My main point is that we can pick whatever biased metric we like, which you demonstrate so perfectly by cherry picking your date, and you choose some unknown “crime rate” figure which depends on your assumptions which you haven’t stated.
Here’s a graph which shows homicide rates vary, but there definitely is no trend that demonstrates a decline of half: https://www.pewresearch.org/wp-content/uploads/2019/08/FT_19...
The usual biased opinions (regurgitated without thought) choose some figure based on “serious crime”, but as a community we also care about lesser crimes, like bikes getting stolen (California). The easiest way to reduce “crime” is simply to stop enforcing it and the number of convictions will drop.