Knowledge that Transforms

This paper studies the speed at which workers’ pretax earnings respond to tax changes along the intensive margin. We do so in the context of Germany, where a large notch in the tax schedule induces sharp bunching in the earnings distribution. We analyze earnings responses to two policy reforms that shift this notch outward and find clear evidence that frictions delay the earnings responses of more than 38% of workers. We propose that heterogeneity in firm labor demand plays a key role in generating the observed differences in the speed of workers’ earnings responses and provide supporting evidence.

This paper obtains comparable estimates of the effect of unemployment insurance (UI) benefits on labor supply throughout the unemployment spell and over the business cycle using a regression kink design and 20 years of administrative data from California. For a given unemployment duration, the behavioral effect of UI benefit levels on labor supply does not vary with the business cycle from 2002 to 2019. However, due to increased coverage from extensions in benefit durations, the duration elasticity of UI benefits rises during recessions. The behavioral effect during the start of the COVID-19 pandemic is substantially lower at all unemployment durations.

true Throughout most of the twentieth century, economic growth was associated with rising median real wages. However, since the early 1980s, measured median real hourly compensation has been stagnant despite robust productivity growth. To the extent that measured real wage growth has occurred, it has been concentrated disproportionately at the upper end of the wage distribution. Many view the lack of growth of median wages over this time period as evidence that the American middle class has not advanced and as a symptom of declining social mobility. The decoupling of measured median real wage growth and productivity growth has been viewed as a puzzle among both academics and policy makers. During this time period, there has also been a separation of wage growth and othermacroeconomic fundamentals. For instance, theUnited States has had record low levels of unemployment in the years prior to the global pandemic, yet during the prepandemic period there was little accompanying wage growth. This presents an apparent contradiction of the long-standing Phillips curve analysis that negatively relates unemployment towage growth. Researchers have begun to dig into this puzzle of late (see, e.g., Del Negro et al. 2020). Recent explanations involve the possibility that we have mismeasured the amount of slack in the economy (Krueger, Cramer, and Cho 2014; Abraham, Haltiwanger, and Rendell 2020) or that we have a flatter

Increasing mother’s labor supply during a child’s preschool years may reduce time investments, yielding a negative direct effect on midchildhood and teenage outcomes. But as mother’s work hours increase, income will rise. Can income compensate for the negative effect of hours? Our mediation analysis exploits exogenous variation in both mother’s hours and family income. Results suggest a negative, insignificant direct effect from increasing mother’s hours on child test scores. However, the positive mediating effect of income creates a positive total effect on test scores (26% of a standard deviation) for a 10-hour increase in mother’s weekly hours in preschool years.

This paper studies learning among coworkers when incentives change. We use a simple principal-agent model to show that when workers are not fully informed on the global shape of the production function, (1) their effort choice changes over time as information is disclosed and processed and (2) changing incentives can trigger this learning process. We test this prediction using personnel data from an egg production plant in Peru. Exploiting a sudden change in the contract parameters, we find that workers learn from each other over the shape of the production function. This adjustment process is costly for the firm.

We study how active labor market policies affect the exchange of information and support among job seekers. Leveraging a unique social network survey in Ethiopia, we find that a randomized job search assistance intervention reduces information sharing and support between treated job seekers and their active job search partners. Because of lower job search support, untreated individuals search less and, suggestively, have worse employment outcomes. These results are consistent with a model of networks where unemployed individuals form job search partnerships to exploit the complementarities of job search.