[nostrademons]

It would be interesting to get stats on homeless churn - the proportion of the homeless people who were homeless 3 months ago, or who will be homeless in 3 months - and then qualitative data on who those people are and why they're homeless.

The aphorism I've heard is that "People become homeless for all sorts of reasons. People stay homeless for just one: mental illness."

Note that it's not contradictory for a majority of the homeless (by the numbers) to have jobs and be homeless for reasons beyond their control, and yet for you (as someone who works with the homeless community) to see a majority of them as suffering from mental illness or drug use. This would happen if there's a large margin of people living on the edge: they're making it work one month, then they lose their job or their rent goes up and they're homeless, then they get another job or find a cheaper place and they're no longer homeless. From your perspective as a provider of services, these are transient folks who come in once and then you never see them again, and you wouldn't remember them. From the statistics on homelessness, they could very well make up a much larger fraction of the homeless than the chronic mental illness & drug abuse cases, particularly in times of great economic change like now.