3 ms·
This, I think, is the core issue. The typical theft before the changes could have been of trivially low values. As soon as people know that the strongest negati
by gota 5y ago
This, I think, is the core issue. The typical theft before the changes could have been of trivially low values. As soon as people know that the strongest negative incentives only kick in after a total value (e.g. 1000 USD) they may start to "optimize" for that.
I mean - if a person would shoplift item A; they may reason that there is very little additional negative incentive to not steal B, C...as well - to get closer to that threshold.
This makes sense even if you assume that thefts of total value under the threshold are being always reported (as commented elsewhere here). Lower numbers of instances, higher value of theft.
Then again, maybe there are underlying mechanisms in the motivations of shoplifting that make this completely wrong. I'd accept and appreciate the input of someone with expertise in this subject