Yeah, that is the "government failure" theory promoted by Weisbrod & Co, but I am yet to see any empirical support of it in the form of correlating gov't social welfare spending with social heterogeneity. If that view held, homogenous Japan or Ireland would have higher social welfare spending than heterogenous Belgium or Spain. In fact the opposite is true: gov't social spending in homogeneous Japan is about 14% of the GDP, Ireland - 19% of the GDP, heterogenous Belgium - 29% of the GDP, Spain, 21%).
Methinks, what makes the difference is a strong working class movement versus the lack of it (see for example the paper by Lester M. Salamon, S. Wojciech Sokolowski, and Helmut K. Anheier _Social Origins of Civil Society: An Overview_ at http://www.jhu.edu/~ccss/pubs/cnpwork/index.html).
wojtek