“Too big to fail” (TBTF) is a popular, but slippery, concept.  The Roman Empire is gone.  So, too, is the USSR.  Within the realm of economics, the East India Company is gone, and IBM, Bell, Kodak, and U.S. Steel are shadows of their former selves.  The list could go on.  Popular discussion notwithstanding, the reality is that nothing is really “too big to fail”. 

There are some firms, however, that may be too big to fail quickly, and that is a substantive concern with large banks.  If the feared insolvency of a large bank leads depositors and other funders to flee, the collapse of bond and equity values may drag down others.  The inability to redeem overnight lending, short-term debt, and longer-maturity liabilities of a large bank could have a cascade effect that threatens the real or perceived viability of other institutions.  This line of reasoning and the events of 2008 have put the TBTF issue at the top of the policy agenda.

It is an important and complicated set of issues that will likely take substantial time to sort out.  The current discussion, however, contains a few red herrings that threaten to lead policymakers astray.

First, the issue has nothing to do with size, per se.  In principle an extremely large bank that was isolated from the remainder of the financial system could disappear without a concern.  Instead, it is the degree of interconnectedness that matters most.  In the crisis, the failure of some large foreign banks did not cause U.S. policymakers to blink – no interconnectedness.  But the collapse of Lehman Brothers – by no means the largest firm – was notable for the collateral damage caused by its interconnectedness. 

For this reason, the tempting but naïve call to “break up the big banks” is just that: naïve.  Size is not the issue.  The naïve attack on size becomes downright dangerous when it ignores the fact that there are economic fundamentals that demand banks of all sizes, which is why every country has some banks of global reach.  Attacking the U.S. banks will simply shift this economic activity elsewhere to foreign banks.

Second, the notion of TBTF has more to do with the risk aversion of policymakers than the reality of balance sheets.  In a crisis, policymakers fear the worst and behave accordingly.   Faced with even a very small probability of a real downside, they generally intervene in ways that would normally be viewed as unnecessary or even counterproductive. 

The corollary is that policymakers do not need more financial regulatory tools.   They will likely be invented anyway in a crisis, and misused both before and after.  Instead, the goal should be to produce more transparency and market discipline.   If it is clear that a bank has a risky balance sheet, counterparties will shed their risk and the bank will become less interconnected automatically.   

Third, a notion closely related to TBTF – but not quite the same thing – is “bigger than without government policy.”  Government interventions that distort the market are pervasive and real.  But they are hardly confined to the largest banks.  Deposit insurance, preferential tax treatment of some financial institutions, the array of perks provided to Fannie Mae and Freddie Mac all produce institutions that are different than the market would produce on its own. 

At present, a lot of the focus is on the putative TBTF subsidy – the supposedly reduced cost of funds to large banks because of their special policy status.  The statistical pursuit of this subsidy seems to be a fool’s errand.  As noted above, it is interconnectedness that is necessary, not size, and none of the efforts attempt to control for this key feature. 

Next, basic finance theory has a core finding: idiosyncratic risk is not priced.  That is, institution-specific risks will be diversified away and not reflected in the asset prices based on expected returns.  Accordingly, comparisons of large and small banks tell the analyst nothing about characteristics specific to those banks.

Instead, the key is how returns vary relative to the overall state of markets and the economy – so-called “beta”.  Here TBTF would definitely show up: when things go really south, a for-sure government subsidy would raise returns and be priced accordingly.  But better performance in bad times also could be due to greater liquidity, a greater scope of products, and broader global diversification. – all characteristics of large, global banks.  Statistical studies that don’t control for these features, and it is very difficult to imagine doing so adequately, will inevitably intermix these attributes and fail to identify correctly a TBTF subsidy.

TBTF is an important policy debate.  But the exclusive focus on size and reliance on flawed statistical studies threatens to lead the debate astray.