public finance is that generous unemployment insurance (UI) benefits reduce labor supplied by those who are eligible for these benefits. Traditionally, this finding has been interpreted as evidence that UI reduces the net wage associated with finding a new job; it also makes leisure relatively more attractive than working, which reduces economic efficiency and welfare. But in Moral Hazard vs. Liquidity and Optimal Unemployment Insurance (NBER Working Paper No. 13967), NBER Research Associate Raj Chetty questions whether the link between unemployment benefits and the duration of unemployment is attributable only to this traditional net wage effect, or whether it might depend on the cash-on-hand available to unemployed individuals (or, what we call liquidity). Nearly half of job losers in the United States report no liquid wealth at the time they lose their job, suggesting that many households may be unable to weather even a temporary shock to their income. Indeed, Chetty finds that the majority of the increase in the duration of unemployment that is caused by UI benefits actually is attributable to the liquidity effect, which he links to cash on hand at the time of job loss, rather than to the net wage effect. Unlike the net wage effect, the liquidity effect improves economic welfare; therefore, the unemployment insurance program may create a substantial benefit by providing households with much needed liquidity while unemployed.
Using a dataset of more than 4,500 unemployment spells, Chetty first shows that increases in UI benefits have much larger effects on the duration of unemployment for liquidity-constrained households (that is, households with low levels of liquid wealth) than for other households. Then, using data from two surveys that Mathematica conducted for the Department of Labor, he finds that lump-sum severance payments also increase the duration of unemployment substantially among liquidity-constrained households. Because lump-sum severance payments are cash grants that do not distort the individual's net wage, this constitutes direct evidence that liquidity effects are large. Combining the two sets of estimates, Chetty concludes that 60 percent of the increase in unemployment duration caused by UI benefits is attributable to the liquidity effect. Finally, Chetty uses this estimate to show that the UI program yields significant welfare gains by providing liquidity, despite reducing efficiency by making work less attractive.
Although Chetty focuses on unemployment, the theoretical approach that he develops in this paper may have broader applicability, such as in calculating the welfare gains from other social and private insurance policies. For example, one could calculate the value of a health insurance program by estimating the extent to which an individual's medical expenditures would differ if he were paid a lump sum rather than an indemnity benefit that would cover his health expenses. This method does not require data on the outcomes of insurance provision -- such as health, consumption, or job match quality -- which historically have proved to be difficult to measure. More generally, Chetty suggests that developing identification strategies similar to his to analyze optimal policy in contexts beyond insurance should be a high priority for further research.
-- Lester Picker