Paper download
Abstract: This paper analyzes self-employment subsidies for the unemployed, a common policy in Europe. We develop a search model with self-employment to study how startup funding affects individual choices, employment, and welfare. The model identifies three channels: (1) direct unemployment reduction, (2) crowding-out effect on paid employment, and (3) budget effect, whereby startup subsidies can generate fiscal savings if unemployment falls sufficiently. (1) and (3) are increasing with labor market rigidity. A simple calibration quantifies these channels and compares welfare effects across labor market structures. Our findings inform active labor policy design in the presence of generous unemployment insurance and rigidities.
Timing and Time Inconsistency in Search Models, with Yun Pei, Economic Letters, Vol 220, Nov 2022, 110846.
Abstract: We study how timing assumption matters for the commitment problem and time inconsistency in search models with unemployment insurance. We analyze the Markov equilibrium without commitment and the Ramsey policy with commitment under two different timings. The first timing is where consumption takes place before search within each period, and the second timing is where search takes place before consumption. We show that time inconsistency occurs mainly through search disincentive under the first timing, but only indirectly under the second timing. We show theoretically and numerically that the magnitude of time inconsistency is stronger under the first timing.
A Dynamic Model of Fiscal Decentralization and Public Debt Accumulation, with Si Guo and Yun Pei, Journal of Public Economics, Vol 212, Aug 2022, 104692.
Abstract: This paper develops a dynamic infinite-horizon model with two layers of governments to study theoretically and quantitatively how fiscal decentralization affects local and central government debt accumulation and spending. In the model, the central government makes transfers to local governments to offset vertical and horizontal fiscal imbalances. But the anticipation of transfers lowers local governments’ expected cost of borrowing and leads to overborrowing ex ante. Absent commitment, the central government over-transfers to reduce local governments’ future need to borrow, and in the equilibrium both local and central debts are inefficiently high. Consistent with empirical evidence, when fiscal decentralization widens vertical fiscal imbalances, local governments become more reliant on transfers, and both local and central debts rise. Applied to Spain, the model explains 39 percent of the rise in total government debt when the vertical fiscal imbalances widened during 1988–1996, and 18 percent of the fall in debt when the imbalances narrowed during 1996–2006.
A Quantitative Theory of Time-Consistent Unemployment Insurance (Job market paper), with Yun Pei, Journal of Monetary Economics, Vol 117, Jan 2021, pp. 848-870.
Abstract: During recessions, the U.S. government substantially increases the duration of unemployment insurance (UI) benefits through multiple extensions. Benefit extensions increase UI coverage and lead to higher average consumption of unemployed workers, but the expectation of an extension may reduce unemployed worker’s job search incentives and lead to higher future unemployment. We show that benefit extensions in recessions arise naturally when the government forgoes prior commitment and makes discretionary UI policies. We endogenize a time-consistent non-commitment UI policy in a stochastic equilibrium search model, and use the model to quantitatively evaluate the benefit extensions implemented during the Great Recession. Switching to the (Ramsey) commitment policy would reduce the unemployment by 2.9 percentage points with small welfare gains.
Delayed Collection of Unemployment Insurance in Recessions, European Economic Review, Vol 118, Sep 2019, pp. 274-295.
Abstract: Using variations in UI policies over time and across U.S. states, this paper provides evidence that allowing unemployed workers to delay the collection of benefits increases their job-finding rate. In a model with discrete job take-up decisions, benefit entitlement, wage-indexed benefits, and heterogeneous job types, I demonstrate that the policy can increase an unemployed worker’s willingness to work, even though more benefits in general reduce the relative value of employment. In a calibrated quantitative model, I find that allowing delayed benefit collection increases the overall job finding rates and may lower unemployment rate both in a steady state stationary economy and over a transition path during 2008-2012.