Rapid Rehousing Data Analysis II (Requiring Income)

In 2009, when the HPRP (Homelessness Prevention and Rapid Rehousing Program) program began, homeless services providers and policy makers understandably began to try to nail down what the best target population was for Rapid Rehousing services so the money could be spent as efficiently as possible. By 2012, there were many studies available detailing what was learned throughout the period of time the idea of Rapid Rehousing was really being implemented, whether correctly or incorrectly. While there was plenty of information about HPRP and its potential use cases, there was a lot that was unknown. HPRP providers and CoC's were having to build the plane in the air, as the saying goes.

One of the assumptions made by many providers and CoCs was that clients with no income would not do well in Rapid Rehousing. Some users reported a misunderstanding that HUD had said RRH clients needed to have an income at Entry. It may have been a CoC-wide rule, or an agency-wide rule, or it could have just been the way a particular case manager or project manager had decided made sense. The logic went that if a client didn't have an income, then they wouldn't be able to eventually take on the rent payments as expected. However, as I stated in the last blog post on Rapid Rehousing, the whole premise of Rapid Rehousing is that housing is the foundation upon which the client can build or rebuild their lives. So the idea, then, of a client being "housing ready" is really obsolete and means that there is no criteria that should be required for a client to be served with Rapid Rehousing, and that includes income. This is supported by the many studies done around that time period about Rapid Rehousing (and Prevention, but that's not what we're talking about here.)

You may be left wondering, so if the data didn't show that income was a factor in predicting success in RRH, then what are some actual factors in predicting success? Here's HUD's guidance on the matter:

Target Populations

Rapid re-housing is an effective intervention for many different types of households experiencing homelessness, including those with no income, with disabilities, and with poor rental history. The majority of households experiencing homelessness are good candidates for rapid re-housing. The only exceptions are households that can exit homelessness with little or no assistance, those who experience chronic homelessness and who need permanent supportive housing, and households who are seeking a therapeutic residential environment, including those recovering from addiction.

So this would suggest that perhaps those families and singles with income and low barriers are actually NOT a good fit for Rapid Rehousing, because they could likely solve their homelessness on their own, or with very little help. And those with a great number of barriers may be best served in Permanent Supportive Housing, leaving a "majority of households" in the sweet spot middle.

Further, from this 2011 study in Hennepin County, MN, the University of Minnesota found that the following were risk factors for families exiting a Rapid-Rehousing-esque program there:

Families who were identified as chronically homeless were 2.3 times more likely to return to shelter.

Risk Factors:

Large family

Chemical dependency in last 5 years

Personal care attendant in last 5 years

Black or Native American

Arrest within last 5 years

Non-Predictive Factors:

Educational status

During the year prior to entering shelter, 23% of the families changed addresses twice, and 16% changed addresses 3 or more times. In the year following shelter, 18% changed addresses twice, and 15% changed 3 or more times.

You can see from this that income is not a predictor of Rapid Rehousing success. And in fact, according to yet another study, it's the other way around: Rapid Rehousing is a predictor of greater income success!

A Washington State Department of Commerce study found that, when creating a matched comparison group, working-age adults who received rapid re-housing were 50 percent more likely to be employed during the quarter they received assistance. Additionally, they were 25 percent more likely to be employed over the following year and earned about $422 more than their peers who were not rapidly re-housed.

Have I nailed this one down yet? Is there anyone still wondering about income being a requirement for entry into Rapid Rehousing? If you are, click around some of those links to the studies above and then come back. If you're not, then let's look at the way we are looking at this in our CoC.

It has been our CoC's goal to find in the data where there are providers who are only serving or favoring clients with an income, so that we can provide further guidance to that agency about targeting their Rapid Rehousing dollars better.

Instead of creating a single report about it, we actually added it to our Project Evaluation report that helps us rank projects for funding every year. It is one of two measures we use to determine how well a provider is targeting the "hard to serve". (The other one is the percentage of households they serve that are unsheltered or in emergency shelters as opposed to clients from institutions.)

Shows a report titled "Targeting Hard to Serve" with a subheading of "Clients with No Income at Entry", then a date range. Shows a redacted provider's numbers who barely missed the goal of 40% of clients having no income at Entry. Begins to show the detail portion of the report with redacted client IDs, Entry and Exit Dates, plus the "Income at Entry" question's answers to for each client's Entry.

This is a measure that we use on all of our project types, not just RRH, but it originated from the need to discourage especially the RRH projects from targeting only clients with an income for their Rapid Rehousing dollars. We generally are not having this problem that much anymore because we have been doing Rapid Rehousing trainings around the state and providers are beginning to come around to see the data and are hitting the right targets.

This report and another similar one have been the subject of many a blog post here, so if you're interested in coding something similar, you can visit the Proximity to Goal post and CoC Performance Measurements post. I am betting that most folks are not going to be creating super ambitious ART reports these days given the changes coming, but if you are looking to add things to your existing reports or just curious, you should take a look. :)

Next blog I will talk about How Rapid is Rapid? How do you tell how quickly homeless clients are being housed from the date they are being identified?