It's addressed on the second paragraph of page 5.

While the rest of the country, over the same twenty-year period, saw the growth in incarceration that led to our current astonishing numbers, New York, despite the Rockefeller drug laws, saw a marked decrease in its number of inmates. “New York City, in the midst of a dramatic reduction in crime, is locking up a much smaller number of people, and particularly of young people, than it was at the height of the crime wave,” Zimring observes. Whatever happened to make street crime fall, it had nothing to do with putting more men in prison.

NYC is not insulated from the rest of the country, which did lock up dramatically more people. As neighboring cities and states imprisoned large numbers of people, it would have an effect on NYC street crime.

That's the kind of Derridean approach to Criminal Justice that got us to where we are now. You're saying the same thing, though: a cities crime rate is entirely dependent on the policies of surrounding cities and not on itself. Think about what you're suggesting: even if NYC experienced a crime drop while also housing fewer inmates, it's only because their own policy was ineffectual and the surrounding cities picked up the slack (or something).

Indeed, many policies work locally but not globally. Harassing criminals and putting them on busses leaving town appears to work, but only locally.

Anyway, the story seems to say that since there was a period of time in which NYC crime decreased while NYC incarcerations also decreased, the theory that incarceration reduces crime is refuted. Hardly.

I live in a city where people are said to commute in from the outer parts in order to commit crimes, but the assertion is only ever backed up by anecdata.

Yes, I've heard this as well, and it is absolutely terrible.

I am told that they are international in scope and go under the gang-sign of 'Bankers'.

Or perhaps Steve Levitt is right that the drop in crime was due to the legalization of abortion (as laid out in the book Freakanomics). I personally don't like either explanation, but it just goes to point out the challenges in getting from correlation to causation.

He explorers and debunks that very idea in the article.