esoteriic.com

Tail hedging is about buying insurance to protect against catastrophic uncertain events. In this article I will outline why I think it is impossible for everyone to do so. The underlying argument is that someone has to eventually pay the piper (be it the claimant, insurer or re-insurer) and that for catastrophic events there isn't going to be enough money to cover all claims. Some tail events can be insured and "easily" covered in isolation, but when all claimants try to claim at the same time then is likely to fail. Moreover, the cost of such insurance will increase when the general sentiment is that of fear.

Tail Events

Tail events are those which are thought to have almost zero probability but have a huge impact: aka a black swan event. The probability is considered incalculable rather than zero, such that saying almost anything about the so-called probability is meaningless. The name tail does come from the notion of the event being in the tail of a statistical distribution. Generally, such events are connoted with negative outcomes but it isn't always the case.

Financial Crisis

The market crash of 08 / 09 was not seen by many, and even those that predicted could be branded as "doom sayers": if you talk about crashes then eventually you have to get it right. This crash can rightly be called a tail event since it was not expected by most people. Such events are characterised by exuberance and the expectation that good times will last forever. Not only is a crash not expected but many may even try to "prove" that such an event is statistically impossible, and they such events occur with greatly regularity than would be predicted.

Black swan events are not limited to financial crises, although most of the literature is related to finance. This means that the theories and concepts could be applied to all industries and perhaps should be, especially important for people dealing with risk management. Therefore, it is possible in theory to purchase insurance that protects against such events. The catch is that the cost of the damage in such events is impossible to calculate: the damage will be great but it is impossible to say with any certainty how great that damage will be.

This makes it a nightmare for insurance companies to determine how much money is needed to cover such claims and therefore how much of a premium should be charged. The liability is essentially uncapped but the premiums are always a finite amount. Insurance companies also need to be consistently profitable, and that they themselves should survive a crash. That is to say that it would be preferable if the insurance company was external to the black swan event.

Cyber-Security

In the last year there has been has been a lot of cyber criminal activity which has caused a headache for organizations globally. Insurance companies have been trying to get to grips with insuring against breaches but from what I've read there is great uncertainty in how to perform the calculations.

While the general techniques to hack into a company are well known, but the specific exploit used is not necessarily known in advance. It is considered impossible to be 100% secure such that no systems can ever be breached. Once a computer has been illegally accessed it may be possible to further exploit other computers on the same network or retrieve sensitive data. Data can be easily duplicated and shared with great ease too. Naturally, these are all the necessary ingredients for breeding black swan events.

If we are talking about one particular company insuring against a data breach then it should be possible for the damage to be mostly contained. If the company has no insurance then it may go bankrupt, or if it has enough insurance then it can survive. Despite any difficulty in the calculation of insurance premiums it is clear that in this relatively simple scenario the insurance company is external to the damage.

The biggest problem would be if all clients had a massive data breach at the same time, and consequently all made a claim. Not only is that a black swan event for each company but it would be a black swan for the insurance company too (which would no longer be external to the problem). This is the heart of the problem: the insurance companies may not always be external to the damage such that they can survive in all circumstances, and was exactly what we saw in the crash of 08 / 09. Black swan events and thinking about risk has become a hot topic in its aftermath.

Since the crash

Since the bottom of the financial crisis in early '09 there has been a lot of talk about risk in the financial industry. More products and services are being created with a risk focus. This is not a surprise given the proximity of the crisis: after a crash the talk of recession, depression, inflation, etc. Talking about black swans and tail events in the investment world is now 'cool' and 'informed' but generally such conversations have missed the point. Simply purchasing a product off-the-shelf in order to hedge a tail risk is likely to fail to do its job. If we are talking about financial crises then it is clear how interlinked the world is such that if the major markets stutter then they will stutter together and that all companies, including the insurers, will suffer. To that extent it seems a challenge for any product to be able to claim to protect during such times.

If we are talking about hedging specific tail risks which are isolated to a single company then there might be some merit in purchasing such a product. The scope of the product should be obvious at time of purchase. Buying an insurance product to protect against a data breach could be possible, but buying an insurance product to protect when all markets fail then it seems naive to expect an off-the-shelf product to do this. For the informed readers that already know about black swan events and the writing of Nassim Taleb then they may be thinking "surely it is possible to hedge tail events, such as the 08 / 09 crash just as Taleb did".

Yes, it is possible but here's the catch: tail hedging works best whenever everyone else is expecting the boom times to continue. Not only did Taleb protect his investment portfolio but his strategy took advantage of the crash and made a profit. The details of his strategy are not wholly appropriate for this article; however, the crux is that he created an insurance-like that makes money whenever markets become volatile. He said that the boom times could not last forever and that there were deep structural problems in the financial industry. Whenever the crash occurred it was not isolated to a single company nor industry, but all companies and all industries.

Why can't someone copy his strategy and be safe in the knowledge that they are protected? To start with, he didn't purchase an off-the-shelf product but purchased a certain type of "options". Moreover, he purchased these options cheaply whenever everyone else thought he was nuts. The price of the options did not reflect the coming crash: had the options underwriters, and the market as a whole, see the coming crash then arguably his options would have been far more expensive. It could be further argued that the options would have been so expensive that his strategy would have never paid off. This is getting to the rub of the argument: whenever everyone is expecting a crash then is natural for the prices of these particular options to become more expensive.

This has been observed in option pricing since the crash: they appear to price in a lot of future volatility. This is the so-called implied volatility of options: the prediction of volatile prices in the future. Options underwriters are essentially insurers: they provide options in order to make a profit. If the prospect of them underwriting a loss is uncapped then they will clearly expect to charge a high fee. Once more making it very difficult, if not impossible, for everyone to buy a ubiquitous tail hedging product and expect to be fully protected when a crash comes.

Conclusion

This suggests that all risk mitigating products will an increased cost whenever the implied probability is "high": the fear of a near-time crash will push up the price of premiums. I suspect this will be true for insurance in general, but very sure it is true for purchasing options or funds / ETFs that have an options component to protect the underlying assets. I've seen or heard of several ETFs that offer protection on the underlying assets (using options) such that if the price was to drop by 30% then the insurance component of the fund would kick-in and prevent a further fall.

Where it makes sense to purchase tail-event insurance: when the volatility is low, the cost of "insurance" is cheap, and the suspected fall-out is not expected by the masses, or ... of course... if there is a regulatory requirement.

Summary

* Tail events are hard to predict: both on timing and impact.

* Calculating the premium to pay for such events is hard because the impact is not known in advance.

* Insuring / hedging against tail events IS possible for specific cases but not likely possible for systemic wide-spread failure.

* When everyone is thinking and talking about risk, the premiums will increase due to a general feeling of risk aversion.

* Caveat emptor for all off-the-shelf products that offer "protection". Either they fail to protect or the costs mean that they fail to be efficient.

Addendum

Also be aware of anyone selling a product which fully considers all risks, or claims to have all risks factored in. One data provider whose name shall not be mentioned had one of their team tell me that their newly upgraded risk analyser could fully account for all events. I can't remember the exact wording but there was the implication that it would be accurate in tail events. It felt like I was talking to someone who should have known better, and it felt like he was expecting me not to know he was bullshitting. It really wasn't clear if it was a sales pitch or if the person genuinely believe in the product. Tail events are by their nature unpredictable, ergo all such "risk" analysers are doomed to failure.