The current issue of the Journal of Law and Economics has this article by Rasmus Landersø.  Here is the abstract:

This paper studies how longer incarceration spells affect offenders' labor market outcomes by using a reform that increased incarceration lengths by approximately 1 month. I use detailed register data for offenders who predominantly serve incarceration spells of 1-2 months. I analyze the sample for several years prior to and after incarceration and show that the reform led to an exogenous increase in incarceration lengths. I find that the longer incarceration spells result in lower unemployment rates and higher earnings, possibly because marginal increases in short incarceration spells improve conditions and incentives for rehabilitation, while the costs of jail related to these outcomes are unaffected. I show that the estimates are robust to different econometric specifications and further provide evidence that my results are not driven by changes in macroeconomic conditions.

I am skeptical about how that would generalize to different conditions, but it's an interesting result.