WA2 Secondary Data analysis

American Journal of Economics and Sociology, Inc.

Steven Levitt on Abortion and Crime: Old Economics in New Bottles Author(s): Robert Chernomas and Ian Hudson Source: The American Journal of Economics and Sociology, Vol. 72, No. 3 (JULY, 2013), pp. 675-700 Published by: American Journal of Economics and Sociology, Inc. Stable URL: https://www.jstor.org/stable/23526056 Accessed: 11-03-2020 16:21 UTC

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide

range of content in a trusted digital archive. We use information technology and tools to increase productivity and

facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at

https://about.jstor.org/terms

American Journal of Economics and Sociology, Inc. is collaborating with JSTOR to digitize, preserve and extend access to The American Journal of Economics and Sociology

This content downloaded from 206.224.223.249 on Wed, 11 Mar 2020 16:21:38 UTC All use subject to https://about.jstor.org/terms

Steven Levitt on Abortion and Crime:

Old Economics in New Bottles

By Robert Chernomas* and Ian Hudson1

Abstract. This article explores the work of Steven Levitt on abortion and crime as representative of a new wave of economists that are redefining the mainstream of the discipline. Levitt has moved away from formalized modeling of narrow self-interested maximization that

was a hallmark of the "old" mainstream. However, he retains the "old"

mainstream linchpins of a focus on the individual by abstracting from

the social context, and an emphasis on the ability of econometric testing to reveal the truth. These crucial elements that Levitt retains from the

"old" mainstream create some problems of both emphasis and analysis for his explanation for the drop in crime in the United States.

Introduction

This article explores the work of Steven Levitt, winner of the John Bates Clark award as the most outstanding economist under the age of 40 and best-selling popular author, as a representative of a new wave of economists increasingly occupying tenured positions at places like Chicago, Harvard, MIT, and in departments that follow their lead (The Economist 2008). For these economists the discipline is no longer constrained by the Samuelsonian requirements of mathematical formalization based on simplifying assumptions. Like Swedish Bank Prize winning, University of Chicago economist Gary Becker, they apply their analysis to nontraditional areas of society. Unlike Becker they have less of an attachment to the rational, calcu lating, maximizing model of behavior, which were hallmarks of the "old" mainstream in the discipline. And so Levitt is identified as a rogue economist, a scientific sleuth, using data to go where no

'Department of Economics, University of Manitoba, Winnipeg, MB Canada, Email: [email protected].

department of Economics, University of Manitoba, 501 Fletcher Argue Building,

Winnipeg, MB Canada, R3T 5V5, Email: [email protected].

American Journal of Economics and Sociology, Vol. 72, No. 3 tjuly. 2013). DOI: 10.1111/ajes.l2024 © 2013 American Journal of Economics and Sociology, Inc.

This content downloaded from 206.224.223.249 on Wed, 11 Mar 2020 16:21:38 UTC All use subject to https://about.jstor.org/terms

676 Tlje American Journal of Economics and Sociology

economist has gone before to find unique answers to questions that no one has asked before.

We will argue in a more profound sense he fits very comfortably within the orthodoxy of mainstream economics by ignoring the social roots—macrofoundations—of the phenomena being investigated in favor of methodological individualism. In this sense, Levitt and his "new" mainstream colleagues have not escaped the criticism of people like Sheila Dow, Robert Heilbroner, and William Milberg, who argue

that the problem with orthodox economics is its lack of attention to the effects of broader social forces on the individual. Levitt's myopia

with respect to American exceptionalism, his lack of consideration of the unique features of the American socioeconomic framework, underscores this deficiency in his work.

This article will take up these issues with a critical examination of Levitt's most famous work on the relationship between abortion and crime in the United States, which frames political issues in a manner that neglects serious scrutiny of the system's effect on real people and societies. Levitt's retention of crucial elements of the "old" mainstream

creates a narrow view of the causes of crime in the United States that

is less nuanced than those put forward by authors like Loic Wacquant, who explicitly locate crime in a social context. But first we will locate Levitt inside the mainstream of the economics profession.

The New Empiricists

The economics discipline has pretensions to scientific objectivity but its more thoughtful practitioners have rightly argued that ideology is inexorably associated with any economic analysis. For Joseph Schumpeter, "it should be perfectly clear that there is a wide gate for

ideology to enter into this process. In fact, it enters on the very ground floor . . . Analytical work begins with material provided by our vision of things, and this vision is ideological almost by defini tion" (2006: 40). Similarly, Joan Robinson claimed that "[pjolitics involve ideology; there is no such thing as a purely economic problem that can be settled by purely economic logic; political inter ests and political prejudice are involved in every discussion of actual questions" (1977: 1318).

This content downloaded from 206.224.223.249 on Wed, 11 Mar 2020 16:21:38 UTC All use subject to https://about.jstor.org/terms

Steven Levitt on Abortion and Crime 677

Not very long ago, the mainstream of the discipline contained a readily identifiable ideology. According to Dow, mainstream econo mists most frequently assumed that people are self interested, which means that only their own considerations enter into their utility. This

assumption is not universal. For example, Becker allows love to enter into an individual's utility function by positively relating others' con sumption to a person's utility. However, even in the case of Becker's individual, who is capable of moving beyond narrow, individual self-interest, this is considered only in the most superficial manner. This is because Becker, and other mainstream theorists, start from the assumption that tastes and preferences (which they view as non economic factors) can be taken as given. So the broader social and economic context that might influence a person's decision is never analyzed (Dow 2002: 23). Feminist/heterodox analysts, on the other hand, consider the social structure that has led to a set of market conditions and social conventions that influence individual behavior.

The social conditions are prior to the individual. How and why they change is the means to understand the conditions under which individuals make the choices they do. Understanding the social struc ture, social conventions, and the labor market are critical if one is to

understand virtually any supposedly individual decision such as crime, smoking, and divorce (Dow 2002: 24).

Heilbroner and Milberg also point out why focusing on the indi vidual while ignoring the broader social context is problematic:

Humans do not and cannot form intelligence, organize memories or exercise will as isolated beings, entirely self-sufficient in the mental psy chological processes. The development of the human psyche, from the earliest moments of infancy on, takes place through the gradual ingestion and incorporation of the individual's surroundings from its earliest familial influences through its exposure to innumerable influences on other indi viduals directly or indirectly. Thus the concept of the individual—the analytical focus of so much conventional social science—appears most clearly in the form of a unique distillation of social influences. In Marx's profound words, the individual appears as "the ensemble of social relations," a "chaotic conception" unless we penetrate the façade of "the individual" to its social roots. (Heilbroner and Milberg 1995: 86)

Dow also argues that the mainstream approach was well suited to mathematical formalization and econometric testing because it used

This content downloaded from 206.224.223.249 on Wed, 11 Mar 2020 16:21:38 UTC All use subject to https://about.jstor.org/terms

678 The American Journal of Economics and Sociology

very specific simplifications. In contrast, the scope for econometric work is constrained in the heterodox approach because its "emphasis on the social rather than the individual approach" eschews simplistic models that are easy to test (Dow 2002: 24).

The hallmarks of the "old" mainstream were self-interested maxi

mization, the individual as the center of inquiry, and a method that focused on mathematical formalization and econometric testing. The question for this article is to what extent the "new" mainstream, as

represented by Levitt, has deviated from these questionable practices. The subtitle of Freakonomics claims that Levitt is a "rogue eco

nomist," but what exactly separated Levitt from the rest of the economics profession in the late 1990s and early 2000s? In the introduction to Freakonomics Levitt provides some clues. It is about "stripping a layer or two from the surface of modern life and seeing what is happening underneath" (Levitt and Dubner 2005: 12). So part of what makes Levitt a "rogue" is that he questions conventional wisdom, or at least elements of conventional wisdom, and uncovers unpredictable, often "distant, even subtle causes" (Levitt and Dubner 2005: 12). A second component of Levitt's method is that the answers to these questions will "come from the data." He makes some very strong claims in this regard, arguing that in contrast to morality, which is how people "would like the world to work," economics represents how it "actually does work" because its meas urement tools can "reliably assess a thicket of information to deter mine the effect of any one factor" (Levitt and Dubner 2005: 13). Finally, when searching for these subtle, often distant, causes of life's

seemingly perplexing enigmas, understanding how incentives are aligned is the key to the solution.1

This is very revealing about Levitt's economic worldview. Many of the much-objected-to centerpieces of "old" mainstream theory have been jettisoned. What remains is a focus on incentives, methodolo gical individualism, and a strong belief that the truth can be revealed

if only the right questions are asked and appropriate data analyzed in a sufficiently rigorous way.

The cutting edge in economics is, then, basically a toolkit—a set of

data-finding techniques that is bound by neither theory nor subject matter. Levitt argues that this particular set of tools can be usefully

This content downloaded from 206.224.223.249 on Wed, 11 Mar 2020 16:21:38 UTC All use subject to https://about.jstor.org/terms

Steven Levitt on Abortion and Crime 679

applied to almost any area of social enquiry since "no subject, however, offbeat, need be beyond its reach" (Levitt and Dubner 2005: 14). Levitt places great confidence in the truth-revealing properties of

data. "I'd like to put together a set of tools that lets us catch terrorists.

I mean that's the goal. I don't necessarily know yet how I'd go about it. But given the right data, I have little doubt that I could figure out the answer" (Dubner 2003: 28). A skeptic might ask what is particu larly "economic" about this latest direction for the discipline. Indeed, many would say he is doing a version of quantitative sociology.

His own personal favorite areas are crime, cheating, and corruption (Dubner 2003), but he has also delved into politics and race. His publication list spans a wide variety of other social sciences. He has written on schoolteachers helping their classes cheat on standardized tests (Jacob and Levitt 2003), sumo wrestlers throwing matches to improve their rankings (Duggan and Levitt 2002), and the systematic fleecing of their clients by real estate agents (Levitt and Syverson 2008). In the area of crime he has inquired into the income distribu tion within criminal gangs (Levitt and Venkatesh 2000). He has also attempted to analyze the impact on crime of policing strategies (Levitt

1997, 2002), incarceration (Levitt 2004a), and capital punishment (Katz et al. 2003). In the realm of politics he has looked at the impact of interest groups on political decisions (Groseclose, Levitt, and Snyder 1999) and dismissed the importance of campaign spending on election results (Levitt 1994). On the subject of race he has examined the differences in test scores (Fryer and Levitt 2004a), used "black" names to test for discrimination (Fryer and Levitt 2004b), and looked at the economic lives of blacks in public housing (Levitt and Venkatesh 2001).

So Levitt may be a rogue economist in an incidental way. He has ignored the traditional subject areas of the economy—interest rates, stock markets, unemployment, economic growth, and corporate behavior—in favor of areas that have normally been the purview of

other disciplines like sociology and politics. In this he is following in the path of Chicago mainstay Becker. He has also done away with the heavily criticized assumptions of the rational, maximizing individual championed by Becker. Indeed, his economics is particularly devoid of what would traditionally be described as economic theory. In the

This content downloaded from 206.224.223.249 on Wed, 11 Mar 2020 16:21:38 UTC All use subject to https://about.jstor.org/terms

680 Hoe American Journal of Economics and Sociology

words of one critic of Freakonomics, "I could detect very little in the research that depended on economic theory in anything but the most superficial way" (DiNardo 2007: 996). He has also eschewed the complex mathematical modeling of much of the mainstream of the discipline in favor of a more "intuitive" approach to puzzle solving.

Levitt's Influence

Measuring an academic's influence on the discipline is always difficult.

If success is measured in terms of name recognition and book sales, Levitt has succeeded far beyond the wildest dreams of any social scientist. He is best known to the broader public for the best seller Freakonomics (over 3 million copies sold), or perhaps now the slightly less successful sequel Superfreakonomics, but he has also garnered considerable recognition within the discipline. Since 1997 he has been at the University of Chicago, the pinnacle of mainstream economics, where he earned tenure in only two years (Dubner 2003). He has published regularly in the top mainstream economics journals. In 2003 he won the John Bates Clark Medal awarded to the best young (under 40) economist in the United States.

But rapid promotion, professional awards, and blockbuster book sales do not necessarily demonstrate that Levitt is actually influenc ing the academic world. In an effort to determine Levitt's academic influence, we will examine his impact in two areas. The most obvious is the economics profession. In addition, since we will use Levitt's work on abortion and criminality as our illustration of Levitt's methods, we will examine his influence in the area of crime.

The standard measure of academic influence is citation counting. One of the problems with a count of total citations as a measure of influence is that it inflates the value of an author who produces a great

many articles, even if none of them have a particularly large impact. In an effort to capture impact of authors who produce works that generate considerable interest, a measure called the h-index has become a widely accepted indicator. The h-index combines the number of articles with the number of citations for each article so that

a score of 7 indicates that the author has produced 7 articles, each of which is cited at least 7 times. An h-score of 10 would result from an

This content downloaded from 206.224.223.249 on Wed, 11 Mar 2020 16:21:38 UTC All use subject to https://about.jstor.org/terms

Steven Levitt on Abortion and Crime 681

author having 10 articles with at least 10 citations each. We will use both total citations and the h-index to measure Levitt's influence in the

academic world.

Levitt is widely cited. Research Papers in Economics (RePEc) collects citation information on over 1 million working papers and journal articles on its database. As of May 2011, Levitt ranked in the top 5 percent of economists in total citations (389th of over 24,000 authors) and h-index (203rd) (IDEAS 2011a). Although the RePEc citation count provides some indication of author influence, it has some important limitations because of the voluntary nature of the database, which contains a large number of working papers and unpublished works. There are a number of alternative databases that could be used, including Web of Science and Scopus, that focus more on published works. We will use a relatively new program, Harzing's Publish or Perish (HPOP), that provides citation information gathered from Google Scholar. The advantage of HPOP is that it is the broadest citation list, including book chapters and dissertations that are not in the other databases. The indices used in HPOP are also less sensitive

to typographical errors than the Web of Science (Long, Boggess, and Jennings 2011: 109).

In order to compare Levitt to the very top rank of his contempo raries, his academic influence, as measured by citations and the h-index, will be compared to his fellow John Bates Clark winners, Matthew Rabin (awarded in 2001) and Daron Acemoglu (awarded in 2005). These authors were chosen because they have relatively similar career lengths, avoiding the bias in favor of academics with longer careers inherent in citation counting. HPOP citations data up to June 2011 were searched for articles produced by the three authors up to and including 2010. The search was conducted in the areas of business and economics and the social sciences and humanities. The

data for each author were checked to ensure that the articles were not

written by another author with a similar name. The articles that were

duplicated in the data due to spelling mistakes or abbreviated titles were combined into one entry. The results are in Table 1.

As measured by citations and the h-index, Levitt's influence is less than Acemoglu, but is comparable to Rabin. It should be noted that comparing anyone to Acemoglu is setting the bar remarkably high.

This content downloaded from 206.224.223.249 on Wed, 11 Mar 2020 16:21:38 UTC All use subject to https://about.jstor.org/terms

682 The American Journal of Economics and Sociology

Table 1

A Comparison of Levitt with Contemporary Winners of John Bates Clark Medal

Citations Citation/article h-index

Levitt (awarded in 2003) 9,268 72 45 Rabin (awarded in 2001) 14,702 200 33

Acemoglu (awarded in 2003) 29,715 105 73

Calculated from HPOP: Citations of articles written up to and including 2010.

According to the RePEc data base he has the sixth highest h-index in the profession (IDEAS 2011b). Rabin has more citations overall and per paper, but Levitt's higher h-index demonstrates that he has written

more widely cited papers. Levitt's citations may not quite rank up with

the absolute top echelon of the economics world, but he can certainly lay claim to an impressive number of widely cited papers.

Levitt's influence on the economics profession has been more about the broad areas of method and expansion of subject boundaries than prominence in any one particular subject area. This would suggest that citation counting would understate his impact. He was not the first to discover the empirical advantages of clever natural experiments, but his success has influenced the research directions of the young scholars in applied microeconomics who will be the future of the profession. As James Poterba remarked, Levitt's "research has high lighted the importance of devoting attention to well thought-out identification strategies as well as to the econometric tools that are used in applied economic analysis. Steve has inspired others to follow his lead, and in the process his influence has exceeded his own direct research contributions" (Poterba 2005: 197). In The Economist's 2008

survey of economics' rising stars, two (Jesse Shapiro and Roland Fryer)

were "recognisably the intellectual heirs of Mr. Levitt. They share the

same knack for finding ingenious ways to answer unlikely questions, often by plundering forgotten troves of data" {The Economist 2008). In their comment on the economics profession, Roger Backhouse and Steven Dulauf identified Levitt's focus on natural experiments as one of the major streams in modern microeconomics (2009). David

This content downloaded from 206.224.223.249 on Wed, 11 Mar 2020 16:21:38 UTC All use subject to https://about.jstor.org/terms

Steven Levitt on Abortion and Crime 683

Colander surveyed 230 graduate students from seven of the top economics departments in the United States for his book The Making of An Economist Redux (2007). What he found was that "more and more 'Freakonomics' is becoming mainstream economics. That is the research that young economists see themselves doing" (Uchitelle 2006). The top mainstream economics departments in the United States are now littered with young economists emulating Levitt's approach, on a wide range of issues from government fiscal policy to economic development to black underachievement.

If impact in a discipline can be associated with the number of other authors citing your work, Levitt also appears to be influential in the area of crime. If only Levitt's articles on crime are considered, his publication record compares favorably with the top authors in crimi nology and criminal justice. Long, Boggess, and Jennings (2011) used HPOP to examine the citations of the top 10 "stars" in criminology and

criminal justice. The top three authors by citation count were Alex Piquero, Joan Petersilia, and Robert Brame. To compare Levitt with these three authors, his non-crime-related works were deleted, leaving

his articles in the areas of policing, incarceration, the decline in crime,

gang finances, and corruption. The steps that were used in the citation

ranking of economists were replicated for Levitt's crime articles and each of the three criminology and criminal justice authors. The results are in Table 2.

Levitt's citations on crime alone would place him among the most influential of authors in criminology and criminal justice. There is a

Table 2

A Comparison of Levitt with the Top 3 Criminologists and Criminal Justice Authors by Citation

Citations Citation/article h-index

Levitt (crime only) 4,776 106 29

Piquero 5,714 22 42

Petersilia 4,672 27 36

Brame 3,163 41 27

Source. Calculated from HPOP: Citations of articles written up to and including 2010.

This content downloaded from 206.224.223.249 on Wed, 11 Mar 2020 16:21:38 UTC All use subject to https://about.jstor.org/terms

684 The American Journal of Economics and Sociology

difference between disciplines in that Levitt has produced far fewer articles, but each article is cited far more times. Both Piquero and Brame have been publishing articles over a period roughly compa rable to Levitt. Petersilia has had a longer career, biasing the results in her favor. Having said that, in terms of total citations Levitt ranks

just behind Piquero. Using the h-index, Levitt would rank just behind Petersilia. If Levitt had produced no other work besides his research on crime, he would have had a distinguished career in that area.

The extent to which other authors cite Levitt's work suggests that he

has a very respectable impact among the economics profession and an impressive influence in the area of crime. His influence in economics is understated by focusing on citations because he has popularized a new direction in microeconomics, especially for young economists.

Case Study: Abortion and Crime

Levitt's work on the causes of the fall in crime in the United States is

an excellent example of this new direction in mainstream economics. It is outside the traditional economic subject areas. It involves a clever natural experiment in its empirical work and it arrives at a counter intuitive, "surprising" result. There are three versions of Levitt's famous

abortion and crime study that move chronologically from the most empirical and academic to the most popular. The first (with John Donohue) appeared in the Quarterly Journal of Economics (QfE) in 2001, the second in the Journal of Economic Perspectives (JEP) in 2004,

and finally, it was popularized in a chapter in Freakonomics titled "Where have all the criminals gone?" Since the Freakonomics chapter contributes little in the way of new analysis this article will focus on

the first two studies. The United States witnessed a significant, long lived, and completely unexpected decline in crime rates starting about 1991. This understandably caught many commentators by surprise since crime rates, especially for violent crime, had been increasing at an alarming rate in the 1980s. Graphic predictions of "chaos" in U.S. cities, made even by academic commentators (Levitt 2004b: 169), not only failed to materialize but crime rates fell dramatically. Levitt's first

contribution to this debate, which was completely consistent with his desire to find surprising or unexpected causes, was to link declining

This content downloaded from 206.224.223.249 on Wed, 11 Mar 2020 16:21:38 UTC All use subject to https://about.jstor.org/terms

Steven Levitt on Abortion and Crime 685

crime rates in 1991 to the legalization of abortion with Roe v. Wade in 1973.

The logic behind Levitt's hypothesized connection rests on two assumptions. First, aborted babies are unwanted, which would seem relatively uncontroversial. Second, unwanted babies are more likely than average to grow up to be criminals. It is certainly true that the birth rates of groups that Levitt considered "at risk" for future crimi

nal activity dropped more than the average as a result of the legali zation of abortion. The decrease in birth rates was twice as great for teenage and nonwhite mothers as it was for the nonteen, white population. Levitt also noted the same trend comparing black and white women (Donohue and Levitt 2001: 387). Parents with

unwanted children were less nurturing and more likely to resent their children. Children of mothers who were denied an abortion

were more likely to be involved in crime and have poorer life prospects even when controlling for the income, age, education, and health of the mother (Donohue and Levitt 2001: 388). In his own

"back of the envelope" calculations he argues that since "at-risk" groups—blacks, teens and unwed mothers—are both more likely to commit crimes and more likely to have decreasing birth rates due to legalized abortion, the crime rate should fall. To take just the black population as an example, Levitt argues that since homicide by black youths is nine times that of whites and the drop in fertility of black women three times as great as for whites (12 percent to 4 percent), the homicide rate will drop by more than just the cohort effect (the impact of having less high crime 18-24 year olds as a result of the legalization of abortion). In Dubner's New York Times Magazine article, the connection was summarized as: "Unwanted ness leads to high crime; abortion leads to less unwantedness; abor tion leads to less crime" (Dubner 2003).

In support of this hypothesis he produced four pieces of evidence in his QJE article. Most obviously, the timing works out. The crime rate

(Levitt uses three types of crime to calculate the crime rate in this article: property crime, violent crime, and murder) dropped dramati cally starting in 1991, 18 years after the 1973 legalization. Since most criminals' peak crime years are between 18 and 24 the increased number of abortions should start to show up exactly in 1991 and

This content downloaded from 206.224.223.249 on Wed, 11 Mar 2020 16:21:38 UTC All use subject to https://about.jstor.org/terms

686 The American Journal of Economics and Sociology

should have a cumulatively larger impact over time. Second, five states

were early movers, legalizing abortion in 1970, three years before the rest of the nation. Crime rates in these five states started to drop three

years before the rest of the country. Third, Levitt uses panel data to

estimate that states with higher abortion levels in the 1970s and early 1980s had lower crime rates from 1985 to 1997. This connection

remains significant even when other factors such as the level of incarceration, the number of police, lagged state welfare generosity, the presence of concealed handgun laws, per capita beer consump tion, and measures of the state's economic well-being (the unemploy ment rate, income per capita, and poverty rate) are factored in to the model (Donohue and Levitt 2001: 401). In this regression Levitt also finds that incarceration rates and more police lead to fewer murders while a higher state unemployment rate is associated with significant increases in property crime, but not violent crime (Donohue and Levitt

2001: 405). Fourth, Levitt finds that crime rates drop much more rapidly among the cohorts after abortion is legalized than those born prior to legalized abortion (Donohue and Levitt 2001: 410).

The estimated magnitude of the impact of legalized abortion on crime is large. Levitt finds that it is responsible for up to 50 percent of the total drop in crime rates, making it much more important than any of the other, more popular explanations for the decrease in crime (Donohue and Levitt 2001: 382). He even makes a back of

the envelope calculation that societal saving resulting from legalized abortion's effect on crime may be as much as $30 billion annually, a not inconsiderable sum (Donohue and Levitt 2001: 414). In his

conclusion, Levitt does write something of a disclaimer, claiming that these reductions could have been achieved through other methods like providing better environments for children or other forms of contraception (Donohue and Levitt 2001: 415).

The JEP article was not an original empirical contribution as was the

QJE piece. Rather, it combined some of his previous research with a review of other quantitative work to distinguish between the factors that did and did not cause the crime rate to fall in the 1990s. His

reading of the literature is that innovative, New-York-City-style police

tactics, a strong economy, gun control laws, changing demographics, increased use of capital punishment, and laws that permit carrying

This content downloaded from 206.224.223.249 on Wed, 11 Mar 2020 16:21:38 UTC All use subject to https://about.jstor.org/terms

Steven Levitt on Abortion and Crime 687

Table 3

Levitt's Causes of Decline in U.S. Crime (percentage decrease in the crime rate)

Homicide Violent Property

Total Decline 43 33 29 Contribution of Individual Factors

Increased Police 5-6 5-6 5-6

Increased Prison Pop 12 12 8 Decline in Crack 6 3 0

Abortion 10 10 10

Source. Levitt (2004b).

concealed weapons had very little or no impact. On the other hand, the increased number of police, rising prison population, the dwin dling use of crack, and, obviously, legalized abortion all had a significant impact.

Further, Levitt attempts to roughly quantify the magnitude of each of the significant factors in reducing homicide, violent crime, and property crime (see Table 3), as well as speculate about the cost effectiveness of a few of the measures. The number of police increased by 14 percent during the 1990s and Levitt estimates this was responsible for a 5 to 6 percent reduction in all three types of crime. This would create around $20 billion in benefits for an $8.4 billion

outlay on police. The $50 billion spent on incarceration are also outweighed by the benefits of an estimated 8 percent drop in property

crime and 12 percent fall in homicide and violent crime, although Levitt does warn that imprisoning large numbers of people may result in declining marginal benefits and have undesirable social conse quences (Levitt 2004b: 179). It is worth mentioning here that Levitt's Freakonomics explanation for the increase in crime in the 1980s was the decrease in monitoring and penalties in the criminal justice system.

"America was getting softer on crime 'for fear of sounding racist', as the economist Gary Becker has written, since African-American and Hispanics commit a disproportionate share of felonies. So, if you were the kind of person who might want to commit a crime, the incentives

This content downloaded from 206.224.223.249 on Wed, 11 Mar 2020 16:21:38 UTC All use subject to https://about.jstor.org/terms

688 Tloe American Journal of Economics and Sociology

were lining up in your favor: a slimmer likelihood of being convicted and if convicted a shorter prison term" (Levitt and Dubner 2005: 122). The decline in crack and its associated violence had a more modest

effect decreasing homicide by 6 percent and property crime by little or nothing, although Levitt admits that the evidence on this score is

very limited (Levitt 2004b: 181). He estimates a more modest impact for abortion in the JEP article compared to the original QJE piece, claiming it accounts for a 10 percent decline in homicide, violent crime, and property crime (one-third of the total).

There is Such a Thing as Society: The Importance of Economic Context

It should hardly come as a surprise that a paper arguing that safer streets have been caused by abortion should spark considerable controversy. Considering the faith that Levitt places in the truth revealing properties of creative natural experiments, his research has attracted considerable criticism on empirical grounds. Critics of Levitt's

empirical work have taken him to task on a number of fronts, from the inappropriateness of his data to his poor choice of tests (Lott and Whitley 2007; Foote and Goetz 2008; Joyce 2009; DiNardo 2007). The uncertainty created by these criticisms reveals that Levitt's claim

about well-defined research and careful selection of data revealing the undisputed "truth" of an issue, especially one as controversial and difficult to demonstrate as the link between legalizing abortion and crime, is overly simplistic. It is possible that Levitt may not have succeeded even on his own, very narrow terms. Further, given Levitt's abandonment of both theory and broader context, if he cannot demonstrate empirical "truth" he is left with little else.

Levitt's focus on legalized abortion as the main factor in the reduc

tion in U.S. crime in the 1990s is particularly revealing about the way he views the economic world. Crucially, he is very dismissive of broader macro economic factors, or the social context in the reduc

tion in crime, while focusing on a "surprising" explanation. In his QJE article the broader context is treated as a control variable, while in the

JEP article economic factors are more explicitly considered but only in a very rudimentary way.

This content downloaded from 206.224.223.249 on Wed, 11 Mar 2020 16:21:38 UTC All use subject to https://about.jstor.org/terms

Steven Levitt on Abortion and Crime 689

As was mentioned in the introduction, Levitt does not view himself

as much of a theoretician. While moving away from "old" mainstream

economics' complicated mathematical models with their implausibly narrow and unrealistic assumptions has its advantages, jettisoning all economic theory might not. This is nicely highlighted in his discussion

about the relationship between the economy and crime. The most famous "old" mainstream theory of the economics of crime is that of

fellow University of Chicago economist Becker, who uses the assump tions of a rational, self-interested, maximizing individual to argue that

the decision to commit a crime is simply the result of a calculation of the costs and benefits. Crime will be reduced when the costs are

increased (through increased policing or jail times) or benefits decreased. According to Becker's model, higher noncrime income should reduce the incentive for criminal activity. Levitt argues that

while this might be true for crimes with financial gain, it is likely to be

less important for more violent crimes. Further, he argues that even theoretically there is an uncertain connection between the economy and crime because increased incomes might lead to a greater number of activities that are related to crime, like "alcohol consumption,

frequenting nightclubs and owning a car" (Levitt 2004b: 170). As evidence for the lack of importance of economic factors Levitt cites a number of other studies that find only a small, although statistically

significant, connection between property crime and the unemploy ment rate (or other economic measures like the income of low-wage workers) and no relationship at all between economic measures and violent crime. He concludes that the only likely mechanism through which the economy is likely to impact crime is by providing the necessary tax revenue to pay for police and jails (Levitt 2004b: 170).

There are both theoretical and empirical problems with this dis missal. Theoretically, it is not clear why Levitt would dismiss the link between violent crime and a strong economy. In keeping with Beck er's rational model, people with higher incomes and steady jobs have much more to lose from long-term incarceration than those living an unrewarding life outside of the penal system. Less in keeping with Becker's model, life in desperate economic conditions contains far more pressure and tension, which can often lead to violence. Finally, Levitt appears to be ignoring the strong connection between violent

This content downloaded from 206.224.223.249 on Wed, 11 Mar 2020 16:21:38 UTC All use subject to https://about.jstor.org/terms

690 The American Journal of Economics and Sociology

crime and economic crime in many situations. This is especially curious since Levitt does make this connection later in the article when

he argues that the decline in crack sales (a crime for financial gain) has

reduced homicide rates (Levitt 2004b: 181). One problem with Levitt's analysis of the link between the economy and crime is that it lacks a theoretical underpinning, even one as crude and individualistic as Becker's, which leads to some problematic inconsistencies.

Given the lack of theoretical underpinnings it is very difficult to criticize the assumptions on which Levitt's analysis is based. However, his dismissal of the economy as a factor in the reduction in crime is based on a fairly selective review of the empirical literature. Levitt argues that "empirical estimates of the impact of macroeconomic variables on crime have been generally consistent across studies" (Levitt 2004b: 170). Yet, this is not really a consensus opinion. Using an economic model similar to that of Becker, Jeffrey Grogger argues that in the 1980s poor job prospects created an incentive towards the illegal drug trade and away from the legitimate job market. He estimated the impact of young people's wages and youth crime and found that, holding other factors constant, a 1 percent decline in wages would lead to a 1 percent increase in crime. Given that real wages for youth declined by 22 percent in the 1980s, this factor would be responsible for a substantial increase in crime during this decade (Grogger 2006). After surveying the literature, Joel Wallman and Alfred Blumstein have no difficulty finding empirical support for

a connection between an improving economy and decreasing crime in the 1990s. In fact, every study cited by Wallman and Blumstein suggests a strong link between the decrease in crime and an improv ing economy. This may have been because the studies cited by these authors focus more specifically on youth unemployment and wages as opposed to the general unemployment rate. Wallman and Blumstein's conclusion is a more nuanced acknowledgment that in the debate between Levitt and others on the connection between the economy and crime, the jury is still out (Wallman and Blumstein 2006: 337-339).

Although here we are most concerned with the evidence connect ing the broader economic context to crime, it is worth noting that the evidence on one of Levitt's "other factors" is also more controversial

than he lets on in the JEP or Freakonomics. He argues that increasing

This content downloaded from 206.224.223.249 on Wed, 11 Mar 2020 16:21:38 UTC All use subject to https://about.jstor.org/terms

Steven Levitt on Abortion and Crime 691

the number of police had an important (and cost-effective) impact. Yet

others are not as convinced. McCrary (2002) found an important error in Levitt's original 1997 study connecting police numbers to crime. When McCrary tried to replicate Levitt's finding after the error had been corrected, he found that increasing police had no significant impact on crime. Wallman and Blumstein are equally skeptical of Levitt's alternative estimate2 that was supposed to correct for McCrary's

discovery, arguing that "findings that don't make sense should at least

raise questions" (Wallman and Blumstein 2006: 335). They are par ticularly concerned with Levitt's finding that increased police numbers were related to a decrease in homicide but an increase in serious

assault. Wallman and Blumstein argue that common sense would suggest that since murder is less likely to be deterred by police presence than violent crime, Levitt's incongruous findings should be viewed with considerable suspicion (Wallman and Blumstein 2006: 335). Again the point here is not to say that Levitt is unequivocally incorrect in stating that more police reduce crime. Rather, that given

the very mixed existing literature on both policing and the economic context his dismissal of the economy and acceptance of policing seems, at best, somewhat selective.

Levitt's dismissal of the broad macro economic context, and failure to even address distributional issues or social context in which U.S.

crime occurs, is in sharp contrast to other efforts at explaining crime. Perhaps the most well known of the authors who place crime in its social and economic context is Wacquant. Wacquant argues that it was the booming economy of the 1990s that was the largest factor in the decrease in crime, accounting for as much as 30 percent of the decrease in the crime rate. According to Wacquant, while economic growth did not lift many youths from the urban ghetto, since many of

the jobs were temporary and low wage, the Latino population espe cially benefited from the upturn in the unskilled labor market. The young black population also responded to the improved economy by attending postsecondary education in much greater numbers, reducing their involvement in street crime (Wacquant 2005: 101).

In addition to a much more careful examination of the economic

context that leads to crime, Wacquant pays more attention to the social

conditions in high-crime areas. He argues that the increase in female

This content downloaded from 206.224.223.249 on Wed, 11 Mar 2020 16:21:38 UTC All use subject to https://about.jstor.org/terms

692 The American Journal of Economics and Sociology

immigration with greater opportunity in ethnic labor market niches helped reduce the crime rate at least on the borders of the ghettos. In

addition, younger siblings actually took to heart the cautionary tale that was their older brothers' life of crime. The violence associated

with drug dealing also declined as oligopolistic gangs signed peace treaties. Finally, Wacquant argues that some credit should be given to the social pressures of those who organized to rid their neighborhoods

of crime (Wacquant 2005: 102-103). Part of Levitt's lack of attention to the social and economic condi

tions that give rise to crime in the United States is rooted in his failure

to contrast the unique situation in the United States with other nations.

While Levitt is exploring the decline in crime in the United States rather than the difference in crime between nations, focusing on the former while ignoring the latter downplays the extent to which U.S. violent crime is exceptional in the industrialized world, a fact that would encourage an analysis contrasting the broader social and economic context in the United States compared to other, less violent nations. As Wacquant points out, crime in the United States is high, but

not abnormally so. In the 2000 International Crime Victimization Survey (ICVS) the United States tied for seventh highest of the 17 countries in the study in terms of crime incidence (the number of crimes per 100 inhabitants). It actually has a lower incidence rate than Sweden and the Netherlands, is about equal to Canada, but is con siderably worse than Japan and Switzerland (Van Kesteren, Mayhew, and Nieuwbeerta 2000: 38). The big difference is that the United States has a problem with more violent crime. In 1991, when U.S. crime rates

were at their height, the U.S. murder rate for males ages 15-24 (24.4 per 100,000) was approximately 10 times higher than in Canada (2.6), Sweden (2.3), and the United Kingdom (2.0), 25 times higher than in Germany (0.9), and 50 times higher than in Japan (0.5) (Wolff, Rutten,

and Bayers 1992: 116). Even after the significant decline in crime rates

during the 1990s the United States still had much higher murder rates.

Between 1999 and 2001 the United States averaged 5.5 homicides each year per 100,000 people. In Canada this was 1.77, the E.U. average was 1.59 (Barclay and Tavares 2003: 10). This homicide rate is for the whole population rather than the high-crime cohort of youth. In 1998,

after much of the fall in homicide rates, U.S. youth (age 10-29) had a

This content downloaded from 206.224.223.249 on Wed, 11 Mar 2020 16:21:38 UTC All use subject to https://about.jstor.org/terms

Steven Levitt on Abortion and Crime 693

homicide rate of 11 per 100,000. This is over six times the Canadian rate (1.7), 12 times the rate in the United Kingdom (0.9), 18 times the French rate (0.6), and 27 times the Japanese rate (0.4) (WHO 2002: 28-29). It is not just murder that was high in the United States even after the 1990s decline. In 2000 the rate of violent crime in the United

States was more than double that of Canada (500 per 100,000 people vs. 220). The United States had 324 aggravated assaults for every 100,000 population while the Canadian rate was less than half of that at 143 (Statistics Canada 2001: 5).

These international comparisons do not disprove Levitt's crime abortion connection for the United States. He is, after all, attempting to explain the drop in the rate of U.S. crime rather than why its level

is so exceptional internationally. This requires a focus on what has changed in the United States rather than what is different between the

United States and other nations. Yet, his focus on the drop in the U.S.

rate, and lack of attention to the gap in international rates betrays his focus on the individual rather than the context in which the individual

operates. Levitt's focus on the "big surprise" and commitment to methodological individualism causes him to neglect key questions about the social and economic context that makes the United States

such a uniquely violent society. Other scholars have attempted, with some success, to do just this.

Wacquant places the high U.S. rate of violent crime firmly in the unique U.S. social and economic context, arguing that it is due to prevalence of handguns, a deep rooted illegal street economy in the impoverished districts of urban areas, a lack of a social welfare system, and racial segregation (Wacquant 2005: 97-98). Other authors have also established a link between violent crime and the differing eco nomic contexts. To provide one example, Hsieh and Pugh (1993) used a meta study to demonstrate consistent correlations between income inequality and homicide among nations and U.S. states.

Levitt's one comparison with the European Union is to demonstrate that the U.S. drop in crime (rather than its abnormally high level) was

exceptional by showing that U.S. crime rates decreased by far more than

those in the European Union. For example, E.U. homicide rates fell by

4 percent between 1995 and 1999 while they dropped 28 percent in the United States (Levitt 2004b: 166). However, even if one were to focus

This content downloaded from 206.224.223.249 on Wed, 11 Mar 2020 16:21:38 UTC All use subject to https://about.jstor.org/terms

694 The American Journal of Economics and Sociology

on merely the reduction in crime, as opposed to the level, between nations, Levitt does not capture the different contexts. Most obviously,

it is much more unlikely that a country with a very low homicide rate,

like most of the European Union, would be able to achieve the same reductions as one with a high rate. It is also worth remarking that, as Wacquant argues about France, these were not particularly good economic times in much of Europe compared the United States, so if the

economy were a factor it should lead to an improved crime statistics in

the United States compared to the European Union. A better comparison is most likely Canada. Canada managed a

similar crime rate reduction between 1991 and 2001 while the number

of police per person fell 9 percent and the ratio of those in jail to the total population decreased 7 percent (in the United States these increased by 10 percent and 47 percent, respectively) (Wacquant 2005: 100). This is not to suggest that the police were not an important factor

in the United States. Uniformed police in NYC increased from 27,000 to 41,000 from 1993 to 2001, a remarkable increase in enforcement.

By 2001 NYC (population 8 million) had half as many uniformed police as the whole of France (population 60 million), heavily con centrated in particular "problem" urban areas (Wacquant 2005: 106). What this international comparison does suggest is a slightly different cost-benefit analysis than Levitt produced in the JEP. If France can achieve a similar crime rate and a much lower rate for violent crime

with a fraction of the number of police officers used in the United States, and Canada can achieve similar crime reductions while impris onment and police decline, are these not better costs and benefits? Societal and economic changes that would create lower crime with less police, as is the case in Canada and France, is surely more desirable than less crime with more police. Again we see that Levitt's explanation for the decline in U.S. crime rate may be dubious. Further, his focus on the decline in U.S. crime rather than the fact that the

United States has a much higher violent crime rate than elsewhere

ignores the unique magnitude of U.S. violent crime and prevents questioning just what it is about the United States that creates so many violent individuals. This problem is exacerbated by Levitt's focus on a

"surprising" but policy-empty explanation of abortion as the principal cause for the fall in crime in the United States.

This content downloaded from 206.224.223.249 on Wed, 11 Mar 2020 16:21:38 UTC All use subject to https://about.jstor.org/terms

Steven Levitt on Abortion and Crime 695

Conclusion

The assumptions and methodology of the "old" mainstream, repre sented by the self-interested, maximizing individual, mathematical formalization, and econometric testing, have been trenchantly criti cized by those outside the mainstream. This article has argued that the "new" mainstream, as illustrated by Levitt's analysis of crime in the

United States, still runs afoul of many of these criticisms. Further, the elements of the "old" mainstream that are maintained in Levitt's work

create some important problems with how he analyzes U.S. crime. We can examine Levitt's work on abortion and crime by looking at

the extent to which it still uses the three "old" mainstream linchpins:

self-interested maximization; a focus on the individual rather than on the social context; and use of modeling and econometric testing. Unlike much of the "old" mainstream Levitt's analysis does not rely

exclusively on the assumption of individual self-interest. In other works Levitt (Levitt and Dubner 2005) does analyze how social influences, like morality, can influence individual decisions, although as DiNardo (2007) pointed out, he does so in an ad-hoc manner. Levitt does not create formal models that explicitly assume individual self interest and then work though maximizing conditions. On the other hand, Levitt's analysis on crime and abortion certainly stresses the importance of incentives. He even borrows from the self-interested individual when he emphasizes the decreasing costs of crime to the individual in the 1980s because of what he argued was a relaxed attitude toward crime and punishment. Interestingly, in a later article,

Levitt appears to be defending the "old" mainstream version of homo economicus against attacks from behavioral economists and their inconvenient laboratory results that show people are much less self interested than economists had assumed them to be. According to

Levitt, when experiments are moved out of the lab and into the field

people behave in a much more self-interested manner, supporting the assumptions of the "old" mainstream (Levitt and List 2008).

If Levitt does not universally follow the "old" mainstream in explic

itly modeling human actions as self-interested maximization, he does take the level of analysis as the individual. This is not to say that he

completely ignores all social influences in his analysis on crime. He

This content downloaded from 206.224.223.249 on Wed, 11 Mar 2020 16:21:38 UTC All use subject to https://about.jstor.org/terms

696 Tloe American Journal of Economies and Sociology

does test to see if economic conditions like unemployment have an impact (although they are dismissed in the JEP article with a fairly selective reading of the existing literature). Yet, compared to Wac quant, his attention to social context, even within the United States, is

superficial. Further, the extent to which the United States is an outlier

compared to other nations in terms of violent crime is never explored.

Although in a later article on the connection between unemployment and crime Levitt does call for more international comparisons (Levitt 2001: 388), the difference between national social and economic contexts is not the subject of enquiry as it has been for other researchers like Wacquant. Levitt focuses on the criminals instead of the society that spawns them in record numbers.

The econometric critique can also apply to Levitt, who definitely shares the "old" mainstream's commitment to empirical testing. While

his models do not push the boundaries of econometric theory, he has a very firm belief in the ability of his creative testing to reveal the truth.

The difficulty with this dependence in his study of abortion and crime

is that the truth has been very elusive, with other researchers able to cast doubt on his results. Further, Levitt's interpretation of empirical support for what did and did not decrease crime in the United States seems to be a little selective, and results in conclusions that down played the broad economic context while stressing the criminal justice factors like increased policing. These conclusions are not as univer sally shared as he makes them out to be in the JEP article. Dow's criticism about the limits of econometric testing to deal with broad social context applies just as readily to Levitt as it did the "old" mainstream.

Although Levitt has moved from formalized modeling of narrow self-interested maximization, he maintains the focus on the individual

by abstracting from the social context and, if anything, he places an even stronger emphasis on the ability of econometric testing to reveal the truth. He, therefore, still runs afoul of the feminist/heterodox

critique of the "old" mainstream expressed by critics like Dow, Heil broner, and Milberg. The crucial elements that Levitt retains from the

"old" mainstream create some problems of both emphasis and analysis for his explanation for the drop in crime in the United States. His focus

on a "surprising" explanation, devoid of theory or context, has not

This content downloaded from 206.224.223.249 on Wed, 11 Mar 2020 16:21:38 UTC All use subject to https://about.jstor.org/terms

Steven Levitt on Abortion and Crime 697

been unassailable on its own empirical grounds. His lack of serious consideration of social context in the United States leads him to some

misleading conclusions about the reasons for the drop in crime. His choice to focus on the U.S. decline exclusively, while ignoring the much lower levels of violent crime in other countries, black box the

unique social and economic factors that make the United States such an exceptional case. Levitt's surprising empirical works have gener ated considerable acclaim and substantial book sales but his reliance

on some of the "old" economic mainstream conventions creates

an analysis that is inferior to that of Wacquant. However, the kind of analysis done by Wacquant is very difficult for individually focused, econometrically obsessed researchers. An analysis of crime cannot be torn from context and history.

Notes

1. John DiNardo argues that Levitt's definition of incentives is perhaps too flexible. It is basically a "two part tautology: when incentives matter they matter. When incentives don't matter it is because of moral incentives" (DiNardo 2007: 996).

2. Levitt used the number of firefighters as a proxy for the number of police in an effort to get away from the endogeniety problem that connects the number of police to the amount of crime.

References

Backhouse, R., and S. Duflauf. (2009). "Robbins on Economic Generalizations and Reality in the Light of Modern Econometrics." Económica 76: 873-890.

Barclay, G., and C. Tavares. (2003). International Comparisons of Criminal Justice Statistics 2001. London: Home Office.

Colander, D. (2007). The Making of an Economist Redux. Princeton: Princeton University Press.

DiNardo, J. (2007). "Interesting Questions in Freakonomics." Journal of Economic Literature 45(4): 973-1000.

Donohue, J., and S. Levitt. (2001). "Legalized Abortion and Crime." Quarterly Journal of Economics 116(2): 379—420.

Dow, S. (2002). Economic Methodology: An Inquiry. Oxford: Oxford Univer sity Press.

Dubner, S. (2003). "Economist of Odd Questions: Inside the Astonishingly Curious Mind of Steven D. Levitt." New York Times Magazine August 3: 23-28.

This content downloaded from 206.224.223.249 on Wed, 11 Mar 2020 16:21:38 UTC All use subject to https://about.jstor.org/terms

698 The American Journal of Economics and Sociology

Duggan, M., and S. Levitt. (2002). "Winning Isn't Everything: Corruption in Sumo Wrestling." American Economic Review 92(5): 1594-1005.

Foote, C., and C. Goetz. (2008). "Testing Economic Hypotheses with State Level Data: A Comment on Donohue and Levitt." Quarterly Journal of Economics 123(1): 407-423.

Fryer, R. Jr., and S. Levitt. (2004a). "Understanding the Black-White Test Score Gap in the First Two Years of School." Review of Economics and Statistics 86(2): 447-464. . (2004b). "The Causes and Consequences of Distinctively Black Names." Quarterly Journal of Economics 119(3): 767-805.

Grogger, J. (2006). "An Economic Model of Recent Trends in Violence." In The Crime Drop in America: Revised Edition. Eds. A. Blumstein and J. Wallman, pp. 266-287. New York: Cambridge University Press.

Groseclose. T., S. Levitt, and J. Snyder. (1999). "Comparing Interest Group Scores Across Time and Chambers: Adjusted ADA Scores for the US Congress." American Political Science Review 93(1): 33-50.

Heilbroner, R., and W. Milberg. (1995). The Crisis of Vision in Modern Economic Thought. Cambridge: Cambridge University Press.

Hsieh, C., and M. Pugh. (1993). "Poverty, Income Inequality, and Violent Crime: A Meta-Analysis of Recent Aggregate Data Studies." Criminal Justice Review 18(2): 182-202.

IDEAS. (2011a). "Information About: Steven Levitt." Http://ideas.repec.org/e/ ple59.html. . (2011b). "Information About: Daron Acemoglu." Http://ideas.repec.org/ e/pacl6.html.

Jacob, B., and S. Levitt. (2003). "Rotten Apples: An Investigation of the Prevalence and Predictors of Teacher Cheating." Quarterly Journal of Economics 118(3): 843-877.

Joyce, T. (2009). "A Simple Test of Abortion and Crime." Review of Economics and Statistics 91(1): 112-123.

Katz, L., S. Levitt, and E. Shustorovich. (2003). "Prison Conditions, Capital Punishment, and Deterrence." American Law and Economics Review 5(2): 318-343.

Levitt, S. (1994). "Using Repeat Challengers to Estimate the Effect of Campaign Spending on Election Putcomes in the US House." Journal of Political Economy 102(4): 777-798. . (1997). "Using Electoral Cycles in Police Hiring to Estimate the Effect of Police on Crime." American Economic Review 87(3): 270-290. . (2001). "Alternative Strategies for Identifying the Link Between Unemployment and Crime." Journal of Quantitative Criminology 17(4): 377-390.

. (2002). "Using Electoral Cycles in Police Hiring to Estimate the Effects of Police on Crime: Reply." American Economic Revieiv 92(4): 1244-1250.

This content downloaded from 206.224.223.249 on Wed, 11 Mar 2020 16:21:38 UTC All use subject to https://about.jstor.org/terms

Steven Levitt on Abortion and Crime 699

. (2004a). "An Empirical Analysis of Imprisoning Drug Offenders." Journal of Public Economics 88(9-10): 2043-2066. . (2004b). "Understanding Why Crime Fell in the 1990s: Four Factors That Explain the Decline and Six That Do Not." Journal of Economic Perspectives 18(1): 163-190.

Levitt, S., and S. Dubner. (2005). Freakonomics: A Rogue Economist Explores the Hidden Side of Everything. New York: Harper Collins.

Levitt, S., and J. List. (2008). "Homo Economicus Evolves." Science 319: 909-910.

Levitt, S., and C. Syverson. (2008). "Market Distortions When Agents are Better Informed: The Value of Information in Real Estate Transactions." Review

of Economics and Statistics 90(4): 599-611. Levitt, S., and S. Venkatesh. (2000). "An Economic Analysis of a Drug

Selling Gang's Finances." Quarterly Journal of Economics 115(3): 755-789.

. (2001). "Growing Up in the Projects: The Economic Lives of a Cohort of Men Who Came of Age in Chicago Public Housing." American Economic Review 91(2): 79-84.

Long, H., L. Boggess, and W. Jennings. (2011). "Re-Assessing Publication Productivity Among Academic 'Stars' in Criminology and Criminal Justice." Journal of Criminal Justice Education 22(1): 102-117.

Lott, J., and J. Whitley. (2007). "Abortion and Crime: Unwanted Children and Out of Wedlock Births." Economic Inquiry 45(2): 304-324.

McCrary, J. (2002). "Using Electoral Cycles in Police Hiring to Estimate the Effects of Police on Crime: Comment." American Economic Review 92: 1236-1243.

Poterba, J. (2005). "Steven D. Levitt: 2003 John Bates Clark Medalist." Journal of Economic Perspectives 19(3): 181-198.

Robinson, T. (1977). "What are the Questions?" Journal of Economic Literature 15(4): 1318-1339.

Schumpeter, J. (2006). History of Economic Analysis. New York: Taylor and Francis.

Statistics Canada. (2001). Juristat: Crime Comparisons Between Canada and the United States 21(11) Catalogue no. 85-002-XPE.

The Economist. (2008). "International Bright Young Things." The Economist December 30. Http://www.economist.com/node/12851150?story_id= 12851150.

Uchitelle, L. (2006). "Students are Leaving the Politics Out of Economics." New York Times January 27. Http://www.nytimes.com/2006/01/27/business/ 27econ.html.

Van Kesteren, J., P. Mayhew, and P. Nieuwbeerta. (2000). Criminal Victimi zation in Seventeen Industrialized Countries: Key Findings from the 2000 International Crime Victimization Survey. Http://www.wodc.nl/

This content downloaded from 206.224.223.249 on Wed, 11 Mar 2020 16:21:38 UTC All use subject to https://about.jstor.org/terms

700 The American Journal of Economics and Sociology

onderzoeksdatabase/w00187-criminal-victimization-in-seventeen

industrialized-countries.aspx. Wacquant, L. (2005). "The 'Scholarly Myths' of the New Law and Order Doxa."

In Socialist Register 2006. Eds. L. Panitch and C. Leys, pp. 93-115. London: Merlin Press.

Wallman, J., and A. Blumstein. (2006). "After the Crime Drop." In The Crime Drop in America: Revised Edition. Eds. A. Blumstein and J. Wallman, pp. 319-348. New York: Cambridge University Press.

WHO. (2002). World Report on Violence and Health. Geneva: World Health Organization.

Wolff, M., P. Rutten, and A. Bayers. (1992). Where We Stand: Can America Make it in the Global Race for Wealth, Health and Happiness. New York: Bantam Books.

This content downloaded from 206.224.223.249 on Wed, 11 Mar 2020 16:21:38 UTC All use subject to https://about.jstor.org/terms