Part of the mission at Scioto Analysis is to improve public policymaking. Although we always strive for the best possible outcomes, we understand the limitations of our kind of work and we only promise better decisions, not necessarily the best decisions. A theoretically perfect policy would have all benefits at zero cost. As economists like to say, there’s no such thing as a free lunch.
I was thinking about this philosophy recently as I read a new working paper by a team of economists from Stanford. They looked at the impact that homeless shelters had on the population, and they found that shelters were not significantly more effective than other interventions at reducing homelessness.
While shelters may not be effective at reducing homelessness relative to other (presumably less costly) interventions, the researchers did find that there were a wide range of other benefits both for people receiving shelter and the community at large.
One of the main externalities the researchers found was a reduction in crime, particularly violent crime in areas with shelters. Much of this reduction was observed among crimes that happen during the evening when shelters are operating.
Another externality they found was a reduction in the number of emergency room visits. We know that people experiencing homelessness utilize emergency medical services at higher rates compared to the population as a whole, which can lead to negative consequences for others due to overcrowding.
The most significant benefit that accrues to people receiving shelter is a reduction in mortality. We’ve talked before about the value of statistical life, and anyone who is familiar with it knows that there is an extremely high willingness to pay for risk of death reduction in our society. Any program that reduces mortality is likely to have a significant return on investment.
If we only evaluate homeless shelters on their ability to reduce homelessness long-term, we might think that they are not worth it for the public sector. The authors mention that street outreach is about as effective at reducing homelessness, and that doesn’t require nearly as many resources.
However, this isn’t the whole story. Shelters do more than just reduce homelessness that we as a society have a high willingness to pay for. We generally like living in cities with less crime, where the emergency room has openings for people who need it, and where people don’t die prematurely. If we only focus on one outcome, we miss the bigger picture about what shelters can provide for us.
I think this provides a clear example of why Scioto Analysis is focused on better decision making rather than perfect decision making. If your goal is to end homelessness in your city, then shelters might not be the best approach. The evidence suggests that in the long run they will not solve the problem.
But they can make it better. What shelters seem to be is a way to treat the symptoms of homelessness, not its causes. But they are very effective at treating these symptoms.
Policy analysts and policymakers can keep looking for solutions that are even better than shelters for addressing homelessness, but right now it appears to be one of the most efficient tools we have.