by Viral Acharya & Matthew Richardson
In a recent Wall Street Journal article, John Varley, Barclays Chief Executive, was quoted as saying: "We see big banks as diversifiers, not risk aggregators." His comments are part of a chorus of Bank CEOs now questioning various reforms that are aimed at large, complex banks.
He is plain wrong and regulators should be wary of such arguments.
Finance 101 tells us that there are two types of risk - idiosyncratic or firm-specific which is diversifiable, and systematic or market-wide risk which is not. While it is certainly true that the expansion of Barclays, and similar financial firms, into multiple business lines may reduce overall volatility of their asset portfolio, this is not what we most care about.
Because an economic crisis is the realization of market-wide risk, the problem we really care about is whether banks - large or small - can withstand such risk and continue to perform valuable intermediation functions. Or is it that when the economy craters, a bank's loans become impaired, the prices of its holdings of securities - many of them carry trades - fall, their underlying investment banking commerce produces less revenues, and the value of their asset management business plummets. In other words, do banks collapse with everything else?
It is this collapse that we recently witnessed in what can be considered the greatest incidence of systemic risk since the Great Depression. "Systemic risk" can be broadly considered to be the joint failure of financial institutions or freezing of capital markets. History has now taught us beyond doubt that such risk can have catastrophic consequences for the broader economy at large. Without financial intermediation, there is no credit, and without credit there is no commerce.
Of course, some market-wide risk has to be held within the financial sector. But to the extent possible, the last institutions we want holding it are the ones that are vital to the financial plumbing of the economy. This is precisely the reason these institutions are too-systemic-to-fail. It might make sense for Barclays as an individual firm to hold market-wide risk and swing up and down with it, but Barclays' top management, shareholders, and in fact, even its creditors, do not bear the full systemic costs when the pendulum swings down. It is the real economy that feels the burden and society pays the price, whether in the form of bailouts, lost productivity or unemployment. While profits remain privatized in good times, downside risks are socialized.
There is only one way out of this dilemma and that is to force these firms to internalize the systemic costs they impose on the rest of the economy. This way, financial institutions like Barclays will organically choose to become less systemic and much of this risk will be held by non-essential firms like regional banks, private equity and hedge funds, mutual funds, or the proverbial Norwegian village, et cetera.
The economically optimal, and most effective, way to make these firms feel the pain is to tax each financial institution an amount equal to the expected systemic costs of a crisis multiplied by that institution's percentage contribution to financial sector losses.
Can such a "systemic risk tax" be measured? The answer is for the most part yes.
The first term - expected systemic costs - measures the level of the tax. There is evidence on what leads to financial crises (excessive leverage, for instance) and the costs to economies of such crises (output gaps, for instance) beyond the impact of a normal economic downturn. Even when the probability of a crisis is low, these realized economic costs are often as high as 15-25% of the GDP so that the level of the tax will be meaningful.
The second term - percentage contribution of each institution to costs incurred in a financial sector collapse - determines which institutions pay more of the tax. Recent work we have done at NYU Stern, along with colleagues Lasse Pedersen and Thomas Philippon, suggests that this term is easier to measure than one might think. For example, we find that the size and volatility of a firm's assets, its correlation with the market, and its leverage go a long way to forecasting which firms faltered in the financial crisis. In fact, the data prior to June 07 provides a who's who of the financial crisis that followed, in fact with a dozen odd firms representing almost all the systemic risk. No prizes to figuring out who is on the list.
In an attempt to lower such a tax, each systemic financial institution will try to reduce their share of this risk, leading to a system-wide reduction in both asset risk held by these firms and their leverage. Of course, some systemic risk will remain and thus some taxes will still be collected. Banks have supported such taxes but called them as "bailout funds". It is important to point out that the purpose of the taxes is not to foster bailout expectations and aggravate moral hazard. The tax proceeds are NOT meant to bail out failed institutions, but to support the affected real sector and solvent financial institutions when there is a systemic crisis. If deemed more than adequate in good time, the proceeds can simply be returned to the taxpayer.
If regulators believe there is still a possibility they will have to bailout firms and provide explicit or implicit guarantees to creditors in a crisis, then they need to additionally (i) charge for these guarantees explicitly (similar to a premium for deposit insurance), (ii) ring-fence the use of government guarantees to the core financial plumbing activities (the "Volcker" rule), (iii) impose additional prudential regulation like minimum capital requirements, and (iv) force some losses on creditors, either through "contingent" capital debt, that is, forced debt-for-equity conversions, or a credible bankruptcy resolution plan in the form of a "living will".
CEOs of large complex banks argue against such a full frontal attack claiming that their banks in fact reduce risk through diversifying and being large. But the data and our most recent experience tells us a different story. Large banks were in fact aggregating economy-wide risk and taking a one-way bet on it. Regulators should not let up. Take no prisoners.