Cheap on Punishment: Examining the Impact of Prison Population Racial Demographics on State-Level Corrections Spending

Abstract

Research has explored the effects of various state-level characteristics, such as racial composition and economic conditions, on correctional budgetary decisions. However, researchers have yet to consider how the racial makeup of state prison populations themselves may impact subsequent corrections spending decisions. Drawing on work suggesting that people of color are simultaneously over-punished and neglected by criminal justice systems, and utilizing a time-series cross-section analysis of 50 states from 1979 through 2017, we explore differences in state budgetary allocations for correctional expenditures based on the racial demographics of prison populations. We find that the relationship between the Black-to-White incarceration ratio and spending on corrections is curvilinear: once a tipping point of Black-to-White incarceration is reached, spending on corrections decreases. This finding is especially pronounced in Southern and Midwestern states. Overall, our results provide a strong starting point for understanding the ways in which Black Americans are neglected in the incarceration setting.

Keywords:

• Punishment
• incarceration
• race

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1 For example, between 1977 and 2019, local and state spending on corrections increased by 347 percent (from 18 to 82 billion dollars), surpassing growth in education, police, highways and roads, health and hospitals, and housing and community development (Urban Institute, n.d.). In terms of per-person funding, per capita corrections expenditures increased more than twice as fast as per-pupil PK–12 expenditures from the years 1979–80 to 2012–13, at 185 percent compared to 73 percent (https://www2.ed.gov/rschstat/eval/other/expenditures-corrections-education/brief.pdf).

2 To measure Republican legislative percent, Stucky et al. (2007) calculated the percentage of Republican state legislators in both legislative houses in a given state-year.

3 Prior research has also explored the importance of a range of factors associated with corrections spending in the national context (Caldeira & Cowart, 1980; Jacobs & Helms, 1999). One early study by Caldeira and Cowart (1980) examined national-level prioritization of spending by exploring political budget requests in response to crime rate changes from 1935 through 1975. Crime increases were found to lead to requests for increased spending on agencies that performed crime prevention and control, such as the Bureau of Prisons. Moreover, budgetary requests for increases in crime prevention and control were found to be more likely to be successful when a Republican president was in power (Caldeira & Cowart 1980). Similarly, Jacobs and Helms (1999) found that from 1954 to 1990, national-level Republican strength (measured as a dichotomous variable of Republican president multiplied by the percentage of Republicans in the House and Senate, and those that responded as Republican in the national Gallup survey) and crime rates were significantly associated with higher total (federal, state, and local) corrections expenditures per capita. The authors also found evidence in support of the racial threat hypothesis: increases in the non-White population percentages were met with increases in per capita corrections spending (Jacobs & Helms, 1999).

4 Per capita total corrections expenditures consider both capital outlay and operating budgets and allow for the measure to be based on the population size.

5 When focusing in detail on specific line-items that make up correctional budgets, important state differences also come to light that align with Williams and Campbell’s (2021) results and suggest that conservative states in the South may be unwilling to spend on incarcerated populations. For instance, a 2004 Bureau of Justice Statistics report found differences by state on spending per inmate, with Alabama prisons spending about $22 a day per inmate and averaging about seven inmates per employee in terms of inmate-to-staff ratio; compared to Maine, which has operating costs averaging $122 per inmate per day and less than two inmates per employee.

6 Muller’s (2012) definition of Northern states include a variation of both Midwest and Northeast states, as defined by the U.S. Census Bureau.

7 While, relative to other regions, Southern states have generally experienced the highest levels of incarceration rates (Schoenfeld, 2018), other regions emerge as the leaders in the racial gap in Black-to-White incarceration rates (Nellis, 2016). For instance, five of the ten states with the largest gap (measured as a ratio) in Black-to-White incarceration rates are situated in the Midwest. Coupled with the growth in both overall incarceration rates and the Black-to-White incarceration ratios from 1979 through 2017, states in the Midwest outpaced the Northeastern and Southern regions’ growth (percentage change) in the percentage of spending allocated to corrections by nearly 30 points.

8 As Beckett (1999, p. 83) outlined in Making Crime Pay, while support for increasing racial equality has grown since the Civil Rights Movement, there is less support for policies aimed at actualizing racial equality, with “white anxiety about the speed and extent of racial reform [manifesting] itself in opposition to policies aimed at the implementation of that principle” (see also Schuman et al., 1997). This anxiety has been used by Southern Republicans in electoral strategies – most visibly through the Southern Strategy – in an attempt to gain support from Southern Whites and keep Black Americans repressed.

9 Spending on salaries, wages, and benefits of employees is, however, greater in the South (about $6,017,146) than in other regions, which spend less than $5,000,000 on these costs on average.

10 Costs per inmate for medical care in the South averaged $2,025 in 2001, compared to $3,417 in the Northeast, $2,260 in the Midwest, and $3,671 in the West. Costs per inmate for food services were $731 in the South, compared to $1,217, $1,211, and $1,024 in the Northeast, Midwest, and West, respectively. Finally, utilities costs per inmate were $670 in the South in 2001, compared to $1,176 in the Northeast, $826 in the Midwest, and $785 in the West (Stephan, 2004; see also Pfaff, 2017).

11 Following prior research, we log-transformed percentage of total state-level expenditures allocated for corrections to minimize issues of skewness across states (Stucky et al., 2007; Williams & Campbell, 2021). Sensitivity analyses were run using the non-log-transformed measure of the percentage of expenditures allocated towards corrections and results remained stable, with all significant factors remaining significant and in the same direction. In line with prior research, models were also run predicting percentage capital outlays and per capita spending (e.g., Breunig & Ernst, 2011; Spelman, 2009). When predicting percent capital outlay and overall spending per capita, our results for the prison demographic ratios were no longer significant, but the state-level factors (e.g., percentage Black and violent crime) remained significant, lending evidence to the argument that state-level factors outside of prisons impact spending on social control responses (such as building prisons and responding to the growth of people of color), but not how monies are spent based on the racial makeup of the prison population. All of the above results are available upon request.

12 Data were not available for years 2001 and 2003 and were therefore linear interpolated.

13 Following prior research (e.g., Williams & Campbell, 2021), we focus on racial differences and omit ethnic group comparisons due to scattered ethnicity data by states for both incarceration and census populations prior to the 1980s. Incarceration spending differences for ethnic sub-groups remain an important avenue for future research.

14 Sensitivity analyses were run using multiple ways of measuring the Black incarceration variable. We included a ratio of Black-to-White incarceration rates based on prison populations, a ratio of Black-to-White incarceration rates based on state populations, and Black incarceration rates without a ratio. The results from models containing these measures aligned with the current results and demonstrated significant, curvilinear relationships with spending across all main models, with a few exceptions in the region-specific models. These results are available upon request.

15 State-level Citizen Ideology_t = (Citizen Ideology_(Sum)d)/d). Citizen Ideology_d,t = (Proportion of electorate preferring incumbent_d,t)(Ideology score of incumbent_d,t)+(Challenger proportion of the electorate_d,t)(Challenger ideology score_d,t), in which d denotes the district and t denotes the year. Each district’s ideological score is then combined to generate the unweighted average.

16 Scholarship suggests that national-level forces impact state-level budgetary decisions (Schoenfeld, 2018). These can be seen in policies such as the Omnibus Crime Control and Safe Streets act of 1968 and the Violent Crime Control and Law Enforcement Act of 1994, and the federal government may incentivize states with extra financial aid when states more heavily enforce these national-level policies. While it is not possible to tease out why or how these monies are allocated at the state level, including it as a control allows the models to account for any additional funds above the traditional income, in which we may see slightly higher percentages allocated for corrections.

17 As a sensitivity analysis, state-level fixed effects models were run, and the key findings of importance remained stable. These results are available upon request.

18 The linear and quadratic terms of Black-to-White incarceration ratio failed to reach significance in states situated in the Northeast and West regions and are therefore not presented in Figure 2. Graphs for these regions can be found in Appendix B.

19 According to Hoyt and Leroux (2007, p. 25), “forgotten cities” in the Midwest became “home to the lion’s share of unwanted facilities such as incinerators, sewage treatment plants, or prisons, and may even view these activities perversely as an economic development strategy worth promoting,” demonstrating neglect (see also + Russo and Linkon, 2009).

20 While racist rhetoric and behavior has been, at least to some degree, tolerated in the South, political and criminal justice actors in the Midwest may feel pressure to appeal to a wider and more diverse constituency all the while neglecting people of color in practice. We see evidence of more discrete forms of Black neglect, such Black victims’ lower experiences of crime clearance, in recent research examining policing and prosecution in the Midwest (e.g., Vaughn et al., 2023).

21 While our dataset does not include the ethnic makeup of prison populations, we did run several supplementary analyses to investigate the relationship between budget allocations and Black-to-White imprisonment ratio for Sunbelt versus non-Sunbelt states. When we removed region from the models and included a dichotomous control for Sunbelt state (1 = yes; 0 = no), we found a positive and significant relationship between Sunbelt states and percentage of budgets allocated for corrections, with the Black-to-White imprisonment ratio remaining significant and curvilinear. When Sunbelt and non-Sunbelt states were analyzed as “sub-samples” (as in our region-specific models), the Black-to-White imprisonment ratio was no longer significant, perhaps due to Sunbelt states being influenced more by Hispanic population growth than Black population changes. These results are available upon request.