Something mentioned in the article - the social security administration is required by law to double-check half of all approvals, but not denials. So the incentive is definitely to deny as quickly as possible - if you approve everyone you'll be caught, but if you deny everyone that's far less likely.

But it offers no data to support the general idea that this is the case (that denials are through the roof, even if TN is a higher rate than the national level, it is marginal. Also it leaves out the median which would be more interesting, since a few states could greatly bring down the average.) The approvals are reviewed because denials can be appealed. It's a safe guard against fraud. I'm not saying there aren't perverse incentives, but "doctors earn big money approving disability claims" has less of a moral outcry from the reverse.

If you are likely to make errors due to having to review too fast to properly evaluate claims, I'd expect you to lean toward denying claims so that your error would mostly be denying claims that should have been approved instead of approving claims that should have been denied.

That's because if your errors tend toward approving claims that should have been denied, your are going to have a higher than normal fraud rate among your approvals, which might raise suspicion that you are in cahoots with the fraudsters maybe taking bribes to approve them.

Denying claims that should have been approved, on the other hand, probably has little risk of getting you personally is trouble.

Hence, there is an incentive when rushed toward denial of claims.

And in reviewing quickly, they're making mistakes, denying a man dying of cancer the disability money he was entitled to under the law. Perhaps some people who don't meet the qualifications are being approved, but I think most people would view false positives as less of a problem than false negatives in this case, and the fact that the denial rate is higher in Tennessee suggests to me that it's faster to deny than approve anyway.