Looks like someone did the statistical analysis that answers at least how soon they are likely to get recaptured (correlation not causation) at https://www.researchgate.net/profile/Bryce-Peterson/publication/377596204_Catch_'em_if_you_can_examining_how_often_and_how_quickly_people_who_escape_from_prisons_and_jails_are_recaptured/links/65b152979ce29c458baa3402/Catch-em-if-you-can-examining-how-often-and-how-quickly-people-who-escape-from-prisons-and-jails-are-recaptured.pdf

So going off their most comprehensive Model 4, violent vs non-violent offence does not make a statistically significant difference.

Factors predicting a statistically significant shorter time to recapture:

• Being white-skinned (i.e. for once there is a reverse-white privilege here, not sure why).
• Shorter sentence length (i.e. being in for less time actually increases the chances of eventual recapture).
• There being evidence the escape was planned (i.e. planned escapes are less likely to succeed than opportunistic ones - or maybe the good planned ones leave no evidence).
• The escape happening in summer (maybe longer daylight hours?).
• The escape happening from somewhere other than a non-secure / minimum security facility, as opposed to a secure facility or a secure or non-secure outside location like work release.
• The escape happening in a larger group (i.e. those who escape in groups are more likely to be recaptured).
• The escape happening other than from a local jail (i.e. local jail does predict a longer time to recapture).
• The escape happening in the Western region.

So I think to answer your question from that, a short sentence length for a non-violent crime doesn't reduce the time to recapture.