Job Market Paper
Using U.S. data, I document a new fact that the wage gap between those with postgraduate and undergraduate degrees is countercyclical – postgraduates have smaller wage shocks than undergraduates over the business cycle. I argue the reason for this phenomenon is the adaptation costs due to the relatively low productivity of new hires who need time to adapt to their jobs. These adaptation costs reduce the value of workers’ outside options and thus the degree to which firms will offer contracts with smoother wages. I find robust facts that are consistent with the predictions of the theory: 1) Postgraduates need longer time to adapt to new jobs than undergraduates; 2) Postgraduates suffer larger wage loses from job displacement; 3) Postgraduates have lower job finding and job-to-job transition rates; 4) The difference in wage cyclicality between postgraduates and undergraduates is higher for workers with long tenure than new hires. To understand how adaptaion costs affect wage cyclicality, I develop an equilibrium search model with long-term contracts and aggregate shocks. In the model, imperfect monitoring of workers’ effort creates a moral hazard problem that requires firms to pay an efficiency wage and restricts risk-sharing between firms and workers. Workers with higher adaptaion costs have more to lose when they leave the current jobs, so they exert more effort regardless of what the firm offers, which alleviates moral hazard and improves risk-sharing. The estimation shows that adaptaion costs alone can explain the differences both in the labour market turnover rates and in the wage cyclicality between postgraduates and undergraduates. The model also shows that postgraduates will accept relatively lower starting wages, but have faster wage growth. Finally, I find that unemployment insurance (UI) crowds out firm insurance, but the effect is smaller for lower educated workers. Lower educated workers have higher welfare gain than postgraduates, which supports the argument for a lower UI replacement rate for postgraduates.
joint with Richard Blundell, Hamish Low, Soren Leth-Petersen, and Costas Meghir
The second-hand car market is subject to asymmetric information about the quality, which generates an endogenous transaction cost — lemons penalty. In this paper, we model sales and purchases of new and second-hand cars and quantify the lemons penalty. We do this by formulating a stochastic life-cycle general equilibrium model of car ownership in which dealers buy old cars from consumers without knowing their exact quality, fix them and sell them back to consumers. Car dealers are offered cars that on average are of lower quality than similar cars in the population. Dealers, therefore, will not pay the expected value of cars being owned to an offered car. They will ask for a price discount, which is the lemons penalty. We structurally estimate the model using a population-wide Danish administrative data set with complete information about car ownership for the period 1992-2009. The data is linked to longitudinal income-tax records of the owners with information about income and wealth. Our results show that 1-year-old car has the largest lemons penalty, which declines over time. Asymmetric information about the quality delays car replacement and substantially lowers transaction volumes. Then we use the estimated model to study the impact of lemons penalty for cars to act as a self-insurance device in the event of unemployment.
This paper studies how unemployment insurance benefit (UI) changes labour force participation over the business cycle. Standard theory predicts that labour force participation should fall in recessions, as the returns of job search decline. In fact, it is acyclical (moves independently of the business cycle). I argue that UI, being negatively correlated with the business cycle, is the reason. I document a new fact that the Maximum UI Weekly Benefit Amount is relatively higher in recessions than in booms. My theory is that as UI is more generous in recessions, unemployment becomes more attractive than out-of-labour-force (OLF). While some workers drop out of the labour force due to no job opportunities, others participate to enjoy the higher UI. I embed the counter-cyclical UI into a search model with aggregate productivity shocks and endogenous labour force participation. I show that although the cyclical variation in the level of UI is small, it plays a vital role in shaping fluctuations in the participation rate. The model is also able to capture other cyclical movements in labour market stocks and gross worker flows. The model also shows that counter-cyclical UI can stabilise the economy by reducing the variation in employment and GDP.