Sloppy arguments about the "moral hazard trade"

Felix Salmon appears to be deeply confused about what he calls the "moral hazard trade" (otherwise known as the TBTF subsidy). James Surowieki challenged Felix's earlier posts about the moral hazard trade, asking, "If big banks have lower borrowing costs than small banks, he asks, why do we automatically attribute that to moral hazard (the idea that they're much more likely to get bailed out in extremis) rather than the simple fact that they're less likely to default?" In response, Felix cites a recent anomalous spike in the COFI (cost of funds index) which was attributable to a change in the index components, and which essentially showed that Wachovia has a much lower cost of funds than small banks:

This datapoint is telling, because Wachovia — largely because of the Golden West acquisition — was a very rocky bank indeed, and was sold as a highly-distressed asset to Wells Fargo. The fact that its cost of funds was so low clearly had nothing to do with its inherent safety, which means that we have to attribute it instead to the moral hazard trade.

Um, what? All this proves is that big banks tend to have a lower cost of funds than small banks. No one is disputing that. The dispute is over how much of that lower cost of funds is attributable to implicit government support, and how much is attributable to other factors, such as the relative credit risk in the portfolios of big banks vs. small banks. Felix simply asserts, without any evidence, that "clearly" the difference had nothing to do with any other factors, and therefore concludes that the entire difference must be attributable to the moral hazard trade.

Of course, it's not at all "clear" that the entire difference is attributable to the moral hazard trade. For one thing, small banks' balance sheets tend to be heavily concentrated in real estate loans — especially commercial real estate, which is in a free-fall right now. The big banks — which got badly burned by CRE in the early '90s — tend to be significantly less concentrated in CRE. So that's one factor that undoubtedly accounts for some of the difference in small banks' cost of funds.

How much of the difference is attributable to small banks' exposure to CRE, and how much is attributable to the moral hazard trade? I don't know. (And I'm not denying that the moral hazard trade accounts for some of the difference; it definitely does.) My point is simply that there are lots of factors at play here. You can't just assert that there are obviously no other factors at play, and that the entire difference must be attributable to the moral hazard trade.