Education and the Crime Rate
How Education Lowers Crime
Washington State University
Economics 490, SES Capstone
Abstract
The main objective of this paper is to analyze the impact of education on the crime rate. The majority of people would assume that increased levels of education would lower the crime rate in a given area. This paper helps to reinforce this idea based on the conclusions reached from the research. There are also a few unexpected results such as higher levels of income per capita having higher rates of violent crime. In the end though, the data and research cement the idea that education is a good crime prevention tool.
Introduction
Crime rates are a concern for every major city in the United States, as well as the world. Why do some cities have higher crime than others? What needs to be done to lower crime rates? These are a few of the questions asked on a daily basis. A very effective tool to fight crime may very well be education. Most Americans would probably prefer to have their tax dollars go towards improving education instead of having to fund a larger police force. Raising education levels is more of a proactive approach whereas adding to the level of law enforcement is retroactive. Prevention is always better than having to fix a problem after it has occurred.
The goal of this paper is to compare the relationship between education and crime rates. I will do this by using data on high school enrollment and crime rates per year. There are a lot of different educational data available as well as different types of crimes, so maybe I will find that areas with high levels of education have a higher murder rate than a different type of crime. Interestingly, Lance Lochner (2000) found that education reduces all types of crime except for rape. My research will focus on increased levels of education and whether or not the crime rate in the United States decreased in the same period. Many people would assume that the two measures are inversely related and I intend to confirm these assumptions. Education level and income level are usually closely related, but will this be true for their impact on crime rate? The outcome of this relationship may be a little surprising.
Dr. Nelson Ejiro Akpotu (2004) conducted a study of inmates in Nigeria and obtained their educational history. Only 16% of the inmates had completed or even attempted what would equate to high school here in the United States. This is an example of how the model can be applied globally and not just in our country. This is also a reason that Daniel Karpowitz (2000) argues for reinstating Pell Grant eligibility for the incarcerated. He believes that there will be a lot less repeat offenders if the incarcerated are given a chance to obtain higher levels of education.
Background
According to current and recent literature, education is a very effective tool in lowering the crime rate. Lance Lochner (2000, 2004) has written a number of papers on the topic of education and crime. Mr. Lochner’s writings all seem to be trying to persuade a larger investment into education by presenting the potential social savings from having less crime. Together with Enrico Moretti (2000), they conclude that each additional year of school lowers the subsequent probability of incarceration for white men by 0.1 percentage point, Lochner (2000). Lochner even came up with a way to measure the social savings monetarily; his model shows that each unit increase in high school graduation levels would save the country $1.4 billion. Those are the kind of numbers that tax payers like to hear.
As seen in Table 1 on page 11, both violent crime and property crime have steadily been decreasing in every year since 1991. Why did these crime rates decline throughout the decade? The answer is simple, because Bill Clinton was president… That might not be the actual reason, but many argue that when Clinton came into office there was more of an effort by the government to improvement the education system. And actually, in the late 1980s President Bush started to press for better schooling in the nation. During this time stricter regulations and testing were required for public schools. Clinton wanted to reward the top schools and redesign the failing ones, Clinton (1996). Figures 1 and 2 below show the inverse relationship between crime and school enrollment.
Figure 1
[pic]
Figure 2
[pic]
Theory and Hypothesis: Building the motivation for the empirical model
The basic model being explored here is that an increase in education, or school enrollment, will subsequently decrease the crime rate. Obviously, there are more factors than just education level that impact crime rates. One of the other variables I used was per capita income. I had come to the conclusion before starting the research that income and education were very closely related and would in turn have similar coefficients in the empirical model that follows in the next section. Does an increase in school enrollment decrease the crime rate in a given area? This would be tested using the null hypothesis which states that school enrollment has no effect on the crime rate.
H0: School enrollment coefficient = 0 vs. Ha School enrollment coefficient ≠ 0
I also looked at the effect of per capita income, which would be tested the same way:
H0: Per capita income = 0 vs. Ha Per capita income ≠ 0
The Empirical Model
In testing the hypothesis from before, we will use the following empirical model:
Violent Crime Ratei = B0 + B1(Per Capita Income)i - B2(High School Enrollment)i+ ei
Violent Crime Ratei = 1820 + 0.00061(Per Capita Income)i – 0.095(High School Enrollment)i+ ei
Alternatively to this model I could have looked at each independent variable alone. I also can use change the dependent variable and use property crime rate.
Prop. Crime Ratei = 1820 – 0.0904(Per Capita Income)i – 0.24(High School Enrollment)i+ ei
Data Description
There were numerous data available that I could have used to come up with similar models. As discussed in further detail previously, graduation rates would have been my ideal statistic to run against the crime crates. Unfortunately, all that were available in years that crime rates were also available were in fact school enrollment numbers. The inclusion of per capita income was also important to my research because I felt that it would have a large impact on crime rates. The data was all gathered from the census bureau.
The dependent variable is as follows:
• National crime rates in a given year
The independent variables are as follows:
• National Per Capita Income in a given year
• National high school enrollment in a given year
Results
Table 3
|Dependent Variable is Violent Crime Rate |
|Independent Variable |Estimated Coefficient |Standard Error |t stat |
|Intercept |1820.2 |159.4 |11.42 |
|HS Enrollment |-0.09505 |0.01525 |-6.23 |
|Per Capita Income |0.00061 |0.003398 |0.18 |
|R2 |83.2% |
|Number of Observations |54 |
The results from our model are shown in the table 3 above and interpreted as follows:
1. The intercept of 1820.2 implies that without any going to high school or earning an income the violent crime rate would be 1820.2. This is a very literal interpretation and not really a likely outcome in life.
*the overall crime rate for Washington DC in 2003 was 1625, which may lead me to believe that our nation’s capital has poor school attendance and very high incomes.
2. High school enrollment has a coefficient of -0.09505 which implies that for every additional person attending high school the subsequent violent crime rate will see a reduction of 0.09505.
3. Per capita income has a positive relationship with violent crime; for every unit increase of PCI the violent crime rate increase by 0.00061.
4. Our estimated model has a very high R2 of 83.2% which means that the total variation of violent crime rates is explained by the model while 16.8% is left unexplained.
The per capita income coefficient did not turn out as I had expected. I anticipated it to have an inverse relationship with the violent crime rate. I don’t have any explanation for this result other than a simple theory; it may be that as per capita income increases the gap between the classes is actually growing. If this is the case then I can understand a higher per capita income having a positive relationship with violent crime rates.
Shortcomings of this work
Originally I wanted to compare graduation rates and crime rates, but in the end was very hard to find decent data. I wasn’t able to compare the same kind of data over time which makes it difficult to make any absolute assumptions. I was however, able to find crime rates and their coinciding school enrollment rates on a national level dating back to 1980. It can be argued that school enrollment rates are not as good of an indicator as graduation rate. Enrollment rates are affected by population increase, but I am assuming that crime rates will also increase as the population increases. If time were more abundant I would check crime rates against as many factors as I could possibly obtain; for now crime rates increasing with population is strictly an assumption on my part. By running a regression of the crime data and actual graduation rates for 2003[1] I was able to come up with similar coefficients to that of the regression with school enrollment.
Another shortcoming is the assumption that expenditure on law enforcement is a factor that affects crime rates. The only problem is that simple logic would see the connection between areas with high crime rates have the necessity for higher amounts of money spent on law enforcement. Chicago (historically a city with high crime rates) is obviously going to spend more money per capita on law enforcement than Pullman. A comparison like this might be more productive by comparing an area’s crime rate to the expenditure on crime in a previous period. This would take me too far away from my model however.
Conclusion
After collecting some data on my own it is very easy to agree with the majority of the literature available and say that education does lower the crime rate. From my model I would definitely say that education and crime rate have an inverse relationship. It is troubling that education isn’t a higher priority for politicians. In Nigeria it costs more money per year to keep a prisoner than it does to educate a university graduate, Akpotu (2004).
Data Appendix
National Crime Rates: National crime rates for a given year as well as crime rates for each state in 2003. A detailed table of the crime rates is attached at the end of this paper in Table A.
High School Enrollment: Number of students enrolled in grades 9-12 for a given year.
Per Capita Income: National averages for per capita income for a given year.
Graduation Rates: High school graduation rates for each state in 2003.
*all of the above data was obtained at the census bureau via the website:
http://www.census.gov/compendia/statab/law_enforcement_courts_prisons/crimes_and_crime_rates/
Table 2
|Rate per 100,000 population: | | | |
|Alabama |60.7 |429 |36,709 |
|Alaska |63.6 |598 |57,027 |
|Arizona |70.0 |513 |41,995 |
|Arkansas |71.8 |456 |32,983 |
|California |71.0 |580 |51,185 |
|Colorado |72.5 |347 |48,198 |
|Connecticut |79.3 |317 |60,528 |
|Delaware |60.7 |675 |50,315 |
|District of Columbia |58.9 |1,625 |46,574 |
|Florida |57.5 |731 |41,236 |
|Georgia |56.3 |455 |43,037 |
|Hawaii |63.7 |272 |53,554 |
|Idaho |77.8 |246 |39,934 |
|Illinois |76.3 |556 |48,953 |
|Indiana |73.0 |352 |42,195 |
|Iowa |82.5 |278 |41,350 |
|Kansas |75.0 |398 |41,638 |
|Kentucky |69.7 |249 |35,269 |
|Louisiana |60.6 |637 |35,110 |
|Maine |74.0 |109 |42,163 |
|Maryland |74.4 |704 |57,424 |
|Massachusetts |72.1 |473 |55,658 |
|Michigan |66.4 |511 |44,905 |
|Minnesota |79.0 |263 |50,860 |
|Mississippi |60.8 |324 |31,642 |
|Missouri |74.7 |491 |41,473 |
|Montana |75.8 |365 |35,239 |
|Nebraska |77.8 |294 |41,657 |
|Nevada |55.9 |616 |44,646 |
|New Hampshire |77.7 |150 |55,580 |
|New Jersey |84.5 |364 |61,359 |
|New Mexico |56.7 |667 |36,043 |
|New York |62.5 |466 |47,349 |
|North Carolina |66.2 |454 |39,428 |
|North Dakota |83.1 |80 |39,447 |
|Ohio |76.5 |334 |42,240 |
|Oklahoma |71.0 |506 |35,357 |
|Oregon |69.0 |295 |41,794 |
|Pennsylvania |79.1 |398 |42,941 |
|Rhode Island |72.3 |286 |48,722 |
|South Carolina |52.5 |806 |39,837 |
|South Dakota |74.5 |174 |38,472 |
|Tennessee |62.2 |691 |38,794 |
|Texas |66.8 |553 |41,759 |
|Utah |76.7 |250 |47,074 |
|Vermont |81.2 |114 |46,543 |
|Virginia |74.9 |278 |51,689 |
|Washington |68.2 |347 |47,659 |
|West Virginia |72.8 |255 |31,504 |
|Wisconsin |80.6 |221 |45,315 |
|Wyoming |74.0 |262 |44,275 |
|U.S. |69.6 |476 |44,684 |
Works Cited
Akpotu, Dr. Nelson Ejiro. “An Analysis of the Link between Education and Crime: Prison Inmates’ Perception of Nigeria.” Africa Educational Research Network. Volume 4; Number 4 (2004). March 3, 2007. http://www2.ncsu.edu/ncsu/aern/crimpap.html.
Clinton, Bill. Between Hope and History, by Bill Clinton, p. 44 Jan 1, 1996.
Lochner, Lance and Moretti, Enrico. “The Social Savings from Reducing Crime through Education.” Joint Center for Poverty Research. Voume 4; Number 5 (2000). March 3, 2007. http://www.jcpr.org/policybriefs/vol4_num5.html.
Lochner, Lance, "Education, Work, and Crime: A Human Capital Approach”. International Economic Review, Vol. 45, No. 3, pp. 811-843, August 2004 Available at SSRN: http://ssrn.com/abstract=573001
Karpowitz, Daniel. “Education as Crime Prevention: The Case for Reinstating the Pell Grant Eligibility for the Incarcerated.” 2000.
-----------------------
[1] Crime Rate = 1564 - 20.2(Graduation Rate) + 0.00654(Median Household Income)
Graduation rates have a stronger impact on crime than school enrollment[pic]-89:€?‚‹Çš›œ?ž«òâÒ¶ª›Œ€tg[O[OC4ghm*hKQ;CJOJQJaJhKQ;CJOJQJaJhjLBCJOJQJaJhm*CJOJQJaJhm*5?CJOJQJaJhm*CJ0OJQJaJ0hdCJ0OJQJaJ0hdhdCJ$OJQJaJ$h>[2]±hdCJ OJQJaJ hdCJ$OJ, but we still see that both are inversely related. 