Senior Thesis Defense · Department of Economics

Minimum Wage and the Teen Labor Market

Evidence from border-county pairs, 2009–2023. Presented by Elena Kowalski to the Undergraduate Thesis Committee, Spring 2026.

Overview

Roadmap

The employment question, how the design isolates the minimum-wage effect from other local shocks, and what 316 border-county pairs show.

  • 01 The Employment Question
  • 02 Identification Strategy
  • 03 Evidence from 316 County Pairs
  • 04 Contribution & Limitations

Across contiguous county pairs, a 10% minimum-wage increase moves teen employment by 0.3 percentage points — statistically indistinguishable from zero.

Chapter 4 · Border-Pair Estimate
The Identification Problem

Naive Panels Overstate Job Loss

Comparing all treated and untreated states implies a −0.15 teen employment elasticity. Restricting comparisons to adjacent counties that share a labor market narrows that to −0.02 — well within sampling noise.

−0.15
−0.02
Naive State Panel
Border-County Design
Estimated teen employment elasticity
Border-Pair Map, PA–NJ Corridor
Identification Strategy

Matching Counties Across State Lines

The design pairs each minimum-wage county with its contiguous neighbor across a state border, so local labor-demand shocks — weather, retail openings, highway construction — are shared by construction.

Employment comes from QCEW county records linked to state and local minimum-wage schedules, restricted to teen-heavy sectors: retail and food service.

316

County Pairs

15

Years, 2009–23

1.3M

Employment-Months
Contribution

What This Thesis Adds

I

A Longer Border-Pair Panel

Extends the Dube-Lester-Reich design six years past its original sample, through three federal non-adjustment years and 41 state-level increases.

II

Firm-Size Heterogeneity

Splits establishments by employee count. Job loss concentrates in firms under 20 employees; larger employers show no detectable effect.

III

A Dynamic Event Study

Traces employment for eight quarters around each increase, ruling out anticipation effects and pre-existing divergent trends.

Evidence

Employment Converges to the Counterfactual

Border-pair synthetic-control estimate for teen employment in treated counties, indexed to the quarter before each minimum-wage increase (Q0 = 100).

90 95 100 105 Q−8 Q−5 Q−2 Q+1 Q+4 Q+8
Treated County (Actual) Synthetic Control (Counterfactual)
Methodology

From Data to Defensible Estimate

Step One

Assemble the Panel

Link county-level QCEW employment records to every state and local minimum-wage schedule change, 2009–2023.

Step Two

Construct Border Pairs

Match each treated county to its closest untreated neighbor across a state line, following the contiguous-pair method.

Step Three

Estimate the Event Study

Run quarter-by-quarter fixed-effects regressions with pair-by-time controls to isolate the minimum-wage effect.

Step Four

Stress-Test the Result

Re-run under placebo dates, alternate bandwidths, and a synthetic-control specification to check the estimate survives.

Robustness

Four Checks the Result Survives

PT

Placebo Test

No Effect Found
SC

Synthetic Control

Confirms Estimate
BW

Bandwidth Check

±10 Miles
FS

Firm-Size Split

Small Firms Only
Verdict

No Detectable Teen Job Loss — With One Caveat

The Estimate Holds Only for Increases Below $12/hr

Committee: Prof. A. Reyes (Chair) · Prof. M. Okafor · Prof. L. Chen
01 / 10