Half a decade ago, in the wake of the emerging-market crisis that spread from East Asia to Russia to Brazil, the policy circuit buzzed with proposals to reform international finance. Bright ideas were mooted -- by think tanks, market sages and a high-profile congressional commission -- and in 2001 the incoming Bush team declared itself "impatient" with the status quo it had inherited. But the upshot of this brainstorming was modest, incremental change. The grand reformist talk only weakened confidence in the existing system while doing little to improve it.

Today we have a new version of this phenomenon. The big buzz right now surrounds post-conflict reconstruction: Both Afghanistan and Iraq have shown that we aren't good at it. At least three bits of legislation on this topic are floating around Congress; the State Department is mulling schemes to strengthen its response; the Pentagon promises to train African peacekeepers. The Center for Global Development recently produced a report on conflict and failed states, and the Council on Foreign Relations has a high-level task force examining the same challenge.

Some of the new thinking is quite radical. The report from the Center for Global Development, for example, advocates a cabinet-level department in the U.S. government to handle strategy toward developing countries. It calls for an initiative to anticipate civil wars so that preventive action can displace expensive remedial efforts; the British government is implementing a version of this policy. Others advocate a new reconstruction trust fund, which would get around the need to pass the hat when a crisis demands urgent intervention.

Much of this sounds good, and I've done my bit in the past to press versions of this agenda. But it's hard to escape the feeling that we're reliving the debate over emerging markets, when ambitious ideas -- an international bankruptcy court, an early-warning system for crises, an expanded International Monetary Fund that could act as a global lender of last resort -- briefly lit the sky and then rapidly fizzled.

Some of the current post-conflict proposals should be ruled out on grounds of political implausibility. After the bureaucratic teething problems of the Department of Homeland Security, is it realistic to propose another cabinet-level agency? Given the trouble of raising money for existing international programs -- the Global Fund to Fight AIDS, Tuberculosis and Malaria, for example, seems likely to get a fraction of the money it is asking for in this year's appeal -- what is the likelihood that a post-conflict trust fund would attract serious cash from donors? Would the French or German government donate money to such a fund, knowing that the money might be spent on a future crisis that it deemed unworthy, such as cleaning up after the Bush administration in Iraq? Unlikely.

Other proposals are politically easy but practically unhelpful. Early-warning, for example, sounds good, and it's not hard to create a new unit in the State Department to monitor leading indicators of conflict. But civil conflicts, like financial crises, are hard to predict with any certainty, and in any case the private sector is already trying. In the financial world, you can consult a credit-rating agency or an investment bank. In the political world, you can read the excellent reports from the International Crisis Group.

Moreover, suppose you identify a list of 20 tense countries where civil war looks likely. What do you do about it? The central finding in development theory in the past decade is that you probably could not do anything. Foreign aid does not work in dysfunctional countries, where the threat of conflict looms; it works only in countries that are well run, which is why the Bush administration's Millennium Challenge Account will funnel cash to a short list of nations that are poor yet governed competently. If the talk of "conflict prevention" gets too much traction, this virtuous selectivity may be undone. Aid will be diverted to corrupt, autocratic environments that seem headed for civil war. It will probably be wasted.

Rather than reaching for a radical new fix, the best approach to the problem of failed states is to nurture the institutions that we have already. You want a post-conflict trust fund? Well, you already have the World Bank, whose subsidized loans can be used to reconstruct countries -- why not just expand it? You want an institution that can parachute into a broken country and rebuild everything from the judicial system to the electricity grid to the central bank? Well, the World Bank can do that too, and did in Bosnia -- again, why not give it more resources? You want an institution that can mobilize peacekeepers and train new indigenous armies? You already have the United Nations and NATO, so why not bolster them before you dream up a third option?

Above all, don't lose sight of past experience. Exactly 60 years ago, at the Bretton Woods conference in July 1944, the United States and its allies created the World Bank, believing (as Franklin Roosevelt told the delegates) that poverty and hopelessness threatened the world's stability. Six decades later, we've learned some lessons about what works and what doesn't in failed and failing states. Ignoring that history will condemn us to relive it.