Uncovering the Gender Participation Gap,in the Crime Market. Evelina Gavrilova,Norwegian School of Economics,Nadia Campaniello. University of Essex,Discussion Paper No 8982,April 2015. P O Box 7240,53072 Bonn,Phone 49 228 3894 0,Fax 49 228 3894 180. E mail iza iza org, Any opinions expressed here are those of the author s and not those of IZA Research published in. this series may include views on policy but the institute itself takes no institutional policy positions. The IZA research network is committed to the IZA Guiding Principles of Research Integrity. The Institute for the Study of Labor IZA in Bonn is a local and virtual international research center. and a place of communication between science politics and business IZA is an independent nonprofit. organization supported by Deutsche Post Foundation The center is associated with the University of. Bonn and offers a stimulating research environment through its international network workshops and. conferences data service project support research visits and doctoral program IZA engages in i. original and internationally competitive research in all fields of labor economics ii development of. policy concepts and iii dissemination of research results and concepts to the interested public. IZA Discussion Papers often represent preliminary work and are circulated to encourage discussion. Citation of such a paper should account for its provisional character A revised version may be. available directly from the author,IZA Discussion Paper No 8982. April 2015, Uncovering the Gender Participation Gap in the Crime Market 1. There is little research on the gender variation in the crime market We document a gender. gap in criminal activities based on property crimes using data from the U S National. Incident Based Reporting System from 1995 to 2015 We show that there is a gender. participation gap with only 30 percent of the crimes being committed by females We try to. explain the gender participation gap by focusing on incentives to commit crime such as. criminal earnings and probability of arrest We show that on average females earn 13 percent. less than males while they face a 9 percent lower likelihood of arrest We find that males. respond more to changes in illegal earnings with an elasticity of 0 36 while females are less. responsive with an elasticity of 0 23 Both genders respond equally to changes in the. probability of arrest with an elasticity around 0 14 Using a Blinder Oaxaca type. decomposition technique we find that differences in incentives explain about 8 percent of the. gender participation gap while differences in responsiveness to changes in incentives. especially illegal earnings explain about 56 percent of the gap The fact that females behave. differently than males has implications for the heterogeneity in response to crime control. JEL Classification J16 K42, Keywords gender gap in crime crime incentives synthetic panel. Corresponding author,Nadia Campaniello,University of Essex. Department of Economics,Wivenhoe Park,Colchester Essex CO4 3SQ. E mail ncampa essex ac uk, This is a revised version posted in March 2018 The original version circulated in April 2015 included. fewer years in the empirical analysis and omitted important fixed effects. 1 Introduction, In the past years there has been considerable attention given to the gender gaps in wages. and employment Yet almost no notice has been given to the gender gap in an illegal setting. and none on the gender gap in criminal earnings Most research in the economics of crime. has focused on male perpetrators Levitt Miles 2007 Freeman 1999 with the implicit. assumptions that female crime is so little that it is of no consequences or that policy impli. cations have external validity across genders Yet female crime has been growing in the last. decades as we show with the percentage share of women who are incarcerated in Figure 1. and it is not possible anymore to assume it away One of the main reason for which there. is a little research on crime is the data limitation on demographic details about the criminal. perpetrators, In this article we want to fill the gap in research on the gender crime gap In the. economic model of crime criminal participation is intrinsically connected with the possible. illegal earnings and detection probability We use detailed matched offense arrest data on. property including white collar crimes from the the U S National Incident Based Reporting. System NIBRS for the period 1995 2015 These data offer the unique opportunity to link. the participation decision to the likelihood of arrest and the illegal earnings We match the. NIBRS data to population statistics and compute cohort specific crime rates2 where cohorts. are defined based on interval of age race gender and county allowing us to explore the. gender participation decisions in crime in light of the associated illegal earnings and arrest. probabilities, We start off by quantifying the gaps in crime for participation illegal earnings and arrest. probability Only around 30 percent of the crimes are committed by women We find that. females earn 7 percent less than males and face a similar probability of arrest When we. explore the heterogeneity with respect to crime type we find that females participate most in. shoplifting In the absence of this crime we observe a more severe earnings gap of 13 percent. and a 9 percent lower probability of arrest for females We interpret this as possible evidence. that females sort into shoplifting as a way to obtain higher criminal earnings yet they suffer. from the higher detection and arrest rate associated with this particular crime. In our second set of results we try to relate the participation rates to earnings and ar. rest probabilities First we assume that criminals rationally form their expectations on the. criminal incentives3 based past experience This mechanism is similar to the one described in. Lochner 2007 Then we use these predicted values to see how female and male crime rates. respond to changes in expected illegal earnings and in the probability of arrest allowing us. to identify gender specific elasticities, We find strong evidence that past incentives predicted future ones generating a strong. first stage As for the responsiveness we find that males have a higher elasticity of illegal. earnings of 0 36 compared to 0 23 for females Both genders have a similar elasticity of. arrest predicting that one percent increase in the probability of arrest leads to a 0 14 percent. decrease in crimes committed By exploiting a partial Blinder Oaxaca decomposition we try. to measure how much of the participation decision is due to different criminal incentives and. different elasticities We find that if females were more manly with respect to incentives. Henceforth we will refer to these crime rates as participation rates. In the standard economic model of crime both higher earnings and a lower arrest probability can persuade. a criminal to commit a crime Therefore we will refer to these two variables as criminal incentives The. probability of arrest deters criminals generating a negative incentive rather than a positive one In our model. we allow males and females to respond differently to these incentives. and responsiveness to them this would reduce the participation gap by 64 percent. The main implication of our findings concerns the external validity of previous studies that. have been conducted only on a sample of male offenders We show that males and females. respond differently to incentives, Crucially our elasticities with respect to the probability of arrest could be compared to. previous literature In a recent review of the literature Chalfin McCrary 2017 show that. estimated elasticities with respect to an increase in the police manpower are in the range of. 10 to 20 percent Our estimates fall well within this range and are significant at conventional. To the best of our knowledge there are no studies with US data with estimated elasticities. with respect to illegal earnings to which we can compare our estimates Our elasticities are. roughly in the range 25 40 percent as estimated by Draca et al 2015 on UK data Our. estimates imply that policies that aim to reduce the illegal earnings of criminals would impact. males more than females, Our research is of interest to policy makers that want to decrease crime We provide ev. idence for the heterogeneity in response between genders to policies that alter the incentives. to engage in crime More specifically if the policy maker wants to discourage males to partic. ipate in crime this would be most effective with policies that decrease the value of potential. earnings Examples for such policies are black market regulation where pawn shops could be. placed under additional surveillance or additional security for higher value items see D Este. Related Literature We contribute to several strands in the literature Most prominently in. the Handbook of Labour Economics Freeman 1999 acknowledges the gap in studies about. the gender variation in crime and underlines that there are no studies by economists that. analyze the large differences in the participation of males and females Since then there has. been scant response to this apparent gap and we are one of the first to fully investigate this. research question, On one hand there are a few papers that look at the criminal market in isolation The. earliest economic study on female criminals Bartel 1979 investigates the determinants of. female participation in crime through an Ehrlich type model of time division The author finds. that probabilities of conviction and arrest have a deterrent effect on females in some property. crimes Our results are in line with her findings Gavrilova 2017 investigates the incentives. for assortative matching in a criminal pair and finds that females are likely discriminated. against when matching with males, On the other hand there is a small literature that tries to identify the participation decision. by looking at shifts in the opportunity cost of crime Recently Corman et al 2014 find that. the 1996 welfare reform in the U S aimed at incentivizing female work led to a decrease. in female arrests for serious property crimes by 4 4 4 9 Cano Urbina et al 2016 look. at the effect of education on female crime They find that one more year of school for females. reduces on average property and violent crimes by 50 while they do not find any effect on. white collar crime They argue that the effect of education on crime for females is probably. due to changes in marital opportunities rather than labor market opportunities In our paper. we control for employment opportunities and wage that represent the opportunity costs of. being involved in criminal activities, A recent study by Beatton et al 2017 examines the convergence of the gender crime. gap in Australia They find gaps that are in line with the ones we find in the US context. They show that over the course of the last 20 years these gaps have contracted by 30 percent. In descriptive evidence here we show that over almost the same period of time the US crime. gap has contracted by 28 percent again a very similar number Overall female participation. in crime seems to be increasing around the world, By focusing our analysis on the illegal earnings of criminals we are contributing to the. understanding of the most understudied element of crime Draca Machin 2015 Recent. literature has only attempted to approximate the illegal earnings of criminals such as Draca. et al 2015 while we have more precise information on the value of the property stolen. Finally we use NIBRS data on property offenses that include white collar crimes a quite. neglected and understudied typology of crimes in the literature. 2 Descriptives, For our analysis we use the National Incident Based Reporting System NIBRS This dataset. contains records on the universe of crime incidents for a given year for a given law enforcement. agency in the United States The data are not representative for the United States as a whole. as many agencies do not submit reports and the expansion of data collection is on going A. typical observation is a coded report about a criminal incident It contains the number of. perpetrators their demographic characteristics and crime codes We match the report data to. the arrest data in order to see whether a perpetrator was arrested We observe both arrested. and non arrested criminals Criminal earnings are recorded regardless of whether there was. We limit our analysis to property crimes including white collar crimes We do this for. two reasons First property crimes are more common than other crimes The violent crime. rate over the period 2016 is 386 crimes per 100 000 inhabitants while the property crime rate. is 2 451 crimes White collar crimes themselves seem to be even more prevalent according to. survey evidence from Huff et al 2010 Second for property and white collar crimes we can. observe a relevant incentive such as the illegal earnings while for violent crimes the incentives. are difficult to quantify, In Table 1 we detail the types of crimes that we analyze Once we select these crime. incidents we have 45 million observations on criminals over the period 1995 to 2015 We. select individuals between 15 44 years of age of black or white race This selection is mainly. guided by the availability of control variables as we want to approximate the opportunity. cost of crime in the best way we can In addition we exclude the two other races Asian and. Native American that are provided in the data because they comprise a too small fraction 3. percent each in the dataset, In order to understand the criminal participation decision we define pseudo individuals. and construct a synthetic panel see Deaton 1985 We aggregate crimes within cohorts and. divide by the corresponding population thus getting a measure of crime rates and variation. in the participation decision The resulting unbalanced panel is treated as pseudo individuals. that can be tracked over time The cohorts are defined based on interval of age 15 24 25 34. and 35 444 race black and white gender male female and county The panel spans over. a period of 20 years, We take data on population by age gender race year and county from the Wide. ranging Online Data for Epidemiologic Research WONDER Data on average wages and. Data on the general population at county level are just available until the age of 44 and we know that 75. percent of the crimes are committed by perpetrators from the selected age groups. employment rates by age gender race decade and state are taken from the CENSUS. Integrated Public Use Microdata Series IPUMS USA,2 2 Descriptive Evidence. Table 2 shows the summary statistics by gender Across cohorts we observe that there are on. average 20 females committing crimes versus 39 males These numbers correspond to crime. rates of 713 per 100 000 inhabitants for males and 331 per 100 000 inhabitants for females. Both of these crime rates are below conventional levels of property crime rates in the US as. cited above for 2016 the crime rate was 2 451 Likely the difference comes from the fact that. in our sample we keep offenders whose demographic characteristics were well observed in a. certain age group while in the official statistics reports are aggregated for all types of offend. ers both unknown and observed Females have a smaller likelihood of arrest of 35 percent. versus 36 percent for males We also show that females have less earnings at 1724 versus. 1928 for males, In Figure 2 we show how these statistics have evolved over time We observe that the fe. male male ratio in criminal participation increased over the 20 year sample period Concur. rently with this increase in female participation we observe that women on average have. decreased their criminal earnings and slightly increased their probability of arrest with respect. In Figure 3 we present the gender participation gap as a function of the earnings gap top. row and the arrest gap bottom row for the different crime types In panel a we show that. when females relative earnings increase their participation in criminal activities decreases. but this relationship seems to be driven by shoplifting Panel b shows that the relationship. becomes positive as one would expect if criminals respond to incentives when shoplifting is. excluded In panel c we show that females participate more in criminal activities when their. relative likelihood of arrest is lower which is again coherent with economic theory A shown. in panel d shoplifting seems to be in line with the overall trend. Since behind the averages in Figure 2 there might be significant hidden heterogeneity in. Figure 4 we compare the densities of earnings and arrest probability for males and females. We present earnings on the top row and arrest probability on the bottom row The first. column shows the overall data while the second column excludes shoplifting In panel a we. observe that female earnings are bimodal while male ones are more to the right in location. In panel b when excluding shoplifting the two distributions are more similar and comparable. However the male distribution still remains to the right of the female mirroring the averages. from Table 2 and showing that males earn more than females In panel c we observe the. density of the probability of arrest for males and females and we notice the sharp spike in. the right part of the density of females In panel d we show that this spike is driven by. shoplifting and as we exclude it we again arrive at more comparable distributions The fact. that shoplifting can drive the spike in panel c of Figure 4 could be due to the crime being. reported only if an arrest takes place Since this would bias the elasticity of crime with respect. to the probability of arrest we exclude shoplifting in the main analysis but include it in the. robustness checks,3 Model of Crime,3 1 Theory, Following Becker 1968 we assume that an individual compares the expected utility of com. mitting a crime with the expected utility of not committing a crime The expected benefits. from a crime are the illegal earnings The expected cost is the sanction length Both benefits. and costs hinge on the realization of the probability of detection and arrest The expected. utility of not committing the crime is based on the opportunity cost of crime being engaged. in the legal labor market We introduce in the model a gender specific component in order. to test for difference between genders in the subsequent sections of this paper. Therefore an individual from gender g g m f decides to be involved in a criminal. activity if a gender specific function fg of costs and benefits is larger than an individual. idiosyncratic error g g can measure any unobserved determinants of crime such as so. ciocultural and family factors By allowing the function to be gender specific we allow each. gender to respond differently to their own expectations. fg E Yg E Pg E Wg E Lg E Jg g 1, We take the expectation of the criminal over variables like illegal earnings denoted by. Y probability of arrest P legal wage W employment L and sanctions J In order to. arrive to a estimable equation we log linearize the function f with respect to the variables. in the model and note that when we sum over all the cases in which fg g delivers the. crime rate Cg,cg g1 yg g2 pg g3 wg g4 lg g5 jg g 2. where lower case letter denote logged quantities Assuming a specific functional form. for f allows us to gauge some intuition about the importance of the different pieces If. criminals have constant relative risk aversion utility functions the different factors that enter. the utility function are unrelated to each other and criminals can either work or engage in. crime condition 1 becomes 5,fg Y g 1 g P g J g W g L g 1 g 1 g g 3. Log linearizing and assuming that g is uniformly distributed it is easy to show that the. elasticity of crime with respect to criminal earnings is proportional to 1 g while the. elasticity of crime with respect to the probability of arrest would be the same for males and. females This would imply that more risk loving criminals are the more they respond to. changes in the expected illegal earnings, Using a log log specification the coefficients measure elasticities Equation 2 implies that a. gender crime participation gap cm vs cf could be both due to 1 gender specific differences. in key variables such as illegal earnings yg and 2 over the way criminals respond to such. incentives for e g g1 The subscript g on the coefficients g denotes that the effect of the. variable could be different by gender With the subscript g on the variables we model the. fact that the expectation could be formed over the gender specific expected values for the. variables That is the expectations for males are based on the variables like earnings and. arrest of other males 6,Hats indicate criminals perceptions. We use the variable g in subscript here in order to describe better the data in the next subsection when. More formally we explicitly model the expectations of criminals as a function of the. previous realization of the variables Lochner 2007 shows with survey data that indeed. criminals form expectations on the probability of arrest by observing realizations in their own. peer group Similarly we assume that the expectation over the probability of arrest pg and. earnings yg is formed by observing the peer group within the cohort of criminals We define. cohorts at the crime type gender race age group and location level so that the cohort. captures the peer group and this peer group is significantly homogeneous This is similar to. Lochner 2007 who also defines a narrow peer group over which the expectations of criminals. are modeled,3 2 Estimation, In this section we present our empirical methodology There are two main reasons for the. presence of a participation gap First there could be a difference in the incentives that each. gender faces such as probability of arrest or earnings 7 Second even if the incentives were. the same maybe the two genders react differently to them In the next subsection we show. how we want to describe the differences between the two genders in terms of illegal earnings. and arrest two of the main incentives in the crime participation decision Second we show. how sensitive this participation decision is to changes in incentives of probability of arrest and. illegal earnings,3 2 1 Incentives, In Section 2 and Figure 3 we document significant differences between males and females in. terms of illegal earnings arrest and participation To describe how earnings and arrest vary. with gender and other factors we estimate the specifications of the following type for the. criminal earnings and probability of arrest,zgit D g f git Xgit git 4. where zgit is the dependent variable either the log illegal earnings or the log arrest rate. for criminals of gender g in cohort i and year t D g f is an indicator function equal to. one when the gender is female A 0 would imply that the dependent variable does not. vary with the gender X is a vector containing personal cohort traits like race age average. wage and unemployment rate, We include offense fixed effects in order to account for the effects of different offenses on. the unconditional gap For example we expect that a criminal would earn more in auto theft. crimes than in shoplifting and if males specialize in the former while females specialize in. the latter this would earn a high unconditional gap In order to control for county specific. heterogeneity such as police presence in any given year we include county year fixed effects. Furthermore in some specifications we also includes state offense fixed effects to control for. differences in sanctions across different states Finally we cluster the standard errors at the. county level in order to account for correlation of residuals over time within county. we define the estimating equations, The probability of arrest is a deterrent to crime rather than an economic incentive In the interest of. the following exposition we will refer to both earnings and arrest probability as incentives while keeping in. mind that incentives is heterogeneous category,3 2 2 Responsiveness to Incentives. Once we map the log differences between the two genders in criminal earnings and arrest. probabilities we turn to the participation decision Starting from Equation 2 we obtain. cgit g1 ygit g2 pgit Xgit g git for g m f 5, where cgit is the log crime rate defined as the number of crimes committed by criminals. in gender g in a given year t by people within a cohort i The cohort is defined by age group. race and county and the crime rate is divided by the general population in the same cohort. y is the log of expected illegal earnings p is the log of the expected probability of arrest. X is a vector containing sociodemographic variables specific to the cohort such as wage. employment rate race and age group, Similar to Equation 4 in some specifications we include county year fixed effects in order. to control for county specific heterogeneity such as police presence in any given year In. some specifications we include state offense fixed effects to control for differences in criminal. sanctions across different states Finally we cluster the standard errors at the county level. Note that we estimate equation 5 separately for males and females. There are two potential problems when using the contemporaneous values ygit and pgit i. reverse causality due to the potential simultaneity between the incentives and the decision to. commit a crime for example crime congestion might lower the likelihood of apprehension. and ii due to the yearly aggregation criminals expectations might be based on future crimes. introducing additional measurement error, Little is known about how criminals form their economic expectations about the crimi. nal incentives but a few articles have developed theories that we use to set up our empirical. methodology As mentioned in Lochner 2007 Theories developed in Sah 1991 and Lochner. 2004 stress that the probability of arrest is learned from others or through one s own experi. ences Beliefs about the probability of arrest are likely to depend on an individuals own past. criminal behavior and arrest outcomes the criminal and arrest outcomes of others around him. and more general signals that may come from local arrest rates or neighborhood conditions. In other words current expectations are a function of past realizations We assume that. past realizations influence current realizations only through the expectations of criminals once. we account for all state offense specific and county year specific influences Individuals who. have been operating in the criminal markets in the past or who know someone who has been. operating in the past use past experience to form expectations about the future Moreover. if expectations are rational then average realizations at time t should be equal to the ex ante. expectations t 1 and these should be a function of all the information available to criminals. up until time t 1 including previous realizations, In modeling the expectations we use a two step procedure where in the first stage we. obtain pdgit and y, d git and in the second step these measures are used in Equation 5 to determine. whether males and females have different responses to the predicted values of incentives These. equations are an instrumental variable 2SLS regression where ygi t 1 and pgi t 1 are used as. instruments for ygit and pgit 8, Given that the first stage equation is a dynamic specification it implies that the coefficient g will suffer. from a downward bias Nickell 1981 This implies that the estimated cg will be smaller than the true g. if g 0 Therefore the first stage F statistic will be lower than the true one and the instrument will be. stronger than we can observe In addition the identified effects in the second stage will be smaller than what. we observe This bias is alleviated by the length of the panel which is 20 years in our case Nonetheless we. Measurement error Illegal earnings and probability of arrest are likely to be impacted. differently by a measurement error The probability of arrest suffers less from measurement. error because we have precise information on those who get arrested For illegal earnings. measurement error after averaging might come from victim reports or law enforcement policies. thus biasing our results towards zero This would be an issue in our IV specifications if such. measurement error was correlated across years Given the small likelihood of victimization. it is unlikely for a victim to report crimes in consecutive years and so the serial correlation. would have to come from group behaviour i e biased reporting by gender race typology of. crime etc Adding the corresponding fixed effects can flexibly account for these sources of. measurement error As for police policies in our empirical specification we include interacted. year county fixed effects which can account for year to year changes in recording crimes by the. police Moreover measurement error might be an issue if i there is a systematic gender bias. in collecting data on illegal earnings we believe this to be implausible especially for property. crimes where the victims have an incentive to report the crime and the value of the stolen. items and ii males and females select into different crimes that have different measurement. error While this could easily be going on since we have a log specification and we control. for the typology of crimes committed any systematic bias or differential measurement error. in reporting would be absorbed by the fixed effects. 3 2 3 Blinder Oaxaca decomposition, In order to determine the importance of the criminal incentives like earnings y and proba. bility of arrest p on the gender crime gap we use a partial Blinder Oaxaca decomposition. limited to y and p We are interested in how much of the difference between male crime. participation measured by cmit and female crime participation cf it can be explained by. incentives endowments in the jargon of the decomposition and responsiveness the betas. To do this we define two counterfactuals using the estimates obtained from Equation 5. In the first counterfactual we will assign to females the male incentives thereby changing. females endowment In this way we want to observe how the difference in earnings and. arrest probability between the two genders contributes to the observed participation gap. Defining cd b1 b2 0 b to be the predicted female crime rate based on Eq. f it ymit f pf it f Zf it f, 5 we derive the counterfactual in the following way. f it ymit f pmit f Zf it,bf cd b1 b2,f it ymit yf it f pmit pf it f 6. In other words substituting the female endowments yf it and pf it with the male ones. ymit and pmit generates the counterfactual female crime rate The new crime rate cm f it can be. interpreted as the crime rate in a world where females have the same earning and the same. probability of arrest as males Figure 5 shows the cumulative distribution functions of the. counterfactual crime rate the male crime rate and the female one In this figure we want to. observe how the distance between cd m, mit and cf it compares to the initial distance between cd mit. and cdf it The difference in distances would show how much of the initial gap arises due to. differences in the possible earnings and arrest probabilities for each gender. To determine how much of the observed gap is due to differences in the estimated responses. to earnings and arrest the s we define the second counterfactual. c fit cd b1 b1 b2 b2,f it m f yf it y f m f pf it pf 7. comment on its effect in results in Section 4, This counterfactual corresponds to a world where females have the same elasticities as. males have but keep their initial endowments In this expression we adjust the female coef. ficients f in the estimation of the predicted female crime rate by replacing them with their. male counterparts 9, To determine how much of the observed gap arises due to differences in the estimated. elasticities we again plot the cumulative distribution functions in Figure 5 In the second. panel of the figure we again compare the difference between the new predicted female crime. rate c fit and the male crime rate cdmit to the actual gap in crime participation between cdmit. and cdf it This comparison shows how much of the crime participation gap is due to differences. in the way males and females respond to incentives. 4 1 Differences in incentives, In Table 3 we present results for the illegal earnings gap As explained in Section 2 2 shoplift. ing is a special crime therefore we exclude it from the first four columns and include it in. the last column as a robustness check In column 1 we show females have 13 percent lower. criminal earnings In the next three columns we add progressively more fixed effects but the. gap remains stable at around 13 percent Our preferred specification is that of column 4 with. all the controls and the fixed effects By adding these fixed effects we want to account for. jurisdiction specific policing responses for the sanction for a specific crime in each state as. well as availability of criminal targets related to the business cycle In column 5 we select. a subsample of daylight crimes committed between 8 AM and 7 PM We do this robustness. check in order to alleviate concerns on reporting bias it might be the case that if the female. perpetrators were not well observed they would be reported as male During daylight he gap. is still negative and is even larger in magnitude Females earn 15 percent less than males in. daylight crimes which is similar to the previous estimates and shows that reporting biases. are unlikely to drive the gap, Finally in column 6 of Table 3 we include shoplifting as a robustness checks We find that. the gap diminishes to 7 percent meaning that the earnings gap in shoplifting is in favor of. females and is significant enough to attenuate the estimates on the earnings gap in all crimes. In Table 4 we present the results for the arrest gap In the first column the arrest gap. is negative at 5 percent meaning that females are 5 percent less likely to be arrested than. males However the gap increases to 9 percent as we control for more heterogeneity by adding. different fixed effects As before our preferred estimate is from the specification in column 4. where we observe an arrest gap of 9 percent The arrest gap increases in daylight crimes to 15. percent suggesting a reporting bias where females could have been reported as males The. increase in the relative number of new females would drive down the relative probability of. arrest and result in a smaller gap as in the previous columns conditional on there being no. other significant differences in crime participation within the 24 hours of the day We assume. that in the daylight hours most perpetrators are clearly observed. In the last column of Table 4 we show that shoplifting drives the estimates on the average. gap in arrest rates which now decreases to 0 percent This reveals that shoplifting is a crime. in which females face a higher likelihood of arrest than males. Subtracting the minimum values y f and pf normalizes the intercepts and rotates the counterfactual around. its minimum value rather than around the initial intercept where y 0 and p 0. Overall conditional on crime participation we find that females earn 7 to 13 percent less. than males and face a lower likelihood of being arrested. 4 2 Differences in Responsiveness, Table 5 shows our estimates for the first stage and reduced form using alternative specifications. with and without county fixed effects and their interaction with year fixed effects and with. and without state fixed effects and their interaction with typology of offense fixed effects. In the top panel we observe that the lag of the log probability of arrest and the lag of. log illegal earnings are good predictors of respectively the log probability of arrest and. log illegal earnings The F statistics is in all the specifications well above conventional. levels at 187 for Males in our preferred specification and 94 for Females In the bottom panel. of Table 5 it is reassuring that the coefficients have the right sign An increase in the illegal. earnings is associated with an increase in the crime rate and an increase in the probability of. arrest is associated with a decrease in the crime rates. In Table 6 we present the 2SLS estimates on the gender specific response elasticities In. each additional column we control for more sources of heterogeneity such as police presence. sentence length and business cycle effects by adding fixed effects and their interactions We. show in the odd columns with the header Male that a 100 percent increase in the expected. illegal earnings would lead to a 36 percent higher participation of males in crime Similarly. the elasticity of expected criminal earnings for females in our preferred specification in the last. column is 23 percent On average we observe that females have always a significantly lower. elasticity of illegal earnings than males As mentioned in Section 3 1 this could be driven by. females being more risk averse than males 10, With respect to the probability of arrest we find that for both males and females the. elasticity is between 14 and 22 percent In our preferred specification in columns 7 and 8. the male and female coefficients are not significantly different and are around 14 percent. Males and females respond similarly to an increase in the probability of arrest. Correcting for the Nickell bias 11 we find somewhat larger first stage coefficients and we. end up with an elasticity of earnings for males of around 0 26 and an elasticity of earnings for. females around 0 15 The elasticities of arrest decrease to 13 percent All elasticities remain. significant at conventional levels, With respect to the control variables we find that being black is associated with an. increase in the crime rate Offenders aged from 15 24 contribute more to the crime rate. than offenders aged 25 34 compared to the excluded category 35 44 The crime rate is also. associated with the average wage and employment rate in all specifications even though the. coefficients change in the different specifications that we use These results seem to be highly. dependent on the inclusion of the fixed effects In our preferred specification in column 7. and 8 an increase in the wage is associated with an increase in the crime rate whereas the. employment rate is negatively associated with the crime rate. To sum up we find that both males and females respond to incentives They are both. more responsive to changes in the illegal earnings and remain equally deterred by increases in. the probability of arrest Males tend to respond more than females. For an overview of experimental evidence that women appear to be more risk averse than men see Eckel. Grossman 2008, We do that by applying the equation 19 from Nickell 1981 which approximates the bias to the estimated. coefficient Then we take the ratio between the reduced form coefficient and the corrected first stage to arrive. at the bias corrected coefficients which are reported in Table 8. In Table 7 we show OLS estimates We find that the estimates of the elasticities are. smaller which is consistent with the simultaneity issues outlined before. Robustness Checks In Table 8 we perform robustness checks with the aim to see whether. our estimates depend on the particular specification we used First of all in the first set. of column we report our estimates from Table 6 corrected for the Nickell bias which were. commented above Second to be sure that our results are not biased by the smaller dimension. of the cohorts in the sample we cut the sample at the median population and take the cohorts. with larger populations Almost all elasticities increase in magnitude In larger cohorts an. increase in the earnings of 100 percent would lead to an increase in participation by 42 percent. for males and 35 percent for females which is larger than the average response At the same. time the response to the probability of arrest increases for males to 18 percent and slightly. decreases to 12 percent for females Overall when we use larger cohorts criminals seem to. be more responsive to incentives, In the last panel of Table 8 we add shoplifting to the sample We find larger earnings. elasticities and smaller arrest elasticities for both genders We interpret this as evidence. that shoplifting is a crime that is recorded only when arrest takes place which biases the. corresponding coefficients towards zero as already shown in Figure 4. 4 3 Blinder Oaxaca decomposition, In Figure 5 we present the cumulative density functions for the Blinder Oaxaca decomposition. exercise In Panel a we plot the result from equation 6 where we assign to females the male. endowments in terms of earnings and probability of arrest In Panel b we plot the result. from equation 7 where we assign to females the same responsiveness to criminal incentives as. males In both panels we also plot the predicted cumulative distribution functions of male. and female crime rates The distance between them exemplifies the participation gaps that. we observe in Section 2 2, In Panel a we show that if females had male endowments the participation gap would. change by only 8 percent This is shown by the vicinity between CDF of female crimes and. the counterfactual predicted from equation 6 This is driven by the fact that in Section 4 2 we. show that both gender react similarly to increases in the probability of arrest The large gap. that remains is partly due to the fact that males have a larger elasticity of illegal income than. females Here it is important to highlight that one has to take a stand on how to treat the. crime fixed effects Keeping the female fixed effects we are changing the endowments without. allowing females to switch across crimes, In Panel b we show that if females had the same elasticities as males then the gap would. be smaller by 56 percent Given that the two genders respond similarly to increases in the. probability of arrest this shift is driven by an increase in the elasticity of illegal earnings This. result implies that differences in risk aversion may go a long way in explaining differences in. criminal attitudes, Together differences in endowments and in the elasticities explain 64 percent of the gap. 5 Conclusion, We motivate our research with the fact that little is known about female criminals and about. what may be driving the gender crime gap, Over the last 50 years there has been an increase in convergence between the roles of males. and females in the legal market yet there is little evidence on the existence or development. of gaps in the criminal market, From a historical perspective in the 70s concurrently with the women emancipation move. ments there have been concerns about an increase in the female participation in crime Simon. 1976 In the last years the broad social context has been redefining sex roles Weisheit 1984. cites a dominant hypothesis that female participation in crime would increase as social sex. roles converge In line with these theories in the present study that focusses on the United. States over the period 1995 2015 we show evidence of the presence of a gender gap in property. crimes that has been shrinking over time, In our data we observe that in 1995 there was one female criminal for every 3 male ones. in 2015 this ratio increased to one female for every 2 male criminals. In order to be able to understand such convergence it is important to determine why. there is a gender gap in the first place In our paper we unpack the gender gap in criminal. participation in three steps First we quantify the observed difference in illegal earnings. and probability of arrest the so called criminal incentives in the Becker s model of crime. across gender controlling for several of characteristics and for a number of fixed effects and. their interactions The mapping of these differences in criminal incentives across gender is. our first contribution We observe that ceteris paribus women earn 13 percent less than men. when committing a crime but they face a 9 percent lower probability of arrest There are a. host of different explanations for this ability choices effort search costs as well as underlying. risk aversion, We also observe that these differences between male and female criminals shrink over time. which may contribute to a reduction in the participation gap if criminals were responding to. such incentives And this is what we try to establish in the second part of the study We. model the participation decision and provide evidence that males and females increase their. criminal engagement when the expected illegal earnings go up and when the probability of. arrest goes down We also find that females and males respond differently to changes in illegal. earnings which might signal underlying differences in preferences for example a higher risk. aversion among women, Finally in order to measure how such differences can contribute to the gender gap we. use a Blinder Oaxaca decomposition Differences in criminal incentives explain 8 percent of. the gender gap while differences in the elasticities explain a much larger part of the gap 56. percent The remaining 36 percent of the gap could be driven by differences in the disutility of. prison Mastrobuoni Rivers forthcoming differences in discount factors Mastrobuoni. Rivers 2016 differences in the opportunity cost of crime due for example to child rearing. More elaborate differences may also play a role Drawing from the economic analysis of. socially constructed identities Akerlof Kranton 2000 it is easy to make the conjecture. that stereotypes can play a role in the participation decision If crime was a masculine job. then entering females would de value the masculinity image and may thus be ostracized by. their male counterpart This would lead to females being less likely to be initiated into crime. than their male counterparts On the flip side if females think that crime is non feminine. they would be less likely to participate This role of stereotypes coincides with the findings of. Steffensmeier 1980 who notes that males are less likely to choose a female partner because. they consider females to be less trustworthy and more governed by passions Gavrilova 2017. shows that there is a bias against females in criminal group formation If females do not. get initiated into crime in the same way as males then maybe they participate less If this. would be the case then the illegal market offers a setting in which discrimination is a blessing. in disguise A blessing because policy makers and society at large likely prefer lower crime.

Intelligent Production of Quality Apparels ... the integrated management of cloth data ... The second one addresses the laying and cutting sections, ...

#25 Cast Out the Bondwoman March 2019 2 who call themselves Jews (according to the flesh) but are not true Jews (according to the spirit). When Christians believe this way, it is called Christian Zionism. To all who have been and are being persuaded along these lines, Paul would say: Cast out the bondwoman and her son, for the son of the bondwoman shall not be an heir with the son of the

Tanaman Penutup Tanah Kacang-kacangan dan Mulsa Jerami Sumarni, N., E. Sumiati, dan R. Rosliani Balai Penelitian Tanaman Sayuran, Jl. Tangkuban Parahu 517, Lembang, Bandung 40391 Naskah diterima tanggal 8 Mei 2008 dan disetujui untuk diterbitkan tanggal 31 Oktober 2008 ABSTRAK. Penelitian dilakukan di Kebun Percobaan Balai Penelitian Tanaman Sayuran Lembang, dari bulan Juli-Oktober 2004 ...

Penelitian ini bertujuan untuk mengetahui konsentrasi toleran larutan atonik dan PEG 6000 yang resisten terhadap cekaman kekeringan serta mengetahui interaksi antara larutan atonik dan PEG 6000 pada planlet kacang hijau secara in vitro. Penelitian ini dilaksanakan pada bulan November 2016 sampai Januari 2017 di Laboratorium Botani (ruang In Vitro), Jurusan Biologi, Fakultas Matematika dan Ilmu ...

Operation Manual Cutting Machine Product Code: 891-Z01. ... and then study this safety manual for the correct operation of the various ... cloth. Using the machine ...

3 Acknowledgements Current Refs: Royce Porter (Organiser & System creator), Matthew Moore (Ref, Web-Monkey & Cardiff Branch Big Cheese), Bethan Charlton (MK Ref and Guildsman Editor), Daniel Hooper (MK Ref) Training Refs: Adventure Writers (but NOT Refs): None at this time; contact the Ref team if you are interested in writing an adventure. Past Refs (with Great Thanks for their time and .

But what this diagram does not show, is that a couple of the drive belts are left with excessive slack. This causes a rapid series of jerks in the drive between the ...

LOVED ONE EXPERIENCING DOMESTIC VIOLENCE . 2 WWW.SAFEANDTOGETHERINSTITUTE.COM PO BOX 745 CANTON, CONNECTICUT 06019 PHONE: 1 (860) 319-0966 WWW.SAFEANDTOGETHERINSTITUTE.COM This book is not meant to replace official service, orders of protection or intervention orders, or calling the police in emergency situations. Nor does it constitute legal or mental health advice. It is also not intended to ...

form must state the functional and if applicable, non-functional requirements User Acceptance Test Plan Customer Required Required Training Plan Customer/QA Analyst Required Optional Training Plan maybe required for Change Request based on the scope of change requested (large) Project Sign-Off Project Manager Required Required