Research Interests : Macro Labor - Search and Matching, Social Networks, Inequality, and Economic Theory

Papers

"The Cyclical Behavior of a Firm's Optimal Market and Referral Hiring (PDF)

Abstract: Firms recruit through both market channels and referral networks, yet standard search-and-matching models usually treat hiring as a single technology. Using evidence from the CPS, JOLTS, the Survey of Consumer Expectations, and U.S. national accounts, I document that the market hiring share is procyclical, while referrals become relatively more important in downturns. I develop and quantify a stochastic Diamond--Mortensen--Pissarides model in which firms create vacancies and allocate recruiting activity between market and referral channels. The channels differ in scale, cost, congestion, vacancy-filling probabilities, and expected match quality. Market recruiting provides broad access to applicants and a higher vacancy-filling probability, whereas referrals have a lower recruiting-cost coefficient and are more likely to generate good matches conditional on contact, but are limited by network access and the transmission of referral information. A central distinction is that firms' recruiting allocations need not equal realized hiring shares because the two channels have different filling probabilities. Real-GDP-disciplined aggregate activity shocks generate countercyclical unemployment and procyclical vacancies, tightness, aggregate job finding, market recruiting, and the market hiring share. Expansions shift recruiting and realized hiring toward the market channel, while downturns raise the relative importance of referrals. Quantitatively, aggregate activity shocks match key stationary labor-market and hiring-composition moments and reproduce the direction of channel reallocation, but understate unemployment, tightness, and aggregate vacancy-filling volatility while overstating hiring-share volatility. I therefore introduce a referral-effectiveness wedge that captures time variation in the ability of network contacts to become usable referral opportunities. The joint economy closely matches the volatility, persistence, and aggregate-activity comovement of the market hiring share. The results show that hiring-channel choice is a distinct propagation margin and that network-based recruiting frictions matter for both the level and cyclicality of hiring composition.

"Market vs Referral Hiring: Implications on Aggregate Welfare, Match Quality and Wage Inequality with Yun Pei (PDF)

Abstract: Empirical evidence from the Survey of Consumer Expectations (SCE) suggests that, in the U.S., between 2013 and 2021, on average, about 50% of workers were hired by a firm through employee referrals and the market channel during a given survey year. This paper builds and calibrates a deterministic Diamond-Mortensen-Pissarides (DMP) search and matching model to study what governs a firm’s optimal market and referral hiring decisions. At the calibrated steady state, we find that, relative to the market channel, the referral channel is more likely to produce a good match, is associated with a lower hiring cost, but also has a lower probability of a match. These trade-offs eventually imply that each channel is equally profitable to hire for a firm. Furthermore, the model is simulated to show that the market hiring rate decreases and the referral hiring rate increases with labor productivity, implying higher aggregate match quality and expected wage gap, favoring referred applicants. Regarding the policy implications of the model, higher unemployment benefits decrease vacancy postings, which in turn increase the unemployment rate, reduce market tightness, encourage market hiring, and discourage referral hiring.

Work in Progress

"A Dynamic Model of a Firm's Market and Referral Hiring"

"Optimal Unemployment Insurance and Firm's Market & Referral Hiring "