"Changes in incarceration and crime are significantly related during the period under consideration. Increases in state prison committals per 100,000 residents tend to reduce crime the following year, whereas increases in the number of persons released from state prisons per 100,000 reidents tend to increase crime the next year."

The National Research Council's "Understanding Crime Trends Workshop Report," eds. Goldberger, A. and Rosenfield, R.

It's tricky to isolate other social trends in the research though. Steven Levitt did a major study of prison overcrowding legislation, which can suddenly change incarceration rates if the legislation is successful, but not if it fails, a bit of a coinflip not directly connected to social or demographic shifts. He found evidence of a link under these conditions.

Levitt, S. The effect of prison population size on crime rates: evidence from prison overcrowding litigation.

Obviously those aren't the only two studies on the subject, it's been a focus of criminology research since it's been a discipline. Levitt followed up with a good meta-analysis that covered ten different explanations for the reduction of crime over the last few decades:

Levitt, S. Understanding Why Crime Fell in the 1990s: Four Factors that Explain the Decline and Six that Do Not. Journal of Economic Perspectives, Winter 2004. pp 163-190.

He cites John DiIulio (ie, DiiuLio) in that work. DiIulio has written a lot on the subject as well, usually arguing that incarceration rates have an impact on crime rates, but adding the nuance that three strikes laws rarely do much good.

Crime rates are very complex, and tied to many complex factors. Incarceration rate appears to be one such factor.

He means people born in those cities were leaving?