The global financial crisis had nothing to do with “black swans” or “fat tails,”or even “black swans with fat tails.” As the Basel Committee on Banking Supervision said last year, financial regulations were too micro-prudential. They argued rules didn’t think enough about the economic cycle, the credit cycle, liquidity and funding, and didn’t focus enough on systemically important institutions.
That was a strong and admirably quick analysis from regulators and supervisors to identify what went wrong during the financial crisis. However, the problem with blaming the financial system and regulations is that it does nothing to capture the imagination of taxpayers who have been left to fork out trillions to bail out bankers. They do not want to hear about what went wrong with the system. They want retribution!
What we are now seeing is politicians, as they do in every crisis, adopting what I call the “bad apple theory.” This theory says a crisis was not the fault of the system or cycle, but was caused by bad bankers selling bad products to bewildered consumers out of bad jurisdictions.
It’s a popular idea because it gives you someone to blame. Unfortunately, it leads to the wrong analysis and reforms. Yet, if financial crises were caused by the bad apple theory, they would have occurred randomly throughout history. But financial crashes are not random. There is, of course, mischief in finance. But, as Warren Buffet said, “You only find out who is swimming naked when the tide goes out.” People swim naked in finance all the time, but they don’t cause the crash – the crash reveals them.
Yes, mischief should be reduced, should be outlawed, but it will not save us from crashes. Financial crashes always follow financial booms. That tells you something. It tells you that you can’t deal with the crisis unless you deal with the boom. The credit mistakes that lead to crashes are not made in the crash, but during booms.
The fundamental problem of crashes, and this is the 85th financial crash we’ve experienced, is that we underestimate risks in the boom and overestimate risks in the crash.
SO, WERE THE RISK MODELS fundamentally wrong? Yes, but not in the way people think. I call the classic model of risk and risk management the “Morris Minor” theory of risks. This is to remind people that the 1950s, the era of the no-frills Morris Minor car, was the age when the Modern Portfolio Theory was developed. It was when Markowitz and Dantzig were among the few people with computers, so were part of a select few with an international database of risk and return, enabling them to identify an efficiency front to optimize returns for a given amount of risk.
Yet, the 1959 Morris Minor world was a very different one to today. Then, we had segmentation of markets. We had Bretton Woods (a system of rules that governed commercial and financial relations among industrial states) and capital controls. Today, we have moved to a world where most homes have at least one computer so that, in half a second, you can simply go to Bloomberg and get the same information as everyone else.
Yet, it is as if economics is stuck with Newton and hasn’t yet caught up with Einstein, not to mention Heisenberg and the uncertainty principle, in which the very action of observing something changes it. Today, the risk-return trade-off is observable; it is not exploitable.
You can fiddle with risk models, fatten the tail, bring in black swans, but what you are missing is risk created by the system itselfWhen you identify an asset that appears to offer a slightly higher return than past risk trade-off, you can be sure you are not the only one to recognize it. Everyone is identifying it at the same time. The asset becomes overvalued because the crowd is all coming to the party at the same time. The asset is no longer low-risk, but high-risk because as volatility begins to rise and as risk models confirm that volatility is rising, you sell. Unfortunately, everyone else is also selling, creating more volatility.
When your risk model lets you know that the risk is higher than initially perceived, you sell other assets. As you sell other assets, you begin to see uncorrelated assets behaving in a correlated fashion. So volatility goes up, correlation goes up, price goes down and the risk models have a fit and tell you to sell everything.
I don’t think there is an alternative to using traditional risk models, so we need to use the models we have and add on to them. What should be added is a measure of crowding in markets. The more crowding you identify, the more you know your risk model has got it wrong; if there is no crowding in the markets you are in, then your risk model is probably perfect.
So identify markets with a large buildup of investors and assume crowding. Then put a large discount factor on what risk models are telling you: correlations and volatility will be higher and returns lower. Wherever there is no crowding, you can assume your risk models are telling you something of value.
However, risk models find it hard to get it right because the issue they face is not exogenous (external) risk but endogenous (grown from within) risk. You can fiddle with risk models, fatten the tail, bring in black swans, but what you are missing is a fundamental risk. What you are not taking into account is risk created by the system itself – by us all having access to the same information and being forced to follow similar patterns of behavior.
Systematic fragility comes from common behavior. This implies that we need to think carefully about strategic behavior. To be fair, this is actually found in books on risk management, but only in a preface page no one ever reads. Roman numeral I: “These models only work assuming independent behavior.” There is no individual behavior in finance; it is all strategic.