I'm going to dispute "similarly qualified". Being married is a huge indicator of future financial stability, and for most of the cited time period, SSM couples could not get married. There are also many statistics which point to SSM relationships (regardless of marital status) ending sooner.
So if the banks are inputting this to a "go/no go" algorithm, the results do not surprise me at all.
So if the banks are inputting this to a "go/no go" algorithm, the results do not surprise me at all.