Compensating Wage Differentials Compensating Wage Differentials Utility not

  • Slides: 7
Download presentation
Compensating Wage Differentials

Compensating Wage Differentials

Compensating Wage Differentials • Utility, not wages, should be equalized across jobs in a

Compensating Wage Differentials • Utility, not wages, should be equalized across jobs in a perf. competitive market. • Utility is affected by hardship, risk of injury, risk of layoff, working with people, etc. • Affects occupations, regions, countries, . . . • Basic idea from Adam Smith, book 1, chapter 10 of Wealth of Nations. • Hedonic model of Rosen (1974)

Basic Model • Equally productive workers are mobile (across regions, occupations, firms, etc. )

Basic Model • Equally productive workers are mobile (across regions, occupations, firms, etc. ) • Marginal worker has perfect information. • What is the necessary consumption that would compensate me (in terms of utility) for working in a noisy job? • Firm side, equilibrium. • Estimate hedonic (reduced-form) equation.

Equalizing Wage Differential for Workplace Safety (β 3) • ln(wij) = β 0 +

Equalizing Wage Differential for Workplace Safety (β 3) • ln(wij) = β 0 + β 1∙Xi + β 2 ∙Zj + β 3 ∙ riskj + εij • How to measure risk? – Self-reported? Problem if workers differ in perception and ability to deal with risk. – Industry- or occupation-wide accident rate? Disregards workers heterogeneity. • Worker unobservables affect not only wage but also their own risk (Shogren & Stamland JPE 2002): how tolerant / able to cope with risk? )

Workplace Safety and Equalizing Wage Differentials • Simple regressions often give wrong ‘sign’. •

Workplace Safety and Equalizing Wage Differentials • Simple regressions often give wrong ‘sign’. • Solution to this endogeneity? – Use switchers to eliminate time-constant workers’ characteristics (But: are switchers random? IV for switching jobs (plant closure)). – Use a proxy for the unobservables. (Viscusi & Hersch, REStat 2001: Smokers select riskier jobs, get injured more and get lower comp. for risk of injury. Labor markets are segmented. )

An example: Night-Working UK LFS 1 2 3 Coeff on Night -0. 079 (0.

An example: Night-Working UK LFS 1 2 3 Coeff on Night -0. 079 (0. 013) -0. 050 (0. 010) 0. 029 (0. 015) Other Controls No Yes C-S/Panel C-S C-s Fixed effects

Some related applications • VSL: Ashenfelter &Greenstone: road safety • Wage curve studies: how

Some related applications • VSL: Ashenfelter &Greenstone: road safety • Wage curve studies: how do regional wages react to regional unemployment? – Cross-section (steady state comparisons): Compensate workers for higher risk of job loss (higher unemployment rate): β>0 – Time differences (Wrt -Wrt-1=β (Urt -Urt-1)+…): Regional equilibration: lower wages when unempl. goes up to support job creation: β<0