To be fair, "Of those, over half (55%) had lived in San Francisco for 10 or more years" is much harder to game, though just as easy to lie about.

I'm more worried about this thought in the article:

> While cities report inconsistently, San Francisco currently spends around $430 per city resident per year on services and programs for the homeless, compared to $260 in New York and $110 in Los Angeles. Yet the problem in our city is worse, and despite increases in spending, has continued to worsen.

It's not a surprise that the city with more spending on services for the homeless has more homeless. That's exactly what you'd expect under pretty much any model.

Do you believe that subsidizing the homeless means you'll get more homeless? Then higher spending will directly cause more homelessness.

Do you believe that places with bigger problems will try harder to do something about them? Then higher spending will reflect more homelessness.

I wonder what the numbers look like per _homeless_ person in the city.

In a country with open borders between states and extremely cheap transportation, homelessness really should be thought of as, at least partially, a federal problem. And funded accordingly.

> [1] https://www.sfchronicle.com/bayarea/article/S-F-spends-recor...

i laughed when I read that title. "Tracking results" is definitely a form of overhead, so the article is literally complaining that charities are ineffective because they don't have enough overhead.

In general, "minimize overhead" isn't equal to "maximize impact per dollar". The obvious software industry exemplar is the oft-repeated complaint about companies hiring expensive engineers and then refusing to buy a $200 monitor.

Applying the Commerce Clause to vagrancy is darkly funny.

The commerce cause is mostly orthogonal (not irrelevant, but orthogonal) to no-strings-attached social services funding, even under very literal/originalist interpretations.