Last week in the New York Times, David Brooks compared the anti-poverty programs of Barack Obama and John Edwards. Senator Obama’s plan focuses on making impoverished places more successful with funding for public transportation and community centers while Mr. Edwards wants to give housing vouchers directly to a million people. Suddenly, the press coverage of the Democratic primaries rose above puerile posturing about Hillary Clinton’s neckline into a serious debate about anti-poverty policy. Should the government focus on fixing poor places or should it provide poor people with the resources to leave those communities?

Mr. Obama’s poverty proposal includes some person-based policies, like expanding the Earned Income Tax Credit, but he also is awfully fond of place-based strategies, like Community Development Block Grants and a “promise neighborhoods” program, which uses national resources to develop 20 impoverished areas. Mr. Edwards is a bigger fan of person-based strategies like housing vouchers and temporary jobs. I am no supporter of Mr. Edwards, but he is right to focus more on helping poor people than poor places.

Mr. Brooks, however, disagrees. He claims that vouchers are ineffective and lauds Mr. Obama’s “more developed view of social capital.” I think that means that Mr. Brooks likes the Harlem Children’s Zone that is the model for Mr. Obama’s “promise neighborhoods” program. I like HCZ too, but its success says as much about the government’s ability to build communities as Google’s success says about the government’s ability to develop Internet search engines.

The Harlem Children’s Zone is an entrepreneurial nonprofit that receives most of its money from private donors. Saying that the federal government is going to fight poverty by developing its own HCZs is like saying that the government is going to increase GDP growth by starting its own Microsofts.

If we want to understand what future federal place-based policies will look like, we should turn our eyes from private nonprofits to past governmental forays into place-making like urban renewal and the Model Cities program. The track record of these place-based programs gives us plenty of reason to be skeptical about this approach.

Place-based strategies often involve massive infrastructure spending that invites waste and fraud. In cities that are poorer than New York, place-based strategies must fight economic forces that will overwhelm any federal project. The Great Lakes were once a great place to make cars and now they aren’t. Federal aid for the Motor City can’t change that.

Moreover, by subsidizing impoverished areas the government essentially is bribing people to live in economically unproductive areas. Even when these policies do make a place more attractive, it isn’t obvious that the poor will benefit. Does a federally subsidized glossy bank building in downtown Buffalo really help the city’s poor? Building strong communities is critically important for poor children, but lumbering federal bureaucracies are not particularly good at building strong communities.

Mr. Brooks’ view of place-based aid is too rosy and his view of vouchers is too negative. Lawrence Katz and Jeffrey Liebman of Harvard, Jens Ludwig of Georgetown, and Jeffrey Kling of the Brookings Institution have studied the U.S. Department of Housing and Urban Development’s “Moving to Opportunity” experiment, which compared random poor people who received vouchers with those who did not get them at random.

Mr. Brooks alluded to the results that suggested that vouchers aren’t cure-alls, but he failed to mention their more positive effects. The experiment showed that vouchers allowed parents to move out of poor neighborhoods which led to improvements in mental health and declines in obesity. Children of voucher recipients were much less likely to be victimized by crime. And parents who received the vouchers earned slightly more.

This may not be enough for me to sign up for Mr. Edwards’ million-voucher plan, but vouchers still look better than most federal place-based projects. At least, the bulk of the money actually will go to poor people, not to politically connected banks and builders.

We should focus on helping people not places, although we can continue to use some place-based tools, like schools, to help poor children. Better schools and safer neighborhoods are the most important things we can give to poor children, which is neither an easy task nor exclusively a Democratic Party agenda.

President Bush’s signature domestic achievement is the No Child Left Behind Act and Rudolph Giuliani’s signature achievement as mayor was to make New York much more safe. I could easily imagine, and I certainly hope, that the policy debate will be enriched by a powerful Republican vision for a fight against poverty that puts people first, with school vouchers, charter schools, and enough police resources so that every child can be well-educated and safe.

Mr. Glaeser is the Glimp professor of economics at Harvard, director of the Taubman Center for State and Local Government, and a senior fellow at the Manhattan Institute.