Behavioural economists are always quick to tell us that human beings are extremely poor at assessing risk: we underestimate likely dangers and overestimate unlikely ones. Automotive accidents are relatively common, but we’ll risk taking an important call while we’re driving. Shark attacks are extremely rare, but thoughts of an encounter with a great white cross our minds when we step through the surf.
A single British person has died from a rat bite in the past four years, whereas the lifetime risk of dying in a road accident for UK citizens is 1 in 240; but 42 per cent of 25- to 34-year-olds admit to having a phobia of rats. Fear of driving is comparatively small because we are willing to take risks if we think that we can control the outcome. We associate driving with being in control of the vehicle, whereas an encounter with a rat seems relatively unpredictable. After the terrorist attacks of 9/11 1.4 million Americans changed their travel plans for Thanksgiving and Christmas despite the fact that it was far, far more likely that something bad would happen to them if they travelled by car than by airplane.
The evaluation of risk is a mathematical computation, meaning that we should be able to make use of data to make better decisions. But the truth is that, for individuals, risk and emotion are inseparable. We think that we’re being rational even when our feelings are affecting our judgment. But does it work the same way for groups of people? Do large corporations and markets find it easier to assess risk by pooling data and using facts to drive decision-making?
Credit ratings agencies are supposedly entirely data-driven businesses; their area of expertise is to assess risk within financial marketplaces by determining the quality of debt obligations such as securities. Consequently, the agencies hold enormous sway over all financial instruments, including sovereign debt, by determining how that debt should be graded. Using a sliding scale from AAA to CCC the agency comes to its decisions based on how likely a borrower is to default on a loan. A rating of BBB or higher means ‘investment grade’; anything below this is regarded as ‘junk’.
In the US, the big three – S&P, Moody’s Investors Service and Fitch Ratings – have been anointed by the government; if you want to participate in the market, SEC and Federal Reserve regulation means that you have no choice but to work with them.
Despite the trust invested in them by the government, these agencies failed to spot the financial crisis of 2008, continuing to rate toxic mortgage-backed securities AAA until the crisis engulfed the global economy. Lehman Brothers, Washington Mutual and AIG were all rated ‘investment grade’ until September 15th 2008, the day Lehman filed for bankruptcy.
Part of the problem was that, in good times, no one wants to be seen as the person putting the pin in the balloon: bubbles tend to be driven by a kind of euphoria; a sense that this time things are going to be different. And agencies bill the issuers of the structured financial products, not the investors, meaning that they have a relationship with the investment banks: the banks are their clients and, consequently, the source of their income. If a bank wants an AAA rating for a tranche of mortgage-backed securities it’s just issued, it’s in a ratings agency’s interest to give the client what it wants. And, during the boom, there was a third dilemma: the financial instruments the agencies were rating were so complex and so new that the agencies simply didn’t have the information needed to make the right decisions – they were, effectively, relying on other ways of making decisions other than data.
Post 2008 with, effectively, a monopoly and Apple-sized margins, what incentive do the agencies have to reform? Congress has encouraged them to come up with some ideas, among them reviewing their analytical models, more complete disclosure on a security’s collateral and better trained analysts. Ideas such as introducing competition by having more agencies and changing the way they are paid – for instance, by investors in debt rather than its issuers – don’t seem to be on the table.
The agencies’ defence is that they are just like anyone else; they are simply offering opinions, which are constitutionally protected as free speech. Effectively they’re arguing that their judgments are made in good faith but cannot be relied upon; which begs the question why, if they’re not to be relied upon, the government has offered them a monopoly on unreliability? Surely, the guy at The Red Lion would be just as happy to receive hefty fees for equally undependable opinions on securitization?
Ongoing litigation in California is challenging that. When the housing market crashed, California’s main public pension system (CalPERS) lost an estimated $10 billion. Its administrators are involved in litigation with two of the ratings agencies to try and secure compensation for their awarding the highest ratings to three financial products CalPERS invested in. The products were subsequently discovered to consist largely of high-risk sub prime mortgages. The agencies in question, Moody’s and S&P, have argued that the action is an unjustified assault on free speech.
In January a judge in San Francisco ruled that, while he agreed that the agencies are constitutionally protected by free speech, the lawsuit could continue because CalPERS had produced evidence that could potentially establish liability, meaning that both agencies allegedly made factually dubious claims about the investment ratings.
In other words the ratings agencies underestimated likely dangers (the securities they were recommending contained toxic crap) and overestimated the unlikely ones (if we don’t give this client a AAA rating we’ll lose market share). It turns out that even those entrusted with using data to offer copper-bottomed judgment were sticking their fingers in the air – just like the guy at the Red Lion.
This is a slightly longer version of a story that appears in April’s WIRED.