All he's done with his analysis is used two thinly disguised proxy variables to hide his true hypothesis, which may very well be a good hypothesis but is unpalatable to discuss in public.
The overcautious avoidance of race as a subject does exist, but is nothing to do with left/right in my experience. I associate it with people who don't live in mixed areas, they tend to be more awkward about mentioning race and certainly aren't more left wing on average.
For violent crime, there is a direct correlation with black Londoners.
"In London in 2006, 75% of the victims of gun crime and 79% of the suspects were "from the African/Caribbean community."
https://en.wikipedia.org/wiki/Race_and_crime_in_the_United_K...
This does not solely reflect lower socioeconomic standing as Pakistani + Bangladeshi communities have higher levels of poverty, and similarly poor areas of cities in Wales and Scotland do not have the same proportionate level of gun crime that occurs in London.