Both the New York Times and some European observers are predicting that Belgium may soon break up. This circumstance, combined with the persistence of longstanding secession movements in Quebec, Scotland, and elsewhere, have persuaded some, including our own David Bernstein, to doubt the viability of federalist solutions to the problem of ethnic conflict.
Belgium has a long history history of tension between the majority of Flemish speakers (concentrated in the north) and the French-speaking Walloon minority (most of whom live in the south). It is certainly possible that the country will break up soon. However, there are reasons to doubt both the likelihood of breakup and the claim that such an event demonstrates the failure of federalism.
It is far from clear that Belgium really will split up. The country has weathered ethnic tensions and threats of secession for over 170 years. The record suggests that the current crisis, triggered by a coalition power struggle in parliament, may well blow over as many previous ones have.
Even if it does not, a peaceful secession today by either the south or the north would not invalidate the successes of federalism in keeping ethnic conflict in check. After all, a combination of federalist decentralization and powersharing kept Belgium together for 170 years with a relative minimum of intergroup violence. Flemish and French-speaking Belgians may not like each other much (though the New York Times is surely wrong to claim that they "cannot stand each other"), but they have coexisted successfully for many decades. During that time, the country achieved a high level of political and economic development. The same is true in Canada, which has longstanding issues with its own secession movement, and even more so in Switzerland, whose federalist system has successfully managed conflicts between four different ethnic groups for centuries. Ethnic federalism is no panacea and won't work everywhere; David is probably right to suggest that it wouldn't work under current conditions in Israel/Palestine. But, overall, it has been a great success in Switzerland, Belgium, Canada, and a number of other countries. By allowing each group to have control of those regions of the country where it is in the majority, while respecting basic minority rights, it prevents the kind of zero-sum power struggle between groups that is likely to occur in an ethnically divided society where all the power is in the hands of the central government.
However, the Belgian example is a warning signal on one important issue: Much of the secessionist impulse by Flemish Belgians stems from "deep resentment in Flanders that its much healthier economy must subsidize the French-speaking south, where unemployment is double that of the north." Federal transfers between different regional governments create a zero-sum game between regions and stimulate resentment in regions whose inhabitants believe they are being fleeced for the benefit of parts of the country dominated by other ethnic group. They also reduce the incentive of regional governments in poorer regions to adopt pro-growth policies likely to improve their economic fortunes. These problems arise even in federal systems that are not divided on ethnic lines (as John McGinnis and I explain here). But they are heightened when ethnic grievances come into play.