Explaining State Black Imprisonment Rates 1983-1999

This paper addresses the problem of rising Black/White disparities in imprisonment and seeks to determine factors predicting which states will have the highest Black imprisonment rates for various offenses. Annual state imprisonment rates for 1983-1999 are calculated from the Correctional Populations of the United States (CPUS) and National Corrections Reporting Program (NCRP) data. Dependent variables are the logged Black and White rates of being "inprison" (CPUS) and of sentences to prison, and the first differences of the logged rates, disaggregated by offense group. Multivariate pooled time-series analysis using panel-corrected standard errors is employed to identify the correlates of Black and White imprisonment rates, for Blacks, with and without a control for the White imprisonment rate. Factors examined include Black and White poverty rates, homicide rates, and unemployment rates, as well as the percent Black, the percentage change in the percent Black, the average level of metropolitan segregation, and the presence of a Republican governor. All independent variables are lagged one year to clarify issues of causal order. Results show a much more complex interplay of inter-racial dynamics than previous theory has predicted. Low White poverty is the most consistent predictor of high Black imprisonment rates. Blacks generally have higher imprisonment rates where the percent Black is lower, while White imprisonment rates are lower where the percent Black is increasing. Imprisonment rates are higher where homicide rates are higher. Predictors of Black drug sentences are different from those for other kinds of sentences. ©8/13/04. Prepared for the 2004 American Sociological Association Meeting. This is a draft in process. Comments and suggestions are appreciated. Table 1. Classification of states by %Black and Dissimilarity (N= in NCRP sample) Percent Black Dissimilarity Index (Sorted) State N N State DissN State DissN < 8% Black 8-20% Black Delaware (10) 0.300 Louisiana (22) 0.640 Arizona (4) 0 Arkansas (5) 0 Idaho (16) 0.310 Georgia (13) 0.641 California (6) 1 Connecticut (9) 0 New Mexico (35) 0.350 Florida (12) 0.641 Colorado (8) 1 Delaware (10) 0 New Hampshire (33) 0.360 Rhode Island (44) 0.660 Idaho (16) 0 Florida (12) 1 Wyoming (56) 0.370 Arkansas (5) 0.660 Indiana (18) 0 Illinois (17) 1 South Dakota (46) 0.390 Nebraska (31) 0.661 Iowa (19) 1 Michigan (26) 1 Montana (30) 0.420 Massachusetts (25) 0.671 Kansas (20) 0 Missouri (29) 1 Utah (49) 0.451 Connecticut (9) 0.670 Kentucky (21) 1 New Jersey (34) 1 South Carolina (45) 0.481 Kentucky (21) 0.671 Massachusetts (25) 1 New York (36) 1 Arizona (4) 0.480 Maryland (24) 0.691 Minnesota (27) 1 Ohio (39) 1 North Carolina (37) 0.491 Missouri (29) 0.701 Montana (30) 0 Pennsylvania (42)1 Nevada (32) 0.501 New Jersey (34) 0.731 Nebraska (31) 1 Texas (48) 1 Washington (53) 0.521 Pennsylvania (42) 0.741 Nevada (32) 1 Virginia (51) 1 Oregon (41) 0.571 Ohio (39) 0.741 New Hampshire (33) 0 Oklahoma (40) 0.580 Indiana (18) 0.760 New Mexico (35) 0 20% + Black Colorado (8) 0.581 Wisconsin (55) 0.781 Oklahoma (40) 0 Alabama (1) 1 Kansas (20) 0.580 New York (36) 0.791 Oregon (41) 1 Georgia (13) 1 Virginia (51) 0.601 Illinois (17) 0.801 Rhode Island (44) 0 Louisiana (22) 0 West Virginia (54) 0.601 Michigan (26) 0.821 South Dakota (46) 0 Maryland (24) 1 Iowa (19) 0.601 Utah (49) 1 Mississippi (28) 1 Mississippi (28) 0.611 Washington (53) 1 North Carolina (37) 1 Minnesota (27) 0.611 West Virginia (54) 1 South Carolina (45) 1 Texas (48) 0.621 Wisconsin (55) 1 Alabama (1) 0.621 Wyoming (56) 0 California (6) 0.631 Table 2. Correlations Among Variables in Static Analysis Black CPUS White CPUS Black Non Drug Black Drug White Non Drug White Drug % Black D % Black Diss Rep Gov White Pov Black Pov Whi Hom Blk Hom Blk Unem Whi Unem Year Black CPUS 1.00 White CPUS 0.56 1.00 Black Non Drug 0.57 0.06 1.00 Black Drug 0.57 0.38 0.47 1.00 White Non Drug 0.19 0.56 0.31 0.31 1.00 White Drug 0.24 0.52 0.18 0.65 0.65 1.00 % Black -0.29 0.07 -0.36 0.05 0.27 0.33 1.00 Change % Black 0.17 -0.18 0.36 0.04 -0.36 -0.29 -0.42 1.00 Dissimilarity -0.22 -0.43 -0.13 -0.06 -0.37 -0.18 0.01 -0.07 1.00 Repub Gov 0.23 0.07 0.16 0.18 -0.04 0.04 -0.03 0.05 0.04 1.00 White Poverty -0.23 0.25 -0.31 -0.46 0.08 -0.12 -0.01 -0.21 -0.31 -0.15 1.00 Black Poverty -0.33 -0.15 -0.13 -0.44 -0.17 -0.35 0.12 -0.01 0.12 -0.01 0.37 1.00 White Homicide -0.05 0.54 -0.06 0.02 0.40 0.33 0.17 -0.52 -0.22 -0.19 0.39 -0.08 1.00 Black Homicide 0.03 0.01 0.22 0.13 0.03 0.07 -0.12 0.13 0.37 0.03 -0.15 -0.04 0.15 1.00 Black Unemployed -0.04 -0.06 0.11 -0.31 -0.33 -0.47 -0.41 0.24 -0.13 0.10 0.25 0.32 -0.15 -0.07 1.00 White Unemployed -0.08 0.04 -0.03 -0.18 -0.15 -0.20 -0.49 -0.07 0.17 0.01 0.30 0.11 0.26 0.21 0.41 1.00 Year 0.59 0.45 0.18 0.65 0.26 0.37 0.04 0.16 -0.23 0.20 -0.20 -0.47 -0.14 0.04 -0.26 -0.22 1.00 Year Squared -0.11 -0.02 -0.15 -0.36 0.04 -0.12 0.02 -0.08 0.00 0.01 0.24 -0.01 -0.09 -0.26 -0.03 -0.08 -0.01 Note: Correlations between independent and dependent variables are bolded if their magnitude is greater than .2 (R>.04). Table 3. Regression of White Imprisonment and New Sentences on Independent Variables. All are Logged Rates Per 100,000 White CPUS In Prison White New Prison Sentences (NCRP) All NCRP NonDrug NonDrug Drug Drug Drug Year 0.051** 0.052** 0.028** 0.012** 0.061** 0.044** 0.042** [0.001] [0.001] [0.001] [0.002] [0.006] [0.005] [0.006] % Black 0.015** 0.008** 0.009** [0.004] [0.003] [0.003] Change % Black -0.027** -0.085** -0.129** -0.122** -0.152** -0.075** -0.092** [0.006] [0.007] [0.010] [0.007] [0.024] [0.022] [0.024] Dissimilarity -1.074** -1.053** -1.864** -1.059** -1.353** [0.043] [0.058] [0.139] [0.183] [0.219] White Poverty -0.077** -0.074** [0.014] [0.014] Black Homicide 0.008** 0.005** [0.001] [0.002] White Homicide 0.120** 0.108** 0.096** 0.082** [0.003] [0.004] [0.012] [0.013] Black Unemployed -0.019** -0.033** -0.033** [0.003] [0.010] [0.010] Constant 5.270** 5.359** 3.703** 4.827** 1.803** 3.137** 3.216** [0.035] [0.048] [0.010] [0.077] [0.048] [0.170] [0.182] Observations 675 612 453 453 453 453 453 Number of States 44 39 29 29 29 29 29 R2 0.65 0.67 0.23 0.45 0.29 0.45 0.45 Model DF 4 4 2 5 3 7 8 Standard errors in brackets + significant at 10%; * significant at 5%; ** significant at 1% a)Total R for all independent variables for full sample = .66 b)Total R for all independent variables for White non-drugs = .45 and for drugs = .48. Table 4. Regression of Black Imprisonment and New Sentences on Independent Variables. All are Logged Rates Per 100,000 Black CPUS In Prison Black Sentences (NCRP) All All NCRP Non Drug Non Drug Non Drug Non Drug Drug Drug Drug Year 0.055** 0.047** 0.049** 0.008** 0.121** 0.123** [0.002] [0.002] [0.002] [0.003] [0.004] [0.004] Year Squared -0.014** -0.012** [0.001] [0.001] % Black -0.014** -0.015** -0.015** -0.009** -0.010** -0.015** -0.015** [0.000] [0.000] [0.001] [0.001] [0.001] [0.002] [0.002] Change % Black 0.071** 0.089** 0.055** 0.053** [0.008] [0.009] [0.011] [0.010] Dissimilarity -0.497** -0.642** -0.749** -0.648** -1.488** [0.081] [0.071] [0.088] [0.087] [0.410] Repub Gov 0.057** 0.051* 0.091** 0.078** 0.303** [0.020] [0.021] [0.018] [0.019] [0.077] White Poverty -0.037** -0.044** -0.071** -0.081** -0.083** -0.146** -0.199** -0.235** [0.005] [0.006] [0.007] [0.011] [0.011] [0.017] [0.018] [0.047] Black Poverty 0.008** 0.011** -0.029** [0.003] [0.003] [0.011] Black Homicide 0.005** 0.004** [0.002] [0.002] White Homicide 0.048** 0.046** 0.051** 0.104** 0.068** [0.006] [0.010] [0.010] [0.010] [0.024] Black Unemployed -0.006** 0.001 -0.027** [0.002] [0.003] [0.010] White Unemployed -0.045** -0.045** [0.011] [0.010] Constant 7.297** 8.018** 8.106** 5.544** 6.006** 6.453** 6.337** 5.998** 6.011** 8.086** [0.011] [0.092] [0.082] [0.020] [0.057] [0.088] [0.090] [0.163] [0.154] [0.573] Observations 675 675 612 453 453 453 453 453 453 453 Number of State 44 44 39 29 29 29 29 29 29 29 R2 0.45 0.51 0.52 0.18 0.31 0.40 0.41 0.61 0.66 0.37 Standard errors in brackets + significant at 10%; * significant at 5%; ** significant at 1% Total R for all independent variables for CPUS full sample = .53, for non-drugs = .41 and for drugs = .67 Table 5. Disparity: Black Rates Regressed on White rates and Independent Variables. All are Logged Rates Per 100,000 Black Imprisonment (CPUS) Drug Sentences (NCRP) Non-drug Sentences (NCRP) White Rate 0.478** 0.314** 0.334** 0.607** 0.274** 0.491** 0.491** 0.871** 0.583** 0.633** [0.049] [0.012] [0.011] [0.020] [0.034] [0.018] [0.025] [0.088] [0.044] [0.043] Year 0.038** 0.039** 0.018** 0.099** 0.087** [0.002] [0.002] [0.002] [0.005] [0.004] Year Squared -0.015** -0.012** [0.001] [0.001] % Black -0.015** -0.015** -0.013** -0.015** -0.012** [0.001] [0.001] [0.001] [0.001] [0.003] Change % Black 0.097** 0.098** [0.009] [0.010] Dissimilarity Repub Gov 0.076** [0.019] White Poverty -0.057** -0.057** -0.079** -0.139** [0.005] [0.004] [0.008] [0.013] Black Poverty 0.008** [0.002] Black Homicide 0.003** 0.002 [0.001] [0.001] White Homicide -0.054** 0.036** [0.006] [0.009] Black Unemployed 0.012** [0.002] White Unemployed -0.041** [0.008] Constant 4.708** 5.549** 5.603** 4.909** 4.519** 4.344** 4.165** 2.726** 3.572** 4.864** [0.256] [0.060] [0.050] [0.103] [0.124] [0.078] [0.116] [0.206] [0.085] [0.136] R2 0.31 0.46 0.57 0.70 0.10 0.52 0.60 0.42 0.71 0.78 Net R2 after White 0.15 0.26 0.39 0.42 0.50 0.29 0.36 Standard errors in brackets. Observations and states the same as for White and Black tables. + significant at 10%; * significant at 5%; ** significant at 1% Total R for all independent variables for imprisonment = .72, for non-drugs = .60 and for drugs = .80 Table 6. Whites: Regression of first difference of log imprisonment rates on independent variables -65 x -3 -4 -6 White CPUS White Drug Sentences White Non-Drug Yearly Mean 0.936** 0.887** 1.008** [0.041] [0.107] [0.066] % Black -0.001** -0.000** -0.001+ -0.002* -0.003** [0.000] [0.000] [0.001] [0.001] [.001] Change % Black -0.002** -0.001* [0.001] [0.001] Dissim Rep Gov -0.009* -0.007* [0.003] [0.003] Black Unempl 0.000* 0.000+ 0.009* 0.008+ 0.005 [0.000] [0.000] [0.005] [0.005] [0.005] D Black Unempl 0.017** 0.008 Constant 0.057** 0.011** -0.03 -0.001 -0.023 0.044 [0.004] [0.004] [0.054] [0.061] [0.053] [0.011] Observations 631 631 424 424 424 424 Number of State 44 44 29 29 29 29 R2 0.03 0.07 0.04 0.04 0.09 0.11 Model DF 4 5 1 2 3 4 a) No independent variables are significant predictors for White non-drug sentences. R2 for mean alone is .08, for time alone is .02. b) For White CPUS imprisonment, R2 mean alone is .05, R2 time is .00, R2 all independent variables is .10. c) For White drug sentences, R2 mean alone is .06, R2 time alone is .02, R2 all iindependent variables is .14. + significant at 10%; * significant at 5%; ** significant at 1% Standard errors in brackets Table 7. Blacks: Regression of first difference of log imprisonment rates on independent variables Black in Prison CPUS Black Drug NCRP Black Non-Drug NCRP Yearly Mean 1.099** 1.000** 0.787** [0.058] [0.043] [0.096] % Black Change % Black -0.006 0.019* [0.004] [0.008] Dissim 0.056** 0.055* 0.577** [0.018] [0.021] [0.134] Black Poverty 0.001** 0.004 0.002* 0.002* [0.000] [0.003] [0.001] [0.001] D Black Poverty