Weekend Long Reads: Philadelphia's Basic Systems Repair Program

by Kevin Schofield

This weekend's "long read" is a study on Philadelphia's Basic Systems Repair Program — and the surprising impact that it had.

The Basic Systems Repair Program is a grant program run by the City that makes awards of up to $20,000 to low-income homeowners for structural repairs of electrical, plumbing, heating, and roof damage to their homes. To enroll in the program, homeowners must apply, meet the income qualifications (the same as for the U.S. Department of Housing and Urban Development's (HUD) Section 8 vouchers), and then be placed on a waiting list — currently for up to three years.

Between January 1, 2006, and April 30, 2013, the period of the study, 13,632 homes received grants from the program. As a basic housing and public-health intervention, that alone is a significant accomplishment. But what the study found was even more amazing: When a house on a blockface (i.e., one side of a city block) was repaired using a program grant, it resulted in a modest but measurable drop in violent crime there.

This is the kind of study where skeptics (like me) wag their finger and exclaim, "You're confusing causation with correlation!" That is, did the house repair cause the drop in crime? Or did a drop in crime lead to a situation where homeowners decided to apply for a repair grant? Or did both happen because of some third event? Trying to design a study that establishes causation is much harder than it seems.

One technique, and the one applied in this study, is called "difference in differences." Crime is not a static statistic; it's always moving up or down based upon a host of factors. If you simply looked at the change in crime rate on a block where a house was repaired via a grant, you wouldn't be able to tell how much of the change was attributable to the house repair versus all the other confounding factors. A "difference in difference" study like this one picks a comparable set of blocks that didn't have a house repair, and compares the change in crime rate over time there with the change on the block with the repaired house. That, at least in theory, eliminates the other factors. But it's not a perfect strategy, because it doesn't control for everything: Is there something special about the blocks where homeowners applied for a repair grant compared to all the other blocks? So the study designers narrowed it down even further: They compared the blocks where a house was repaired with the blocks where there is a house on the waiting list for a repair grant. That cleverly eliminates everything except the actual house repair, including the homeowners themselves.

The study went further and found that when more than one house on a blockface was repaired, there was a further (but smaller) decline in crime, up to about 5 or 6 homes. That adds further credence to the belief that there is a cause-and-effect relationship between a house repair and a drop in crime.

The research report has an interesting discussion about why repairing homes might lead to lower crime rates, tying in some of the concepts from the decades-long debate over "broken windows" theory.

There are limits to the study: It's one city in one region of the country so there is no guarantee that the results would be the same in other cities. But it's a tantalizing — and fairly inexpensive — intervention with two big benefits.

Association Between Structural Housing Repairs for Low-Income Homeowners and Neighborhood Crime

Featured image is attributed to F Delventhal (under a Creative Commons, CC BY 2.0 license).