AMERICAN JOURNAL OF CRIMINAL JUSTICE, VOL. XVII, NO. 1, 1992

INCREASING IMPRISONMENT: A F U N C T I O N OF C R I M E OR SOCIOECONOMIC FACTORS?

Thomas M. Arvanites VUlanova University

INTRODUCTION Over the past three decades there has been a dramatic increase in the number of crime control bureaucracies in the United States. This increase is evident in a several of areas. In 1988, federal, state and local governments spent nearly 61 billion dollars on the criminal justice system (BJS, 1990a). This amount represents a 35 % increase since 1985, or an average annual increase in excess of 10%. Not only has this increase exceeded the rate of inflation, which was less than 4 percent annually over the same period, but the money spent on the criminal justice system as a percentage of all government expenditures increased from 2.9 % to 3.2 % (BJS, 1990a). The number of police per 100,000 population is another component of the crime control apparatus that has been increasing. Since 1965 when there were 2.3 officers per 1,000, the number has increased to 2.5 in 1975 and to 2.7 by 1988 (BJS, 1990b). The most dramatic increase, however, has been in the number of persons confined in federal and state prisons. Between 1980 and 1988 the number of people incarcerated in these facilities increased from 329,821 to 627,588 (BJS 1989), an increase of 90%. More important, the imprisonment rate increased from 154 per 100,000 population in 1981 (BJS, 1982) to 244 in 1988 (BJS 1989), an increase of 58%. The increase in imprisonment rates is frequently attributed to increasing crime. There is evidence, however, which suggests that the level of crime alone is insufficient to explain variations in imprisonment. As is evident in Figure I, the increase in the number o f persons in prison between 1983 and 1988 far exceeded the increase in the number of persons arrested for violent crimes. The trend was especially dramatic between 1986 and 1988 when arrests for violent offenses were relatively stable, yet the prison population continued to increase. Clearly, the increasing number o f drug arrests have contributed to the increase in prison populations, but,

20

O~ cO ! 0 cO

i i

i i

i J

i ~

i

I

i'.

i \ ,

!

i

,.

i

o) co

i

co co "o~

~

I

~)

I

i i

i

-~

i

z 0

L\

g

'-

!"

I

E

~>-

e

i

i

8 n

l i

i il

i

i

,.

a

ii

i

iI

i

P

i:j Ji

0 00 0I.~ 0 0 6 ~0 6 1.0 0 (spuesnoq•

1.0

~N

i

i

9i

J

0

I

iz

I

I0

-~

i~in I

o

60 6 1.~0

,0

-o~

i

I

J

21 as is evident below, the increase in imprisonment exceeded the increase in the total crime as well. A comparison of the increase in imprisonment rate versus the increase in the total crime rate reveals a similar trend. Between 1983 and 1988, the incarceration rate increased 36%, from 179 to 244 (BJS, 1984 and 1989). The index offense crime rate increased 10% (from 5,175 to 5,664 per 100,000 population) and the number of crimes increased 15% (from 12,108,600 to 13,923,086) CGCR 1989). Bridges et al. (1987) reported that violent crime rates have a "surprisingly limited influence on rates of imprisonment" (1987:352). Michalowski and Pearson (1990) reported that between 1970 and 1980, the rate of imprisonment increased 50% while the official crime rate increased only 39 %. Clearly, factors other than crime and arrest rates are contributing to the increasing level of imprisonment. Rather than being a rational response to the crime problem, the increase in imprisonment may be the result of a new "get tough" response towards crime (Box, 1987). There is little doubt that the U.S. has become considerably more punitive during the last decade. Throughout the 1980s approximately 75 % of the public responded that the penalties imposed on criminals were "not harsh enough" (Sourcebook, 1990). Support for capital punishment has increased from 60% in 1978 to almost 80% in 1988 (Sourcebook, 1990). At the end of 1990, there were 66 % more people on death row than in 1984 (2,356 vs 1,420) (BJS, 1990c). Politicians frequently adopt a tough stance on crime on the grounds that the public demands it. This is not to imply, of course that, the entire population supports the "get tough" approach, but the "Willie Horton" factor in the 1988 presidential campaign demonstrated quite clearly that candidates for office must be sensitive to the label that they are "soft on crime." This "law and order" approach has had a significant influence on sentencing. Beginning with Maine in 1975, many states have eliminated parole, and legislators are enacting more mandatory sentences and longer prison sentences. As of 1982, 37 states have mandatory prison sentences, and 27 impose determinate sentences, 15 of which use mandatory sentences exclusively (Duffee, 1989). The abandonment of the "rehabilitative" model in favor of the "just deserts" model has resulted in an increase in the length of the average prison sentence (Greenberg & Humphries, 1980). In addition to the proliferation of the "get tough" approach, voters in this country have increasingly supported candidates who adopt a more conservative platform. The impact of this new conservatism has not been confined to the criminal courts. Tax reform of the 1980s has clearly benefitted the upper class more so than the lower classes. The federal government appears less committed to affirmative action, and last year President Bush described the Civil Rights Act of 1991 as a "quota bill." At

22 the state level, abortion rights are being curtailed and welfare benefits are being reduced. Conflict theorists have long argued that crime alone was insufficient to explain the level of crime repression. It is often suggested that crime control, like any social policy, is more responsive to the interests of the economically and politically powerful (Quinney, 1977; Turk, 1969). Specifically, conflict theorists have argued that the criminal justice system will be used to control those groups of persons (the poor, unemployed and racial minorities) who threaten, or who are perceived to threaten the interests of the powerful. Although the research is inconsistent, there isconsiderable evidence that these extra-legal variables, such as race, poverty rates, and unemployment, do affect incarceration rates independent of crime rates. Given this new conservatism, it is plausible that these variables may become even more important determinants of incarceration rates. This study examines the increase in imprisonment rates. It does so for two reasons: first, and more obviously, because of the significant increase in the rate; and second, because with the exception of the death penalty, imprisonment is one of the most coercive forms of control.

PRIOR RESEARCH Race Conflict theorists argue that subordinate and culturally dissimilar groups are perceived to threaten the interests of the economically and politically powerful. There is little doubt that, in terms of economic and political power, nonwhites represent a subordinate group. Further, they are frequently viewed as a criminal threat. Swigert and Farrell (1976) reported that whites and social control authorities often hold criminal stereotypes of nonwhites. The presence of nonwhites is viewed as an indicator of a crime problem (Lizotte & Bordua, 1980), and the fear of crime is positively related to the presence of nonwhites (Liska et al., 1982). In analyzing the number of police officers per capita, Liska et al. (1981) and Greenberg et al. (1985) reported that when controlling for crime, the presence of nonwhites had a direct and significant effect. Support for the premise that race has an independent effect on imprisonment is mixed. Joubert et al. (1981) reported that percent black was the most important determinant of imprisonment rates. Carroll and Doubet (1983) have criticized this study for a number of reasons. Using violent crime rates instead of the total crime rate and including region and education variables they report that percent black has no effect. Examining state imprisonment rates in 1970 and 1980, Michalowski and Pearson (1990) reported "exceptionally high correlations between region and race (.823 in

23 1970 and .809 in 1980)" that raise some questions about findings Usuch as those by Carroll and Doubet" (1990:63). As a result, Michalowski and Pearson included region and not race in their 50 state analysis. They argued that the history of the south was conceptually more important than race. When they restricted their sample to non-southern states, percent black males was included in the analysis and was positively related to imprisonment. They reported, however, that the high correlation between percent black males and violent crime (.769 in 1970 and .628 in 1980) made it impossible to Udetermine whether race does or does not have an independent effect on imprisonment" (1990:67). Examining incarceration rates that included jail data as well as federal and state level data, Arvanites and Asher (1991) reported that percent nonwhite was positively related to the imprisonment rate. As with Michalowsld and Pearson, they found a positive effect in nonsouthern states as well as across the entire U.S. They also reported that utilizing total rather than violent crime rates significantlyreduced the correlation between crime and percent nonwhite, which was a problem in the above studies. Further, they demonstrated how the presence of high levels of multicollinearity can "mask" the effect of race. Bridges et al. (1987) examined the effect of minority population on the imprisonment rates of whites and nonwhites in the state of Washington. They reported that percent nonwhite population was unrelated to imprisonment rates of whites, but it was a significant predictor of the black imprisonment rate.

Unemployment In terms of economic threat, one of the most frequently examined variables has been the rate of unemployment. The choice of unemployment rates as an indicator of economic threat may be traced to the work of Rusche and Kirschheimer (1939), who suggested that the need to control surplus labor would lead to a positive relationship between unemployment and imprisonment, independent of crime rates. More recent theorists have suggested that the number of unemployed persons must be reduced so as not to provide evidence that the capitalistic economic system is not so successful after all (Mathiesen, 1997:77), and that the unemployed threaten the social relationships of production (Quinney, 1977: Spitzer, 1975). A number of studies have reported a direct positive relationship between unemployment and imprisonment (Box & Hale, 1982; Greenberg, 1977; Jankovic, 1977; Inverarity & McCarthy, 1988; Yeager, 1979). Greenberg and Yeager, however, failed to control for crime, so the effect of unemployment independent of crime cannot be determined.

24 The strongest support for the positive relationship between unemployment and imprisonment was found by Inverarity and McCarthy (1988). They compared the relationship between unemployment and prison admissions from 1948 through 1984, controlling organizational factors (e.g., prison capacity) that might drive imprisonment rates and changes in the age structure (an increase in the number of "crime prone age" males) that might be responsible for increases in the imprisonment rate. They reported that, while changes in the age structure do affect imprisonment rates, the direct positive relationship between unemployment and prison admissions remains when age is controlled for. Mixed support was reported for the "capacity hypothesis." Supportive of Berk et al. (1983), prison releases had a greater impact on prison admissions than unemployment. However, when controlling for prison releases, unemployment was positively related to imprisonment. Examining incarceration rates between 1976 and 1981, Galster and Scaturo (1985) reported little support for the hypothesis that increased level of unemployment results in an increase in the number of people incarcerated. They examined the relationship between unemployment and the rate of court commitments to prison, rate of conditional releases, rate of unconditional releases and rate of parole violation returnees. They reported a negative relationship between unemployment and court commitments to prison and no relationship between unemployment and the number of conditional releases. Within the southern states, however, unemployment was positively related to court commitments to prison and the number of parole revocations in 1980 and 1981. Using prison admission and release data from 1974 through 1979, Parker and Horowitz (1986) also reported very little evidence of an unemployment effect. More recent studies have also failed to document an unemployment effect. Both Arvanites and Asher (1991) and Michalowski and Pearson (1990) reported that unemployment was not related to the rate of imprisonment. Colvin (1990), using county data, also reported that unemployment was not related to imprisonment rates. None of these studies disaggregated prison data into admissions and releases. They used imprisonment rate based on one day population counts as the operational definition of imprisonment. The failure to document an unemployment effect should not be attributed simply to the use of aggregated prison data. While lnverarity and McCarthy (1988), Galster and Scaturo (1985) and Parker and Horowitz (1986) all decomposed prison population into prison admissions and releases, only Inverarity and McCarthy reported a positive relationship between unemployment and imprisonment. Of course, another explanation of the discrepant findings might be the differing time spans used. Inverarity and McCarthy's study spanned 36 years while the other studies spanned only six years.

25

Economic Inequality The relative effect of other economic factors (e.g., percent poor or income inequality) on imprisonment has not received the attention accorded to unemployment. Ono reason may be the inconsistent findings that have resulted from studies on other forms of social control such as police department size. Jacobs (1979) reported that economic inequality was positively associated with police per capita, while Greenberg et al. (1985) did not. Jackson and Carroll (1981) reported that economic inequality was not related to expenditures for police. Colvin (1990) found that, controlling for crime, the percentage of persons living in poverty was positively related to imprisonment rates. McCarthy (1990) reported that poverty was a significant predictor of sentences to prison but not to jail. When the analysis was restricted to urban counties, however, this relationship failed to achieve statistical significance. Neither Colvin nor Arvanites and Asher (1991) reported any relationship between per capita income and imprisonment rates at the county level.

DATA AND METHODS Conflict theorists suggest that the criminal justice system can be used as an instrument of class domination. The goal of this research is to examine the relative effects of extra-legal variables on the interstate variations in the incarceration rate. This research tests the following hypotheses. Controlling for crime, interstate variations in the imprisonment rates are: HI: positively related to thepercent of the population that is black. H2: positively related to the poverty level. H3: positively related the unemployment rate. Further, it is expected that the effect of these extra-legal variables will be greater in 1988 than in 1980.

Independent Variables This study utilizes a cross-sectional analysis of all 50 states and the District of Columbia for 1980 and 1988. Crime data were extracted from the Uniform Crime Reports and are expressed in rates per 100,000 population. The impact of crime was assessed using two measures. Several

26 studies (Michalowski & Pearson, 1990; Bridges et al., 1990; and Galster & Scaturo, 1985) included the violent crime rate only, while Colvin (1990) and Arvanites and Asher (1991) used total crime rate. In studies where only violent crime was included, it was argued that violent crime is the most relevant factor because it is the most serious crime, thus the most likely to result in state imprisonment. There are, however, theoretical as well as methodological reasons for using total crime rate. While it is true that violent offenders are more likely to be imprisoned, a significant portion of state inmates are incarcerated for non-violent offenses. In 1986 for example, 31% of state prison inmates were incarcerated for property offenses, 9 % for drug offenses and 6 % for public order and other crimes (BJS, 1988). Arvanites and Asher (1991) clearly demonstrated that the high level of multicollinearity present when violent crime is included in the model reduces the independent effect of race. Two types of extra-legal variables were examined. First was the racial composition of each state. The percent of population comprised by blacks was collected from the 1980 and 1988 Statistical Abstracts of the United States. Second, two economic indicators were examined. It is not the relative wealth of a state that presents a "threat" to the economic order, but rather the degree of income inequality that exists within a state. Conflict theory suggests that it is the "threat" to the economic elite that influences crime control. Thus, other things being equal, imprisonment rates should be higher in states with greater economic inequality. The percentage of the population living below the official poverty line was used as an indicator of economic inequality. Poverty data for 1980 were not available so 1979 data were used. These data were obtained from the Statistical Abstracts of the United States (1980). The 1988 poverty rate was obtained from the Institute for Research on Poverty at the University of Wisconsin-Madison. The effect of unemployment was also examined. This was expressed as the percentage of persons 16 years old and over who are out of work. These data were also obtained from the Statistical Abstracts (1980 and 1988).

Dependent Variables There has been some debate as to whether imprisonment data should be aggregated (e.g., rates or census) or disaggregated into admission and release data. Several researchers (Galster and Scaturo, 1985; and Inverarity & McCarthy, 1988) have followed the Berk et al. (1983) suggestion that prison data should be disaggregated. This argument is grounded on the premise that organizational constraints (i.e., capacity) are critical factors in explaining imprisonment rates. Recently, however, other researchers (Arvanites & Asher, 1991; Colvin, 1990; Michalowski and Pearson, 1990) have used aggregated data. Michalowski and Pearson argued that admission

27 and release data "are less representative of the size of the prison system as a proportion of the population than data regarding stock (one-day) populations" (1990:62). This argument seems plausible because it is possible for one system with larger numbers of admissions and releases than another system to actually confine a smaller percent of their population. Thus, prison census (measured in one-day counts), which represent a more accurate indicator of the states' actual level of incarceration, were used. The level of incarceration was operationalized as the rate of incarceration (IRATE) per 100,000 residents between the ages of 18 and 65. The rate was based on these ages because they are ones most at risk of imprisonment. IRATE was calculated by dividing the raw number of persons incarcerated in state and federal prisons by the number o f persons between thoSe ages, then multiplying by 100,000. These data were collected from Prisoners in 1981 (BJS, 1982) and Prisoners in 1988 (BJS, 1989).

RESULTS The primary purpose of this study was to test whether or not structural or socio-economic variables contribute to the explanation of imprisonment rates across states after controlling for crime rates. To perform that test, multiple regression techniques were utilized, with crosssectional data by state for the years 1980 and 1988, to distinguish the separate effects of crime rates and such structural variables as racial composition and the percentage of population living below the official poverty line. The zero-order correlations are reported in Table 1. The values below the diagonal are for 1980 and the ones above are for 1988. The values within the inner box are correlations among the independent variables. Several of these correlations should be emphasized. First, for both years, percent black had the strongest correlation with IRATE of all independent variables. Second, as expected, the correlation between the percent black and violent crime was quite high (.680 in 1980 and .710 in 1988). The correlations between total crime and percent black were significantly lower (.268 and .395). Table 2 presents four estimated equations using data from all fifty states plus the District of Columbia. Equations 1 and 2 control for the total crime rate and Equations 3 and 4 control for the violent crime rate. Equations 3 and 4 are presented for comparative purposes and because the violent crime rate has been used as an independent variable in other studies. All four equations have the same extra-legal variables identified in the methods section. Equations 1 and 3 explain 1980 IRATE while Equations

28 2 and 4 explain 1988 IRATE. The coefficients presented are standardized beta coefficients, and the numbers in parentheses are t-statistics. All four equations provide strong support for the hypothesis that percent black is positively related to IRATE, after controlling for crime. In each of the equations the beta coefficient for percent black exceeded the beta for either total crime or violent crime. The relative contribution of

T A B L E 1. -- Zero O r d e r Correlation Coefficients, All States

(a)

(b)

(c)

(d)

(e)

(t)

IRATE

Total Crime

Violent Crime

Percent Black

Poverty

Unempl oyment Rate

.764

.770

.063

.126

.806

.395

-.073

.050

~;~i~i~i~i~i~i~i~i~i~i~i~!~i~i .710

.040

.100

!ii;ililiiii!iiiililili~i!iiiii!i!i!iii!i!iil .295

.153

i>:+>>:+>:+:+:+:+:+:+:+:.1

(a) .::.!.!.i.::~.::.::.i~:: ! ii.::::i.::.::.i., .575 :::::::::::::::::::::::::::::::::::::::::::::::

(b)

.493

!~!~!~!~!~!~!~!~!~!~!~! !~!~i!~!

(c)

.694

.786

,..................................................1

(d)

.762

.268

. 80

(e)

.455

-.191

.198

.563

(13

.130

.029

.150

.165

iiiiiiiiiiiiiiiii~iiiiii~iiiiiiiiiiiii!iiiii!iiii .748 .156

iiiiiiiii!i!iii!i!iiiiiiiiiii~i!i!i!i!ili!i!!!!!ili! 9 I.;.i.:.;.i*:.i.i,;.i.:.:.:.l.:.:.i.i*]*i*:.:.:.i.:.:.l

Note: Values below the diagonal are for 1980; those above the diagonal are for 1988.

29

liiiiiiiii!iii!ii~,~ ~

~

~

'~iiiiiiiiiiiiiiiiiiiiiiiiiii ~ ~ ~

~ ~

~. ~

~ ~

~

iiiiiiiiiiiiiiiiiiiiiiiiiiiiiiii

iiii!ili!i~i~ililililili~i

0

.~

z

~

0

Z

0

30 percent black was greater in the models utilizing total crime (Equations 1 and 2) than the models with violent crime (Equations 3 and 4). This result may be the effect of the high correlations between violent crime and percent black (.680 in 1980 and .710 in 1988). While all the betas for crime and percent black were significant, note that in 1988 the beta for percent black was nearly three times greater than the beta for total crime (Equation 2), while in 1980 the relative contributions of percent black and crime are much closer (Equation 1). This result suggests that the percentage of the population that is black has become a more important determinant of imprisonment in 1988 than in 1980. The most surprising result involved the effect of poverty. It was positively related to IRATE in equation 1 but not in equation 3. In 1988, however, poverty was negatively related to imprisonment in both equations (2 and 4). This complete reversal of the effect of poverty would not be predicted by conflict or consensus theorists. Also contrary to expections, there was no support for the hypothesis that unemployment was related to imprisonment. Although the unemployment betas were substantially greater in 1988 than in 1980 they did not attain statistical significance. Because of the high correlation between unemployment and poverty (.748) it seemed plausible that one of these variables was "stealing" some of the effect of the other, thus rendering it insignificant. As a result, equations 2 and 4 were re-estimated, once including unemployment and excluding poverty, and once including poverty and excluding unemployment. Again, neither of these economic variables were significant. Liska et al. (1981) and Galster and Scaturo (1985) have reported that the effect of structural variables differs by region (south versus non-south). Table 3 presents the zero-order correlations for the nonsoutheru states ~. Again, the correlation between percent black and violent crime is much too high (.831 and .842) to draw any meaningful conclusions about the independent effect of race on imprisonment. Table Four presents the same four equations as Table 2. Once again, the highest betas were for percent black. Equation 8 should be interpreted cautiously. The results appear to suggest that violent crime is not related to the rate of imprisonment. The correlations between violent crime and percent black (noted above) render interpretation of this equation unreliable. The major discrepancy between the nonsouthern states and the entire country involved poverty. In the nonsouthern states poverty was not related to IRATE in either direction for either year. Again, the findings do not support the hypothesis that unemployment is related to imprisonment.

31

TABLE

3. -- Z e r o Order Correlation Coefficients, Non-Southern States

(a)

(b)

(c)

(d)

(e)

(f)

IRATE

Total Crime

Violent Crime

Percent Black

Poveay

Unempl oyment Rate

.801

.874

.012

.109

[]iiiiiiiiiiiiiiiiiii~liiiiiiililiiiiiiii] .814 ::::::::::::::::::::::::::::::::::::::::::::::

.567

.010

.107

.842

.045

.118

i::i:!:~:~:!:!::! !~.:.:+:-:.:.:.:.:.:.:.J

(a)

~i~i~i~i~~i~ii~i~ i~i~i~

(b)

.682

.657

~0,

.~03

.~,6

.~1 !ili~iii! ii~ -~8

(e)

.372

.020

.346

.383

(f)

.211

.122

.210

.197

_.0,4

~i~i~i~i!i~:ji:::i:i~i:~iii~:?:~ij:j::i:i:i:iil ~:~:~s~i~!~s~i .691 !ii:i:iii:i~i:iiii:ii:i:i:i:i:i:i:?:i:i:ii~i .106 i:~:~:~:~i:i~:~:~:~:~:~:!~i~i~i:~:

Note: Values below the diagonal are for 1980; those above the diagonal are for 1988.

32

~w

~D

E~

33

DISCUSSION AND CONCLUSIONS The primary purpose of this study was to examine the factors that explain interstate variations in the level of imprisonment. Due to the inconsistent ~ r c h findings, the question continues to be controversial. As prison census and rates set new records each year the question is an extremely important one. The simple explanation that increasing prison populations are the result of increasing crime is easily discounted by the fact that the increase in imprisonment in the U.S. has far exceeded the increase in crime. Conflict theory suggests that socio-economic variables are important determinants of imprisonment. The results of this study provide further support for the conflict view that racially dissimilar populations exert a positive effect on the level of incarceration when controlling for crime. More important, the effect of minority populations increased during the 1980s while the impact of crime decreased. All of the results, however, were not supportive of conflict theory. There was no support for the hypothesis that unemployment was related to imprisonment. The effect of poverty was the most perplexing. In 1980 it was positively related to the imprisonment rate, which is the relationship predicted by conflict theory. In 1988, however, it was negatively related to imprisonment. This finding may be due to an error in the specification of the model. Theoretically, it may be possible that crime increases imprisonment, which ultimately reduces poverty by incarcerating a segment of the population who tend to be poor. Thus, they are not included in the official poverty statistics. However, since many of these inmates have dependents, whom they cannot support, this does not seem very plausible. The findings reported here demonstrate that the level of crime is not the most important factor in explaining interstate variations in the level of imprisonment. Michalowski and Pearson (1990) reported a similar finding. More important, it appears the effect of crime has declined. Clearly, other factors contributes significantly to increasing levels of imprisonment. Berk et al. (1983) have suggested that prison capacity is a critical determinant in the level of imprisonment. It is plausible that prison population is more a function of prison space than actual crime. Throughout the 1980s prison expansion programs were occurring in the majority of states. In 1991, 44 states were expanding their prison systems. Since imprisonment increased at a greater rate than crime through most of the 1980s, the critical question is what is driving prison expansion. It might simply be a "get tough" attitude that has resulted in the building of more prisons, thus enabling the criminal justice system to incarcerate more people for longer periods of time, rather than a response to actual increases in crime. Surveys of the general population throughout the 1970s and the 1980s report that the majority of people believe that criminals are not punished harshly enough.

34 This perception may explain the political commitment to the construction of new prisons. Once built, at a cost averaging between $60,000 to 90,000 per cell, they will certainly be filled. The declining impact of crime suggests that availability may be an important variable. Of course, the overcrowding that exists in many prisons today suggests that society will imprison regardless of the available cell space. The most alarming finding was the increasing effect of percent black on the imprisonment rate. This discovery suggests that society may be becoming less tolerant of nonwhite populations and/or are feeling more threatened by them. An interesting question is whether the racial composition of the prison population has changed in the last ten years. It may well be that society's new punitive response is now, more than ever, disproportionately affecting blacks. This disparity would not be unique. Christianson (1981) documented quite well how the state prison system has become increasingly black. The work of Bridges et al. (1987) suggests that this increase may be due, in part, to the size of the black population. They reported that "nonwhites, but not whites, are particularly likely to be sentenced to prison in counties with relatively large minority populations" (1987:345). Other recent government policies (e.g. a limited retreat from its commitment to affirmative action, reducing entitlement programs and limiting appeals for death row inmates) have disproportionately affected black citizens. Finally, these results demonstrate that total crime rate is a better crime variable than violent crime for two reasons. Since a substantial portion of state prison inmates (approximately 45 %) are incarcerated for nonviolent offenses, the use of total crime is preferable because it is a more complete measure of crime for which people are imprisoned for. Statistically, the multicollinearity that is present when both percent black and violent crime are included in the regressions makes it impossible to determine the separate effects of race and crime. While this is a statistical finding it has theoretical implications. Models with high levels of multicollinearity can produce results that are unreliable. REFERENCES Arvanites, T. M. & M. A. Asher (1991). Factors influencing the level of coercive control in the U.S.. Presented at the American Society of Criminology. San Francisco, CA. Berk, R.S., S. Messenger, D. Rauma & J. Bereeothea (1983). Prisons as selfregulating systems. Law and Society Review 17:547-86. Box, S. (1987). Recession, Crime and Punishment. Totowa (NJ):Barnes and Noble.

35 Box, S., & C. Hale (1982). Economic crisis and the rising imprisonment rate in England and Wales. Crime and Social Justice 17:20-35. Bridges, G.S., R. D. Crutchfield, & E. E. Simpson (1987). Crime social structure and criminal punishment: White and nonwhite rates of imprisonment. Social Problems 34(4):345-360. Carroll, L., & M. B. Doubet (1983). U.S. social structure and imprisonment. Criminology 21 (3):449-556. Christianson, S. (1981). Our black prisons. Crime and Delinquency 27 (3):36475. Colvin, M. (1990). Labor markets, industrial monopolization, welfare and imprisonment:Evidence from a cross-section of U.S. counties. Sociological Quarterly 31 0):441-57. Galster, G., & L. Scaturo (1985). The U.S. criminal justice system: Unemployment and the severity of punishment. Journal of Research in Crime and Delinquency 22 (2): 163-89. Grcenberg, D. (1977). The dynamics of oscillatory punishment process. Journal of Criminal Law and Criminology 68:643-51. Greenberg, D., R. Kessler, & C. Lol~in (1985). Social inequality and crime control. Journal of Criminal Law and Criminology 76:684-704. Institute for Poverty Research (1991). State poverty rates for whites, blacks and Hispanics in the late 1980s. Focus 13 (1) University of Wisconisn Madison. Inverarity, J., & D. McCarthy (1988). Punishment and social structure revisited: Unemployment and imprisonment in the United States, 1948-1984. Sociological Quarterly 29 (2):263-279. Jackson, P. I., & L. Carroll (1981). Race and the war on crime: The sociological determinants of municipal expenditures in 90 non-southern U.S. cities. American Sociological Review 46:290-305. Jacobs, D. (1979). Inequality and police strength. American Sociological Review 44:913-24. Jankovic, I. (1977). Labor market and imprisonment. Crime and Social Justice 8:17-31 Joubert, P., J. S. Picou, & W. A, Mclntosh (1981). U.S. Social Structure, Crime and Punishment. Criminology 19 (3):344-59.

36 Liska, A. E. (1987). A critical examination of macro perspectives of crime control. Annual Review of Sociology (13):67-88. Liska, A. E., J. Lawrence. & M. Benson (1981). Perspectives on legal order: The capacity for crime control. American Journal of Sociology 87 (2):41326. Liska, A., J. Lawrence, & A. Sanchirico (1982). Fear of crime as a social fact. Social Forces 60:760-71. Lizotte, A., & D. Bordua (1980). Firearms ownership for sport and protection. American Sociological Review, 45:229-43. Mathiesen, T. (1974). The Politics of Abolition. London: Martin and Robertson. McCarthy, B. (1990). A micro-level analysis of social structure and social control: Intra-state use of jail and prison confmement. Justice Quarterly 7 (20):32540. Michalowski, R., & M. Pearson (1990). Punishment and social structure at the state level: A cross-sectional comparison of 1970 and 1980. Journal of Research in Crime and Delinquency 27(1):52-78. Parker, R. N. & A. Horwitz (1986). Unemployment, crime and imprisonment: A panel approach. Criminology 24(4):751-73. Quinney, R. (1977). Class, State and Crime. New York: McKay. Rusche, G., & O. Hirchheimer (1939). Punishment and Social Structure. New York: Columbia Spitzer, S. (1975). Towards a Marxian theory of crime. Social Problems 22:368401. Swigert, V., & R. Fan'el (1976). Murder, Inequality and the Law: Differential Treatment in The Legal Process. Lexington, Mass.: Health. Turk, A. (1969). Criminality and Legal Order. Chicago: Rand McNally. Yeager, M. (1979). Unemployment and imprisonment. Journal of Criminal Law and Criminology 70:586-588. U.S. Census Bureau (1980). Statistical Abstracts of the United States. U.S. Government Printing Office: Washington, D.C. (1988) Statistical Abstracts of the United States. U.S. Government Printing Office: Washington, D.C.

37 U.S. Department of Justice (1982). Prisoners in 1981. Bureau of Justice Statistics, Washington, D.C. (1983). Crime in the United States. Federal Bureau of Investigation. (1984). Prisoners in 1983. Bureau of Justice Statistics, Washington, D.C. (1988a). Crime in the United States. Federal Bureau of Investigation. (1988b). Proftle of State Inmates, 1986. Bureau of Justice Statistics, Washington, D.C. (1989). Prisoners in 1988. Bureau of Justice Statistics, Washington, D.C. (1990a). Justice Expenditure and Employment, 1988. Bureau of Justice Statistics, Washington, D.C. (1990b). Source book of Criminal Justice Statistics. Statistics, Washington, D.C. (1990e). Capital Punishment, Washington, D.C.

Bureau of Justice

1990. Bureau of Justice Statistics,

38

ENDNOTES .

Southern States were defined as Alabama, Arkansas, Florida, Georgia, Kentucky, Louisana, Maryland Mississippi, North Carolina, South Carolina, Texas and Virginia.

INCREASING IMPRISONMENT: A FUNCTION OF CRIME OR SOCIO-ECONOMIC FACTORS?

Between 1980 and 1988 the prison census increased 90% while the imprisonment rate in the U.S. increased 58 %. While these increases are frequently attributed to increased crime, they have far exceeded the increase in crime. Conflict theorists predict that extra-legal variables directly affect imprisonment rates, independent of crime. Utilizing state level data from these two years, this study investigates whether racial and economic variables influence the level of imprisonment. Percent black was not only the strongest predictor of imprisonment rates in both years, but its impact increased bewteen 1980 and 1988. Unemployment was not related to imprisonment in either year. The poverty rate was positively related in 1980 but negatively related to the rate of imprisonment in 1988. A separate analysis was conducted on data from nonsouthern states, and the results were the same. Thomas M. Arvanites earned his Ph.D. at the State University of New York at Albany. He is a member o f the Department of Sociology at Villanova University. His research interests include determinants of the level of coercive control and the interrelationship between the criminal justice and mental health systems.