The literature is genuinely divided. Quasi-experimental studies since the 1990s find small, context-dependent effects in both directions, and reasonable economists disagree on their size and sign. This is a well-evidenced stalemate, not an absence of evidence.
Moderate minimum-wage increases reduce employment among low-wage workers.
Foundational natural-experiment work found no measurable disemployment from a moderate increase Card 1994, and later bunching-estimator designs reach similar conclusions Cengiz 2019. Other careful studies find the opposite — modest negative effects on employment or hours Neumark 2008, Meer 2016.
A recent systematic review concludes the effect is small and highly context-dependent, varying with the bite of the increase and local labour-market conditions Dube 2019. The weighted evidence sits near the midpoint: substantial mass on both sides, little net lean.
The score is computed deterministically from the citations below, following the methodology in the manuscript (§5). Nothing is hand-tuned per entry — change the evidence and the number moves.
| Citation | sᵢ | mᵢ | tᵢ | e^(−tᵢ/T) | mᵢ·e^(−tᵢ/T) | sᵢ·mᵢ·e^(−tᵢ/T) |
|---|---|---|---|---|---|---|
| Card 1994 | −1 | 8.0 | 32y | 0.202 | 1.62 | −1.62 |
| Cengiz 2019 | −1 | 8.0 | 7y | 0.705 | 5.64 | −5.64 |
| Neumark 2008 | +1 | 7.0 | 18y | 0.407 | 2.85 | +2.85 |
| Meer 2016 | +1 | 6.0 | 10y | 0.607 | 3.64 | +3.64 |
| Dube 2019 | 0 | 7.0 | 7y | 0.705 | 4.93 | +0.00 |
| Σ over citations | 36.0 | 18.67 Mₚ | −0.77 Xₚ·Mₚ |
Reading it: the verdict is Mixed evidence. Direction comes only from Xₚ (tanh is always positive); certainty comes from tanh(Mₚ/W₀) = 0.78. Because Cₚ = 50 when Xₚ = 0 regardless of mass, a balanced stalemate and an unstudied question both read as 50 — which is why Mₚ is always shown alongside the score.
| Revision | Score Cₚ | Mass Mₚ | Center Xₚ |
|---|---|---|---|
| Current revision | 48 | 18.7 | −0.04 |
| 19 May 2026 | 48↓ 1 | 18.7 | −0.04 |
| 24 Apr 2026 | 49 | 16.2 | −0.02 |