To assess the impact of Minnesota’s reform on its Black-White disparities in incarceration, we created a panel dataset that captures Black incarceration rates, White incarceration rates, and Black-White incarceration rate ratios from 2001-2015 at the state level. To create this data set, we drew upon prison and jail incarceration data from Vera Institute of Justice’s Incarceration Trends Dataset, and state level population data by race/ethnicity from the U.S. Census Bureau. Then, we employed a quasi-experimental synthetic control approach for each of the three outcomes, where we matched the treated state (Minnesota) to the weighted combination of states that most closely matched its pre-treatment trends in the outcome of interest. To assess potential mechanisms that may explain any association or lack thereof, we reviewed text of racial impact statements to identify whether a bill was projected to exacerbate disparities, as well as whether the bill in question became a law.
We find no evidence that Minnesota’s racial impact statement reform had any impact on Black-White disparities in incarceration. We also failed to find evidence that it impacted Black incarceration rates or White incarceration rates. Our review of racial impact statement content and legislative outcomes of those bills suggests that these null effects were not driven by legislators doubling down on racist beliefs in response to the statements. Instead, we find suggestive evidence that these null effects are in part driven by the fact that racial impact statements are responses to legislation that has already been proposed, rather than analyses that, at least in Minnesota, shape what legislators choose to propose.
Racial impact statement reforms are useful because they provide information to legislators and, in theory, members of the public. However, at least in Minnesota, this approach has not made a dent in Black-White disparities in incarceration. More proactive approaches, like concerned interest groups proposing model bills to state legislatures that are likely to reduce disparities, may be more effective.