It seems like you maybe thinking this is saying police do nothing, it isn't.

No consistent association means the data doesn't back up higher or lower funding having an impact on crime. It doesn't say anything about rates when the funding is zero or when funding is very high.

I think it means can't pay to reduce crime, or not pay and expect crime to go up.

Testing for zero would be extremely difficult, because we only have one Toronto sized city in Canada.

I'm guessing here but I suspect that there's a significant number of places with zero police presence that have very little crime. And this article suggests that there are very well funded police presences where crime still happens.