A generation ago, the Supreme Court upended the voting rights world. In the breakthrough case of Thornburg v. Gingles, the Court held that minority groups that are residentially segregated and electorally polarized are entitled to districts in which they can elect their preferred candidates. But while the legal standard for vote dilution has been clear ever since, the real-world impact of the Court’s decision has remained a mystery. Scholars have failed to answer basic empirical questions about the operation of the Gingles framework. To wit: Did minorities’ descriptive representation improve due to the case? If so, did this improvement come about through the mechanisms—racial segregation and polarization—contemplated by the Court? And is there a tradeoff between minorities’ descriptive and substantive representation, or can both be raised in tandem?

In this Article, I tackle these questions using a series of novel datasets. For the first time, I am able to quantify all of Gingles’s elements: racial segregation and polarization, and descriptive and substantive representation. I am also able to track them at the state legislative level, over the entire modern redistricting era, and for black and Hispanic voters. Compared to the cross-sectional congressional studies of black representation that form the bulk of the literature, these features provide far more analytical leverage.

I find that the proportion of black legislators in the South rose precipitously after the Court’s intervention. But neither this proportion in the non-South, nor the share of Hispanic legislators nationwide, increased much. I also find that Gingles worked exactly as intended for segregated and polarized black populations. These groups now elect many more of their preferred candidates than they did prior to the decision. But this progress has not materialized for Hispanics, suggesting that their votes often continue to be diluted. Lastly, I find a modest tradeoff between minorities’ descriptive representation and both the share of seats held by Democrats and the liberalism of the median legislator. But this tradeoff disappears when Democrats are responsible for redistricting, and it intensifies when Republicans are in charge. In combination, these results provide fodder for both Gingles’s advocates and its critics. More importantly, they mean that the decision’s impact can finally be assessed empirically.