Card pioneered the systematic use of natural experiments in economics. He exploited policy changes such as minimum-wage hikes, sudden immigration shocks, and school funding differences to identify causal effects. In the famous minimum-wage study, a border-pair difference-in-differences design showed almost no job loss. His Mariel boatlift work revealed that a large influx of low-skill workers had limited impact on native wages. These findings reshaped debates on labor-market policy.

Card’s strategy centers on exploiting exogenous shocks to identify causal parameters. In the New Jersey–Pennsylvania fast-food study, a border difference-in-differences estimator produced near-zero employment elasticity with respect to minimum-wage changes. The Mariel boatlift analysis treated the sudden 7 % labor-supply increase in Miami as a local shock; city-by-city comparisons and synthetic-control robustness checks supported negligible wage effects for low-skill natives. Education work used state-level school quality as an instrument, revealing sizeable marginal returns to schooling. Collectively, these findings highlight monopsony power, price-pass-through mechanisms, and nonlinear labor-demand schedules, integrating rigorous econometrics into policy evaluation.

Card (1992, 1995) institutionalized causal inference via natural experiments. The NJ–PA design employed county-pair fixed effects, time trends, and cluster-robust SEs, showing ε_emp,wage ≈ 0. Synthetic-control replications and additive heterogeneous treatment models confirmed robustness. His Mariel work combined city d-d with covariate reweighting to address composition changes. Schooling returns exploited IVs such as pupil–teacher ratios and term length within a 2SLS framework to mitigate ability bias. These studies spurred models of monopsonistic competition, asymmetric price adjustment, and complementarities in labor demand.