What Government Does

By DAVID BROOKS

We in the commentariat spend a lot of time reporting on the White House and the U.S. Capitol, but relatively little time reporting on the dozens of federal buildings around them where public policy actually gets executed. We engage in debates about the size of government but spend little time directly observing what government is and isn’t good at.

To help compensate, I spent some time this week at the Department of Housing and Urban Development. It is headquartered in a dispiriting 1960s building that is brutalist on the outside and dreary within — 10 stories of basement, as the employees say.

The secretary, Shaun Donovan, was trained as an engineer and is a numbers guy. He learned from his experience as New York City’s housing commissioner that you can’t fight social problems like crime and homelessness unless you have good data. So he helped create a program called HUDStat, which tracks homelessness among veterans and the results of the various efforts to combat it.

I observed a strategy meeting led by Donovan and Scott Gould, the deputy secretary of the Veterans Administration, with about 30 career personnel and political appointees. The purpose of the meeting was to see which regions were doing a good job of getting the veterans treatment and housing vouchers, and which weren’t. (Democrats seem to feel comfortable using vouchers to address housing problems but not education and health care problems.)

The career workers at the meeting were impressive. They made short, highly informed presentations and answered arcane questions about legislative history. They had achieved a herculean task of getting two government agencies to agree on a single data set, a single methodology and a single progress report.

Homelessness touches many government services, from housing to education, and the federal workers presented complicated flowcharts trying to organize overlapping programs into one coherent system. How do you set up services for a homeless female veteran who has a drug addiction, psychiatric problems and is a victim of domestic violence? If a federal agency issues housing vouchers, how should it alert the local housing authority that more residents are on the way?

The HUDStat report is blunt about which state and local departments are efficiently moving veterans into housing (Indiana and Ohio) and which is lagging behind (California).

The career people treated the political people with almost military deference. The career people often spoke about managing the organizational structures and establishing clear rules for case-workers; Donovan and Gould spoke more about the experience of the veterans on the street and probed for ways to move everything faster:

Can we use money from other voucher programs to get the veterans security deposits? Probably not, the accounting issues are too complex. How long does it take between a homeless veteran’s first contact and actually moving in? The norm is 127 days. Can we reduce that wait time to 10 days? No, but maybe 90 is possible. Even though the 2011 budget was passed late, can we use some of that money quickly so legislators will be pleased when budget time comes up again in 2012? Yes.

Unlike some political appointees, Donovan and Gould are deeply involved in the intricacies and are powerfully driving policy. Many government efforts are designed to minimize failure and avert scandal. In this program, each region has a clear numeric definition of success. There are clear standards for how quickly veteran homelessness should be reduced year by year. So far, the program is surpassing its targets by 46 percent.

The big question I had was this: How large is the gap between the neatness of data on a bar chart and the messy reality on the street?

For example, 75 percent of veterans in the program have psychiatric, drug or alcohol problems. Under the old policy, social workers tried to get the veterans treated first and offered free housing as an inducement. Now the Housing First approach prevails: Get them the stability of an apartment, then treat their drinking, drug and mental issues.

That produces good homelessness data, but are ill and addicted veterans off their meds and menacing apartment buildings? Does the approach work as well for the severely ill? Does it work as well in sparsely served areas?

Donovan notes that research supports the counterintuitive Housing First approach. Some studies do, indeed, show modest benefits, but I was struck by the vast difference between the way a government sees the world — numerically and organizationally — and the gritty and unpredictable way the world sometimes looks to, say, a crime reporter or a homeless veteran himself.

Over all, visiting HUD was  tremendously useful. Amid the hot-rhetoric government wars, it was important to see the talent and commitment of real-life government workers running a successful program — and to see the limitations inherent in government planning.