Solution
David answered on
Dec 22 2021
The Effect of the Maximum Length of Incarceration on the Degree of Crime: An Empirical Investigation
1. Introduction
The United States (US) has seen varying degree of crime rate over time. According to the Criminal Victimization Survey (2011), crime rates in America generally rose after World War II, and peaked between the 1970s and early 1990s; however, crime rate declined after 1990. Studies (Levitt, 2004; Donohue and Levitt, 2000; Kleiman, 2009) have cited several reasons for the decrease in the rate of crime in the US. Among these factors, the increase in the number of police officers, the increase in the number of imprisonment, changing demographics of the aging population are some of the main causes behind the decrease in the rate of crime in the US. The US has the highest rate of incarceration in the world. In 2006, 7 million people were on probation or parole, of which 2.2 million were incarcerated (Walmsley, 2011; Lee 2011). Eide (2000) suggests that an increase in the probability of punishment through enforcement has a greater negative effect on the rate of crime than the severity of punishment.
Historically, the criminal justice system has used imprisonment to deal with offenders, in anticipation that imprisonment will lower the rate of crime through a combination of dete
ence and incapacitation effects; however, existing literature show little connection between sentencing and crime rate. A large number of studies find no co
elation between the severity of sentencing and the level of crime. Hoel and Gelb (2008) argue that offenders are often guided by their impulses, rather than prudent foresight, and hence, the idea of tougher sentences to deter crime may not be valid. On the other hand, Kelaher and Sarafidis (2011) show that criminal justice system can exert much greater influence on crime than what past studies have suggested. According to them, the risk of apprehension and conviction are more effective in reducing crime than increasing the severity of punishment. They argue that the lack of consensus that exists among the studies on the impact of the criminal justice system on crime is mainly due to simultaneity; dynamic misspecification and omitted variable bias. Erhlich (1973) shows that crime varies inversely with the probability of prison and the time served in prison. Durlauf and Nagin (2011) suggest that the reduction in the rate of crime can be achieved through shortening of sentences (imprisonment period) and to use the cost savings to support more and better policing. Hellman and Alper (2000) show that the increased cost of crime in terms of higher sentences or higher probabilities of conviction are associated with reduced crimes. Wright (2010) in evaluating certainty versus severity of punishment show that increase in the certainty of punishment, opposed to severity of punishment are more likely to produce dete
ent benefits. She argues that punishment may be expected to affect dete
ence in one of the two ways: first through the increase in the certainty of punishment, which will deter the potential offenders through the risk of apprehension; and second through the severity of punishment. Research also finds that lower risk offenders are more likely to be negatively affected by incarceration. Wright (2010) argues that unlike long sentences, when term of imprisonment is short then offenders can often maintain their ties with the society (family, employers, and community), which later help them to successfully re-enter the society.
Dete
ence theory assumes that human beings are rational, and therefore, they will always consider the consequences of their behavior before committing a crime. However, people under the influence of drug or alcohol often fail to think rationally and get involved into crimes and even violent crimes even with a past record of conviction and incarceration.
Findings from the past studies suggest that punishment through incarceration can have both positive (through adverse effect of punishment) and negative effects (through dete
ence effect of punishment) on the rate of crime.
In order to examine whether imprisonment has any effect on the rate of crime, in this study, I analyze the effect of the length of imprisonment on the severity of crime (degree of crime). Past studies use the dete
ence effect of punishment on the rate of crime; however, no study has examined the effect of incarceration on the degree of crime, hence, the results from my study will add to the existing literature in a significant way. Using data from the Pennsylvania Sentencing 1996, I find that in general the maximum length of incarceration increases the severity of crime; however, when I use different lengths or periods of incarceration, I find both positive and negative effects of incarceration on the degree of crime. Degree of crime is likely to decrease with the period of incarceration being less (less than a year, 1 to 5 years, and 6 to 10 years); however, the degree is likely to increase if the period of imprisonment is more (more than 20 years).
I organize the paper as following: in Section 2, I describe the data. In Section 3, I explain the empirical model. In Section 4, I present the results from the empirical estimation and finally in Section 5, I conclude.
2. Data
In this paper, I use the Pennsylvania Sentencing data 1996. The data is collected on the basis of the sentencing reports as presented by the judges to the Pennsylvania Commission (legislative agency) when they use the sentencing guidelines as provided by the Commission to sentence felony and misdemeanor offenses. This data reflect all the felonies and the misdemeanor that were sentenced during the calendar year 1996. The data provide information on each offender’s demographic characteristics, prior and cu
ent offense history, the Offense Gravity Score assigned by the Commission, and the sentence given to the offender. The raw data consist of 1, 01,823 observations. After deleting all the observations with missing values for the key variables, I obtain a sample of offenders with 45,747 observations. Table 1 presents the definition of the variables and the descriptive statistics for the variables in my model.
In order to analyze the effect of incarceration on the degree of crime, I use the Offense Gravity Score (variable GRADE in the raw data) to define the dependent variable for my analysis, where the variable CRIMEDEG takes on the value 1 if the crime committed is severe or of higher degree (i.e., felony 1 or 2; murder) and 0 if the crime committed is less severe or of lower degree (felony 3; misdemeanor 1, 2, and 3; unclassified felony and misdemeanor). In my sample, 19% of the total crime is of higher degree, whereas 81% is of lower degree.
The main independent variable in my analysis is INCAR, which defines the maximum length of incarceration in months (variable INCMAX in the raw data). The maximum length of incarceration varies between a maximum period of less than 12 months in sentence and a period of more than 20 years in sentence (50 years of sentence, life sentence, and death sentence).
I further extend my analysis by dividing the maximum period of incarceration into five different binary variables in order to examine their effects on the degree of crime variable (CRIMEDEG). The variable PLT12M takes on the value 1 if the maximum length of incarceration is less than 12 months, 0 otherwise. Similarly, the variables P1T5Y, P6T10Y, P11T20Y, and PDT20Y are equal to 1 if the maximum length of incarceration is 1 to 5 years, 6 to 10 years, 11 to 20 years, and more than 20 years, respectively; and 0 otherwise. Most of the offenders were sentenced for 1 to 5 years (84%) followed by the maximum period of incarceration for 6 to 10 years (7%), less than 12 months (6%), more than 20 years (2%), and 11 to 20 years (0.2%). In my analysis, I treat P11T20Y as the reference category.
Table 2 reports the summary of maximum length of incarceration (months) by the degree of crime. It shows that the average length of incarceration is approximately 2 years if the degree of crime is low (felony 3; misdemeanor 1, 2, or 3; unclassified felony and misdemeanor), whereas it is nearly 7 years if the degree of crime is high (felony 1 or 2; or murder).
Table 3 presents the frequency of crime degree by the maximum length of incarceration in months. It shows that higher degree of crime is associated more with 6 to 10 years (1,996), 11 to 20 years (1,014) and more than 20 years (113) of incarceration terms, as compared to that of the lower degree of crime (1,233, 66, and 1, respectively). Whereas lower degree of crime is associated more with the incarceration terms of less than 12 months (2,757) and 1 to 5 years (33,210) as compared to that of the higher degree of crime (121, 5,236, respectively).
I also define other variables for my model as follows: (a) Age during the time of offense may be important to understand the maturity level of the offender. Degree of crime may vary with the offender being a youth or an adult. Therefore, I include the offender’s age during the time of offense (AGE), calculated as the difference between the date of offense and the date of birth of the offender in my model. The average age of an offender during the time of offense is almost 31 years; (b) Offense may vary by the gender of the offender. Crimes may be associated more with males than with females; therefore, I include the gender (MALE=1 if the offender is male, 0 otherwise) of the offenders in my model. In my sample, 88% offenders are male. (c) Crimes may also vary by race/ethnicity; therefore, I include three binary variables indicating race/ethnicity of the offenders. The binary variable WHITE equals 1 if the race of the offender is white, 0 otherwise. The binary variables BLACK and HISP take the value 1 if the offender is black and Hispanic, respectively, 0 otherwise. The reference group consists of Asian, American Indian, and others (ORACE). 59% of the offenders in my sample are white, 29% are black, 7% are Hispanic, and the rest 4% belong to other races/ethnicity. (d) Past record of crime is also included in the analysis. Past record of any crime may discourage or encourage an offender to commit another crime; therefore, I define the variable PCRIME which takes the value 1 if the offender was convicted in a court of law for any prior crime, 0 otherwise (includes adjudication). 46% of the offenders in my sample have committed some prior crime for which they were convicted in a court of law. In addition, I define three interaction terms: WMALE (WHITE*MALE), BMALE (BLACK*MALE), HMALE (HISP*MALE) in order to identify the effect of these interaction terms on the degree of crime.
In the next section, I discuss the empirical model.
3. Empirical Model
In Model I, I analyze the effect of the length of incarceration on the degree of crime using the following probit model:
Model I
CRIMEDEG = α0+β1*INCAR+β2*AGE+β3*MALE+β4*WHITE+β5*BLACK+β6*HISP
+β7*PCRIME+ β8*WMALE+ β9*BMALE+ β10*HMALE+ε1 ……. (1)
I use several specifications to a
ive at the baseline model. In specification 1 (SPECIF1), I regress CRIMEDEG on INCAR. In specification 2 (SPECIF2), I add AGE with SPECIF1. Subsequently, I add MALE with SPECIF2 and WHITE, BLACK, and HISP with SPECIF3 to obtain specification 3 (SPECIF 3) and 4 (SPECIF 4), respectively. In specification 5 (SPECIF5), I add the variable PCRIME with SPECIF4. Finally, in specification 6 (SPECIF6), I add the interaction terms (WMALE, BMALE, and HMALE, respectively) with SPECIF5.
In Model II, I further analyze the effect of the length of incarceration on the degree of crime using the binary independent variables PLT12M, P1T5Y, P6T10Y, P11T10Y, and PGT20Y in the place of INCAR in Model I (Equation 1). Here, I use the following probit model:
Model II
CRIMEDEG = δ0+γ1*AGE+γ2*MALE+γ3*WHITE+γ4*BLACK+γ5*HISP+γ6*PCRIME
+γ7*WMALE+γ8*BMALE+γ9*HMALE+γ10*PLT12M+γ11*P1T5Y
+γ12*P6T10Y+γ13*PGT20Y+ε2 ……. (2)
Punishment or sentencing in the form of incarceration may discourage (“dete
ence effect”) an offender to commit future crimes (or reduce the degree of crime). Alternatively, punishment or sentencing may be associated with increased number of crimes or more severe crimes (“adverse effect”) depending upon the severity (impact of sentencing on a person’s psychology) and the length (more time spent in a judicial custody may impact a person adversely) of sentencing. Therefore, my first empirical hypotheses for the two models are as following:
Hypothesis E1(a): in Model I, the effect of INCAR on CRIMEDEG may be positive or negative, ceteris paribus, thus β1 can be less than or greater than 0.
Similarly, different lengths of sentencing (incarceration) may have different effects on the degree of crime, thus,
Hypothesis E1(b): in Model II, the effect of the binary independent variables (PLT12M, P1T5Y, P6T10Y, PGT20Y) on the degree of crime may be positive or negative, all else constant; therefore, γ10 to γ13 can be less than or greater than 0.
I use the empirical model and the data to examine the true effect of incarceration (sentencing) on the degree of crime, that is, to find whether the dete
ence effect or the adverse effect of incarceration dominates.
A young person or less matured person may be more prone to committing crimes due to their vulnerability in getting addicted in intoxicated substances (such as drugs, drinking, etc.), or may be more prone to getting involved in...