One way to get a sense of what happened to mortgages, where the government doesn't track default rates by race, is to look at default rates for college loans. From a 2007 Education Sector article by Eric Dillon: "A Closer Look at Student Loan Default Rates:"
Black students who graduated in 1992–93 school year had an overall default rate that was over five times higher than white students and over nine times higher than Asian students. The differences for Hispanic students are not as large, but are still substantial. Hispanic students' overall default rate was over twice that of white students and four times higher than Asian students. And these differences cannot be fully explained by differences in borrowing patterns or salaries. The 1994 percentage of monthly income going to student loan payments—an indication of both how much debt a student has and their earnings—was actually lowest for Hispanic students and only slightly higher than average for black students. ...
The median monthly loan payment as a percentage of monthly income in 1994 for recent 1992–93 graduates was 6.7 percent for the
as a whole, 7.4 percent for Asian students, 7.2 percent for black students, 5.3 percent for Hispanic students, and 6.8 percent for white students. United States
In other words, income didn't have much to do with default rates. (On the other hand, whites and Asians would be more likely to have relatives who could help them out repaying their loans during hard times. That's something that's overlooked in thinking about mortgages, as well.)