There is a long tradition in sociology, going back to Durkheim, that treats social solidarity -- the process of gluing a society together -- as fundamental. One factor in this process that gets a lot of attention is income inequality. Sociologists and economists have developed lots of different measures, the easiest to conceptualize are measures of "income share"; the percentage of national income received by a particular group (e.g., the top 10% of wage earners take home 30% of all the income).
While there is little agreement on precisely what level of difference is most appropriate (or if there is even such a thing as a single best appropriate level), there is broad agreement that tilting to the extreme in either direction can be problematic. Too little income differential and economic productivity suffers as the rewards and motivation for productivity are removed. Too much income differential and social cohesion suffers as feelings of exploitation or hopelessness spread. In an age of cheap and powerful weapons, the results can be tragic and equally disruptive to economic productivity as many countries in Latin America and Africa have discovered.
Slate contributor Timothy Noah has recently completed the fascinating series The Great Divergence, available in its entirety as a pdf file here. While the increase in income disparity within the US has been the topic of discussion for many years, the series is notable both for placing the debate in the context of a longer historical frame and for the the breadth of the research and explanations it examines.
As the above above graph shows, income disparity within the US has been increasing rapidly since the late 1970's. The difference between rich and poor is now as great as it was at the end of the Robber Baron era. More interesting to me, particularly in light of the current debate about taxes in the US, is the data below taken from the work of Princeton political scientist Larry Bartels (author of Unequal Democracy).
The graph compares the rate of income growth for individuals of different income levels during years of Republican and Democratic presidential administrations. Three things are worth noting. First, income growth at all income levels has been higher when Democrats were president (blue lines) than when Republicans were president (red lines). (Technically, the growth rates at the 95th percentile are equal, even though the graph shows higher growth for the Democrats. The difference for the 95th percentile, unlike for the others, is within the margin or error.)
Second, the length of the blue lines are roughly the same. In other words, "equal economic growth for all" seems to be a more accurate characterization of the outcome under Democrats than "economic redistribution favoring the poor." While the top percentile does do a bit worse than the rest under the Democrats, there is little difference in the rate of income growth among the bottom 80% of the population.
Third,not only does everyone do worse, but the disparities in income growth are dramatically larger during Republican administrations. The poor (20th income percentile) average only .4% income growth under Republicans, while the richest (95th percentile) see their income grow at a substantially higher rate of 1.9%.
Significantly for Canadians, this is one area where information about the US has little relevance for Canada. While poverty and inequality certainly remain issues here, the same trend toward massive increases in inequality aren't present.
For more information, see the Resilience Science post Inequality in the USA -- driven by politics? which summarizes and links to several other related works.