The probability of being caught deters far more reliably than the severity of the punishment that follows.
Beccaria argued it in 1764 and modern deterrence research has largely borne it out: increases in the perceived likelihood of apprehension produce measurable reductions in offending, while increases in sentence length produce small and often undetectable ones. Severity is the cheaper lever politically and the weaker one empirically, which is most of the story of sentencing policy.
In practice
- A policy that doubles sentences and leaves clearance rates unchanged has bought very little deterrence.
- An expenses regime with a light penalty and a real chance of audit works better than a severe one nobody expects to be applied.