The NYT reported on a Brookings study that examined inequality by city. The study showed that the largest and fastest growing cities tended to have the most inequality. In fact, the study understated the differences in inequality for two reasons.
First its focus was on the 95th percentile of the income distribution compared to the 20th percentile. While the 95th percentile had gotten at least an even share of the country's income growth over the last three decades, the big gains have been concentrated in the one percent. For this reason, the analysis would be missing most of the impact of rising inequality since 1979.
The other factor, which is likely more important, is that it doesn't take account of differences in housing costs by area. While the price of food doesn't differ much from city to city, housing prices vary enormously. And housing will almost certainly take up a much larger share of a lower income household's budget in an expensive city than in a cheaper one. For example, the median rent in San Francisco is $2,146 a month. By comparison, the median rent in Wichita, Kansas, one of them equal cities in the study, was just $793 a month. While a moderate income household can reasonably expect to afford housing in Wichita, there is no way that a moderate income household in San Francisco would be able to afford a housing unit there. An analysis that accurately measured differences in income would have to factor in the differences in housing costs by city.