Unforunately, the inspector general stays mum on how it arrived at its conclusion:
Federal Reserve Board officials provided institution-specific data and an analysis of the state of the U.S. economy that was deemed sensitive, confidential, and, restricted. As a result, we could only generalize from the data obtained.
That's a shame because a new paper by Harvard's Ryan Taliaferro arrives at a slightly different conclusion: Banks did appear to use some of their TARP funds to increase lending. To arrive at his conclusion, Taliaferro compared different metrics of otherwise similar TARP and non-TARP banks in the months after the financial crisis began:
Based on a matched sample of participating and non-participating banks, of each dollar of new government equity they received, participants used roughly fifteen cents to support increased lending, while they used roughly sixty cents to increase their regulatory capital ratios.
This chart from the paper, displaying the fraction of TARP funds used for different purposes, shows that the average bank also directed some of their new capital towardsof their capital to boost earnings:
Of course, a 15% boost in lending probably isn't going to quell TARP critics. And this is where I'm somewhat sympathetic to Mulligan. It's unfortunate that for political reasons TARP had to be sold as a measure to boost lending, when in fact it was really intended to stabilize the banking system. So it's not surprising that banks used the majority of the funds to increase capital and calm fears over possible insolvency. (I'm only somewhat sympathetic to Mulligan because my recollection is that he wanted to allow banks to fail. More to my liking are the views of fellow U of Chicago professor Douglas Diamond, who argued for government intervention.)
UPDATE: Tim Fernholz calls out Mulligan for cherry-picking from the SIGTARP report.