Powerful though the Fed may be, it can’t cure all negative externalities all of the time. FRB Gov. Tarullo is fond of talking about negative feedback loops as he ponders an end to all the harm that big banks can do as they topple, but rules are at least as inter-connected as GSIBs. A case in point: TLAC. U.S. GSIBs aren’t TBTF anymore post-TLAC, but it could well make them riskier and thus more likely to fail – the only difference this time is that there won’t be any taxpayer backstop. The reason: TLAC costs a lot, costs have consequences, and markets are merciless.

The inter-connection between TLAC and insolvency risk is laid out in the in-depth report on the FRB’s proposal sent to you shortly before the holiday. In it, we noted how much tougher the FRB proposal is than the FSB’s, detailing not just the terms and conditions stipulated for top-tier TLAC, but also other new funding restrictions on the parent holding company and how new funding flows resulting from TLAC would affect the company as a whole and its critical operating subsidiaries.

In our analysis, this TLAC conclusion is based on a fundamental fact the FRB has either failed to consider or discounted: when BHCs raise funding, they have to use it for something and that something has to yield more than its funding costs for the bank not to go bust. This is the essence of financial intermediation and, as long as GSIBs are banks, it’s the fundamental fact of their life or death.

It will be hard for GSIBs to find places to put TLAC because TLAC will cost a lot – long-term, unsecured debt issued by a GSIB will cost more, and more will be a lot more when – God willing – the yield curve normalizes and term premiums realign to something like sensible. In any market that’s got its head on straight, long-term unsecured debt costs more than short-term debt and lots more than core deposits. Unless GSIBs balloon their balance sheets, lower-risk, lower-return assets will be set aside in search of higher-yielding ones that preserve at least a prayer of net interest margin.

Earlier this week, we also provided you with an analysis of a new study released by the Bank for International Settlements validating our fears about how risky it can be to set liquidity rules too high. In short, this study used econometric models to conclude that banks with high liquidity requirements put as much funding as they can into the riskiest assets they can get away with to keep their shareholders’ noses as much above water as possible. The study looked expressly at the liquidity coverage ratio, but its conclusions are as germane to TLAC as to the LCR or, for that matter, the Fed’s forthcoming net-stable-funding-ratio rule: high-quality, liquid assets squeeze out high-quality, less-liquid assets combining into an inexorable vise of higher funding costs for lower-return assets that forces banks to take more risk with whatever funds they can find.

The study does posit one way to break the cycle between stringent liquidity requirements and greater solvency risk: lots more capital. It importantly notes, though, that banks can’t raise capital when they want – the greater the market stress and the larger the need for it, the harder equity is to find and the more it costs. FDIC Vice Chairman Tom Hoenig has posited one solution to this in concert with his complaint about the Fed’s TLAC rule: force GSIBs to raise lots more leverage capital at all times so that, he believes, GSIBs can never become insolvent.

But, leverage capital also has costs, costs we’re already seeing in the distorted balance sheets at several U.S. GSIBs. As a recent FedFin study demonstrated, these distortions so far aren’t risky – quite the contrary. Banks are harboring trillions in excess reserves because they can’t find any place else to put them under current market and economic circumstances. This might make Mr. Hoenig happy, but it’s stymying Fed monetary-policy execution and making asset management a lot more risky, but at least it’s safeguarding the banks.

When markets normalize, economic growth picks up, and – most importantly – all of the rules combine with TLAC to restructure both sides of the U.S. GSIB balance sheets, the new structure of the redefined GSIBs will not only be different, but almost surely a lot more risky.

Some of this risk will go off the books – GSIBs are already fleeing to asset management and other fee-based businesses which don’t generate assets that then trigger TLAC, the liquidity rules, leverage capital, and all the other standards designed to shrink negative externalities into a happy, positive sense of big-bank well-being. This fundamental restructuring may solve for some of the costs of all the new rules without hiking GSIB risk, but it’s no solution.

The more GSIBs flee financial intermediation, the less they are banks. The less they are banks, the more other companies will try to be banks. The more the other companies that try to be banks aren’t banks, the less all of the GSIB rules matter in terms of overall financial-system stability. FRB Vice Chairman Fischer said as much yesterday even as he defended all of the rules imposed on all of the U.S. GSIBs. But, like Mr. Tarullo, he can’t have it all. Either we have functional big banks that can perform effective financial intermediation or we have a new-age U.S. financial system which poses risks both to long-term economic stability and near-term financial-market resilience.