From: Social
Problems Vol. 39 No. 4 November 1992 Pp.
421-446
Labor Surplus and
Punishment: A Review and Assessment of Theory and Evidence
THEODORE G. CHIRICOS, Florida State University
MIRIAM A. DELONE, University of Nebraska at Omaha
Since the pioneering work
of Rusche and Kirchheimer (1939), the theoretical links between labor surplus
and punishment have seen extensive development. Eleven of those links described here
are mediated by economic, political, and ideological factors such as the value of labor, the systemic needs of capitalism, and the ideology of judges and their communities.
The sophistication of theorizing about labor surplus and punishment is matched with
skepticism about the corresponding empirical evidence which has been termed
"elusive" and "contradictory" at best. The results from 44
empirical studies are systematically assessed. The evidence suggests that
independent of the effects of crime, labor surplus is
consistently and significantly related to prison
population, and to prison admissions when time‑series and individual
level data are used. The relationship of labor surplus and punishment
appears slightly stronger when age, race, and gender specific measures are
employed. The limitations of existing research, in light of theoretical developments, are discussed.
The
role of state punishment in controlling labor surplus was described
historically in the seminal work of Rusche and Kirchheimer (1939). Since then, a variety of accounts (Adamson 1984; Foucault 1979; Melossi
and Pavarini 1981) have
elaborated that historical assessment. More recently, the link between labor
surplus and punishment has received extensive empirical attention. In fact, at
least 44 studies have been published in the past 20 years that provide
empirical evidence for the labor surplus‑punishment relationship.
This
paper's objective is to review that empirical evidence in light of theoretical
formulations that have evolved from the original Rusche and Kirchheimer
thesis. The goal is to assess not only the results of research on labor surplus
and punishment but to assess the "empirical plausibility" (Spitzer 1980) of those theoretical formulations. In a number of ways, it
appears that research lags considerably behind theory in this regard and we
hope this discussion will highlight directions for further research.
Theoretical Linkages: Labor Surplus and
Punishment
Before summarizing
the results of empirical research on this issue, we briefly describe several
theoretical models used to account for the relationship of labor surplus to
punishment. They are not intended to exhaust the theoretical issues attending
that relationship, but rather to highlight key concepts and linkages that may
be more or less well reflected in the research reviewed below.
Theorizing
about punishment and surplus labor has clustered around three principal issues:
the value of labor, systemic needs of capitalism, and judicial action. Specific
theoretical contributions have sometimes involved more than one of these, but
generally they have emphasized that the link between surplus labor and "harsher punishment"' is mediated by (1) the diminished
value of labor, (2) the structural or systemic needs of capitalism, or (3) the interaction of ideology and the motivated action of
judges and others. A visual summary of these theoretical linkages, which we
will discuss in detail, is provided in Figure 1.
These explanatory
sketches are not mutually exclusive. Instead, they complement one another by
reflecting different levels and issues of analysis. From top to bottom, the
first is principally economic, the second political, and the third ideological.
Despite the importance of dialectic relationships in the intellectual roots of
much of this theorizing, these linkages have generally been developed in
unidirectional terms; this is reflected in Figure 1.
One of
Rusche and Kirchheimer's central arguments was that during periods of labor
surplus, as in Europe during the first half of nineteenth century, harsh
punishments replaced the exploitation of labor by the state. Several of the
reasons for this shift involve the diminished value of labor. First, "the
profit which had accrued to prison managers when men were scarce and wages high
disappeared" and with it, both the motive and means for maintaining
reasonable prison (work) conditions (1939:108). Thus,
harsher conditions of prison punishment were a direct by‑product of
prison labor's lower value.
A
second factor relates to the value of free labor and the necessity of
reproducing it. For example, during the first half of the nineteenth century,
wages were "frequently lower than the minimum necessary to reproduce the
labor power of . . . workers" (1939:108). For
prisons to avoid undermining the compulsions of free wage labor, the principle
of "less eligibility" required that "the upper margin for the
maintenance of the prisoners" be kept "below the living standard of
the lowest classes of the free population" (1939:108).
Finally,
Rusche and Kirchheimer noted that conditions of labor surplus not only impoverished
the working class but increased the motivation for crime. Citing Engels'
observation that "want leaves the workingman the choice between starving
slowly, killing himself speedily, or taking what he needs where he finds
it," the authors noted that "the poorer the masses became, the
harsher the punishments in order to deter them from crime" (1939:18).
The
intervening mechanism of criminal motivation is a direct consequence of the
diminished value of labor during conditions of labor surplus. However, several
authors have described the deterrent role of punishment in relation to
unemployment without explicit mention of labor value, but its role is readily
inferred. For example, Greenberg notes:
Persons who are unemployed
can be assumed to have a greater incentive to steal .... If this reason‑
ing is correct . . . judges
will respond to this by imposing prison sentences more frequently
(1977:648).
Similarly, Jankovic
hypothesized that "a rise in unemployment will lead to an increase in
prison commitments" because "the policy of deterrence dictates an
intensification of punishment in order to combat the increased temptation to
commit crime" (1977:20). In
short, labor surplus is presumed to reduce the value of labor, which makes
prison labor less profitable, prison conditions less palatable, and criminal
motivation more likely. Each of these contributes to harsher punishments.
A
second line of theorizing is grounded in structural accounts of the state's
roles in reproducing capitalist productive relations (Carnoy 1984; Poulantzas
1978). As
several theorists have noted, those roles principally address systemic needs of
accumulation, legitimation, and control. To some degree, the profitable
exploitation of prison labor discussed above, and the principle of less
eligibility, both involve the state directly in the process of accumulation.
But a
number of authors have argued that surplus labor creates problems of legitimacy
and control that the state's punitive apparatus helps to address. The control
of surplus populations is considered a means to preclude questions about the legitimacy of a
system of productive relations that continually makes human workers redundant.
Spitzer
noted that surplus populations, while increasingly characteristic and necessary
in state‑monopoly capitalism, can nevertheless "undermine the ideology
of equality so essential to the legitimation of [capitalist] production
relations" (1975:643).
He noted further that such populations "must be neutralized or
controlled if production relations and conditions for increased accumulation
are to remain unimpaired" (1975:645).
In this context, the legal system helps to control that portion of
surplus labor that is young, active, and potentially most threatening‑what
Spitzer termed "social dynamite" (1975:646).
Similar
views have been developed by Wallace (1980)
and Adamson (1984).
Wallace noted that "legitimization" of the capitalist order
is assisted by the activity of the criminal justice system, "whose
function, in part is to contain and maintain labor power" (1980:59). Adamson
attributed harsh regimes of punishment, in part, to the "potential
political threat" posed by surplus labor during the troughs of business
cycles (1984:437).
The
need to control surplus labor has also been described without specifically
raising the question of legitimacy. For example, Lynch noted that
"marginalized workers" are not subject to traditional "work‑place
controls" and so "incarceration becomes one method for controlling
marginalized labor" (1988:322).
Similarly, Jankovic suggested that "imprisonment can be used to
regulate the size of the surplus labor force" (1977:20) and Quinney contended that
"criminal justice is the modern means of controlling surplus
population" (1977:131).
Resonating
with Spitzer's concept of "social dynamite," several authors have
sharpened the focus of systemic control needs. For example, Myers and Sabol
wrote that "prisons seem to siphon off the most superfluous class of
workers, such as young black men" (1987:48), while Box and Hale hypothesized
that the relationship between unemployment and imprisonment will be strongest
for young males (1982:26).
Melossi has written that "dangerous classes" have come to
be "defined by a mix of economic and racial, ethnic and national references" so
that unemployed young black males have likely become "a privileged target
group for imprisonment" in the United States and England (1989:317).
A third
approach to the relationship between labor surplus and punishment has emphasized
the human "agency" and ideology of criminal justice personnel,
primarily judges. Greenberg (1977)
was one of the first to take this approach when he argued that to
explain the strong empirical link between rates of unemployment and
incarceration it is
plausible to assume that judges
are less willing to grant probation to offenders when they are unemployed, or
that unemployment affects levels of community tolerance toward offenders, to
which judges respond in sentencing (1977:650).
Similarly,
Box and Hale (1982) looked at the "everyday micro‑processes of
interaction between the accused and the accusers and the perceptions of
judicial decision‑makers" (1982:21) to address what they described
as "lacunae" in the structural approaches discussed above. In their
view, structural explanations that remain at the level of system needs invite
an implicit "conspiracy account in which the powerful deliberately attempt
to fragment and . . . discipline the unemployed by increasing the rates of . .
. imprisonment" (Box and Hale 1982:21).
Box and
Hale's approach attempts to provide "agency" without conspiracy by
focusing on "unintended consequences" flowing from the aggregated
responses of individual judges who "routinely" respond to defendant
attributes such as unemployment.
If, as
it is likely, many of them believe the orthodox view that unemployment causes
more crime . . . then extending the use and severity of imprisonment . . . will appear
to them as nothing less than a normal and rational response. In the aggregate
these decisions . . . shore up an economic and social system threatened by its
contradictions. But clearly, this objective consequence was not intended by the
individuals whose decisions brought it about (Box and Hale 1982:23).
Both
Box (1987) and Hale (1989b) have independently developed elements of the foregoing
position. Box contended that various crime control actors, from the police to
the judiciary, each make an "unintended and unwitting" contribution
to "reducing anxieties created by the existence of population
surplus" (1987:133). These anxieties, in combination with the belief that
“unemployment causes crime” presumably structure the outcomes of judicial
action.
Reflecting
on the background and training of magistrates, Hale observed that
Their
natural constituency is "conservative" and especially the
preservation of private property. Consequently rises in the level of
unemployment is [sic] likely to be a source of deep anxiety to them since they
believe that the unemployed are weak and amoral and therefore more likely to be
criminal (1989b:347).
Melossi
(1985a, 1985b, 1989) has also reacted to the limits of purely structural
explanations. Regarding the "great synchrony" of movement that
characterizes unemployment and imprisonment, Melossi stressed that what
frequently "pass for explanations are some magic structuralist formulas
about 'the needs of capital' or the
'need for social control'
"(1985b:205). The problem, according to Melossi, is that such explanations
"hypostatize collectivities' behavior in a way which is independent from
the motivated actions of the actual persons involved" (1985b:205).
Melossi
(1989:319) argued that neither the
state nor the motives of
individual agents of control need be invoked to account for the labor
surplus/punishment relationship. He suggested that explanations consider what
he terms a "discursive chain" that links business cycles with the
conditions of punishment. From this perspective,
In
periods of economic decline, a "discursive chain" of punitiveness and
severity spreads across society, linking the attitude of "moral
panic" expressed by business leaders and "moral entrepreneurs"
to the ways in which citizens, police, courts and correctional authorities
perceive behavior as deviant and/or criminal (1985a:183).
The
three principal types of explanation for the link between labor surplus and
punishment have emphasized (1) the economic value
of labor and the motive for economic crime; (2) the political needs
of capital, especially legitimation and
control; and (3) the ideological components of agency,
especially beliefs about the causes of crime and perceptions of moral panic. The question we address in the
remainder of this paper is whether and to what extent the accumulated evidence
of empirical research supports or even addresses the theoretical considerations
articulated above.
As
noted above, the theoretical links between labor surplus and punishment have
been clarified extensively over the past 20 years. Less clear, however, is the
nature of empirical support for the labor surplus‑punishment thesis. Even
among those researching this issue, which is most often operationalized as
unemployment and imprisonment, there is some uncertainty over what the
accumulated evidence really shows.
At one
extreme, Parker and Horwitz concluded that "the available evidence
suggests a genuine lack of a relationship" (1986:769). At the other,
Inverarity and Grattet asserted that "the key finding of this
[accumulated] research is that unemployment rates directly affect imprisonment
holding crime rates constant" (1989:351).
More
typical are assessments that describe this evidence as "conflicting"
(Inverarity and McCarthy 1988) "elusive" (Melossi 1989), or
"contradictory" (Michalowski and Pearson 1990). Some have suggested
that the relationship is likely contingent on estimation technique (Inverarity
and Tedrow 1988) or scope conditions (Inverarity and McCarthy 1988) for the
analysis. Whatever the merit of these assessments, none makes reference to more
than a dozen prior studies.
This
paper examines the findings of 44 studies reporting empirical assessments of
the relationship between labor surplus and punishment. Thirty‑four of the
studies have been published since 1980 and all but one since 1976. Our initial
objective is to assess under what conditions the relationship is most often
positive and significant. In this respect, we are exploring the contingencies
of the relationship as suggested by Inverarity and his colleagues.
The
objective here is similar to that accomplished by "meta‑analysis"
(Glass, McGaw, and Smith 1981; Hunter, Schmidt, and Jackson 1982). This
technique is not used at this time because most of the studies reviewed make
use of regression; Hunter, Schmidt, and Jackson caution that "slopes and
covariances are comparable across studies only if exactly the same instruments
are used to measure the independent and dependent variables in each study. It
is a rare set of studies in which this is true" (1982:33‑34).
Appendix
A summarizes the methodology and findings of aggregate level research and
Appendix B does the same for studies using individual level data. The first
column of Appendix A identifies the author, place studied, and the number of
estimates from that research included in this review. Each row in Appendix A
reports a discrete estimate of the relationship of surplus labor to punishment
from a particular study. Sometimes the differences between estimates are
subtle and may not be apparent until one reads across to the dependent
variables.
If the
only difference between estimates from a given study is the number of
independent variables included, all are excluded except that estimate which (1)
the authors considered the "best fitting" or (2) had the largest
number of independent variables. This was done to keep the number of estimates
from a given study within reasonable limits.
The
next six columns of Appendix A describe methodological conditions of the
research, such as unit of analysis, year(s) of study, the type of analysis or
estimation technique, the number of variables in the estimating equation, the
labor market indicator used, and whether or not the estimate controls for
crime. The symbols used for labor market indicator and type of analysis are
identified in the legend at the end of the table.
The columns on the far right of Appendix A specify the measure of punishment included in each estimate‑prison admissions or prison population. Within these columns, the direction, (+) or (‑) of a particular estimate, is shown as well as its statistical significance (*). Letters or superscripts adjacent to the (+) or (‑) symbols further specify conditions of the relationship (e.g., male admissions, female prison population, etc.).
Appendix
B provides a similar description of methodology and results from studies using
individual level data. The first column indicates author and place studied. The
next eight columns describe the year(s) of data collection; type of sample
studied; method of analysis or estimation technique; labor market indicator;
subsamples examined; number of variables in
the equation; and whether or
not the analysis controlled for the defendant's race, gender, crime
seriousness, or prior record. The final two columns of Appendix B show the
direction,
(+) or (‑), and
significance (*) of the relationship between a defendant's labor market status
and either admission to incarceration or the severity of sentence.
Table
1 • Summary of Direction and
Statistical Significance for Relationship between Labor Surplus and Punishment
%
Positive & % Negative &
(N) % Positive % Negative Significance Significance
All Studies 262 87 13 57 3
Prison Admissions 147 86 14 60 5
Prison Population 95 93 7 64 3
Prison Severity 20 70 30 25 5
National 90 93 8 66 1
Regional 56 82 17 60 9
State 43 85 15 34 4
County 09 100 0 78 0
Individual 54 70 30 25 4
Time‑Series 147 90 10 71 4
Cross‑Section 61 85 15 44 3
Control Crime 174 83 17 55 6
Not Control Crime 88 95 5 66 0
Social Dynamite 51 94 6 63 0
Not Social Dynamite 211 85 15 56 5
Table 1
reports the frequency of positive and
significant positive relationships between labor surplus and the several measures
of punishment found in Appendices A and B. It also reports those frequencies
for findings under different methodological conditions. The 44 studies
reviewed produce a total of 262 estimates of the labor surplus‑punishment
relationship.
It is important to note that such an accounting of
results gives equal
"weight" to each of the relationships produced by research of highly diverse quality and sophistication. No attempt is made here to assess the
relative merits of the individual studies reviewed. We recognize the value of
such a qualitative assessment and consider it to be a fruitful "next
step" in the study of this important question. While the numerical calculus
of research results does not "speak for itself " it does help to
establish the parameters of "empirical plausibility" for the
relationship and the limits of existing research in relation to theory.
For all
studies included in this review, Table 1 shows the relationship of labor
surplus to punishment is more than six times as likely to be positive (87
percent) as negative (13 percent) and nineteen times as likely to be
significant and positive (57 percent) as significant and negative (3 percent).
While this finding offers substantial support for the link between labor surplus
and punishment, our objective is to explore the conditions under which this
relationship may be enhanced or diminished.
Table 1
compares the evidence for three different measures of punishment: prison admissions, prison populations, and severity of prison sentences. The strongest support for the
Rusche and Kirchheimer thesis is found for prison population with positive (93
percent) and significant and positive (64 percent)
results slightly higher than for prison admissions, (86 and 60 percent). Severity of prison sentence, involving only
individual level data and a small number of cases, has the lowest frequency of
positive (70 percent) and significant (25 percent)
findings. However, even for these, the weakest results in Table 1, the labor
surplus‑punishment relationship is significant five times more often
than could be expected by chance alone.
If
prison population is, in fact, more
responsive to labor surplus than prison admissions, this
may suggest a cumulative impact of unemployment that is greater than its
immediate impact More important, it may indicate that both the "front
door" and "back door" of prisons play a meaningful role in
controlling surplus population as several have suggested (Inverarity and
Grattet 1989; Wallace 1980).
Support
for the relationship of labor surplus to punishment appears strong at every
level of aggregation. Excluding county level data, which involve too few
estimates to be conclusive, Table 1 shows that studies using national data
have the greatest likelihood of positive (93 percent) and significant (66 percent) relationships. Regional estimates produce
positive (82 percent) and significant (60
percent) estimates slightly less often. Even the "weakest" support
for the relationship of labor surplus to punishment shows positive
relationships more than 70 percent of the time for individual data, 85 percent of the time for state data, and significant
results that are five and seven times more likely than chance.
The
strength of national estimates of the relationship of labor surplus to
punishment is somewhat surprising in light of the fact that both labor markets
and punishment policies are more local than national. Certainly, regional,
state, and county estimates could be regarded as more appropriate levels of
aggregation. The results for state level data are due largely to very weak
results for cross‑sectional estimates of prison admissions (Table 2) that
compose most of the state level findings. Though based on too few estimates,
the results using county data suggest the desirability of aggregation below
that of the state.
Much
has been made of the relative merits of time‑series and cross‑sectional
approaches to this question. For example, Galster and Scaturo (1985) dismissed
time‑series results altogether. It is true that time‑series
estimates at the national level can be expected to produce significant effects
due to the canceling of measurement error between subunits and the tendency for
national series to rise together over time. But as Inverarity and Tedrow (1988)
have argued, there can be compelling problems with either time‑series or
cross‑sectional approaches.
More
than two‑thirds of the aggregate estimates of the labor surplus‑punishment
relationship involve time‑series. While cross‑sectional results
offer somewhat more support for the thesis that they are related than some have
allowed (Inverarity and Tedrow 1988), results for time‑series are more
often positive (90 percent vs. 85 percent) and substantially more often
significant (71 percent vs. 44 percent). Almost all (90 percent) of these time‑series
involve national or regional data and as noted above, can be expected to
produce significant findings more often than cross‑sectional estimates.
One
measure of how meaningful the strong time‑series results found here may
be is provided by comparing them with similar results using other national
level data. If, as some argue, national time‑series are predisposed to
significant findings, the question is whether the labor surplus/punishment
results are stronger than for other series. A recent review of unemployment
and crime rate research (Chiricos 1987) provides the most salient comparison.
That review reported national time‑series estimates that were 75 percent
positive and 36 percent significant. The present time‑series estimates
are 90 percent positive and 71 percent significant, suggesting a relationship
that is even more consistently confirmed and may well be meaningful.
Controlling for Crime
One of
the most important theoretical issues involving the relationship of labor
surplus to punishment is the extent to which it is mediated by crime. While
prison admissions and populations could be expected to expand as crime is
stimulated by unemployment (Box 1987; Chiricos 1987), a key element of
theorizing is that labor surplus has a direct effect on punishment independent of its
indirect effect on crime. Not controlling for crime confounds the direct and
indirect effects of labor surplus, while controlling for crime isolates its
direct effects.
Table 1
compares findings that do and do not control for crime. Relationships without controls for crime have
positive (95 percent) and significant (66 percent) results more often than
those with controls for crime (83 percent and 55 percent). These results
suggest that while labor surplus probably has an indirect impact on punishment
through its influence on crime, there is clearly a direct and substantial labor
surplus‑punishment link that is independent of the mediating influence
of criminal behavior.
Social Dynamite
A final
contingency of labor surplus‑punishment research addresses the hypothesis
that the state's punitive apparatus is mobilized to control "social
dynamite." As noted previously, "social dynamite" refers to
surplus labor that is young, active, and potentially threatening, both
materially and symbolically, to established interests (Spitzer 1975). For this
purpose, we compared the relationship of labor surplus to punishment for
estimates that involve young and/or male and/or black subsamples. The
relationship is expected to be stronger for such groups than for others. The
data from Table 1 moderately support that expectation. Estimates that are
specific to categories of "social dynamite" are almost always
positive (94 percent) and are significant almost two‑thirds (63 percent)
of the time. Results not limited to these subsamples (but certainly also
including them) are only slightly less supportive of the labor surplus-punishment
thesis.
Prison Admissions and Populations Compared
Table 2
describes the relationship of labor surplus to punishment separately for prison
admissions and prison population under varying methodological conditions. What
is striking about the data in Table 2 is the remarkable consistency of the
findings. With the outstanding exception of cross‑sectional estimates of
prison admissions, the frequency of positive labor-surplus punishment results
never drops below 85 percent for admissions and 90 percent for population.
Making the same single exception, significant results are achieved for both
measures of punishment at least 50 percent of the time.
Table 2 • Summary of Direction and
Statistical Significance for Labor Surplus arid Prison
Admissions and Population Under Varying
Methodological Conditions
%
Positive & % Negative
&
(N) %
Positive % Negative Significance Significance
Prison Admissions 147 86 14 60 5
Time‑Series 88 89 11 72 5
Cross‑Section 25 72 28 4 8
Individual 34 91 9 52 3
Control
Crime 86 85 15 57 3
Not
Control Crime 61 89 11 54 7
Prison
Population 95 93 7 64 3
Time‑Series 59 92 8 69 3
Cross‑Section 36 94 6 56 0
Control
Crime 60 90 10 65 3
Not
Control Crime 35 97 3 63 0
A closer look at
cross‑sectional findings in Table 2 is revealing. For prison admissions, positive (72 percent) and significant positive (4
percent) results, though based on a small number of estimates, are easily the
weakest in this review. But cross‑sectional estimates of prison population produce positive (94 percent) and significant (56
percent) results that are much more consistent with those from time‑series.
All of the cross‑sectional estimates use data for states. For that level
of aggregation, it is clear that cross‑sectional methodologies have not
described the kind of relationship found in all other methodological contexts
of this review.
One explanation
for the weaker prison admission results may be that annual admissions involve
more random fluctuation due to transitory factors that are difficult to measure
and model (e.g., local political campaigns or highly publicized crimes). For
prison populations such "noise" is likely averaged out over time
resulting in more stable estimates. It is also worth noting that one half of
the cross‑sectional findings for admissions are produced by a single
study (Galster and Scaturo 1985) which
is responsible for seven of the eight negative results in this category and
only one significant positive result. Inverarity and Tedrow (1988) have raised questions about the approach used in
this study.
It is
worth noting that individual level results, including the severity of
punishment described in Table 1, were relatively weak (70 percent positive; 25 percent
significant). However, individual results for labor surplus‑punishment
involving only admissions in
Table 2 were comparatively strong, with
91 percent positive and 52 percent significant.
Discussion
and Implications
This
review is not intended to "prove" anything about the link between
labor surplus and punishment; the limitations of our method, and of social
science generally, see to that. Instead, we have examined what Spitzer (1980:184) calls the "empirical plausibility" of a
relationship that has stimulated a rich diversity of theoretical insights but
only equivocal assessments of empirical support.
What is
empirically plausible about the relationship of labor surplus to punishment on
the basis of this review? The following possibilities are suggested:
1. Labor
surplus is consistently and significantly related to prison populations.
2. The
relationship appears to be direct and independent of the mediating influence of
crime.
3. The
relationship holds for cross‑sectional as well as for time‑series
estimates.
4. Labor surplus is consistently and significantly related
to prison admissions when aggregate time series and individual level data are
used.
5. The relationship also appears to be direct and
independent of the mediating influence of crime.
6. Labor surplus appears insignificantly related to prison
admissions when cross‑sectional aggregate methodologies are used.
7. Support for the relationship is consistent at all levels
of aggregation except the state level, which has been used predominantly in
cross‑sectional estimates of prison admissions.
8. For individual level data, labor surplus is consistently
and significantly linked with prison admissions, but much less so with
severity of prison sentence.
9. For the young and/or blacks and/or males, the
relationship is slightly more consistent and significant than for others.
In
short, it is empirically plausible to argue that the relationship of labor surplus
to punishment is not as "contradictory" or "elusive" as
some have suggested. Indeed, when compared with other reviews of large bodies
of research (e.g., Chiricos 1987; Hagan 1974; Kleck 1981) the apparent
consistency and significance of the findings are impressive. Even more
impressive is the fact that such results are produced by research using
measures of labor surplus and punishment that capture only a portion of the
underlying phenomena.
In this
regard, it is important to remember that 96 percent
of the findings of a relationship are based on official measures of
unemployment which, as is widely known, underestimate the amount of labor
surplus by as much as one‑half (Perlo 1988; Sorrentino
1979). Moreover, these official
measures do little to capture the dimensions of "social dynamite"
(Spitzer 1975) so central to this issue. In
fact, only 7 percent of the relationships
actually involve age, race, or gender specific measures of even official rates
of unemployment. It is reasonable to
hypothesize even stronger support for
the thesis with measures that
either reflect labor surplus more inclusively or can be effectively
disaggregated to specify "social dynamite."
At the
same time, more than 90 percent of these relationships focus entirely on prison
as the measure of punishment. Only three of the studies at the aggregate level
(Jankovic 1977; Laffargue and Godefroy 1989;
McCarthy 1990) include jail in
their measures of incarceration and there is some indication from individual
level research (e.g., Chiricos and Bales 1991) that jail may be more responsive
to unemployment than prisons. Moreover, as the state's fiscal crisis (O'Connor
1973) promotes cost‑saving alternatives, the use of punishments other
than prison has proliferated (Austin and Tillman 1988; Scull 1977). These
relationships have yet to find their way into the labor surplus/punishment
equation.
The
consistency and significance of the relationship of labor surplus to punishment
is remarkably strong, especially in view of the limits on the measurement of
labor surplus and punishment. However, it is one thing to affirm the
"empirical plausibility" of a relationship and quite another to
explain it. For all the consistency and significance of empirical results, the
research has left many if not most of the key theoretical issues unexamined.
As
noted earlier (Figure 1) at least 11 types of theoretical linkage have been
used to explain the relationship between labor surplus and punishment. Generally,
each presumes that the relationship of labor surplus to punishment is
independent of an actual increase in crime due to labor surplus. Tables 1 and 2
clearly show that the general presumption is correct. The link between labor
surplus and punishment is frequently independent of the mediating effects of
crime. Thus, it is clear that the state's punitive apparatus plays a direct and
significant role in the control of surplus labor.
Beyond
that however, it is not clear that any of the specific theoretical linkages
described in Figure 1 are addressed by the accumulated research. In short,
because research on labor surplus and punishment has been limited primarily to
aggregate measures of unemployment and imprisonment, the
rich diversity of theorizing on this question remains largely ignored. There are, for example, no direct and
independent measures of the value of labor, judicial anxiety, moral panic, or
punitive ideology in any of this research. The structural needs of capital are
more easily described than measured. Operational indicators of "social
dynamite" are limited. Individual level studies do address judicial agency
in the form of sentencing outcomes, but there is no way to infer from existing
research whether and how judicial anxiety, moral panic, or community
intolerance influence those outcomes.
To
examine these issues, other measures of labor surplus and the value of labor
are needed, especially age, race, and gender specific measures. Other indices
of punishment, especially jail and community controls, are also needed. Levels
of aggregation below those of nations and states, which come closer to the
structural level of labor markets and punishments, need greater attention.
Individual level research, which generally has not been informed by theorizing
on labor surplus and punishment, should be integrated with aggregate analyses
of labor market conditions. Closer attention to the interaction of punishment
with other mechanisms for controlling labor surplus is required. The mediating
roles of community intolerance, judicial anxiety, and punitive ideology should
also be directly assessed.
Our suggested areas of research
are not simply a "wish list" for future research. They describe, in
part, the current limits of "empirical plausibility" in our efforts
to explain the complex role of the state in controlling surplus and marginal
labor. That role, as noted above, involves the state in the often contradictory
activities of repression, legitimation, and accumulation. Linking unemployment
and imprisonment rates barely describes those activities. What is called for,
in short, is an imaginative research
strategy which will begin to match our theoretical understanding of
labor surplus and punishment.
Appendix A
• Summary of Aggregate Studies Involving Labor Market Conditions and
Imprisonment
Crime Dependent Dependent Total Dependent
Unit
of Variables
in Labor Market Rate Variable: Variable:
Study/Equation(s) Place Analysis Years of
Study Type of Analysis Equation Indicator Control Admission Population
Berk et
al. (1981)
1. California years 1851‑1970 SE/OLS 7 DEP No (+)*
Berk et
al. (1983)
1. California years 1860‑1975 SE/OIS 7 DEP No (+)(m)*
2.
California years 1860‑1975 SE/OLS 7 DEP No (+)(f)
Box and
Hale (1982)
1. England & Wales years 1949‑1979 OLS 4 UPLRT Yes (+)*
2.
England & Wales years 1949‑1979 OLS 5 UPLRT Yes (+)
3.
England & Wales years 1949‑1979 OLS 4 UPLRT Yes (+)(m)*
4.
England & Wales years 1949‑1979 OIS 5 UPLRT Yes (+)(m)*
5.
England & Wales years 1949‑1979 OLS 4 UPLRT Yes (+)(f
)*
6. England & Wales years 1949‑1979 OLS 5 UPLRT Yes (+)(f )*
7. England & Wales years 1949‑1979 OLS 4 UPLRT(my) Yes (+)(my)*
Box and
Hale (1985)
1. England & Wales years 1952‑1979 OLS 4 UPLRT No (+)*
2. England & Wales years 1952‑1979 OLS 4 UPLRT No (+)(m)*
3.
England & Wales years 1952‑1979 OLS 4 UPLRT No (+)(my)*
Brenner
(1976)
1. US states years 1935‑1973 OLS 6 UPLRT No (+)*
2. US
states years 1935‑1965 OLS 6 UPLRT No (+)*
3. US
states years 1935‑1973 OIS 4 UPLRT No (+)*
4.
Northeast states years 1935‑1973 OIS 4 UPLRT No (+)*
5. Mid‑Atlantic
states years 1935‑1973 OLS 4 UPLRT No (+)*
6. East North
Central states years 1935‑1973 OIS 4 UPLRT No (+)*
7. West North Central states years 1935‑1973 OLS 4 UPLRT No (+)*
8. South Atlantic states years 1935‑1973 OLS 4 UPLRT No (+)*
9. East South Central states years 1935‑1973 OLS 4 UPLRT No (+)
10. West
South Central states years 1935‑1973 OIS 4 UPLRT No (+)*
11.
Mountain states years 1935‑1973 OLS 4 UPLRT No (+)
12.
Pacific states years 1935‑1973 OIS 4 UPLRT No (+)*
13. US
states years 1935‑1965 OLS 4 UPLRT No (+)*
14. Northeast
states years 1935‑1965 OLS 4 UPLRT No (+)*
15. Mid‑Atlantic
states years 1935‑1965 OLS 4 UPLRT No (+)*
16. East
North Central states years 1935‑1965 OLS 4 UPLRT No (+)*
Appendix A • Summary of Aggregate
Studies Involving Labor Market Conditions and Imprisonment (continued)
Total Crime Dependent Dependent
Unit
of Variables
in Labor Market Rate Variable: Variable:
Study/Equation(s)
Place Analysis Years of Study Type of Analysis Equation Indicator Control Admission Population
17. West
North Central states years 1935‑1965 OLS 4 UPLRT No (+)*
18.
South Atlantic states years 1935‑1965 OLS 4 UPLRT No (+)*
19. East
South Central states years 1935‑1965 OLS 4 UPLRT No (+)*
20. West
South states years 1935‑1965 OLS 4 UPLRT No (+)*
21.
Mountain states years 1935‑1965 OLS 4 UPLRT No (+)*
22.
Pacific states years 1935‑1965 OLS 4 UPLRT No (+)*
Carroll
and Doubet (1983)
1. US states (46) states 1970 PA 10 UPLRT Yes (+)
Dobbins
and Bass (1958)
1. Louisiana years 1941‑1954 PM(r) 1 UPLRT No (+)(m)*'
2. Louisiana years 1941‑1954 PART CORR 2 UPLRT No (+)(m)
3. Louisiana years 1941‑1954 MULT CORR 2 UPLRT No (+)(m)
Galster
and Scaturo (1985)
1. US states states 1976 PA 3 UPLRT Yes (-)
2. US states states 1977 PA 3 UPLRT Yes (-)*
3. US states states 1978 PA 3 UPLRT Yes (-)
4. US states states 1979 PA 3 UPLRT Yes (-)
5. US states states 1980 PA 3 UPLRT Yes (-)
6. US states states 1981 PA 3 UPLRT Yes (+))
7. US states states 1976 PA 3 UPLRT Yes (+)*'
8. US states states 1977 PA 3 UPLRT Yes (+)
9. US states states 1978 PA 3 UPLRT Yes (+)
10. US
states states 1979 PA 3 UPLRT Yes (+)
11. US
states states 1980 PA 3 UPLRT Yes (+)
12. US
states states 1981 PA 3 UPLRT Yes (+)
Grabosky
(1980)
1. US states & federal years 1930‑1970 REG(u) 7 UPLRT No (+)
Hale
(1989a)
1. England & Wales years 1953‑1984 REG(u) 3 UPLRT Yes (+)*
2. England & Wales years 1953‑1973 REG(u) 3 UPLRT Yes (+)
3. Males 16‑65 years 1974‑1984 REG(u) 3 UPLRT Yes (+)*
‑
Appendix A • Summary of Aggregate Studies Involving Labor Market Conditions and Imprisonment
(continued)
Total Crime Dependent Dependent
Unit
of Variables
in Labor Market Rate Variable: Variable:
Study/Equation(s)
Place Analysis Years of
Study Type of Analysis Equation Indicator Control Admission Population
Hale
(1989b)
1. England & Wales years 1954‑1984 REG(u) 7 UPLRT Yes (+)*
UPLRT (‑)
2. England & Wales years 1954‑1984 REG(u) 3 UPLRT Yes (+)*
Inverarity
and Grattet (1989)
1. US states years 1948‑1985 REG(u) 4 UPLRT Yes (+)*
2. US states years 1948‑1985 REG(u) 7 UPLRT Yes (+)*
3. US states years 1948‑1985 REG(u) 7 UPLRT Yes (+)*
Inverarity
and McCarthy (1988)
1. US states years 1948‑1984 GLS 4 UPLRT Yes (+)*
2. US states years 1948‑1984 OLS 7 UPLRT Yes (+)*
3. US states years 1948‑1984 2SLS 4 UPLRT Yes (+)*
4. US states years 1948‑1984 OLS 7 UPLRTM Yes (+)*
5. US states years 1948‑1984 OLS 7 UPLRTC Yes (+)*
Inverarity
and Tedrow (1988)
1. US states states 1975 OLS 5 UPLRT Yes (‑)
2. US states states 1976 OLS 5 UPLRT Yes (‑)*
3. US states years 1974‑1985 PD/OLS 6 UPLRT Yes (+)*
4. US states years 1974‑1985 PD/FEB 6 UPLRT Yes (‑)
5. US states years 1974‑1985 PD/FEW 6 UPLRT Yes (+)*
6. US states years 1974‑1985 PD/RE 6 UPLRT Yes (+)*
7. US states years 1974‑1985 PD/GLS 6 UPLRT Yes (+)*
8. US states years 1974‑1985 RE 6 UPLRT Yes (+)*
9. US states years 1974‑1985 RE 6 UPLRT Yes (+)
Jankovic
(1977)
1. US states & federal years 1926‑1974 REG(u) 2 UPL Yes (+)* (+)*
2. US states years 1926‑1974 REG(u) 2 UPL Yes (+)* (+)*
3. US federal years 1926‑1974 REG(u) 2 UPL Yes (‑) (‑)
4. US states & federal years 1947‑1974 REG(u) 2 UPL Yes (+)* (+)*
5. US states years 1947‑1974 REG(u) 2 UPL Yes (+)* (+)*
6. US federal years 1947‑1974 REG(u) 2 UPL Yes (+) (+)
7. US federal years 1960‑1974 REG(u) 2 UPL Yes (+)* (+)*
8. US states 6 federal years 1926‑1974 REG(u) 4 UPL Yes (+)*
9. US states & federal years 1932‑1974 REG(u) 3 UPL Yes (+)* (+)*
Appendix A • Summary of Aggregate Studies Involving Labor Market Conditions and
Imprisonment (continued)
Total Crime Dependent Dependent
Unit of Variables
in Labor Market Rate Variable: Variable:
Study/Equation(s)
Place Analysis Years of Study Type of Analysis Equation Indicator Control Admission Population
r
10. US
states years 1932‑1974 REG(u) 3 UPL Yes (+)* (+)*
11. US
federal years 1932‑1974 REG(u) 3 UPL Yes (+) (+)
12. US
states b‑ federal years 1947‑1974 REG(u) 3 UPL Yes (+)* (+)*
13. US
states years 1947‑1974 REG(u) 3 UPL Yes (+)* (+)*
14. US
federal years 1947‑1974 REG(u) 3 UPL Yes (+) (+)*
15. Sun
Co. Jail years 1969‑1976 REG(u) 2 UPL Yes (+)*
16. Sun
Co. Jail years 1969‑1976 REG(u) 3 UPL Yes (+)*
17. Sun
Co. Jail years 1969‑1976 REG(u) 3 UPL Yes (+)*
Johnson
et al. (1990)
1. US
states states 1982 PM(r) 1 UPLRT No (+)
2. US
states states 1986 PM(r) 1 UPLRT No (+)*
3. US
states states 1982 REG 6 UPLRT Yes (+)*
4. US
states states 1982 REG 5 UPLRT Yes (+)*
5. US
states states 1982 REG 4 UPLRT No (+)
6. US
states states 1986 REG 6 UPLRT Yes (+)*
7. US
states states 1986 REG 5 UPLRT Yes (+)*
8. US
states states 1986 REG 4 UPLRT No (+)*
9. South states 1982 REG 7 UPLRT Yes (‑)
10. Non‑South states 1982 REG 7 UPLRT Yes (+)*
11.
South states 1986 REG 7 UPLRT Yes (+)
12. Non‑South states 1986 REG 7 UPLRT Yes (+)*
Laffargue
and Godefroy (1989)
1.
France years 1929‑1938 ML 3 UPLRT Yes (+) (+)
2.
France years 1952‑1985 ML 2 TENSION1 Yes (+)
3.
France years 1952‑1985 ML 3 TENSION1 Yes (+)(m)*
4.
France years 1952‑1985 ML 3 TENSION2 Yes (+)(m)*
TENSION3 Yes (+)(m)
5.
France years 1952‑1985 ML 4 UPLRTDV Yes (+)(m)*
52/69 Yes (+)(m)*
6.
France years 1952‑1985 ML 3 TENSION1 Yes (+)(m)*
7.
France years 1952‑1985 ML 2 TENSION1 Yes (+)(m)*
8.
France years 1952‑1985 ML 3 TENSION1 Yes (+)(m)*
9.
France years 1952‑1985 ML 4 UPLRTDV Yes (-)
Appendix A • Summary of Aggregate Studies
Involving Labor Market Conditions and Imprisonment (continued)
Total Crime Dependent Dependent
Unit
of Variables
in Labor Market Rate Variable: Variable:
Study/Equation(s) Place Analysis Years of Study Type of Analysis Equation Indicator Control Admission Population
Lynch
(1988)
1. US states years 1950‑1980 OLS 4 UPLRI Yes (+)*
SVRT (+)*
Marenin,
Piscotta and Juliana (1983)
1. US states years 1958‑1978 PM(r) 1 UPLRT No (+)* (+)*
2. US states years 1958‑1978 PART CORR 3 UPLRT Yes (+)* (+)*
3. US states years 1958‑1978 REG(u) 2 UPLRT Yes (+) (+)
4. New England states years 1958‑1978 REG(u) 2 UPLRT Yes (+) (+)*
5. North Central states years 1958‑1978 REG(u) 2 UPLRT Yes (+) (+)*
6. Midwest states years 1958‑1978 REG(u) 2 UPLRT Yes (‑)* (‑)
7. Marginal South states years 1958‑1978 REG(u) 2 UPLRT Yes (‑) (+)
8. Deep South states years 1958‑1978 REG(u) 2 UPLRT Yes (+)* (+)*
9. Mountain states years 1958‑1978 REG(u) 2 UPLRT Yes (‑)* (+)*
10. West
Coast states years 1958‑1978 REG(u) 2 UPLRT Yes (‑) (+)
11. 80% ‑f‑
of pop. in SMSA years 1958‑1978 REG(u) 2 UPLRT Yes (-)* (+)
12. 60‑79%
of pop. in SMSA years 1958‑1978 REG(u) 2 UPLRT Yes (+)* (+)*
13. 40‑49%
of pop. in SMSA years 1958‑1978 REG(u) 2 UPLRT Yes (‑)* (‑)*
14.
_< 39% of pop. in SMSA years 1958‑1978 REG(u) 2 UPLRT Yes (‑) (+)*
McCarthy
(1990)
1. California, jail countries 1981 PM(r) 1 UPLRT No (+)*
2. California, prison countries 1981 PM(r) 1 UPLRT No (+)
3. California, jail countries 1981 REG(u) 5 UPLRT Yes (+)*
4. California, prison countries 1981 REG(u) 5 UPLRT Yes NS
5. MSA, jail countries 1981 REG(u) 5 UPLRT Yes (+)*
6. Non‑MSA, jail countries 1981 REG(u) 5 UPLRT Yes (+)
7. MSA, prison countries 1981 REG(u) 5 UPLRT Yes (+)*
8. Non‑MSA, prison countries 1981 REG(u) 5 UPLRT Yes NS
Melossi
(1987)
1. Italy years 1896‑1965 SE 3 BUSCYCLE Yes (+)
Michalowski
and Pearson (1990)
1. US states (48) states 1970 PM(r) 1 UPLRT No (‑)
2. US states (48) states 1980 PM(r) 1 UPLRT No (+)
3. US states (48) states 1970 REG(u) 5 UPLRT Yes (+)*
4. US states (48) states 1980 REG(u) 5 UPLRT Yes (+)*
Appendix
A • Summary of Aggregate Studies Involving
Labor Market Conditions and imprisonment (continued) x
Total Crime Dependent Dependent
Unit of Variables in Labor Market Rate Variable: Variable:
Study/Equation(s)
Place Analysis Years of Study Type of Analysis Equation Indicator Control Admission Population
r
5. US Non‑south (35) states 1970 PM(r) 1 UPLRT No (+)
6. US Non‑south (35) states 1980 PM(r) 1 UPLRT No (+)
7. US Non‑south (35) states 1970 REG(u) 5 UPLRT Yes (+)
8. US Non‑south (35) states 1980 REG(u) 5 UPLRT Yes
Myers
and Sabol (1987)
1. US northern (14) years census yrs GLS 6 UPL No (+)(m)*
2. US northern (14) years 1890‑1980 GLS 7 UPLwm No (+)*
UPLbm No (+)
3. US northern (14) years census yrs GLS 4 UPLwm No (+)(wm)*
4. US northern (14) years 1890‑1980 GLS 6 UPLwm No (+)(wm)*
5. US northern (14) years census yrs GLS 4 UPLbm No (+)(bm)*
6. US northern (14) years 1890‑1980 GLS 6 UPLbm No (+)(bm)*
Nagel
(1978)
1. US states states 1975 2SLS 5 UPLRT Yes (+)
Parker
and Horowitz (1986)
1. US states states 1974 PM(r) 1 UPLRT No (+) (+)
2. US states states 1975 PM(r) 1 UPLRT No (+) (+)
3. US states states 1975 PM(r) 1 UPLRT No (+)
4. US states states 1976 PM(r) 1 UPLRT No (+) (+)*
5. US states states 1976 PM(r) 1 UPLRT No (+) (+)*
6. US states states 1977 PM(r) 1 UPLRT No (+) (+)*
7. US states states 1977 PM(r) 1 UPLRT No (+) (+)*
8. US states states 1978 PM(r) 1 UPLRT No (+) (+)*
9. US states states 1978 PM(r) 1 UPLRT No (+) (+)*
10. US
states states 1979 PM(r) 1 UPLRT No (+) (+)
11. US
states states 1974‑1979 SE 2 UPLRT No (-) (-)
Sabol
(1989)
1. England & Wales years 1946‑1985 SE 5 UPLRT No (+)*
Wallace
(1980)
1. US states years 1971‑1977 REG(u) 16 LFPRT Yes (+)
LFPRT
(76) (‑)*
Appendix A • Summary of Aggregate Studies Involving Labor Market Conditions and Imprisonment
(continued)
Total Crime Dependent Dependent
Unit
of Variables
in Labor Market Rate Variable: Variable:
Study/Equation(s)
Place Analysis Years of Study Type of Analysis Equation Indicator Control Admission Population
Yeager(1979)
1. US states years FY 1952‑1978 PM(r) 1 UPLRT No (+)* (+)*
2. US states years FY 1952‑1978 REG(u) 3 UPLRT No (+)* (+)*
Notes:
Types of Analysis UPL = unemployment totals
FEB = Fixed effects between UPLRT = unemployment rates
FEW = Fixed effects within UPLRTM = unemployment rate
for the monopoly sector
GLS = Generalized least
squares UPLRTC = unemployment rate
for the competitive sector
ML = Maximum likelihood
MULT CORR = Multiple
correlation coefficient Others
OLS = Ordinary least squares DV52/62
= dummy variable for 1952 or 1962
PA = Path analysis b
= black
PART CORR = partial
correlation coefficient f
= female
PD = Pooled data m
= male
PM(r) = Product moment
correlation w
= white
RE
= Random effects estimates y = youth
REG(u) = Regression with
unidentified estimation technique
RDC = Rank difference
coefficient Footnotes
SE = Structural equations 1.
Lag of 1 year.
2SLS = Two stage least
squares 2.
Lag of 2 years 1942‑1945.
3.
Analysis excludes the "war years" 1942‑1945.
Labor Market Indicators 4. Differentiated rate
variable.
BUSCYCLE = business cycle 5.
Natural log.
LFPRT = Labor force
participation rate 6.
Results for 18+ population.
LPPRT 76 = Labor force
participation rate dummy variable, 1 = 1976 0 = non1976 7. The actual result of the analysis is negative, but it is
indicated as positive to show
SVRT = rate of surplus value that
the results are supportive of the expected relationship.
TENSION I = supply/demand for
jobs 1852‑85 8.
The test of significance does not conform with classical test theory. The
authors
TENSION2 = supply/demand for
jobs 1952‑68 indicate
that the model specified with one year lag terms and a trend variable fits no
TENSION3 = supply/demand for
jobs 1969‑85 better
than a fully saturated model and the signs of the coefficients are
inconsistent.
Appendix B • Summary of Individual
Level Studies Involving Labor Market Conditions and Imprisonment
Labor Variables Dependent Dependent
Type
of Market Subsample in Controls Variable: Variable:
Author(s)/Place Years Sample Description Analysis Indicator Description Equation RGSP Admissions Severity
M
Bernstein
et al. (1977)
1. New York City 1974‑1975 Sample of male defendants,
New DUMREG EMPL6 None 6 RSP (-)
York,
Dec. 1974 to March 1974
Chiricos
and Bales (1991)
1. 2 Florida Counties 1982 Sample
of 2773 felons Logit EMPLI None 9 RSP (+)*
2. 2 Florida
Counties 1982 and misdemeanors OLS EMPLI None 9 RSP (+)
3. 2 Florida Counties 1982 Logit EMPLI Pub
Ord 9 RSP (+)*
4. 2 Florida Counties 1982 Logit EMPL1 Drug 9 RSP (‑)
5. 2 Florida Counties 1982 Logit EMPLI Prop 9 RSP (+)*
6. 2 Florida Counties 1982 Logit EMPLI Violent 9 RSP (+)*
7. 2 Florida Counties 1982 Logit EMPLI Male 9 RSP (+)*
8. 2 Florida Counties 1982 Logit EMPL1 Young
Male 9 RSP (+)*
9. 2 Florida Counties 1982 Logit EMPLI Y
Black Male 9 RSP (+)*
9 RSP
Clarke
&
Koch
(1976)
1. Mecklenberg Co, NC 1971 798
burglary and PCS/CT EMPL5 Prison
x empl 1 *
2. Mecklenberg Co, NC 1971 larceny
defendants PCS/CT EMPL5 add
offense 2 S NS
3. Mecklenberg Co, NC 1971 PCS/CT EMPL5 add
income 3 S NS
4. Mecklenberg Co, NC 1971 PCS/CT EMPL5 add
crim hist 3 SP NS
Farrington
and Morris (1983)
1. Cambridge City 1979 Sample
of 408 from PCS/CT EMPL7 All 1 GP *
2. Cambridge City 1979 Magistrate
Court, Jan PCS/CT EMPL7 Male 1 P *
3. Cambridge City 1979 to
July 1979 PCS/CT EMPL7 Female 1 P NS
4. Cambridge City 1979 REG EMPL7 All 5 P NS
5. Cambridge City 1979 LOGIT EMPU All 6 P NS
6. Cambridge City 1979 REG EMPL7 Male 3 P NS
7. Cambridge City 1979 REG EMPL7 Female 5 P NS
Frazier
and Bock (1982)
1. 6 Florida Counties 1972‑1973 Florida judicial REG FMPLI CORR 1 (+)*
2. 6 Florida Counties 1972‑1973 district, June 72 to REG EMPLI Basic 8 RGSP (+)*
3. 6 Florida Counties 1972‑1973 May 73, N=229, key REG EMPLI ADD:
Prob 9 RGSP (+)*
4. 6 Florida Counties 1972‑1973 comparison individual REG EMPLI ADD: J1 10 RGSP (+)
5. 6 Florida Counties 1972‑1973 judges REG EMPLI ADD: J2 10 RGSP (+)
6. 6 Florida Counties 1972‑1973 REG EMPLI ADD:
J3 10 RGSP (+)
Appendix B • Summary
of Individual Level
Studies Involving Labor Market Conditions and Imprisonment (continued)
Labor Variables Dependent Dependent
Type
of Market Subsample in Controls Variable: Variable:
Author(s)/Place Years Sample Description Analysis Indicator Description Equation RGSP Admissions Severity
7. 6 Florida Counties 1972‑1973 REG EMPL1 ADD:
J4 10 RGSP (+)
8. 6 Florida Counties 1972‑1973 REG EMPL1 ADD:
J5 10 RGSP (+)
9. 6 Florida Counties 1972‑1973 REG EMPL1 ADD:
J6 10 RGSP (+)
10.
6 Florida Counties 1972‑1973
Sent. disp. by judge REG EMPL1 Basic/SA 10 RGSP (+)
11.
6 Florida Counties 1972‑1973
Background Char. REG EMPL1 Basic/AGE 10 RGSP (+)
12.6
Florida Counties 1972‑1973 REG EMPL1 Basic/YRS 10 RGSP (+)*
Kruttschritt
(1980)
1. City in Northern CA 1972‑1976 Female sample, Dec 72 REG EMPL4 Dist Peace 4 R (+)
2. City in Northern CA 1972‑1976 to 76, N=1034 REG EMPL4 Assault 4 R (‑)
3. City in Northern CA 1972‑1976 REG EMPL4 Petty
Theft 4 R (‑)
4. City in Northern CA 1972‑1976 REG EMPL4 Forgery 4 R (+)*
5. City in Northern CA 1972‑1976 REG EMPL4 Drugs 4 R (+)
Lotz
and Hewitt (1977)
1. King County, WA 1973 1832
convicted felons REG WORK None 27 RGSP (+)*
Miethe
and Moore (1985)
1. Minnesota 1978‑1979 1226 convicted felons REG EMPL1 PRE 15 RGSP (+)*
2. Minnesota 1980‑1981 1280 convicted felons REG EMPL1 POST 15 RGSP (+)*
3. Minnesota 1978‑1979
1226 convicted felons REG EMPL1 PRE 15 RGSP (+)
4. Minnesota 1980‑1981 1280 convicted felons REG EMPL1 POST 15 RGSP (+)
Miethe
and Moore (1986)
1. Minnesota 1977‑1978
2329 felony offenders, REG EMPL1 Alt DV 12 RGSP (+)*
2. Minnesota 1977‑1978
sentenced July 1977 to REG EMPL1 Alt DV 13 RGSP (+)
3. Minnesota 1977‑1978
June 1978 REG EMPL1 White 12 RGSP (+)*
Black 13 RGSP (+)*
4. Minnesota 1977‑1978 REG EMPL1 White 12 RGSP (+)
Black 13 RGSP (+)
Moore
and Miethe (1986)
1.
Minnesota 1980‑1981 1523 felony cases, July REG EMPLI None 18 RGSP (‑)
1980
to June 1981 EMPLST
(‑)*
2. Minnesota 1980‑1981
(focus on guidelines REG EMPL1 None 19 RGSP (+)
outcome) EMPLST (+)
Myers (1979)
1.
Marion County, IN 1974‑1976
980 felony cases, Jan. REG EMPL2 None 6 RGSP NS
1974
to June 1976
Appendix B • Summary of Individual Level Studies Involving Labor
Market Conditions and Imprisonment (continued)
Labor Variables Dependent Dependent
Type of Market Subsample in Controls
Variable:
Variable:
Author(s)/Place Years Sample Description Analysis Indicator Description Equation
RGSP Admissions Severity
Myers
(1987)
1. Georgia 1976‑1982 Random sample of felony REG EMPL1 Alt DV
22 RGSP (+)*
2. Georgia 1976‑1982 convictions, Jan 1976 REG EMPL1 Alt DV
22 RGSP (‑)*
to
June 1982 N=15,270
Myers
and Talarico (1986)
1. Georgia 1976‑1982 Random sample of felony REG EMPL1 Basic
17 RGSP (‑)
2. Georgia 1976‑1982 convictions, Jan 1976 REG EMPL1 Add: BUR &
25 RGSP (‑)
to
June 1982 N=17,217 CON
Smith
(1986)
1. 6 US Cities 1978 Males charged with Logit EMPL3 Pled Cases
23 R (+)*
2. 6 US Cities 1978 robbery and burglary; Logit EMPL3 Tried Cases
20 R (+)*
comparison
of 1,533
pled
and 387 tried cases
Unnever
(1982)
1. Miami, FL 1971 Convicted male drug Logit EMPL1 CORR
1 (+)
2. Miami, FL 1971 offenders, N=313 Logit EMPL1 Race & Empl
3 RGSP (+)
3. Miami, FL 1971 Logit EMPLI ADD: pr & see
7 RGSP (+)*
4.
Miami, FL 1971 Logit EMPLI ADD: Lawyer 8
RGSP (+)
5. Miami, FL 1971 Logit EMPLI ADD: bail out
10 RGSP (+)
Unnever
et al (1980)
1. 6 Florida counties 1972‑1973 Florida judicial Logit EMPLI None
8 RGSP (+)*
district,
June 1972 to
May
1973, N=229
Notes:
Labor
Market Indicators EMPL5
= unemployed, employed, unknown
EMPI. =
unemployed, employed EMPL6
= time employed: unemployed 6+ months, unemployed
EMPL1 =
unemployed, employed, self‑employed <
6 months, employed < 6 months, employed 6+ months
EMPL3 =
unemployed, irregular, part‑time, full‑time EMPL7
= school, employed, unemployed
EMPL4 =
employed, temporarily unemployed, welfare/not looking, retired, EMPLST = employment
stability
housewife, student WORKHIST
= work history