1. Measurement / data quality concerns (the most famous issue)
What Card & Krueger did
They studied fast-food employment in:
- New Jersey (raised minimum wage)
- Pennsylvania (did not)
They collected employment data via a telephone survey of restaurants before and after the policy change.
Criticism
A. Survey response error
Later researchers argued:
- Some employment figures reported by phone were inaccurate or inconsistently reported
- Managers may not precisely track employment headcounts in real time
- Nonresponse and reporting differences could bias results
B. Reconstructed payroll data tells a different story (partial contradiction)
A follow-up study using administrative payroll records (instead of phone surveys) found:
- A small negative employment effect, unlike the original paper’s zero/positive result
This became one of the strongest critiques:
“When you use more reliable data, the result weakens.”
2. Sampling issues
A. Small and non-random sample
- Only a few hundred fast-food restaurants
- Not a random sample of all employers or all low-wage workers
Critics argue:
- Results may not generalize beyond fast food
- Industry may behave differently than others (e.g., retail, manufacturing)
B. Chain composition changes
Some critics noted:
- Changes in which types of restaurants responded before/after
- Potential compositional bias (not exactly the same establishments across periods)
3. Timing and short-run vs long-run effects
Card & Krueger measured relatively short-run effects.
Criticism:
-
Even if employment doesn’t fall immediately, firms might adjust later via:
- automation
- slower hiring
- reduced expansion
So critics argue:
“No short-run effect ≠ no long-run effect”
4. Specification and interpretation disputes
Even among economists who accept the data, there is disagreement about interpretation:
A. “Employment levels vs employment quality”
Critics argue the study may miss:
- hours reductions
- reduced benefits
- slower job growth
- increased prices passed to consumers
So even if headcount is stable:
labor market adjustment may still be occurring elsewhere
No comments:
Post a Comment