Labor market and imprisonment

LABOR MARKET AND IMPRISONMENT Author(s): Ivan Jankovic Source: Crime and Social Justice, No. 8 (fall-winter 1977), pp. 17-31 Published by: Social Justice/Global Options Stable URL: http://www.jstor.org/stable/29766015 Accessed: 24-03-2017 15:39 UTC JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected]. Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at http://about.jstor.org/terms Social Justice/Global Options is collaborating with JSTOR to digitize, preserve and extend access to Crime and Social Justice This content downloaded from 213.198.240.67 on Fri, 24 Mar 2017 15:39:14 UTC All use subject to http://about.jstor.org/terms LABOR MARKET AND IMPRISONMENT* Ivan Jankovic** I. INTRODUCTION The prevalent conception of punishment treats it as an epiphenomenon of crime, a reaction by the sociological literature on punishment is concerned with its deterrent effects. At times, this theoreti? cal concept of punishment is also reflected in the design of empirical studies, as when punishment (e.g., number of prison admissions) is used as an state to the criminal's breach of the legal order. index of crime. This conception was expressed by Immanuel Kant Among the first to successfully break the supposed bond between crime and punishment by (1&&7, 2:194): "Juridical punishment can never be administered merely as a means for promoting demonstrating how penal policies are shaped by economic and political considerations, were Georg another good, either with regard to the criminal himself or to civil society, but must in all cases be Rusche and Otto Kirchheimer (1933, 1939). Their imposed only because the individual on whom it is theory of punishment, according to which "every system of production tends to discover punishments inflicted has committed a crime." Alternately, which correspond to its productive relationships," punishment is seen as a measure to prevent crime and as a weapon in the war against crime," but it is provides the starting point for the present essay. always a part of a dyad, one side of the coin whose We shall try to further transcend the bond which is supposed to exist between crime and punishment, other face is represented by crime. and to examine the latter as an independent The underlying assumption here is that forms and intensity of punishment must be determined by phenomenon in its manifold relationships to the forms and magnitude of crime, and that the social and economic structure. primary function of punishment is to revenge, prevent, contain and decrease crime. In sociology, this is reflected by the fact that the largest part of To state that punishment is not a simple consequence of crime is not to deny that most judges, when passing a sentence, sincerely believe themselves to be reacting to the defendant's crime. Nor is it to deny that theories of punishment reflect a sincere concern over crime and a belief meetings of the Pacific Sociological Association in that punishment is a necessary consequence of Sacramento, April 21, 1977. Both versions are adapted from an unpublished Ph.D. dissertation, "Punishment and crime. The statement implies that theories, as the Postindustrial Society: A Study of Unemployment, well as practice, of punishment reflect prevailing * An earlier version of this paper was presented at the Crime and Imprisonment in the United States," Depart? ment of Sociology, University of California, Santa Barbara. ** Ivan Jankovic is presently preparing and translating Marxist writings in Western criminology for the journal Marxism Today, which attempts to keep the Yugoslav public informed about Marxist scholarship abroad. Jankovic is scheduled to join the faculty at the University of Kragujevac in 1978. ideologies which are, in turn, partly determined by economic requirements of concrete systems of material production. So, for example, societies plagued by shortages of labor (e.g., sixteenth century Germany) develop ideologies which emphasize man's duty to work (the Protestant ethic); those faced with an oversupply of labor (e.g., nineteenth century England) resort to ideologies which make work one's right, to be 17 / Crime and Social Justice This content downloaded from 213.198.240.67 on Fri, 24 Mar 2017 15:39:14 UTC All use subject to http://about.jstor.org/terms fought for in the labor market (the laissez faire of liberal capitalism). All theories of punishment must apply to con? For felonies and misdemeanors, a fine is by no heimer applied theirs to Western societies from the means a typical punishment. In fact, the use of fines seems to be declining. Less than 1% of all felons sentenced in California superior courts in 1974 were fined (California Bureau of Criminal Statistics, 1976:14). Consistently, my analysis of imprisonment in advanced capitalist countries. The use of imprisonment could be explained by condi? were imposed as a condition of probation.) More crete penal systems in specific historical periods and socioeconomic systems. Rusche and Kirch Middle Ages to the 1930s, but encountered some difficulties in explaining the continuing use of tions of general labor shortages, when convicts could be profitably exploited. But it defies such explanation in capitalist societies which are not only faced with a permanent oversupply of labor but in which freedom of labor is the essential condition of its productivity, and, therefore, con? vict labor cannot be profitably exploited. Ulti? mately, Rusche and Kirchheimer were forced to dismiss imprisonment as an "irrational" penal meas? ure in developed capitalist countries. The persistent use of imprisonment in the most advanced capitalist societies in the final quarter of the twentieth century, however, demands a better explanation. If the proposition that "every system of production tends to discover (and use) punish? ments which correspond to its productive relation? ships" is true, then imprisonment must be meeting some needs of advanced capitalist economies. The main task of this essay will be to explore the applicability of the Rusche-Kirchheimer theory to contemporary Western societies. Specifically, the potential connections between use of imprison? ment and the conditions of the labor market will be examined. II. A CRITIQUE OF THE RUSCHE-KIRCHHEIMER THEORY In their treatment of modern penal systems, Rusche and Kirchheimer have opened themselves to two major criticisms. The first is that they fail to provide an explanation for the continued use of imprisonment - a punishment which does not seem to correspond to productive relationships of ad? vanced capitalism. The second is that they overem? phasize the use of fines as the typically capitalist punishment. The fine is, no doubt, better suited to a capitalist than to any other economy and it is most widely used in capitalist countries. However, it has not turned, as Rusche and Kirchheimer believe, into the dominant punishment of the twentieth century - at least with reference to felonies and serious misdemeanors. The early twentieth century trend toward an increase in the frequency of fines has been replaced by increased use of probation. 2,250 sentences (mainly for misdemeanors) in Sun? shine County, California (see below), showed only 18% to be fines. (In another 20% of all cases, fines interestingly, the use of fines in Sunshine County has been steadily declining over the years, from 24% in 1970 to 12% in 1974. The magnitude of fines has decreased as well, in spite of inflation, from a mean of $160 in 1970 to $130 in 1974. The dominant (most common) punishments in postindustrial societies - if we are to judge from contemporary American data - are incarceration (jail and prison) and probation, in that order, and often in various combinations. The most frequent single sentence passed by California superior courts in 1974 was probation with a jail term as a condition of the probation order (46.7%). Custodial sentences (with or without probation) were awarded to 81% of the persons sentenced, while probation (with or without incarceration and/or fine) accounted for 69% (California Bureau of Criminal Statistics, 1976:14). In the Sunshine County sample, the most common single punishment was jail (25%). All custodial sentences (with or without probation and/or fines) accounted for 45% of the total, and probation (with or without jail) accounted for 50%. Probation orders in California superior courts grew from 52% of all sentences in 1966 to 69% in 1974 (California Bureau of Criminal Statistics, 1976:14). Similarly, the percentage of probation sentences in Sunshine County increased from 23% in 1970 to 42% in 1974. With regard to imprisonment, Rusche and Kirchheimer have argued that its early forms were introduced in order to provide needed forced labor. Forced labor has no economic justification, how? ever, within a capitalist system of production, where freedom of labor is the conditio sine qua non of its productivity. It might be noted, parentheti? cally, that prison labor, initiated solely as a source of profit, assumed a purely punitive character in the nineteenth century, and is justified in the twentieth century by its alleged educational and therapeutic value. Reaching this impasse, Rusche and Kirchheimer are forced to explain solitary confinement, a typical feature of the nineteenth century prison, as an irrational punitive resp? onse - evidence of "a mentality which, as a result of surplus population, abandons the attempt to find a rational policy of rehabilitation and conceals this fact with a moral ideology (i.e. penitence)" (1939:137). Fall-Winter 19771 18 This content downloaded from 213.198.240.67 on Fri, 24 Mar 2017 15:39:14 UTC All use subject to http://about.jstor.org/terms By accepting the "irrationality" of imprison? information about the probationer, a task for which ment, Rusche and Kirchheimer appear to have cybernetic information-processing systems are underestimated the heuristic value of their own theory. In particular, they have not pursued two hypotheses which are implicit in their work and ideally suited. By keeping the probationer in the community, probation does not interfere with his or her employment and so does not disrupt the production process. In fact, maintenance of a steady employ? essay. ment is a standard condition of probation orders. In The first of these is that there is a negative relationship between economic conditions and this way, a part of the labor force is monitored and severity of punishment: when economy is bad, the controlled by the state, while being actively punishments are more severe. The second hypo? engaged in the production process. Finally, proba? tion is cheaper than imprisonment, an argument thesis deals with the relationship between the labor market and forms of punishment. When labor is which has historically carried much weight in all scarce, Rusche and Kirchheimer note, punishments movements for penal reform. The cost to the state per probationer in California in 1974 was $1,150, as are attempted which make greatest use of convict which are formulated in more detail in the present compared with $4,112 per jail inmate and nearly $10,000 per prison inmate (California Bureau of ments which are wasteful of labor (e.g., death Criminal Statistics, 1975:15, 17). labor (the house of correction, transportation, etc.). Conversely, when labor is abundant, punish? penalty) may be used. It generally so happens that punishments which preserve labor are also less severe than those which waste it, but this is only incidental to the primary purpose of exploitation of labor. This hypothesis works well when applied to precapitalist societies in which labor could be Rusche and Kirchheimer's difficulties in explaining the use of imprisonment in advanced capitalist societies stem from their insistence on the condition of exploitability of convicts' labor. The workhouses and some early forms of imprison? ment are easily explicable as attempts on the part forced and yet productive, but it breaks down when of the state to mitigate periodic shortages of labor market and imprisonment. "except as a punishment for crime whereof the party shall have been duly convicted." But ad? vanced capitalist societies are characterized by a applied to advanced capitalist countries. What Rusche and Kirchheimer have failed to do is to provide an alternative connection between the III. PUNISHMENT IN POSTINDUSTRIAL SOCIETIES The two questions which have yet to be an? swered within the framework of the Rusche Kirchheimer theory are: 1) in what ways does probation correspond to the productive relation? ships of advanced capitalism? and 2) given the persistence of incarceration, what functions, if any, does this sanction serve for advanced capitalist economies? Little more than speculation can be offered in response to the first question. Postindustrial societies are characterized by a decisive shift from manufacturing to service and information-process? ing activities, with the concomitant development of appropriate technologies (Levitan et al., 1976:1). Increasing use of probation is consistent with both these developments. The relationship between a probationer and a probation officer is the epitome of a service relationship. In modern correctional literature, probationers are typically referred to as "clients," probation officers as "agents," and proba? tion as "service" (cf. Remington et al., 1969:793 814). At the same time, probation supervision requires gathering and monitoring of extensive labor. Similarly, the United States Constitution in 1865 prohibited slavery and involuntary servitude permanent oversupply of labor. In addition, convict labor cannot be exploited because forced labor cannot produce profit in capitalist economies (on the question of forced labor, see Evans, 1970). It is possible, however, that the exploitability of labor is not the crucial intervening variable in the relationship between contemporary, postindustrial economic conditions and punishment. One of the most striking features of capitalist economies is that they are always faced with an oversupply of labor. A leading British economist, Lord Beveridge (1930:70), noting this peculiarity, asked a rhetorical question: "Why should it be the normal condition of the labor market to have more sellers than buyers, two men to every job and not at least as often two jobs to every man?" His own answer had to do with the fragmentation and organizational imperfections of the market itself. But Marx had earlier given a structural answer which seems superior to Beveridge's. In order to survive, Marx noted, capitalist economies need to maintain a permanent reserve of surplus labor which can at short notice be coupled with newly invested capital. This reserve army of labor is created not by fiat or from any ulterior motives but by the technological processes essential to capitalism. Thus, mechanization and automation of established industries increase the capital-to-labor 19 / Crime and Social Justice This content downloaded from 213.198.240.67 on Fri, 24 Mar 2017 15:39:14 UTC All use subject to http://about.jstor.org/terms ratio in favor of capital, minimizing the contribu? tion of labor. But labor is the only source of newly created value and, hence, of profit. Accordingly, capital has a tendency to move into new areas of production, likely at first to be labor-intensive, offering higher potential profits. When these areas are mechanized, the cycle is repeated. The reserve army of labor, then, is a necessary condition for the rapid movement of capital (Marx, 1867:628-40). A reserve army of labor also places an effective limit on economic demands of employed workers. The very existence of an army of unemployed persons reminds the employed laborer of his expendability. An economist who argues that "minimum" unemployment is unavoidable in a "free" society, formulates this point with a disarm? ing frankness: "The labor market should never be so tight that workers have no incentive to be on their toes" (Copeland, 1966:3). There is widespread agreement among econo? mists that the "officially" unemployed represent required to maintain social harmony - to fulfill the state's 'legitimization' function" (O'Connor, 1973:7). These projects and services are "social expenses" of the state. Two main components of the state's effort to support, and thus control, the surplus population are the social welfare system and the criminal justice system. Given the persistence and the magnitude of the surplus population in ad? vanced capitalist countries, imprisonment may serve to contain a fraction of it and to manipulate its size. IV. PRESENT RESEARCH A. Hypotheses The first hypothesis to be tested is that imprisonment and unemployment co-vary directly. The independent variable is unemployment, and the expectation is that a rise in unemployment will lead to an increase in prison commitments and prison population. This is a restatement of Rusche and Kirchheimer's "severity" hypothesis: when the economy is bad, punishments are more severe. Unemployment is taken as an index of the state of the economy, and imprisonment as an index of severity of punishment. Imprisonment data indicate severity of punish? ment in the following ways. First, imprisonment is taken to be the most severe form of punishment in contemporary American society. Second, the data on annual prison admissions indicate the frequency of punishment, which is one component of the LNS only a part of the reserve army of labor. There is disagreement, however, as to how large this part is. Some believe that the officially unemployed are severity dimension. Finally, the data on number of prisoners incarcerated on a given day every year serve as a very rough index of the magnitude of punishment. Holding admissions constant, an in? crease in prison population indicates that the only the tip of an iceberg of surplus labor, which includes" the sporadically employed; the part-time average sentences served are longer. workers, form a reserve for 'female occupations'; the armies of migrant labor, both agricultural and industrial; the black population with its extraordi? narily high rate of unemployment; and the foreign reserves of labor" (Braverman, 1974:386). Others refer to "discouraged workers" but tend to be vague about their numbers (Levitan et al., 1976:119). Thus, under conditions which make it profitable to maintain a permanent oversupply of labor - what Marx (1967:630-31) called the "surplus population" or "reserve army of labor" - imprisonment can be used to regulate the size of the surplus labor force. deterrence dictates an intensification of punish? ment in order to combat the assumed increased employed; the mass of women who, as house Punishment is expected to be more severe during economic crises because the policy of temptation to commit crime. Furthermore, intensi? fied punishments help preserve the socioeconomic order, threatened at times of economic crises, regardless of the stated penal policies. The same hypothesis could be deduced from This surplus population depends directly on the other theoretical models. Most notably, as we have seen, the widely held assumption that high unem? ployment (and economic crises in general) results in increased criminal activity, also implies that high unemployment will result in more punishments, including imprisonment. It will be necessary, there? network of "projects and services which are activity (as indicated by number of crimes known state for its economic welfare. It is supported by a fore, to control for the influence of criminal Fall-Winter 1977/ 20 This content downloaded from 213.198.240.67 on Fri, 24 Mar 2017 15:39:14 UTC All use subject to http://about.jstor.org/terms to the police and number of arrests) on the relationship between unemployment and imprison? way regardless of the conscious goals of the state's policies. Officials who commit persons to prison ment. This observation leads to the following may be motivated to reduce crime, provide bed and refinement of the first hypothesis: board for destitute criminals, provide skilled (1) Unemployment and use of imprisonment co vary directly, regardless of the volume of crime. Or: (la) The correlation between unemployment and imprisonment is significantly greater than zero, regardless of changes in the volume of criminal activity. As stated, the hypothesis implies that imprison? ment could increase even if crime were decreasing, provided that unemployment is rising. (The reverse is implied as well.) The second hypothesis to be tested is that increased imprisonment functions to reduce unem? ployment. This "utility" hypothesis asserts that the effect of changing penal policies is reflected in changes in the conditions of the labor market. It is derived from Rusche and Kirchheimer's theory, which suggests that each socioeconomic system invents and uses punishments which correspond to its productive relationships. When applied to the use of imprisonment in advanced capitalist coun? tries, this theory suggests the hypothesis that imprisonment may be used to remove a part of the surplus population from the labor market. Recently, the same hypothesis has been indepen? dently advanced in at least two essays. (2) In order to test this hypothesis it is first necessary to determine the magnitude of the potential impact of incarceration on unemploy? ment. This is an exploratory question, designed to provide some idea of the relationship between the two variables. The next step is to reverse the implied causal order between imprisonment and unemployment and to test whether the size of prison population (and admissions) at time ti has a negative effect on unemployment rates at times t2, t3, etc. This hypothesis can be stated as follows: (2) The size of the prison population co-varies inversely with lagged unemployment rates. Or: (2a) The inverse correlation between the size of the prison population and lagged unemployment rates is significantly greater than zero. It should be emphasized again that the present argument is not concerned with the motivation of state officials who impose and administer punish? ments, or with their understanding of and rationali? zations for specific penal policies. What matters is the effect that penal policies may have on the national economy. If this effect is a reduction of unemployment rates, the finding will give credence to the idea that imprisonment as a punishment in capitalist economies functions, in part, to regulate the labor market. Imprisonment may function this mechanics for work in a particular prison industry or something else. Nevertheless, the effect of their actions may be the regulation of the labor market. B. Data Two different sets of data were used in testing the two hypotheses. The first set is nationwide statistics on imprisonment rates for the United States, 1926-1974, and on unemployment rates and certain other demographic data for the same period. The second set is statistics obtained in a mid-sized California jurisdiction, called here Sun? shine County. These include unemployment and imprisonment rates by month for an eight-year period (1969-1976). The national statistics were obtained from the U.S. Statistical Abstracts and from the publication of the Bureau of Prisons, National Prisoner Statis? tics: Prisoners in State and Federal Institutions for Adult Felons. They include: 1) number of persons detained on a given day (usually December 31) each year; 2) number of persons entering prisons each year either after being sentenced to prison by a court, or as violators of an earlier conditional release from prison; 3) number of persons released from prison during each year either conditionally (mainly on parole), or unconditionally (mainly at expiration of sentence). Unemployment data were drawn from the U.S. Statistical Abstracts and include size of the civil ian labor force (in thousands), number of persons unemployed (in thousands), and unemployment rate (percent unemployed in total labor force). The unemployment rates are annual averages, based on seasonally adjusted monthly rates. Population figures were also taken from the U.S. Statistical Abstracts. They refer to resident civilian population. Imprisonment rates (per 1,000) were computed on the basis of this population. The size of the armed forces was obtained from the same source. Crime and arrest data came from the FBI publication Uniform Crime Reports. Crime data include numbers reported annually to the FBI of "seven major crimes" (murder, assault, rape, rob? bery, burglary, theft and vehicle theft). From 1937 to 1957, the figures are based on reports from 353 cities with 25,000 or more inhabitants; since 1957 they reflect a wider reporting base, with rural and suburban, as well as urban communities included. No nationwide crime data exist prior to 1937. Arrest data from 1932 to 1952 are based on examination of fingerprint cards filed with the FBI 21 / Crime and Social Justice This content downloaded from 213.198.240.67 on Fri, 24 Mar 2017 15:39:14 UTC All use subject to http://about.jstor.org/terms and, since 1952, on reports submitted to the FBI by local law enforcement agencies. Again, no nation? wide arrest data are available prior to 1932. In Sunshine County, monthly labor force statis? tics were obtained from the County Employment and Training Department. This department collects weekly information on unemployment claims and on numbers of employed persons, and tabulates them by month. On the basis of the employed and unemployed categories, the size of the total (civilian) labor force and unemployment rates are computed. The same department provided monthly estimates of the total county population. Crime and arrest data were not available on a monthly basis for the entire county. Only one law enforcement agency in the county, the Sheriff's Department, could provide monthly counts of reported crimes and arrests made within its juris? diction for the period under study. The Sheriff's Department accounts for over one third of all crimes reported and arrests made in the county annually, and its share is fairly constant over time. However, the data base changed in June, 1971, when one of the cities policed under contract by the Sheriff's Department acquired its own police force. The effect of this change in reporting base was examined by use of a dummy variable, and was found to be nonsignificant. Information about the Sunshine County jail population was obtained from jail records, and it consists of average daily population for each month, from January, 1969, to December, 1976, inclusive. Except for the first year (1969), the total population was broken down into sentenced and unsentenced categories. These statistics were supplemented with data previously gathered by this author for a study of sentencing patterns in Sunshine County. The sentencing study examined dispositions of six cate? gories of offenses, of which four were misdemea? nors and two felonies. The total sample included 2,250 cases disposed of in the Sunshine County Municipal Court and Superior Courts over a five year period (from January, 1970, to December, 1974, inclusive). The sample was random, strati? fied by year (450 cases for each year) and by offense. The offenses included: robbery (N=75), burglary (N=625), misdemeanor drunk driving (N=500), drunk and disorderly (N=450), being under the influence of narcotics (N=450) and possession of marijuana (N=250). imprisonment and unemployment values. The rou? tine procedure was to solve regression equations in which the dependent variable was some index of imprisonment and the independent variables were unemployment and population. In this way, it was possible to measure the impact of unemployment on imprisonment, holding population constant. The same technique was used to control for the impact of crime on imprisonment. Crime was added to the regression equation as an independent variable, which made it possible to assess its relative contribution to an explanation of variance in imprisonment. The multiple regression technique is described by Blalock (1972:429-68) and by Nie et al. (1975:320-42). It is applicable to the present data, which are all represented by ratio-level variables. In the present context, it is used as a descriptive technique, intended to decompose and summarize the assumed linear dependence of imprisonment on unemployment and crime. At the same time, inferential statistics (F ratios and T tests) are used to assess the statistical significance of observed linear relationships. Since the present data represent time series, it was necessary to consider the autocorrelation effect (see Pindyck and Rubinfeld, 1976:106-20). In the present study, the Durbin-Watson d statistic was used to test for the presence of serial (auto-) correlation. This statistic tests the null hypothesis that no serial correlation is present (Pindyck and Rubinfeld, 1976:113). Whenever serial correlation was established, the standard procedure was to purge its effects by using an AR(1) model and to report the adjusted regression results. D. Findings and Analysis The first step in the analysis was to determine the correlation between unemployment and various indices of imprisonment, controlling for population growth. Six different regressions were run to examine the effect of unemployment on different indices of prison population. Results for the United States, 1926-1974, are summarized in Table I, which suggests that as the total number of unem? ployed persons increases, the total number of persons present in and admitted to prisons also increases. The partial correlation coefficient between total prison population and unemployment, with population held constant, is .43. The unstandardized regression coefficient indicates that an increase of C. Methods The basic technique used in the present study was that of multiple linear regression. Regression analysis was used, with population included as an independent variable, in order to standardize the 1,000 unemployed persons corresponds to an in? crease of 2 prison inmates. The coefficient of determination suggests that 32% of the variation in prison population is accounted for by the joint action of unemployment and civilian population. Fall-Winter 19771 22 This content downloaded from 213.198.240.67 on Fri, 24 Mar 2017 15:39:14 UTC All use subject to http://about.jstor.org/terms All these results are significant at p=.01 or better. Essentially the same results are obtained when the due to the extraordinary shortage of labor created by the mobilization of the nation for the war effort. In order to eliminate the effects of the Great actually admitted to prisons. When the prison population is broken down into Depression and World War II, a set of regressions were run on the time series of 28 years after the its two major jurisdictional components (state vs. federal institutions), it appears that the above war (1947-1974). The results, reported in Table 2, correlation is due to the relationship between show that prison admissions were particularly unemployment and state prison population, and that responsive to changes in unemployment. For every no significant relationship exists between unem? 1,000 additional unemployed persons, there were ployment and federal prison population. In fact, 4.36 additional admissions. The partial correlation the regression and correlation coefficients asso? coefficient between state prison admissions and ciated with the impact of unemployment on federal unemployment, with population held constant, was prisoners are negative, ranging from -.03 to -.05. .49. Unemployment and population growth, opera? The overall positive correlation is due to the fact ting jointly, accounted for 50% of the variance in that federal inmates account for only one tenth of total prison admissions. Similar but somewhat the total prison population. weaker results were obtained for prisoners present in state institutions (R2=.26) and all institutions Analysis of different subperiods within the dependent variable is the number of persons national sample showed that the hypothesized (R2=.27). positive relationship did not obtain during the For the federal prison population, the regression period of the Great Depression (1930-1940). All and correlation coefficients were higher than the regression and correlation coefficients for this coefficients for the entire 49-year period. period were negative (-.02 to -.42), although none Although not statistically significant, they have the expected (positive) signs and approach the lower was statistically significant. During World War II, unemployment and imprisonment both decreased, producing high positive correlations, most likely limits of statistical significance. Further investiga? tion revealed that, with the passage of time, the 23 / Oime and Socw/ Justice This content downloaded from 213.198.240.67 on Fri, 24 Mar 2017 15:39:14 UTC All use subject to http://about.jstor.org/terms relationship between unemployment and the popu? lation of federal prisons became much stronger. As shown in Table 3, for the 15-year period since 1960 (1960-1974), the correlation between unemploy? ment and total federal prisoners, and the correla? tion between unemployment and federal prison admissions, are both very high. Unstandardized regression coefficients indicate that an increase of 1,000 unemployed persons corresponds to an increase of 1.69 and 1.31 federal prisoners present and received, respectively. Partial correlation coefficients are .81 and .82. The R(2) statistic indicates that almost 70% of the variation in federal prison population and federal prison admis? sions is attributable to the joint effect of unem? ployment and population. In order to establish the effect of changing patterns over time on the overall results, a regression was run with two dummy variables designed to distinguish the prewar and war years in comparison with the postwar period. The results (Table 4) show that the differences, which are in the expected direction, are not statistically signifi? cant. This is consistent with the finding that regression results for the 1926-1974 and 1947-1974 periods are not significantly different. The second step in the analysis was designed to test the effect of the volume of crime on the imprisonment-unemployment relationship. Two in? dices of crime were available: number of reported ral of the periods under study. For example, throughout the decade of the 1960s (see Appen? dix 1), prison population was declining (as was unemployment), while arrests were sharply increasing. Similar results were established for the postwar period, 1947-1974 (Table 6). Here the federal prison population, too, correlates with unemployment (p=.05). However, some regressions presented in Table 6 have low determinants and standardized regression coefficients larger than 1 (for total prison population), indicating presence of serious multicolinearity. (This multicolinearity is due to strong correlation between arrests and population growth.) In addition, the low Durbin-Watson statist? ics point to a strong serial correlation effect which was not fully purged by the AR(1) model. The foregoing applies to regressions which use prison population as dependent variables. Regressions which use prison admissions show better results. Examination of the national sample thus demon? strates that the expected positive relationship between imprisonment and unemployment holds regardless of the volume of criminal activity. The third step in the analysis correlated monthly Sunshine County fluctuations in unemploy? ment and fluctuations in the total population of the county jail. Table 7 shows that an increase of 1,000 in either population or unemployment corresponds to an increase of 4 jail inmates each month. The were preferable because they provided a longer partial correlation coefficient between unemploy? ment and jail population, with civilian population held constant, is .23 (significant at p=.01). When controls are introduced for the changes in arrests, unemployment and population as appears that neither crimes nor arrests influence the relationship between unemployment and im? prisonment. Both these variables have nonsignifi? crimes, 1937-1974, and number of arrests, 1932 1974. Both were used in two sets of regressions, and the results were very similar. However, arrests period for analysis. The basic procedure was to run multiple regressions with total prison inmates and total prison admissions as dependent variables, and independent variables. Results of these regressions, summarized in Table 5, indicate that the unemployment-imprison? ment relationship obtains regardless of the volume of crime. Except for the two regressions in which the volume of criminal activity (Table 8), it cant regression coefficients and their partial correlation coefficients (with jail population) do not exceed .1 (p greater than .05). Results obtained from the local sample are fully consistent with those from the national sample. dependent variables are represented by federal Thus, both monthly and annual correlations regression coefficient associated with arrests fluctuations in the volume of crimes and arrests. The fourth step in the analysis was to study the prisoners, regression coefficients associated with unemployment are all significant at p=.01, while no attains statistical significance. Partial correlation coefficients between unemployment and state prison population, obtained when the effect of arrests on both variables was accounted for, are between unemployment and incarceration are posi? tive, statistically significant and unaffected by two apparent exceptions to the unemployment imprisonment relationship - the 11-year period of the Great Depression and the federal prison still around .5 and significant at p=.01. Interestingly, regression and correlation coeffi? cients associated with arrests are almost all nega? population prior to 1960. prison population. This relationship between arrests and prison population actually existed during seve decade of the 1930s sent unemployment rates well tive. A literal interpretation suggests that an increase in arrests corresponds to a decrease in 1. The Great Depression (1930-1940) The deep crisis which struck America in the above any "normal" or "acceptable" limits. The Fall-Winter 1977/ 24 This content downloaded from 213.198.240.67 on Fri, 24 Mar 2017 15:39:14 UTC All use subject to http://about.jstor.org/terms unemployment rate first rose to 8.9% in 1930, then skyrocketed to 25.2% (37.6% for nonfarm laborers) in 1933, and stayed above or around 15% (20% for nonfarm laborers) until 1941. In absolute numbers, almost 13 million persons were unemployed in 1933, and for most of the decade this number stayed close to or above 10 million. Neither prison population nor prison commit? ments kept pace with unemployment. Coefficients of correlation between imprisonment and unem? ployment rates for this period are negative, ranging from -.02 to -.42, although none are statistically significant. Arrest had a high negative correlation with unemployment rates (r=-.87; N=9; p=.001). These findings do not contradict the labor market-imprisonment hypothesis. The hypothesis postulates that imprisonment may be used to absorb a part of the surplus labor, but this is expected to hold under normal economic condi? tions. Under normal conditions, the number of persons unemployed varies, but usually stays within definable limits (between 3% and 6% of the labor force). Under such conditions, the prison population maintains a relatively stable relationship with the number of unemployed persons. Expressed as a proportion of the unemployed labor force, prison population normally varies between 4% and 9%, with a mean of 5.2% for the years 1926-1974. If the same relationship had continued during the Depression years, the number of prisoners would have varied between ^50,000 and 770,000, with a mean value of approximately 600,000. This figure refers to state and federal prisons only. If local jails and institutions for juvenile delinquents were included, the total would have exceeded one million persons incarcerated on any one day during the year. The capacity of prisons, which is admit? tedly an uncommonly flexible concept, could not possibly have absorbed such a mass of inmates. (3) There is a general consensus that prisons became overcrowded during the Depression, and that the overcrowding was due to significant cuts in correctional budgets and personnel (Sellin, 1937). between 1931 and 1940, while the average for 1926-1930 was about 113,000 and the average for 1941-1950 was about the same as in 1931-1940. This difference between admissions in 1931 and total population in subsequent years suggests that fewer people were admitted to prisons in the Depression years, but those who were committed stayed in prison for longer periods than formerly. Longer sentences thus contributed to prison over? crowding, thereby stemming the 1930-1931 trend toward greatly increased admissions. Similar results were reported by Stern (1940), who undertook a largely descriptive study of prison commitments and sentences in Pennsylvania from 1924 to 1933. Stern found that prison commitments rose in 1930 and peaked in 1931, after which time they started to decline. He also found that average sentences, especially those meted out to recidi? vists, were much longer in the Depression years (1930-1933) than in the pre-Depression years (1924 1929). Stern concluded that "allowing for the fact that for years there has been a feeling in this country for more severe punishment of criminals, it can be said the data show that in cases of recidivists committed for serious property crimes there was a definite tendency to greater severity. (This) study shows that the economic situation apparently influences policies of court and penal administration" (1940:711). After 1932, criminal justice expenditures declined at all levels of government, and all criminal justice agencies suffered reductions in personnel (see Historical Abstracts of the United States, series 1012-1027). Although it has been argued (Smith, 1935) that these reductions had a beneficial long-term effect on criminal justice agencies (by pruning the police force of the least fit members, etc.), their immediate result was a decline in arrests, prosecutions and prison commitments. 2. Federal Prisons The immediate response of criminal justice The second set of data which does not conform to the hypothesis is represented by federal prison in the first two years (1930 and 1931). Then, difficult to say why federal penal policies would agencies to the Depression was repressive - there was an increase in prison admissions and population apparently, considerations of the costs of imprison? ment produced a trend toward more moderate sentences, thus reducing both prison population and prison admissions to levels below those expected on the basis of the 1930-1931 policies. In 1931, prison commitments reached a peak unsurpassed throughout Depression years. Except for 1940, when about 73,000 persons were admitted to prisons, the number of admissions did not again reach the 1931 peak until 1953. The number of prisoners present, however, averaged about 150,000 population and admissions prior to 1960. It is produce an exception in these years and why they would cease to produce it after 1960. The apparent anomaly may be due to a different composition of the federal prison population, which might include larger numbers of offenders with higher socio? economic status (tax law violators, embezzlers) who are not typical of the marginal, reserve labor force. This explanation also implies that, since 1960, the characteristics of federal prisoners have been becoming increasingly similar to those of state prisoners. However, the data presently 25/ Crime and Social Justice This content downloaded from 213.198.240.67 on Fri, 24 Mar 2017 15:39:14 UTC All use subject to http://about.jstor.org/terms available do not make it possible to test this or other plausible explanations of the exception. Imprisonment and Unemployment Considering that numbers of unemployed persons are usually in the millions and that prison population rarely exceeds 200,000 in any one year, what impact, if any, would incarceration have on unemployment rates? Prison population accounts for approximately one quarter of 1% of the total labor force, based on the 1926-1974 average (X = .26; S.D. = .02). tionalized would be in the labor force if they were not under institutional care. This population includes prison and jail inmates, some mental patients, some elderly citizens and some inmates of other institutions (e.g., unwed mothers). If one half of the total institutional population were in the labor force, and if all of them were unemployed, the unemployment rates would rise by an average of 1.3%. If only one half of those in the labor force (one quarter of the total institutionalized popula? tion) were unemployed, the impact on unemploy? ment rates would be, on the average, seven tenths of 1% (.7%). Expressed as a percentage of the unemployed labor force, it accounts for roughly 5% (X = 5.2; S.D. = 3.4). The potential impact of prison population on unemployment rates during the years 1926-1974 UNEMPlOYMENr IHSURAMCfc was considered under three different models: 1) all prisoners are, before incarceration, in the labor force and all are unemployed; 2) all prisoners are, before incarceration, in the labor force and 50% are employed; 3) all prisoners are, before incar? ceration, in the labor force and all are employed. Assumptions 1) and 3) are somewhat extreme, but assumption 2) is probably not far removed from reality. In other words, if all prisoners were to be released, about half of them would be unemployed. In the first case, the unemployment rates from 1926 to 1974 would have increased by an average of .25% (one fourth of 1%). In the second case, the increase would have been .11%. Finally, in the third case, there would have been a decrease in unemployment rates of .02%. All three models, however, substantially under? estimate the potential impact of incarceration on unemployment rates. No reliable data are available concerning national jail population on an annual basis. The Sunshine County data indicate that jail population behaves in the same way as does prison population in relation to unemployment. There? fore, if both jail and prison populations were considered simultaneously, the effect would be o > z The recession is 01; "Go home, I tell yt Very similar findings emerge when Sunshine County data are examined. Expressed as a percen? tage of the county labor force, jail population in 19691976 varied from .17% to .36%, with a mean value of .25%. When compared to the unemployed labor force, the jail population ranged between 2% and 5.3% with a mean of 3.6%. Assessment of the potential impact of jail population on unemployment rates was based on the same three alternegative models used with refer? ence to the national data. The estimated change in almost doubled. The same argument should hold for the entire institutionalized population, but, because of lack of data, it is here made in a tentative and speculative fashion. The total institutionalized population in would be, respectively: 1) .21%; 2) .09%; 3) -.03%. Our second ("utility") hypothesis predicts that imprisonment will have a delayed negative effect statistics in a dramatic way. After all, a rise in the one year and correlated with prison variables in the the U.S. of about two million in 1970 would, if included in the labor force, affect labor force unemployment rate by even one half of 1% is in some circles a cause for alarm and for "action." But, of course, it is not reasonable to expect all, or even a majority of the presently institutionalized population to be economically active if not institu? tionalized. It may be not unreasonable, however, to assume that 50% of all those persons now institu unemployment rates under the three conditions on unemployment. The results of a test of this hypothesis with the national data are shown in Table 9. When labor force variables are lagged by preceding years, the relevant correlation coeffi? cients have the expected (negative) signs, but fall short of statistical significance. For example, the coefficient of correlation between number of in? mates present in prisons and the unemployment rate in the year following is -.22. This result is not improved when the relationship is tested in a Fall-Winter This content downloaded from 213.198.240.67 on Fri, 24 Mar 2017 15:39:14 UTC All use subject to http://about.jstor.org/terms The findings from the Sunshine County sample are even less encouraging. When unemployment is Sunshine County, California, from January 1969 to December 1976. The findings were consistent with the hypothesis. The relationship between unemploy? ment and imprisonment was positive and statisti? between jail population and unemployment rates criminal activity. There were two exceptions. This regression run with population as a standardizing variable. lagged by up to six months, all correlations are still positive. Thus, the hypothesis that imprisonment acts to reduce unemployment rates was not supported by the present study. One possible explanation is that prison population, taken by itself, is not suffi? ciently large to produce an observable effect on cally significant, regardless of the volume of relationship did not obtain during the Great Depression (1930-1940), and federal imprisonment rates did not correlate with unemployment rates prior to 1960. It was suggested that the extent of unemployment during the Depression and the conci? unemployment, and that future tests should inc? lude - at the minimum - population of local jails liatory policies of the New Deal prevented the positive correlation between imprisonment and unemployment. No satisfactory explanation was found for the aberrant behavior of the federal populations on unemployment. prison population, but it was suggested tentatively and test the combined effects of jail and prison V. CONCLUSIONS Our main purpose was to test the applicability that, prior to 1960, federal prisoners included larger numbers of white-collar offenders, atypical of the marginal labor force members who populate the state prisons. One's faith in these findings is bolstered by the fact that results from the national and local of the Rusche-Kirchheimer theory of punishment to contemporary postindustrial societies. The focus of samples correspond very closely. For instance, the proportion of incarcerated population to the total conditions in the United States from 1926 to 1974. identical (.26% and .25%, respectively). research was on the imprisonment and economic civilian labor force in the two samples is almost The utility hypothesis was not supported by the Two distinct hypotheses were identified as either explicitly or implicitly contained in the data. The relationship between imprisonment and Rusche-Kirchheimer theory, which is best summa? rized with the phrase, "Every system of production lagged unemployment rates in the national sample first hypothesis refers to severity of punishment not be established. The failure to confirm the utility hypothesis detracts from the significance of the confirmation of the severity hypothesis, for the meaning of the demonstrated relationship between unemployment and imprisonment remains somewhat ambiguous. It is suggested, however, that the available data base was too narrow, and that the expected impact of was negative as predicted, but the coefficients tends to discover (and use) punishments which were not statistically significant. In the local correspond to its productive relationships." The sample, the predicted negative relationship could and predicts that criminal punishments will be more severe at times of economic crises. This was operationalized in terms of the relationship between use of imprisonment and unemployment rates, and it was suggested that prison population will increase as unemployment increases, regard? less of the volume of recorded criminal activity. The second hypothesis refers to the utility of punishment and predicts a functional relationship between different punishments and the productive relationships inherent in different socioeconomic systems. For our purposes, this hypothesis was operationalized in terms of the impact of imprison? ment on unemployment, and it was predicted that the size of the prison population will be negatively related to lagged unemployment rates. The hypothesis seemed reasonable on the assumption that postindustrial societies are permanently faced with an oversupply of labor, and that imprisonment removes a part of the surplus population from the labor market, thereby reducing unemployment rates. incarceration policies on unemployment rates might yet be demonstrated if the entire incarcer? ated population (not only the prison inmates, but also the inmates of local jails and juvenile homes) were used as an index of imprisonment. Despite their shortcomings, the present findings do lend some support to the Rusche-Kirchheimer severity hypothesis. Their greatest immediate con? tribution, however, may be that of putting current speculations about imprisonment and unemploy? ment rates on a solid foundation. By providing some empirical estimates of the magnitude of the poten? tial impact of imprisonment on unemployment rates, the present study may help to remove discussions such as Quinney's (1977) and Reasons and Kaplan's (1975) from an almost metaphysical The severity hypothesis was tested on two samples of data: the national statistics for the realm and place them on firmer empirical ground. Another contribution made by the present study U.S., 1926-1974 and the monthly statistics for 27 / Crime and Social Justice This content downloaded from 213.198.240.67 on Fri, 24 Mar 2017 15:39:14 UTC All use subject to http://about.jstor.org/terms is in its demonstration that the relationship between unemployment and imprisonment is a direct one, independent of the changes in criminal activity. TABLES Table 1. Prison population regressed on number of unemployed (U) and resident civilian population (P), U.S., 1926-1974. unstandardized standardized Table 2. Prison population regressed on number of unemployed (U) and resident civilian population (P), U.S., 1947-1974. unstandardized standardized regression regression partial regression regression partial coefficient T-test coefficient coefficient coefficient T-test coefficient coefficient Total prisoners Total prisoners present present a .43 .51 U 2.02 3.23a .39 U 3.42 2.92a .51 .46 P .76 3.53a .43 .03 P .06 .17 .03 admitted r=.56 R2=.32 s.E.=5.75 F=10.79a DW=1.28 determinant=.99 U 1.71 3.08a .38 .41 admitted P .78 3.50a .43 .46 r = .52 r2=.27 S.E. = 4.35 F=4.51b DW=1.25 determinant=.97 U 4.58 2.77a .43 .49 P .64 2.62a .41 .47 r=.56 R2=.31 s.E.=5.13 F=10.40a DW=1.03 r =determinant=.99 .66 r2=.44 S.E.=5.89 F=9.63a DW State prisoners present U 2.04 3.58a .43 .47 State prisoners present U 3.12 2.92a .51 .51 P .72 3.64a .43 .47 P .005 .01 .002 .003 r=59 R2=.35 s.E.=5.25 F=12.27a DW=1.35 r =determinant=.99 .51 R2=.26 S.E.=4.06 F=4.39b admitted U 1.66 3.39a .40 .45 admitted P .87 3.60a .43 .47 U 4.36 2.81a .42 .49 P .64 3.15a .47 .54 r=59 R2=35 S.E.=4.53 F=12.18a DW=1.06 determinant=1.0 r = .70 R2=.50 S.E. = 5.46 F=12.09a Federal prisoners present U -.04 -.35. -.05 -.05 P .06 2.27b .32 .32 Federal prisoners present U .34 1.62 .30 .31 P .05 1.59 .29 .30 r=.33 R2=,11 S.E. = 1.13 F=2.79 DW=1.63 determinant=.98 admitted U -.03 -.29 -.04 -.04 P .06 .32 .002 .04 admitted U P r=46 R2=.21 s.e.=.75 f=3.38 .33 1.64 .32 .31 -.03 -.83 -.16 -.16 r = .07 R2 = .004 S.E. = 1.12 F=.11 r DW=1.82 determinant=.97 = .33 R2=.11 s.e. = .74 F=1.48 DW a=p<.01 b=p<.05 a=p<.01 Table 3. Federal prison population regressed on number of unemployed (U) and resident civilian population (P), U.S., 1960-1974. unstandardized regression coefficient standardized partial regression coefficient T-test coefficient Table 4. Prison population regressed on number of unemployed (U) and resident civilian population (P), with dummy variables to distinguish prewar (D1) and war (D2) periods from the postwar period, U.S., 1926-1974. unstandardized regression Federal prisoners present 4.86d U 1.69 P -.09 -3.27a .80 -.54 admitted U 1.31 P -.002 5.043 .83 -.01 S.E. = .99 F = 13.89i r=.83 R2=.69 r=83 R2=.69 s.e. a=p<.01 -.08 .81 -.68 1.27 determinant^ .92 = .74 F= 13.60' DW=1.44 .82 -.02 determinant^ .92 b = p<05 coefficient Total prisoners admitted U 1.68 P .77 D1 .64 D2 -.39 standardized regression partial T-test coefficient coeffic 2.78a 3.29a .08 -.07 .37 .43 .01 -.01 R=.56 R2=.31 s.E. = 5.25 F=5.03a DW=1.03 determinant = .42 .39 .44 .01 -.01 a=p<01 Fall-Winter!977/ 28 This content downloaded from 213.198.240.67 on Fri, 24 Mar 2017 15:39:14 UTC All use subject to http://about.jstor.org/terms Table 5. Prison population regressed on number of unemployed (U), resident civilian population (P) and number of arrests (A), U.S., 1932-1974. unstandardized standardized regression regression partial coefficient coefficient coefficient T-test Total prisoners present 3.32 U .42 3.87 d .53 P 1.16 .50 .88 3.56a A -4.77 -.21 -.33 -1.53 R = .74 R2 = .54 S.E. = 5.81 F = 15.383 DW = 1.51 determinant=.19 admitted U P A 2.57 .76 -.33 3.22a .43 .33 1.86 -.10 .46 .29 -.01 Table 6. Prison population regressed on number of unemployed (U), resident civilian population (P) and number of arrests (A), U.S., 1947-1974. unstandardized standardized regression regression partial coefficient T-test coefficient coefficient Total prisoners present 4.26 U 1.19 P A -7.03 3.12 1.18 -5.52 4.05a 4.05a .43 .99 .43 -1.74 .54 .54 -.27 R=.76 R2 = .58 S.E. = 5.26 F= 17.65a DW= 1.56 determinant=.18 admitted U P 2.33 .86 -.80 3.77 a .43 .38 -.04 2.29 -.28 .46 .34 -.04 admitted admitted U P 1.20 .36 .73 .17 .12 .25 .19 .05 .11 R=42 R2=.17 S.E. = 1.01 F=2.75 DW=1.15 determinant=.18 .003 .015 -.13 .02 .28 .003 .14 -.11 -.22 .003 .04 -.03 State prisoners present admitted 2.79a 1.65 -.60 -.22 .50 .32 -.12 3.75 1.05 -6.24 2.64a 2.36b .42 1.20 -.84 .48 .44 -.32 4.50 1.03 -3.27 .45 .61 -1.63 S.E.=4.86 F: = 7.22a DW=.91 2.87?? 1.92b -.76 determinant=.07 .43 .73 -.29 .51 .37 -.32 .36 1.19 -.87 .40 .34 -.24 R = .71 R2 = .50 S.E. = 5.49 F = =7.96a DW=.82 determinant=.13 Federal prisoners present .58 U .14 P A -.87 admitted U P A 2.11 ? R = .65 R2 = .43 1.78 b -1.21 S.E. = .93 I = 5.79a DW= 82 1.49 .31 -.09 .45 -1.11 R=36 R2 = .13 .79 .29 -.40 .29 S.E. = .75 I = 1.18 DW=1.09 determinant = .04 .29 -.22 .16 determinant=.26 b=p<.05 a=p?.01 b = p<.05 -.31 4.70 .98 -2.78 R=.69 R2=.48 R = .05 R2=.002 S.E. = 1.08 F=.34 DW=1.44 determinant=.10 a=p<.01 .48 .44 r = .67 R2=.44 S.E.=5.94 F = 6.25a DW=.82 determinant=.16 r=63 R2=.40 S.E.=4.59 F=8.75a DW=1.09 determinant=.50 Federal prisoners present .17 U P .02 A .45 -1.61 .41 1.23 -.84 r=70 R2 = .50 S.E. = 5.53 F= 7.67a DW=.84 determinant=.07 -.01 r=.61 R2=.37 S.E.=5.17 F=7.59a DW=1.08 determinant=.47 State prisoners present U P A 2.63a 2.37b Table 8. Jail population regressed on number of unemployed (U) and total civilian population (P), controlling for arrests (A) and crimes (C), Sunshine County, January 1969-December 1976. Table 7. Jail population regressed on number of unemployed (U) and total civilian population (P), Sunshine County, January 1969-December 1976. unstandardized regression coefficient T-test Jail population U P .004 .004 standardized regression partial coefficient coefficient 2.31 b 6.12a R = .60 R2 = .36 S.E. = 21.59 F = 25.88? a=p<.01 b = p<.01 .19 .52 .23 .53 1 DW=1.53 determinant = unstandardized regression coefficient T-test Jail population U P A .004 .004 .000 U P c .004 .004 .010 R = .62 R2=38 R = .64 R2 = .41 standardized regression partial coefficient coefficient a .05 .19 .54 .00 .23 .55 .006 2.20 a .18 .22 2.28 6.42 a S.E. = 21.71 F = 19.00a DW = 1.52 determinant=.94 6.81a .57 .99 .06 .58 .10 1.64 determinant=.90 S.E. = 21.60 F = 21.23a DW = a=p<.01 29 / Crime and SocialJustice This content downloaded from 213.198.240.67 on Fri, 24 Mar 2017 15:39:14 UTC All use subject to http://about.jstor.org/terms Table 9. Pearson's product-moment coefficients of correlation between prison population and lagged labor force variables, U.S., 1926-1974. 1 2 3 4 5 6 7 8 1 total prisoners present .99 .79 .75 .85 .79 -.07 -.22 2 total prisoners admitted .74 .76 .84 .79 -.06 -.21 3 state prisoners present .99 .83 .76 -.03 -.18 4 state prisoners 3. An enlightening definition of prison overcrowding was given by a veteran prison administrator (Clemmer, 1957:283): "A prison may be said to be crowded when it is unable to perform its statutory duties of safeguarding, providing decent care and engaging in the training and instruction of inmates by virtue of the fact that the obstacles thereto are brought about by insufficient living and work space, or insufficient budget or insufficient personnel, according to acceptable minimal standards as, or when prescribed by the American Correctional Association and as interpreted by a responsible administrator." admitted .81 .74 -.01 -.16 5 total labor force .97 -.23 -.38 6 number REFERENCES employed -.46 -.59 7 number unemployed .98 Beveridge, William H. 1930 Unemployment: A Problem of Industry. 2nd ed. London: Longmans, Green. 8 unemployment rate critical values of r for two-tailed test of statistical significance: p= .05 .01 r= .28 .37 Blalock, Hubert M. 1972 Social Statistics. 2nd ed. New York: McGraw Hill. Braverman, Harry 1974 Labor and Monopoly Capital: The Degradation of.Work in the Twentieth Century. New York: Monthly Review Press. California Bureau of Criminal Statistics 1976 California Comprehensive Data Systems. Criminal Justice Profile, 1975. Sacramento, Ca.: Statistical Analysis Center. Clemmer, Donald FOOTNOTES 1957 "Some Aspects of Crowded Prisons." Proceed? ings of the 87th Annual Congress of Correc? tions of the American Correctional Associa? tion, 1957. New York: American Correctional Association: 280-91. Copeland, Morris A. 1. Recently, Quinney (1977) has developed a more 1966 Toward Full Employment in Our Free Enter? general theory of criminal justice based on the prise concept Economy. New York: Fordham University of "surplus population control." His book was published Press. too late for detailed consideration in this essay. Evans, Robert "Some Notes of on Coerced Labor." Journal of 2. Reasons and Kaplan (1975) included 1970 "reduction unemployment rates" among their eleven "latent func? Economic History 30:861-66. tions of prisons," but they did not advance or refer to any evidence. At the same time, some Kant, of their Immanueleleven functions seem positively frivolous (e.g.: prisons 1887 Thefunction Philosophy of Law. Vol. 2. Edinburgh. as a means of birth control), so that it is not clear whether they have advanced the imprisonment-unem? Levitan, Sar A., Garth L. Magnum and Ray Marshall 1976 Human Resources and Labor Markets: Labor ployment hypothesis in good faith or in jest. Quinney claims that "hundreds of thousands find economic supportin American Economy. 2nd ed. and Manpower through criminal offenses and economic security while New York: Harper and Row. being confined in prison, at the same time lowering the unemployment rate of the society" (1977:129; emphasis Marx, Karl added), but he does not cite any evidence forA this 1967 Capital: Critique of Political Economy. statement. Vol. 1. New York: International Publishers. Fall-Winter 1977130 This content downloaded from 213.198.240.67 on Fri, 24 Mar 2017 15:39:14 UTC All use subject to http://about.jstor.org/terms Nie, Norman H., C. Hadlai Hull, Jean G. Jenkins, Karin Steinbrenner and Dale H. Bert 1975 SPSS: Statistical Package for the Social Sciences. 2nd ed. New York: McGraw-Hill. Rusche, Georg 1933 "Arbeitsmarkt und Strafvollzug: Gedanken zur Soziologie der Straf justiz." Zeitschrift fur Sozialforschung 2:63-78. O'Connor, James 1973 The Fiscal Crisis of the State. New York: St. Martin's Press. Rusche, Georg and Otto Kirchheimer Pindyck, Robert S. and Daniel L. Rubinfeld Sellin, Thorsten 1976 Econometric Models and Economic Forecasts. New York: McGraw-Hill. 1939 Punishment and Social Structure. New York: Columbia University Press. 1937 Research Memorandum on Crime in the Depression. New York: Social Science Research Council. Quinney, Richard 1977 Class, State and Crime: On the Theory and Practice of Criminal Justice. New York: David McKay. Reasons, Charles E. and Russell L. Kaplan 1975 "Tear Down the Walls? Some Functions of Prisons." Crime and Delinquency 21:360-72. Remington, Frank J., Donald J. Newman, Edward L. Kimball, Marygold Melli and Herman Goldstein 1969 Criminal Justice Administration: Materials and Cases. New York: Bobbs-Merrill. Smith, Bruce 1935 "Police Service." In Clarence E. Ridley and Orin F. Nolting (eds.) What the Depression has Done in Cities. Chicago: International City Managers Association. Stern, Leon Thomas 1940 "The Effect of the Depression on Prison Commitments and Sentences." Journal of Criminal Law, Criminology, and Police Science 31:696-711. 31 / Crime and Social Justice This content downloaded from 213.198.240.67 on Fri, 24 Mar 2017 15:39:14 UTC All use subject to http://about.jstor.org/terms