Group government

Steve Goldsmith, the innovative mayor of Indianapolis from 1992 to 1999, spends two days a week teaching government at Harvard and the rest of the week in Washington. He headed up the domestic policy team during George W. Bush's 1999-2000 campaign but has not been involved with the White House in any major way since then.

Mr. Goldsmith is co-author (with William D. Eggers) of Governing by Network: The New Shape of the Public Sector (Brookings, 2004). The essential argument is that government officials need to move away from command-and-control production by their own employees in their own buildings, and instead emphasize delivery of services by networked private and nonprofit groups operating on government contracts.

WORLD: Why not move directly to citizen control by use of vouchers and tax credits?

GOLDSMITH: Networks and vouchers are not mutually exclusive goals. Clearly vouchers empower choice, and they are worthwhile and important. But there needs to be a broad transformation in how we think about government. In education, for example, vouchers have an important effect on children's opportunities, but charter schools are also better than government schools.

WORLD: Can the parts of a successful network-say, a religious and a secular program-have entirely different methodologies but the same public-policy goal?

GOLDSMITH: In a welfare-to work-program I could easily imagine a young woman with a drug problem enrolled in both Teen Challenge and a workforce training program. Both have the goal of equipping her to be a more fulfilled human being with confidence in herself and therefore the ability to get a job.

WORLD: What about when both government and church want to fight teen pregnancy, but the church group has the goal of helping a young woman gain faith in God, in the belief that a teen thus transformed is less likely to get pregnant. Is that a problem for government, as long as success for the program is measured not by conversions but by few pregnancies?

GOLDSMITH: That's a good example of how folks with disparate approaches can create a blended message. Churches would preach that premarital sex is wrong . . . government officials could recognize the public good in that. It all depends on government's willingness to acknowledge a vital role for religion in producing a good for the community.

WORLD: How far can such a partnership go?

GOLDSMITH: It's easy and traditional for government and faith leaders to work together on housing when the issue is merely housing. The harder question is how the faith-based provider chooses who gets the house and what kind of counsel to offer. It's easy when a shelter is just a hotel. It's harder when at the shelter someone is trying to transform lives.

WORLD: What have been the successes and failures of the faith-based initiative?

GOLDSMITH: The president's leadership has mainstreamed and legitimized the faith-based solution more than before but also provoked an outcry against it. We have a long way to go at every level of government before people understand the transformational nature of religion. There remains a residual suspicion on the part of the professionals that religious solutions aren't at their level of "professionalism."

WORLD: Is your networking approach an alternative to decentralization through vouchers and tax credits?

GOLDSMITH: Vouchers and tax credits are always preferable, and networks should never take the place of those.

WORLD: How does the network model affect spending levels?

GOLDSMITH: They are independent issues: Networking will give you more per dollar spent, but the number of dollars may stay the same or even go up.

WORLD: Why, then, emphasize networks?

GOLDSMITH: This is a way of thinking about the changes that is a little less threatening. . . . On education, we can help officials to think not, "How do I provide government schools?" but, "My job is to make sure young adults are educated." The official needs to ask, "Why do I want to run a bus system? Why do I want to run a custodial service?" Most are a long way from that level of understanding.