Essay:

Essay details:

  • Subject area(s): Business
  • Price: Free download
  • Published on: 21st September 2019
  • File format: Text
  • Number of pages: 2

Text preview of this essay:

This page is a preview - download the full version of this essay above.

Introduction

“Sports offer an important opportunity for building life skills of at-risk youth

that allow them to better cope with daily life challenges and move away from involvement in violence, crime or drug use.”

(United Nations Office on Drugs and Crime, 2018)

More than eight million American high school students participate in some form of school-sponsored athletic program (National Federation of State High School Association, 2017) and an additional 500,000 students participate in college athletics (National College Athletic Association, 2017). Sports participation has been linked to substantial private human capital benefits, including increased educational attainment (Darling et al. 2005; Rees and Sabia, 2010), improved health (Kaestner & Xu, 2010; Sabo et al., 2005), and better labor market outcomes (Perisco et al., 2004; Stevenson, 2010) .

In addition to these private benefits, a number of policymakers have argued that there are important external social benefits of sports participation that are generated by the crime-reducing effects of sports. For instance, several branches of the United Nations highlight sports as an important tool in peacekeeping, in fighting the war on drugs, and in violent crime prevention. In the United Nations Declaration of the 2030 Agenda for Sustainable Development, the role of sports in facilitating social progress is highlighted:

"Sport is also an important enabler of sustainable development. We recognize the growing contribution of sport to the realization of development and peace in its promotion of tolerance and respect and the contributions it makes to the empowerment of women and of young people, individuals and communities as well as to health, education and social inclusion objectives."

(The UN 2030 Agenda for Sustainable Development, 2018)

In the United States, the Violent Crime Control and Law Enforcement Act of 1994, the crime bill for which Secretary Hillary Clinton came under fire during the 2016 presidential campaign, allocated $50 million in federal funding for “midnight basketball” programs that were designed to provide alternatives to criminal behavior for youths as well as provide legitimate networking channels to enhance youths’ social capital accumulation.   Despite policymakers’ claims that public initiatives to encourage sports participation is an effective anti-crime tool, there is relatively little causal evidence on this subject.

A number of studies have empirically examined the relationship between sports participation and crime, but nearly all have treated participation as exogenous to crime (Nichols, 2007; Crabbe, 2000; Smith & Waddington, 2004; Hartmann & Massoglia, 2007; McGinn et al., 2008; Caruso, 2011; Ekholm;2013; Spruit et al., 2016). If those who engage in sporting activities are positively selected on difficult-to-measure characteristics (such as ability, discipline, and social consciousness), then estimates of the beneficial effect of sports on crime will be biased upward.  On the other hand, if sports programs attract those most at risk for delinquent behaviors, then estimates of the crime-reducing effects of sports participation will be biased toward zero.

This study improves upon the prior literature by exploiting a natural experiment to estimate the causal effect of sports participation on crime.  We exploit the enactment of Title IX, which mandated that public schools increase their female sports participation proportion to male sports participation levels, to generate plausibly exogenous variation in female sports participation to identify its’ effect on criminal activity.  Using county-level crime data between 1980 and 2000 from the Uniform Crime Reports (UCR), reduced form and instrumental variables (IV) estimates show that Title IX is associated with a decrease in crime among females. Specifically, we find that a 10-percentage point increase in female sports participation caused by Title IX is associated with a 2 percent decrease in overall arrests and a 3.6 percent decrease in violent crime arrests. These reductions are driven by burglary, larceny, robbery, motor vehicle theft, and aggravated assault. Our findings are robust to controls for state-level economic conditions, county-specific time trends, and male crime rates, suggesting that the effects we estimate are not driven by spurious corrections. Our estimates suggest that the approximately 25-percentage point increase in sports participation induced by Title IX generated $600 million in benefits of crime avoidance.

Background

2.1. Title IX Legislation

Signed by President Richard Nixon in 1972, Title IX was added to the Education Amendments in that year. This legislation required educational institutions to meet requirements that banned gender discrimination within six (6) years. Specifically, Title IX mandates that:

‘‘No person in the United States shall, on the basis of sex, be excluded from participation in, be denied the benefits of, or be subjected to discrimination under any educational program or activity receiving financial assistance.’’

(20 U.S. Code § 1681 – Sex)

Title IX applies to institutions that receive federal financial assistance from the Department of Education. These institutions include approximately 16,500 local school districts, 7,000 postsecondary institutions, as well as charter schools, for-profit schools, and education agencies of the 50 states and the District of Columbia (U.S. Department of Education Office for Civil Rights, 2015).

The U.S. Department of Education Office for Civil Rights enforces the legislation by using the three-part test developed by the Department of Education to examine whether an institution is meeting its Title IX obligations. An institution is in compliance with the legislation if it meets any one of the following parts of the test:

“(1) The number of male and female athletes is substantially proportionate to their respective enrollments; or

(2) The institution has a history and continuing practice of expanding participation opportunities responsive to the developing interests and abilities of the underrepresented sex; or

(3) The institution is fully and effectively accommodating the interests and abilities of the underrepresented sex. ports.”

(U.S. Department of Education, 1979)

While Title IX did not apply only to sports funding, athletics was a funding area with large disparities and, in some cases, official gender segregation. Thus, one of Title IX’s largest impacts was on female sports participation.   Between 1971 and 1978, the number of females participating in high school sports increased by more than 600 percent (Gavora, 2002).

It is important to note that male sports participation did not fell after the legislation. Stevenson (2007) documented some compositional changes among sports, such as soccer, expanded, while others, like wrestling, became less prevalent; however, overall male sports participation did not decrease after the policy was implemented. Title IX might have influenced compositional shifts in male sports, but these shifts were small, and not systematic.

2.2. Impacts of Title IX

A number of studies have examined the impacts of Title IX on female sports participation, human capital acquisition, and labor market outcomes (Stevenson, 2010; Kaestner & Xu, 2010; Clark & Ayres, 2014; Schulkind, 2016). With regard to 'first stage' effects on female sports, Stevenson (2010) exploits heterogeneity across states in the bindingness of Title IX to identify its effects on female high school sports participation.  Specifically, she argued that states with higher male sports participation rates prior to the adoption of Title IX saw larger increases in female sports participation. Stevenson’s (2010) results show that a 10-percentage point increase in pre-Title IX male sports participation is associated with a 3 to 4-percentage point increase in female sports participation. This finding suggests that Title IX had significant impacts on high school sports participation among women.

Using this approach to identify exogenous variation in high school sports participation, Stevenson (2010) used an instrumental variables (IV) approach to estimate the impact of sports on the educational attainment and employment of women ages 25-to-35. Using data from the U.S. Census from 1980 and 2000, Stevenson (2010) found that a 10-percentage point increase in female sports participation, induced by Title IX, generated an increase in the average educational attainment of 0.39 years, and a 1.3-percentage point increase in the probability that females would attain at least some postsecondary education. In addition, Stevenson (2010) explored whether Title IX increased the subsequent employment of women; she found a 1.9-percentage point increase in the probability of being employed.  

There is also evidence that Title IX increased health capital accumulation among women.  Kaestner and Xu (2010) revealed that Title IX-induced increases in female sports participation were associated with diminished body mass index (BMI), reduced rates of obesity, and increased physical activity in adult women. They found that a 20-percentage point increase in female sports participation between 1971 and 1978 (which they attributed to Title IX) led to a 0.586 unit decrease in BMI and a 2.8-percentage point decrease in the probability of their being obese.

Furthermore, Schulkind (2016) linked Title IX-induced female sports participation increase to the intergenerational transmission of health. Using a similar identification strategy as Stevenson (2010), she found that a 10-percentage point increase in female sports participation translated to 1.5 fewer low birthweight births per 1,000 births, 0.34 fewer very low birthweight births per 1,000, and a 0.33 standard deviation improvement in Apgar scores, an indication of the health of a neonate.

Finally, Clarke and Ayres (2014) examined the effect of female sports participation on adult women’s social lives. They found that a 10-percentage point increase in state-level female sports participation generated a 5 to 6 percentage point increase in the rate of female secularism  and a 5-percentage point increase in fertility rates. Clarke and Ayres (2014) also found that Title IX might have affected marriage rates among new mothers by showing that Title IX resulted in a 6-percentage point increase in the proportion of mothers who were unmarried.  

2.3. Mechanisms to Explain the Link Between Sports and Crime

There are a number of channels through which sports participation, particularly in high school, can affect crime. Time availability, human capital acquisition (including social capital acquisition), and labor market outcomes (employment and earnings) are three primary channels through which sports might influence criminal activity. For a majority of crimes, we would generally expect sports participation, via the aforementioned channels, to reduce crime. Each of these mechanisms is discussed in more detail below.

Time spent in sports participation could be important, in terms of limiting the time available for participating in criminal activity. More time spent in school, and in extracurricular activities, is associated with reductions in criminal activity (Barron et al., 2000; Jacobs & Lefgren, 2003; Lochner & Moretti, 2004; Machin et al., 2011; Beatton, 2016; Bell, 2016) . For instance, Anderson (2014), using Uniform Crime Reports (UCR), showed that the minimum dropout age requirements have a significant and negative effect on juvenile property and violent crime arrest rates. Similarly, Jacobs and Lefgren (2003) find that property crime increases significantly in areas where teens have days off school. Each of these studies found short-term criminal impacts, but it is possible that short-term effects on youth criminal behavior may have long-term effects on future criminal activity if these incapacitation effects prevent youths from heading down a path to a criminal career.

Second, sports participation could impact human capital acquisition, which could affect future criminal activity. Human capital acquisition may take the form of increased educational attainment, social networking skills, discipline, and better health.  Of course, it is also possible that social networks formed while participating in organized sports could generate negative peer effects that increase the probability of crime.  There is, in fact, strong evidence that educational attainment is negatively related to future crime (Lochner and Moretti 2004; Buonanno & Leonida, 2007; Machin et al. 2011; Hjalmarsson, 2015; Lochner & Urbina, 2017).  Lochner and Urbina (2017) used changes in compulsory schooling laws as an instrument for educational attainment and, using data from the 1960 to 1990 Uniform Crime Reports, found that increases in education reduced female arrest rates.

Several studies have explored the impact of sports on academic achievement, educational attainment, and life skills (Eide & Ronan, 2001; Perisco et al., 2004; Pfeifer & Cornelissen, 2010; Rees & Sabia, 2010) . For instance, Eide and Ronan (2001) examined the link between sports participation and educational attainment, accounting for unobservables through the estimation of instrumental variables. They found that the effects of sports participation varied by race and gender. Specifically, they documented that sports participation had a positive effect on the educational attainments of black males and white females. They also found a negative effect on the educational attainment of white males. Similarly, Rees and Sabia (2010) used the height of a student as an instrument, and found that sports participation may increase aspirations to attend college; however, they found no effects on academic achievement or attending college. And, as discussed above, Stevenson (2010) found strong evidence that Title IX-induced increases in female sports participation is positively related to educational attainment.

Finally, sports participation may also influence criminal activity through its effects on employment and wages. Many studies document a positive relationship between high school sports participation and better employment opportunities later in life. (Barron et al., 2000; Eccles et al., 2003; Kuhn & Weinberger 2005; Ewing, 2007; Lechner, 2009; Stevenson, 2010). Lechner (2009) found that sports participation has a positive long-term impact on earnings and wages. Kuhn and Weinberger (2005) also documented that higher wages are attributable to the leadership skills developed through athletic participation.

In summary, if sports participation increases human capital acquisition and improves labor market outcomes, sports participation may reduce crime.

2.4. Sports Participation and Crime

Previous studies that have examined the relationship between sports and crime have been largely cross-sectional in nature, and have treated sports participation as exogenous to crime (Cameron & MacDougall, 2000; Crabbe, 2000; Smith & Waddington, 2004; Hartmann & Massoglia, 2007; McGinn et al., 2008; Nichols, 2010; Ekholm, 2013; Spruit, 2015). Findings in the literature have been mixed, owed to the identification strategy used (cross-sectional versus panel fixed effects models), as well as the set of controls available to researchers. A meta-analysis by Spruit et al. (2015) analyzed 51 published and unpublished studies, with 48 independent samples, containing 132,366 observations on sports participation and delinquent behavior. Overall, they were unable to find any significant association between sports participation and delinquency. Their findings suggest that adolescent athletes are neither more nor less delinquent than non-athletes.  However, this meta-analysis did not differentiate among identification strategies nor did it explore long-run versus short-run impacts on crime.

Hartmann and Massoglia (2007) found that the effect of sports participation on deviant behavior depended on the type of behavior. They find that shoplifting decreases with sports participation, while drunk-driving increases, suggesting both positive and negative social consequences. Moreover, they highlighted that these effects extended later in life. Using a fixed effects model, Caruso (2011) also documented a robust negative association between sports participation and adult property and juvenile crimes in Italy. A 10 percent increase in sports participation is associated with a reduction in property crimes by approximately 3 percent and a reduction in juvenile crimes by approximately 8 percent.

Furthermore, a number of studies have examined the effects of sports-based program interventions on crime (Winther and Currie, 1987; Hawkins, 1998; Hartmann & Depro, 2006; Carmichael, 2008; Kelly, 2013; Parker et. al., 2017). For instance, Winther and Currie (1987) argued that the sports program (Northern Fly-In Sports Camps) that focused on building teamwork, character and self-esteem among Aboriginal Canadian young offenders and potential young offenders in Manitoba resulted in a reduction in youth crime in the communities that offered the program. Similarly, Hartmann and Depro (2006) examined the effect of midnight basketball programs and found that cities adopted official midnight basketball leagues experienced sharper decreases in property crime rates.

However, these programs were not randomly assigned to youths and studies have taken their placement in communities as exogenous to criminal behavior (as well as treated take-up as exogenous to crime).

Data and Methods

3.1. Data

We compile county-level, age-specific arrest counts for females from the Federal Bureau of Investigation’s (FBI) Uniform Crime Reports (UCR) from 1980 to 2000.  County crime rates are measured as female arrest counts per 1,000 of the age-specific female population, where county population data is generated using the National Cancer Institute’s Survey of Epidemiology and End Results (SEER).

Arrest data is collected via voluntary reports from more than 16,000 city, county, and state agencies. This data may understate actual levels of crime because not every crime is reported to law enforcement (Gould et al. 2002). However, there is a high level of correlation between arrest reports from UCR and actual crimes committed when the latter is measurable (Lochner and Moretti 2004).  Moreover, measurement error in UCR is unlikely to be correlated with our measure of Title IX (See below).  

To further guard against possible idiosyncratic measurement errors in crime, we follow the method described in Anderson (2014) in which counties that have ten or fewer years of arrest data are removed from our sample. Additionally, age-specific county-year arrest counts greater or less than two standard deviations from the county-specific mean are excluded. Finally, we control for the number of agencies reporting within a county for any given year to address county-specific time-varying changes in reporting behavior.

Our primary analysis measures violent crimes (e.g., murder, rape, robbery, aggravated assault, and simple assault), property crimes (e.g., larceny, burglary, motor vehicle theft, and arson), drug crimes (including sale and possession), and prostitution. In our analysis, we focus on females ages 25 to 39 years old between the years 1980 and 2000.  This roughly maps to the ages and years window explored by Stevenson (2010).

In Table 1, we show the weighted means of female arrests per 1,000 population. As expected, rates of property crime arrests were higher than violent crime and drug-related arrests. Property crime arrests were 4.56, violent crime arrests were 2.9 and drug-related arrests were 2.49 per-thousand . Figure 1 also reveals crime rates per 1,000 for females ages 25-to-39 between 1980 and 2014.  We document an increase in female arrest rates between 1980’s and 1990’s. Additionally, we observe a decline in female arrest rates beginning from late 1990’s.

Next, we obtain data on gender-specific state sports participation from the National Federation of State High School Associations (NFSH). Each state and the District of Columbia have their public high school associations and every year they collect information on sports participation. This state-level information is gathered and compiled by the NFSH. Annual sports participation survey data has been compiled since the 1969-1970 academic year .

Following Stevenson (2010), and Schulkind (2016), we measure the gender-specific sports participation rate as the ratio of the number of high school students participating in a high school sport to the gender-specific state population ages 14-to-17.  In Figure 2, we present the gender-specific high school sports participation rates from 1970 to 2014.  Consistent with prior studies, we observe a sharp increase in female sports participation rate between 1971 and 1978, after which female participation rate approaches the male sports participation rate. This is consistent with the six-year compliance window of Title IX.

3.2. Empirical Strategy

Our empirical approach follows that of Stevenson (2010). First, we draw county-by-month data from the UCR and estimate the following reduced form equation (1):

〖Arrests〗_icst=  〖 β〗_0+ 〖 β〗_1 (〖Post Title IX 〗_t× 〖Boys^' Participation〗_s^1971 )+κ_c+ψ_t+τ_i+δ^' X_st+ε_icst  

(1)

where i, c, s, and t, indexes birth cohort, county, state, and year respectively. The dependent variable Arrests is the number of female arrests per thousand population of 25-to-39-year-olds in birth cohort i in a county c in state s at year t. The interaction term (Post Title IX * Boys’ Participation1971) represents boys’ sports participation in the academic year 1971 interacted with dichotomous variable represents birth year after 1957. This variable captures the individuals that enrolled the high school after Title IX was enacted. κ_c,ψ_t and τ_i represents county, year and birth cohort effects respectively. In addition, the vector Xcst includes controls for state-level unemployment, personal income, inflation, and interest rates at the age of 18 and at the year of the arrest. Data for state and year specific controls are obtained from Federal Reserve Economic Data (FRED), The Local Area Unemployment Statistics (LAUS), Bureau of Labor Statistics (BLS) and Bureau of Economic Analysis (BEA).

Following Stevenson (2010), our coefficient of interest,1, is identified from cross-state variation in the pre-Title IX (1971) male sports participation rate interacted with whether the female was at least 18 years old in 1975 (and thus fully affected by Title IX).  Note that because crime counts are reported for five-year age intervals, we include only fully affected age groups. As a sensitivity check, we also code the variations of the policy variable to capture those “partly covered” by Title IX when they are in high school. Results using this alternate coding are quantitatively similar and available in the Appendix Table 2.

In Figure 2, we show the cross-state distribution of male sports participation rates in 1971 and document large variation in male sports participation among states in the pre-Title IX period. For instance, Rhode Island, North Carolina and District of Columbia had the lowest male sports participation among states. Their participation rates were roughly 20 percent. Whereas, North Dakota, Nebraska and Minnesota had the highest sports participation rates. Almost every male high school student participated in athletics .

 A key threat to identification would be if crime trends among younger cohorts of 25-to-39-year-old females trended differently in states that were more (or less) bound by Title IX.  To address confounding of this form, we take a number of tacks. First, we follow the approach of Kaestner and Xu (2010) and add a control for the crime rate for men ages 25-to-39:

〖Arrests〗_icst=  〖 β〗_0+ 〖 β〗_1 (〖Post Title IX 〗_t× 〖Boys^' Participation〗_s^1971 )   + β_2 〖Male Arrests〗_icst+κ_c+ψ_t+τ_i  +δ^' X_it+ε_icst

(2)

where the 〖Male Crime Rate〗_icst added to the right-hand side of the estimating equation will control for any unmeasured crime trends that are associated with Title IX’s implementation across states that are unrelated to Title IX’s impact on women.  Second, we add controls for county-specific linear time trends to the right-hand side of equation (3):

〖Arrests〗_icst=〖 β〗_0+〖 β〗_1 (〖Post Title IX 〗_t× 〖Boys^' Participation〗_s^1971 )+κ_c+ ψ_t+τ_i+v_c*t +δ^' X_it+ε_icst

(3)

where v_c*t denotes county-specific linear time trends. This approach should control for anything trending linearly at the county-level that could have been correlated with crime rates of the affected cohort.  Moreover, we also experiment with controlling for birth cohort-specific linear time trends. Results for those estimates are quantitatively similar and available upon request.

In addition to the reduced form specification outlined above, we also use a two-stage least squares (2SLS) approach to estimate the impact of Title IX-induced changes in female sports participation on female crime.  In the first stage, we estimate equation (4):

〖Female Participation〗_icst=  〖 ∝〗_0+ 〖 ∝〗_1 (〖Post Title IX 〗_t× 〖Boys^' Participation〗_s^1971 )+κ_c+ ψ_t+τ_i+v_c*t +∂^' X_st+ε_icst

(4)

The instrument for female sports participation is, as above, the interaction of the state level male sports participation in 1971 and an indicator variable equal to one for the post Title IX cohort. Thus, the instrument is set equal to zero for females’ that attended high school in a pre-Title IX world and it is equal to the male sports participation in 1971. The instrument is expected to be correlated with female sports participation due to the nature of the policy, but uncorrelated with unmeasured determinants of female arrests. The second stage equation is then estimated as equation (5):

〖Arrests〗_icst=  〖 φ〗_0+ 〖 φ〗_1 (Female Participation) ̂_ist+ κ_c+   ψ_t+τ_i+v_c*t +θ^' X_st+ε_icst

(5)

Our 2SLS estimate (〖 φ〗_1 ) can then be interpreted as the impact of Title IX-induced increases in female sports participation on female crime rates.

Results

For ease of presentation, tables focus on reduced form estimates of β1 as well as instrumental variables estimates〖 φ〗_1. Estimated coefficients of the control variables are available in Appendix Table 1. Similar to Stevenson (2010), standard errors are clustered at the state-year level. Estimates are weighted based on county-level age- and gender-specific populations obtained from the SEER. Unweighted estimates are qualitatively similar to those presented and available upon request.

Reduced Form Estimates

Table 2 presents reduced form estimates of the relationship between Title IX and female arrests. It shows the results for total, violent, property and drug-related arrests for the full sample of 25- to 39-year-old females between 1980 and 2000. The first three columns report estimates with county, year, birth cohort fixed effects, and economic controls. Similar to Stevenson (2010), in the second column, we include macro-economic controls for the age of 18. Furthermore, we include macro-economic controls at the year of the arrest. Columns (4) and (5) present estimations with male crime rates and county specific linear time trends as additional controls to further account for any trends in crime that should be unaffected by Title IX.

Across specifications, there is strong evidence that Title IX is associated with a reduction in criminal arrest rates. Without any economic controls on the right-hand side of estimating equation, there is evidence of a negative relationship between male sports participation arrests in the pre-Title IX period and criminal arrests. Given that our right-hand side variable ranges from 0 male sports participation to approximately 1, the marginal effects reported in Table 2 represent the effects of moving from zero female sports participation to full (100 percent) participation.  Given our descriptive findings in Figures 1 and 2, we discuss the magnitudes of the effects as the more policy relevant impacts of a 10 percentage-point higher male sports participation rate.  The results in column (1) show that a 10 percentage-point higher male sports participation rate (in 1971) is associated with a .39 per 1,000 reduction in the total crime rate, a 0.082 per 1,000 reduction in the violent crime rate, a 0.07 per 1,000 reduction in the property crime rate, and a .23 per 1,000 reduction in the drug crime rate. The magnitudes of these effects suggest that each 10 percentage higher male sports participation rate is associated with a 2 to 5 percent reduction in female crime rates.  These findings are robust to the inclusion of macroeconomic controls in columns (2) and (3).  

 As noted above, controlling for male crime rates should capture any unmeasured crime trends that are associated with Title IX’s implementation across states that are unrelated to Title IX’s impact on women. However, given that it is theoretically possible for male crime rates to be affected by Title IX (for instance, through marriage markets and labor market spillovers), these findings should be interpreted as lower bound estimates.  Following Kaestner and Xu (2011), when we add “male crime rates” on the right-hand side of the estimating equation, estimates of 〖 β〗_1 become smaller but the findings are generally still statistically distinguishable from zero at conventional levels.  We find that a 10 percent higher male sports participation rate is associated with a 2 percent decline in total crime, a (statistically insignificant) 1 percent reduction in property crime, a 3 percent reduction in violent crime, and a 4 percent decrease in drug-related crimes.

In column (5), we control for county-level linear time trends which take into account anything trending linearly at the county-level that could correlate with crime rates. The results continue to show that Title IX is associated with reductions in female crime. This suggests potentially important external social benefits of Title IX to those who were not directly affected by the policy change.

In an effort to better understand female crime behavior, we break down the arrests into: larceny, burglary, motor vehicle theft, robbery, arson, aggravated assaults, other assaults, and prostitution. In Table 3, we present reduced form estimates of the relationship between Title IX and specific crime arrests for the full sample of 25- to 39-year-old females between 1980 and 2000. In the first column, we add macro-economic controls for the age of 18 and at the year of the arrest. Controlling for those had a little impact on our estimates. A 10 percent higher male sports participation is associated with a 15 percent decrease in burglary, a 0.5 percent decrease in motor vehicle theft arrests, and a 1.8 percent decrease in other assaults.

In the second column of Table 3, we include county-level linear time trends to the right-hand side of the estimating equation. Controlling for linear time trends has a little impact on our estimates. A 10 percent higher male sports participation is still associated with a 9 percent decrease in aggravated assault, a 12 percent decrease in burglary, a 0.5 percent decrease in larceny, a 0.5 percent decrease motor vehicle theft, and a 1.1 percent decrease in other assaults arrests.

Instrumental Variable Estimates

Estimates of the relationship between state-level female sports participation rates and states’ Title IX compliance are presented in Table 4. Consistent with the evidence from Stevenson (2010), our first stage estimates (Equation 3) find that Title IX compliance is associated with a statistically significant 26 - 32 percentage point increase in the state-level female sports participation rates. The F-statistics in the first stage easily satisfy the instrument relevance standards proposed by Stagier and Stock (1997) with values ranging from 22.46 to 31.13.

Table 5 presents 2SLS estimates of the relationship between predicted female sports participation and the female crime arrests. The 2SLS estimates on the relationship between female sports participation and total crimes are substantially more negative than the reduced form estimates using Title IX compliance, implying that the local average treatment effect captured by this estimation is higher than the average effect. This consistent with the hypothesis that Title IX induced sports participation reduces the criminal activity. In other words, exogenous increase in sports participation reduces crime more.  

In the first column of Table 5, we include the aforementioned fixed-effects and macro-economic controls. A 10-percentage point increase in female sports participation is associated with a 5.6 percent decrease in total, an 8 percent decrease in violent crime, and a 9.4 percent decrease in drug-related arrests.

When we include county-specific linear time trends on the right-hand side of the estimating equation, a 10-percentage point increase in female sports participation is associated with a 2 percent decrease in total, a 3.6 percent decrease in violent crime and a 4.5 percent decrease in drug-related arrests. Furthermore, we restrict the arrests to property crimes, a 10-percentage point increase female sports participation is associated with a 0.8 percent decrease in property crime arrests. However, this association is not statistically significant.

In Table 6, we present 2SLS estimation results for specific crime arrests. The first column includes 2SLS in the absence of geographic-specific time trends. We find significant negative associations between female sports participation and burglary, motor vehicle theft, and assaults. A Title IX-induced 10-percentage point increase in female sports participation is associated with a 5.7 percent decrease in assaults, a 49 percent decrease in burglaries, and a 1.3 percent decrease in motor vehicle theft arrests.  These findings are somewhat larger than, but consistent with OLS estimates described above.

Finally, in the second column of Table 6, we include county-specific linear time trends on the right-hand side of the estimating equation. Adding linear time trends has a little impact on our estimates. A 10-percentage point increase in female sports participation is associated with a 4.4 percent decrease in assaults, a 46 percent decrease in burglaries and, a 1.7 percent decrease in motor vehicle thefts.

We will undertake a number of revisions in the next iteration of this paper.  First, we will explore the effects of Title IX-induced increases in female sports participation on a more detailed set of specific crimes, including additional minor crimes, to get a sense if there is heterogeneity in impacts of Title IX.  In addition, we will empirically explore several mechanisms that could be at work, specifically, improvements in female employment, educational attainment and earnings.  And finally, we will explore whether increases in participation in particular sports is driving the reduction in crimes we observe.

Conclusion

Recent research suggests that, Title IX has substantially increased high school female sports participation and may lead to increased educational attainment and improved labor market outcomes.  However, little work has explored whether there are positive spillover effects to society that might justify, on welfare grounds, policies, and programs to increase sports participation.  In making their case for increased public funding for sports, many prominent national and international organizations, including the United Nations Office on Drugs and Crime, have made the case that increasing sports participation may reduce crime.  While several cross-sectional studies have posited a link between sports participation and criminal behavior, most of these studies have failed to adequately account for selection on unobservables or reverse causality.  

This study uses the introduction of Title IX to estimate the causal impact of female sports participation on crime.  Our results document that Title IX compliance induce 25 to 35 percentage point increase in female sports participation. Turning to crime, using data from Uniform Crime Reports and National Federation of State High School Association, 2SLS estimates provide some evidence that increases in sports participation induced by Title IX lead to lower arrests in several types of crimes. A 25-percentage-point increase in female sports participation is associated with a 5 percent reductions in total arrests, a 11.25 percent decrease in drug-related arrests, and a 9 percent reductions in violent crime arrests among 25-to-39-year-old females. In addition, our crime-specific estimates provide evidence of heterogeneous effects. Burglary, robbery, larceny, aggravated assault, and drug crimes experience larger statistically significant effects.

The magnitudes of our estimates suggest that a decrease in crime associated with Title IX induced sports participation may help avert some social costs. It has been estimated that the per-offense social cost of crime is approximately $20,500 in 2017 dollars (Miller, Cohen, and Wiersema, 1996). Therefore, the estimates we obtain suggest back-of-the-envelope averted crime costs were $600 million in 2017 US dollars. The findings from this paper provide some evidence that sports participation not only has positive private benefits but also has external social benefits.

...(download the rest of the essay above)

About this essay:

This essay was submitted to us by a student in order to help you with your studies.

If you use part of this page in your own work, you need to provide a citation, as follows:

Essay Sauce, . Available from:< https://www.essaysauce.com/essays/business/2018-3-1-1519912358.php > [Accessed 02.03.21].