UC Berkeley

Growing inequality and stagnating wages at the bottom of the earning distribution are the most striking social phenomena of the last 30 years. Moreover, the 2009 Great Recession surged unemployment and created unprecedented tension between rich and poor in most developed countries. These circumstances renewed the interest of politicians, policy makers, and economists toward public policies aimed at alleviating inequality. In this thesis, I empirically assess the effectiveness of two prominent public policies in helping the poor: the minimum wage and unemployment insurance.

Minimum wage is the most radical policy tool for elevating the wages of the bottom economic bracket. However, despite several decades of microeconometric evidence for increases, the minimum wage remains a highly controversial policy. The first two chapters of this dissertation are devoted to assessing the economic effects of an unusually large and persistent increase in the minimum wage instituted in Hungary in 2001. The minimum wage to the median wage increased from the current U.S. level (35%) to the level of 55%, which is equivalent with an (~60%) increase in the minimum wage in real terms.

In the first chapter, my co-author Péter Harasztosi and I study the employment effects of this unique minimum wage reform. We propose a new approach to estimating the employment effects of a minimum wage increase that exploits information on the distribution of wages before and after the policy change. We infer the number of jobs destroyed by comparing the number of pre-reform jobs below the new minimum wage to the excess number of jobs paying at (and above) the new minimum wage. The evolution of the earning distribution in Hungary shows that this ratio is close to one, suggesting that most firms responded to the reform by raising wages instead of destroying jobs. We confirm this conclusion using comparisons across subgroups of workers with larger and smaller fractions of worker affected by the minimum wage change. Our group-level estimates, again, imply that the higher minimum wage had, at most, a small negative effect on employment, and with the standard errors we can rule out larger than -0.3 employment elasticities with respect to wages.

In the second chapter, my co-author Péter Harasztosi and I study the economic incidence of the minimum wage polices. If minimum wage increase has a small negative effect on employment and a large effect on wages, the total remunerations allocated to low-wage workers must increase. Using a large panel of firms and the Hungarian minimum wage increase, we show that this is indeed the case: firms highly exposed to the minimum wage experienced a large increase in their total labor cost. However, this raises a question: who pays for this cost increase? We show that firms' profits are not affected in response to the minimum wage, suggesting that firm-owners do not bear the incidence of the minimum wage increase. Instead, we document that total revenue of low-paying employers increased considerably, indicating that firms passed the effect of the minimum wage to consumers. Consistent with that explanation, we show that firms facing more elastic output demand, and so less ability to pass-through the effect of the minimum wage, experienced larger employment losses and lower increase in their total labor cost.

In the third chapter, Stefanno DellaVigna, Balázs Reizer, Johannes Schmieder and I scrutinize the job search behavior of the unemployed. We propose a model of job search with reference-dependent preferences, where the reference point is given by recent income. Newly unemployed individuals are faced with a loss because their recent past income is higher than the unemployment benefit they receive, and so they search hard. However, over time they get used to lower income, and thus search less. They search harder, again, in anticipation of a benefit cut, only to ultimately get used to the change. The model fits the typical shape of the exit from unemployment, including the spike at the UI exhaustion point. The model also makes unique predictions for the response of benefit changes. Second, we provide evidence using a reform in the unemployment system in Hungary. Most unemployment insurance programs have constant replacement rate for a fixed period, typically followed by lower benefits under unemployment assistance. In November 2005, Hungary switched from this standard single-step UI system to a two-step system, with unchanged overall generosity. We show that the system generated increased hazard rates in anticipation of, and especially following, benefit cuts in ways the standard model has a hard time fitting, even when allowing for unobserved heterogeneity. We structurally estimate the model and estimate a weight on gain-loss utility comparable to the weight of the standard utility term, and a speed of adjustment of the reference point of eight months. The results suggest that a revenue-neutral shift to multiple-step UI systems can speed exit from unemployment.