Research

Research Papers


In this paper, I use an equilibrium search model of the labor market to quantify the importance of two classic explanations for why the job-finding rate strongly falls with unemployment duration: (i) unobserved worker heterogeneity and (ii) loss of skills. In the model, unobserved heterogeneity has a large impact on job-finding rates but not on reemployment wages. Wages depend almost exclusively on workers' skill levels. I use the available causal estimates of the effects of unemployment duration on callback rates and reemployment wages to calibrate the model parameters for unobserved heterogeneity and skill loss. Quantitative analysis of the model reveals that unobserved heterogeneity accounts for almost two thirds of total duration dependence in unemployment, while skill loss for the remaining one third. Next, the effects of training and subsidized employment programs on duration dependence are examined through the lens of the model. Subsidized employment raises job-finding rates for all workers but helps short-term more than long-term unemployed workers. Training raises the job-finding rates of the long-term unemployed but lowers job-finding rates for workers with short unemployment spells. The interaction of unobserved heterogeneity with skill loss due to endogenous vacancy creation is crucial for these results.


In an attempt to mitigate the negative effects of clientelism, many governments around the world have adopted meritocratic hiring of public employees. This paper shows that meritocratic government hiring can have unintended negative consequences on macroeconomic aggregates. In many countries, public employees enjoy considerable job security and generous compensation schemes; as a result, many talented workers choose to work for the public sector, which deprives the private sector of productive potential employees. This, in turn, reduces firms’ incentives to create jobs, increases unemployment, and lowers GDP. To quantify the effects of this novel channel, we extend the standard Diamond-Mortensen-Pissarides model to incorporate workers of heterogeneous productivity and a government that fills public sector jobs based on merit. We calibrate the model to aggregate data from Greece and perform a series of counterfactual exercises. We find that the adverse effects of our mechanism on the economy’s TFP, GDP, and unemployment are sizable.


A salient feature of over-the-counter (OTC) markets is intermediation: dealers buy from and sell to customers as well as other dealers. Traditionally, the search-theoretic literature of OTC markets has rationalized this as a consequence of random meetings and ex post bargaining between investors. We show that neither of these are necessary conditions for intermediation. We build a model of a fully decentralized OTC market in which search is directed and sellers post prices ex ante. Intermediation arises naturally as an equilibrium outcome for a broad class of matching functions commonly used in the literature. We further explore, both analytically and numerically, how the extent of intermediation depends on the nature of frictions and model primitives. Our numerical exercises also contrast the model’s equilibrium implications to those of a benchmark model with random meetings and ex post bargaining.


We develop an equilibrium search model with a labor force participation decision, job-to-job transitions, and endogenous separations. The calibrated model perfectly matches the observed labor market flows in US data. We use the model to simulate the effects of an extension of unemployment insurance benefits to 99 weeks. The reform leads to a decrease in employment, an increase in the labor force participation and unemployment rate, while it leaves labor productivity roughly constant. Using a model-based decomposition, as well as comparisons with alternative simplified models, we show that modeling workers' participation decisions, job-to-job transitions, and endogenous separations together is crucial for a complete and accurate analysis of UI reforms.

Older version that includes analysis with workers' life-cycle


We describe how labor market disparities between white prime-age Hispanic and non-Hispanic men have evolved over the last 50 years. Using data from the Census, the ACS, and the March CPS, we examine several employment and earning outcomes distinguishing between US born and immigrant Hispanics. US born Hispanics have experienced gains to employment, having a 2% disadvantage prior to 1990 and a 1% advantage after 2010 compared to non-Hispanics. In terms of earnings, the results are less positive. Hispanics face a substantial negative earnings disparity of between 10% and 20% that shows only modest improvement over time. Most of the employment gain is driven by those with less than a high school degree, while the earnings disparity increases with education. Immigrant Hispanics who completed their education in the US have experienced an even great increase in employment, reaching a positive 10% differential compared to non-Hispanics, but their earnings persistently remain 15-25% lower than non-Hispanics. English language proficiency accounts for a substantial part of the earnings disparity between Hispanic immigrants and non-Hispanics.

Older version that includes racial disparities for comparison purposes.



A large literature in macroeconomics concludes that disruptions in financial markets have large negative effects on output and (un)employment.  Though diverse, papers in this literature share a common characteristic: they all employ frameworks where money is not explicitly modeled. This paper argues that the omission of money may hinder a model's ability to evaluate the real effects of financial shocks, since it deprives agents of a payment instrument that they could have used to cope with the resulting liquidity disruption. In a carefully calibrated New-Monetarist model with frictional labor, product, and financial markets we show that the existence of money dampens or even nearly eliminates the real impact of financial shocks, depending on the nature of the shock. We also show that the propagation of financial shocks to the real economy depends on the inflation level: high inflation regimes magnify the real effects of adverse financial shocks.



The first step in a worker's career is often particularly hard. Many firms seeking workers require experience in a related field, so a vicious circle is created, whereby an entry level job is required in order to get an entry level job. Consequently, entrant workers have lower job-finding rates and longer unemployment durations than the unemployed who have looked for a job in the past. To study the welfare implications of these observations, we consider a version of the DMP model where firms who match with entrant workers have to incur training costs. As a result, firms are biased against entrant workers, who, in turn, stay unemployed for a prolonged period of time, exposing themselves to a persistent skill loss shock. We use a calibrated version of the model to quantitatively assess the effectiveness of four government interventions, whose common goal is to reduce bias against entrant workers. We find that the most effective intervention takes the form of a subsidy that induces firms to rank entrants higher than experienced workers and that this policy brings the economy very close to the constrained efficient outcome.



Most financial trade occurs in decentralized “over the counter” (OTC) markets which are plagued by frictions. The pioneering theoretical work of Duffie, Garleanu, and Pedersen (2005) uses search theory to model these frictions and study price formation in OTC asset markets. In this paper, we conduct the first experimental evaluation of the main prediction of their model. As predicted, we find that when the asset is plentiful, prices are lower when traders have more meeting opportunities, while when the asset is in short supply, prices are higher when traders have more meeting opportunities. We also find that, contrary to the theory, meetings between traders often do not result in successful transactions and that these trade failures create additional frictions affecting asset prices.