The Vox link you previously posted shows the large spike (+~30%) being 2020 related.

The 21.8/100k you cite for St Louis in 2003 was also abnormally and suspiciously low, as 2002 and 2004 were in the 30s/100k. I wouldn't baseline anything off a single local outlier that's 50% off from other years. Perhaps a statistical / data collection anomaly? You can see in 2004 the rate was about 35/100k, while in 2013 it was 38/100k, a scant change.

You can see that data in context within a chart in this NYT article from 2015 [2] which discusses some possible factors in St Louis' persistently high crime. 2014 and onwards is likely skewed considerably by the Ferguson unrest, and it sounds like St Louis might still be dealing with that alongside other factors.

Outside the local context, none of this data at a macro scale looks like the trend we see in the 70s-80s that the lead-crime hypothesis seeks to describe though. Crime nationwide has indeed been growing slightly since 2015... to hypothesize about why we're seeing this, I'd guess it's largely due to loss of opportunity and rising income inequality leading to poverty and desperation, which our recent economic policies have only been worsening. But I'm absolutely no expert in the subject, just trying to figure things out like you are.

[1] https://www.macrotrends.net/countries/USA/united-states/murd...

[2] https://www.nytimes.com/2015/02/11/us/st-louis-puzzles-over-...