The Great American Decline

By Franklin E. Zimring
Oxford University Press, New York 2007

There is much to be learned from Frank Zimring’s highly readable and penetrating examination of the U.S drop in crime in the 1990s.  Zimring does a masterful job of illustrating how when crime was rising in the 1970s the message that nothing can effectively prevent it was widely accepted.  Skip ahead to the period following the 1990s crime drop and every policy from more and better policing to greater reliance on incarceration to more gun control – or even more guns – has some champion for the view that they are particularly effective crime fighters.  As is often the case, the public commentary tends to be simplistic and misleading, but my own view is that the rough contours of why crime fell are clearer than Zimring believes.

Zimring begins by showing the crime drop in the U.S. was broad and sustained:  in the 1990s, five of seven index crimes fell by around 40 percent, and the remaining two dropped by more than 20 percent.  Zimring also shows that Canada had drops in five of seven index crimes that were 70 percent or more of the American decline (p.126).  Zimring documents that homicide trends in Canada and the U.S. tend to move in lockstep, even though Canadian rates are much lower.  For the 1990s, when the U.S. homicide rate fell by 39 percent, Canada was not far behind with a 34 percent drop. 

A major theme of his book is that the Canadian experience can be used to undermine the importance of a number of factors that have been offered as major explanations of the U.S. crime drop.  For example, while many have argued that the booming economy and the robust growth in police and prison inmates explain the U.S. crime drop, Zimring contends that none of these factors were present in Canada in the 1990s.  In fact, Zimring asserts that the only common explanatory factor shared by the two countries was the roughly 25 percent drop between 1980 and 2000 in the proportion of the high-crime risk 18-29 year old age cohort.  This drives Zimring to conclude that “the same lack of known causes that holds for the majority of Canadian crime declines may hold for 40 percent of the U.S. crime drop.” (p.134.)  Zimring then attributes the large unexplained portion of the U.S. and Canadian crime drops to “cyclical effects [that cannot] be easily assigned to discernable [sic] causes” (p.209). 

I read the data somewhat differently.  Clearly, demographics were important in both countries, but I think the U.S. and Canada shared some other identifiable similarities in the 1990s. First, contrary to what Zimring argues, Canada did have a strongly improving economy (albeit starting from a worse position) in the 1990s.  Canadian unemployment fell 4.6 percentage points from 10.7 percent in 1992 to 6.1 percent in 2000 (while the U.S. unemployment rate fell 3.5 percentage points from 7.5 percent to 4.0 percent over this same period).¹ While Canada didn’t have the prison explosion or large growth in the number of police that were observed in the U.S., it did get similar boosts from the economy and demographics.  If one accepts the estimate that 40 percent of the American crime drop came from police and more prisoners, those factors may explain why the U.S. did somewhat better in reducing crime in the 1990s than Canada did.² 

Second, while Zimring de-emphasizes the influence of crack, I suspect that the reversal of the late 1980s crack-induced crime boom is an important explanation of the falling U.S. and Canadian urban crime rates in the 1990s.  Zimring rightly notes that crack consumption did not fall sharply, but this leads him to downplay the crack explanation, while it leads me to believe that the important behavioral change in reducing crime came on the part of crack suppliers, not consumers.

Third, if legalization of abortion can explain the remaining portion of the 1990s crime drop in both countries, then the story is complete.³  Introducing crack and abortion legalization into the story of the Canadian and U.S. crime declines has two advantages.  Their combined effect would resolve the unexplained crime drop, and unlike the other factors mentioned above – demographics, the economy, police, and incarceration -- which were gradually and persistently working to dampen crime, the crack reversal and abortion legalization are more plausible catalysts for the abrupt crime shift of the early 1990s (with Canada moving towards abortion legalization in 1969, and the U.S. starting down the path in 1970 before going the entire way in 1973). 

While Zimring largely opts for a continuing mystery, I think the four common factors in the U.S. and Canada and the two added elements for the U.S. alone pretty much tell the story.  There is still some dispute over the relative shares:  Levitt (2004) concludes that the crack and abortion factors account for roughly one-third of the total crime drop in the U.S., and Sen (2007) argues that Canadian abortion legalization alone explained roughly 40 percent of the smaller Canadian decline.⁴  But if you put the entire package together, it would seem to explain both the late 1980s crack-led crime surge and the subsequent sharp crime declines, as well as the greater drops in the U.S. (which had the advantage of the police and prison increases missing in Canada).  Four factors pushing in favor of lower crime were enough to lead to large drops in Canada (demographics, the economy, the end of the crack crime epidemic, and legalized abortion); two more factors pressing in the same direction in the U.S. caused an even greater decline (more police and more prisoners).

¹ Bureau of Labor Statistics, “Comparative Civilian Labor Force Statistics, 10 Countries, 1960-2004.” Updated 13 May 2005. <www.bls.gov/fls/flslforc.pdf> (cited 22 May 2007).

² Levitt, Steven. 2004.  “Understanding Why Crime Fell in the 1990s.” Journal of Economic Perspectives. 18:1, pp. 163-190.

³ Donohue, John and Steven Levitt. 2001. “The Impact of Legalized Abortion on Crime.” The Quarterly Journal of Economics 116:379–420.  One intriguing fact:  the country that enjoyed by far the greatest percentage drop in homicide in the 1990s, dwarfing that in the U.S. or Canada, was Italy, whose murder rate fell by more than 50 percent.  This made it interesting to learn from Zimring that Italy was the sole European nation to move from complete prohibition to full abortion legalization in the 1970s.  Indeed, after legalization in 1978, Italy’s increase in abortions was so great that it briefly had the highest abortion rate of the five European countries that Zimring discusses (p. 213).

⁴  Levitt, supra note 2.  Anindya Sen, “Does Increased Abortion Lead to Lower Crime?  Evaluating the Relationship Between Crime, Abortion, and Fertility,” B.E. Journal of Economic Analysis & Policy (2007).