My job market paper, Ranking Firms Using Revealed Preference, exploits the intuitive notion that workers move towards more preferred firms. I use this revealed preference approach combined with a standard search-theoretic model to estimate the value of working at essentially each firm in the United States. This approach imposes sufficient structure to map the 1.5 million by 1.5 million matrix of worker flows across all firms in the economy into estimates of firm value.
I then combine the revealed-preference based estimate of firm values with earnings data to decompose the variance of firm-level earnings into a rents and compensating differentials component. This comparison is informative about the role of rents in the labor market because dispersion in earnings that corresponds to dispersion in utility implies that frictions prevent competition from forcing all firms to offer the same level of utility. On the other hand, this comparison is informative about compensating differentials because a high-paying firm that is low-value must have bad nonearnings characteristics, while a low-paying firm that is high-value must have good nonearnings characteristics.
Using the U.S. Census Bureau’s Longitudinal Employer Household Dynamics (LEHD) dataset, I find that both rents and compensating differentials explanations are operative, but compensating differentials are more important than rents. I use my estimates of compensating differentials to show that the presence of compensating differentials increases earnings inequality. If amenities were removed and earnings changed to compensate, then the variance of earnings would fall. The importance of compensating differentials also helps explain why search models cannot generate the observed extent of earnings dispersion in the labor market: some of it does not reflect utility dispersion.