A professor of mine has a recent blog post where he wonders why executives take risks that academic finance has known for decades don’t add value:
Whenever I meet a senior executive from a company that got in trouble because of too much leverage, I ask (nicely, I hope) whether they know MM, which implies that investment decisions should not be made based on capital structure. Most say no, but they are very interested to hear about it.
This is stuff that has been known for a long time.
This got me thinking about the value of the knowing things. My visceral reaction is that learning through study and research has to be valuable, but sadly my post-MBA experience has largely shown otherwise. And when I say valuable, I mean in a career sense since to me knowledge has high intrinsic value.
There are several reasons while I think “knowing things” is not valuable as it should be outside of academia:
1) People prefer (over) confidence
Knowing something could be unwise or dangerous gets in the way of being supremely confident. Those who have no inkling that a particular action could be problematic can advocate for it without expressing any doubt. The persuasive ability of complete confidence has always seemed like a cognitive bias to me but I’ve seen it work over and over.
2) Tail risk pays off in the short run. Or maybe more bluntly put: The incentives reward stupidity
Let me start by explaining tail risk. Tail risk is the negative impact from a “rare” event in the probability distribution. The formal definition is a 3 standard deviation event. It’s easy to make money most of the time on tail risk. One can write a way out of the money call or put option and pocket the premium. Until the rare extreme swing to the stock happens, this strategy produces a steady stream of money. When the rare event happens, the loss can be enormous.
Overlevering a company has the same effect. Most of the time a company can service its debts. But a particularly bad downturn or other unexpected event can push the company to bankruptcy court. Since in most years we do not experience a bad downturn, the company’s returns look good.
In reality, that bankruptcy risk is reflected in needing higher returns on the stock. The Modigliani-Miller theorem says the “savings” on the “cheap” debt is exactly offset by higher required return on the equity to compensate for the increased risk in other words: on a risk adjusted basis, there is no extra return for loading up on debt. Like all finance theories, there are assumptions like frictionless markets and no taxes that aren’t met in the real world, so in reality it’s not exactly offset and tax treatment favors moderate debt.
But in the short run, the equity returns look higher and these managers are likely to be rewarded for that.
3) It’s downer
Knowing what I just explained is a downer since most of the time everything is going well and voicing concern about over leverage, is not going to be unpopular. After all the risks pay off now most of the time, and explaining there’s no real value being created and those extra returns are not actually better in the long run just annoys those who don’t understand this. All this worrying about disaster sounds like Chicken Little.
The stories I’ve heard about the mortgages security practices in the great real estate bubble are even worse than this. The loan issuers and resellers knew the loans were extremely risky and a bad long run bet, but were making so much money issuing or trading them that they didn’t care. The resulting disaster was so bad that many of the businesses in this mortgage backed security chain went bankrupt in the bust due to their exposure to these assets they knew were bad. But when everyone is making huge bonuses taking these risks, anyone who points out the risks is probably making a career limiting move.
4) Over emphasis on prior experience
Both in hiring and in the decision process, most people over emphasize their past experience. The problem with this approach is that the experiences of a career span is not a full picture or as we might say in more technical language, a representative distribution. Even 30 years experience in a field is a short sample of possible risks. Understanding the full range of possible outcomes and thus risk exposure requires study, often of the academic kind.
The problem is even worse for rare events that don’t show up in the historical data since history based quantitative methods do not account for those risks either. To understand those kinds of risks, one must take an approach based on theory rather than experience.