Only theoclassical economics constantly recycles variants of its worst ideas that have proven disastrous when they have influenced policy. Other fields advance because they embrace the scientific method. Theoclassical economists repeat their worst errors because they embrace anti-governmental dogmas that blind them to the inherent weaknesses of the corporate form and limited liability. This represents a dramatic regression in understanding from over 200 years ago when classical scholars like Adam Smith were warning that corporations were inherently criminogenic and likely to produce what we now label "control frauds."
I have just experienced two related examples of this regression in understanding by economists of the inherent risks of corporations and control fraud in the context of financial corporations. Both of the economists admit that corporate governance is flawed but implicitly assume control fraud out of existence and explicitly assume that the problems with governance are not inherent and can be remedied by moderate tweaks. As I will show, the tweaks are actually radical, but so badly designed that they would be even more criminogenic than the highly criminogenic status quo. The first proposal, "contingent capital" was presented by a more junior scholar who will remain nameless. The second proposal, limiting limited liability for bank shareholders, was suggested by Tyler Cowen. I ran across Cowen's proposal to limit bank shareholders' limited liability while researching my response to his newest claim that our largest financial corporations represent the leading wave of our future as a "hyper-meritocracy."
Contingent Capital: and The Princess Bride
A speaker whose talk I attended stated as a fact that it would be "inconceivable" for any bank to fail if banks adopted "contingent capital." Under "contingent capital" the bank would sell debt instruments that would convert to equity under a specified "trigger." The trigger would be the bank's reported capital falling below some threshold that the speaker did not identify. The theoclassical theory is, as always, the creation of "private market discipline" to prevent bank failures. The most catastrophic bank failures (the ones that drove the global financial crisis) are "accounting control frauds." The speaker's theory ignored fraud (the word never appeared). His thesis was that because the debt holders who would be converted into stockholders and the existing shareholders would suffer severe losses if the conversion were to occur, the stockholders and the convertible debt holders would have powerful incentives to prevent the bank's capital from ever falling to the level that the trigger would force the conversion. The speaker assumed that the problem with banking is moral hazard and implicitly assumed that moral hazard leads only to excessive risk-taking rather than the "sure thing" of control fraud. He further assumed that the CEO controlled the bank and that the shareholders lacked the proper incentives to discipline the CEO's excessive risk-taking. Contingent capital would purportedly address both problems by increasing the amount of bank capital and putting shareholders at a greater risk of loss such that they would have improved incentives to exercise effective discipline over the CEO and prevent excessive risk-taking.
The presentation I attended did not note that we had tried a variant of this strategy, which theoclassical economists assured us would produce superb private discipline, in the last three crises through the use of subordinated debt ("subdebt"). Subdebt means that the purchaser agrees to a lower bankruptcy priority in return for a higher yield. In the event the bank is placed in receivership the subdebt holders receive no payments unless there are sufficient funds to pay all the secured and unsecured general creditors' claims in full. Typically, this would mean that the subdebt holders would be wiped out if there is a receivership. In the S&L crisis, we routinely wiped out the subdebt holders in our receiverships. Neoclassical economists claimed that subdebt holders provided the ideal source of private market discipline for three reasons. First, because they were so likely to be wiped out if the bank failed they had powerful incentives to exert superb discipline. Second, because the typical sub debt holder bought a substantial amount of debt that incentive was amplified. Third, only "sophisticated" (actually, wealthy) investors can purchase subdebt they should also have superior skills in disciplining the bank CEO and preventing him from looting the bank.
Theoclassical economists have long favored capital remedies that can be implemented by private parties over vigorous financial regulation. Their explicit premise is that their capital remedy will eliminate "agency" problems (e.g., excessive risk-taking by the CEO) and therefore means that there should be no, or virtually no, regulatory restrictions.
Theoclassical economists were so successful in pushing subdebt as the ideal source of vigorous and competent private market discipline that they convinced their regulators to allow banks to count subdebt toward meeting the bank's capital requirement. Encouraging banks to issue high cost debt is an expensive policy that should require strong benefits that outweigh the costs. (Sorry, I've engaged in a fantasy. Neoclassical economists demand benefit-cost studies only for acts of regulation, not anti-regulation.)
The problem is that subdebt failed, universally, in providing effective private market discipline at banks. It failed during the S&L debacle, the bank frauds that aided Enron's frauds, and the current mortgage fraud crisis. Subdebt simply helped fund the growth and frauds of accounting control frauds in each of the modern U.S. crises and the ongoing global crisis.
Contingent capital is, conceptually, weaker than subdebt as a source of private market discipline. Like subdebt; it will be expensive for the bank to sell because it will be far riskier than regular debt and will require a significantly greater yield. To count toward capital, subdebt must be relatively long-term debt. The convertible debt could be much shorter term, which would create perverse incentives on the part of the owners of the convertible debt to cover up a failing bank's losses to avoid triggering the conversion prior to the debt maturing and being repaid in full prior to the bank's failure.
Bank stockholders have even lower priority than subdebt holders in the event of a receivership. Theoclassical theory already asserts that shareholders should have the right incentives to exert private market discipline. The reality is that bank stockholders have also consistently failed to exert effective private market discipline in the modern crises.
The speaker stated that it was "inconceivable" that a bank could fail if it had contingent capital. Indeed, he stated that it was inconceivable that any bank would ever hit the trigger that would cause its debt to convert to equity. I, of course, quoted The Princess Bride.
"Vizzini: HE DIDN'T FALL? INCONCEIVABLE.
Inigo Montoya: You keep using that word. I do not think it means what you think it means."
When the CEO causes a bank to become an accounting control fraud he creates three "sure things" if he follows the "recipe."
- Grow extremely quickly by
- Making bad loans at premium yield
- While employing extreme leverage, and
- Providing only grossly inadequate allowances for loan and lease losses (ALLL)
The three sure things are that the bank will report record (albeit fictional) profits, the CEO will promptly be made wealthy by modern executive compensation, and the bank will suffer catastrophic losses. Note that the fraud recipe works by massively overstating asset values. "Capital" is simply an accounting residual (Assets -- Liabilities = Capital). This means that accounting control frauds inflate reported capital and simultaneously reduce actual capital when they inflate asset values by making bad loans. If a material number of bankers mimic the recipe the fraud recipe also becomes the ideal means of hyper-inflating financial bubbles. This further inflates reported (fictional) asset and capital values and increases real losses.