That's not how it works is it? You can't propose and enact policies with no proof they work and then just say 'well you can't prove they dont'. Ok, my new policy is we should arrest you and your family because it will reduce crime. I have no evidence it will work, but you don't have evidence it won't work.
Having said that, there have been studies done on stop and search in London and New York and there is no evidence it has any effect. Google the BLUNT 2 data from early last decade. Exceptionally little to no effect (and lasting in the region of three weeks if there was a fraction of a percentage changed).
Or here
https://www.crimeandjustice.org.uk/news/stop-and-search-has-no-impact-crime-new-report
"We found little evidence of an effect from stop and search on violent crime. The only statistically significant result was the net effect from weapons searches at the weekly level, and the effect here was very small indeed; a 10% increase in stop/search led to a 0.01% decrease in non-domestic violent crime. Similarly we found very little effect as a result of stop/search on weapon enabled non-domestic violent crime, and no effect whatsoever when it came to ambulance incident data for calls related to stab/shot/weapon wounds"
So to have any sort of even marginal effect you would need massive increases in stop and search, many times over what was being done here and New York at the height of its policy. Just to get a half a percent or so effect. And no evidence of even that when it comes to knife crime specifically. All with the increase in alienation and distrust of local communities. Never mind the social effect and politics of it, it's a shockingly crude, wasteful and ineffective policy.