Overview
The speaker evaluates a paper on wage-setting and on-the-job search, focusing on whether cost-of-living shocks can trigger wage-price spirals. They accept the importance of on-the-job search for wage inflation but argue the model’s wage-posting assumption removes key bargaining channels. They contrast three mechanisms for wage-price spirals and present evidence from a job-ladder model and Danish tax reform to assess which mechanisms align with observed data.
Evidence linking quits and on-the-job search to wage inflation
- Quits correlate with wage inflation better than unemployment.
- Hiring flows: 50% of people hired have a job; 25% are recalls (maternity/paternity returns); 25% come from unemployment.
- Implication: Understanding wage inflation requires understanding how wages are bargained by employed workers; the speaker “buys that.”
Model assumptions and wage-setting dynamics
- Frictional labor market with random search and a representative firm that posts wages, implying one wage in equilibrium.
- No wage motive for on-the-job search; workers switch jobs due to exogenous preference shocks. Turnover is separate from wages.
- Firm’s trade-off: higher posted wage vs. lower quit probability. If quit probability changes, the equilibrium wage changes.
Cost-of-living shock in the model
- For wages to rise after a cost-of-living shock, quits must increase.
- Labor supply channel: Workers could search more, but on-the-job search is exogenous in the model, so it does not change.
- Labor demand channel: More vacancies could raise quits, but if the central bank looks through a permanent rise in the price level, aggregate demand and labor demand are unchanged; quits do not change; wages do not change.
- Therefore, on impact only the real wage adjusts; the cost-of-living shock has no propagation in this model.
Three wage–price spiral stories discussed
- Bargaining story: Workers seek compensation for higher living costs from their own employer (Nash bargaining/surplus sharing). The model rules this out by construction.
- Outside offers story: Workers search, obtain higher external offers, and return for counteroffers (sequential auction bargaining). Also ruled out by construction.
- Preemptive wage rises by employers: Firms raise wages to prevent quits. The paper speaks to this, but it presupposes workers may search for higher wages—an element the discussant suggests is missing.
Job-ladder model and Danish tax reform evidence
- Ongoing work: A job ladder model with endogenous on-the-job search and sequential auction bargaining.
- Mechanism: Unemployed accept low initial wages due to absent bargaining power; later, employed workers search, receive outside offers, and may obtain counteroffers from incumbents.
- Empirical test: Danish tax reform using micro data.
- 2012 system: People with less than 4 600 euros faced a 42 percent marginal tax rate; above that, 55 percent (13 percentage point increase).
- 2013 change: The threshold was moved eight percentage points to the right.
- Implication: At the 2012 threshold, workers experienced a 13 percentage point fall in marginal tax; those far left/right were unaffected for any feasible on-the-job wage gain, yielding an inverse V-shaped differential response across the distribution.
- Findings:
- On-the-job search: Inverse V-shaped pattern in both model (red) and data (blue).
- Wage growth for stayers: Same inverse V-shaped pattern.
- Leavers: No differential pattern across years, though average wages rise; consistent with the model.
Implications for spiral mechanisms and model design
- Data reject the Nash bargaining story (would imply a V shape, not inverse V).
- Data are consistent with the outside offers/counteroffer mechanism.
- Data reject preemptive wage rises in wage-posting models: a single posted wage for leavers and stayers cannot explain the observed differential behavior.
- Agreement with Jake’s conclusion: It is unlikely that wage-price spirals are initiated by employers raising wages.
- Nonetheless, even without unions, workers can exert bargaining power via outside offers and renegotiation.
- Potential model refinement: Incorporate a wage distribution that makes search meaningful; directed search could give the model a fair chance to test the preemptive mechanism.
Next steps / actions
- Consider extending the model with directed search and a wage distribution to evaluate preemptive wage-setting under realistic on-the-job search incentives.
- Open to questions.