I'm seeing some dodgy statistics in the comments. The actual number of homeless in a given area is difficult to even estimate, those who compile and track such stats usually include a codicils about that. Like unemployment stats, which really only track the number of people who have applied for unemployment, the methodology is known to be imperfect. I believe homeless counts follow the number of people staying the night at shelters. Thing is, it's estimated that at least 50 percent of homeless folk never, ever, use a shelter. In homeless circles, shelters are perceived as more dangerous than sleeping in the streets. Lots of predators can be found at shelters, the shelter management simply lack the resources to keep them safe. So doubling the "official" count is a more accurate estimate. I was quite surprised to learn of the number of homeless where I previously lived. It's a growing problem here in the US.