No contradiction. It's talking poor neighborhoods, not poor individuals. It's claiming that if you're black, there's about a 66% chance you grew up likely living in a neighborhood which is at least 20% poor. That doesn't contradict the Census data that 26% of blacks and 10% of whites are poor.
For example, using the Census figures, imagine the extreme case that blacks and whites lived completely segregated in their respective racial communities, and the poor within each race were totally uniformly distributed within their enclaves. Then in that case, 100% of blacks would live in neighborhoods that were greater than 20% poor - in fact exactly 26% poor. And 100% of whites would live in what would be deemed non-poor areas -- all white neighborhoods would be exactly 10% poor.
So what we're seeing in Sharkey is merely the secondary effect of a non-uniform distribution, not a change in the underlying poverty rates.
For example, using the Census figures, imagine the extreme case that blacks and whites lived completely segregated in their respective racial communities, and the poor within each race were totally uniformly distributed within their enclaves. Then in that case, 100% of blacks would live in neighborhoods that were greater than 20% poor - in fact exactly 26% poor. And 100% of whites would live in what would be deemed non-poor areas -- all white neighborhoods would be exactly 10% poor.
So what we're seeing in Sharkey is merely the secondary effect of a non-uniform distribution, not a change in the underlying poverty rates.