Why you can’t hedge tail risk

By Felix Salmon
July 1, 2011
Azam Ahmed's big article on tail-risk funds got a lot of play in the NYT yesterday, but I don't know why.

" data-share-img="" data-share="twitter,facebook,linkedin,reddit,google" data-share-count="true">

Azam Ahmed’s big article on tail-risk funds got a lot of play in the NYT yesterday, but I don’t know why — he gives us no evidence to believe that they’re large, important, or even growing particularly fast. There aren’t many hard numbers in the piece, but the ones he does supply seem small to me:

Products linked to an index known as the market’s “fear gauge” total nearly $2.5 billion. And in the last year, the amount of money managed in dedicated tail-risk accounts by the bond giant Pimco has doubled to $23 billion.

Boaz Weinstein, a former trader at Deutsche Bank who lost more than $1 billion of the bank’s money during the financial crisis, began raising money for his own Armageddon fund late last year. It has since grown to $400 million of mostly institutional money, part of the $3.3 billion he has raised for his hedge funds.

The only remotely significant number here is Pimco’s $23 billion in “dedicated tail-risk accounts” — but I’m a bit suspicious of that number, given that Ahmed claims it has doubled in the past year. If you go back to last year’s iteration of this article, from Bloomberg, it was pretty clear that the dedicated accounts hadn’t even launched yet:

The Pimco Tail Risk Hedging Fund 1 will be the first in a potential series of partnerships, according to a private placement filed with the U.S. Securities and Exchange Commission on June 23. The initial fund will be designed to protect investors from a drop of more than 15 percent in a benchmark index that Bhansali declined to identify.

So it looks as though Ahmed is including, here, funds which are designed to go up in value in good times and which simply incorporate a certain amount of tail-risk hedging — funds like the Pimco Global Multi-Asset Fund. That’s emphatically not a dedicated tail-risk account which offers costly insurance against extreme market events.

Ahmed’s article, then, insofar as it purports to be reporting actual news, rests mainly on very vague statements about how “investment professionals have a new pitch,” or what “clients are suddenly realizing,” or that “Wall Street lawyers say money manager clients have approached them in recent months about forming new funds aimed at providing protection.”

And as an analysis of tail-risk funds, it’s weak. There are pro forma to-be-sure grafs about how “protection does not come cheap and occasionally fails to work”, but there’s no indication that tail risk hedging is, conceptually speaking, pretty much impossible. Emanuel Derman gives one reason why: tail risk is not a simple and identifiable thing which can be insured against.

The value of a portfolio can be substantially destroyed by more than one cause. Portfolios can be ruined by equity crashes, credit spread widenings, bond defaults, high interest rates, sustained inflation, increases in volatility, illiquidity, etc. To protect your portfolio against so-called tail risk may require spending money on insurance against all these risks, and there is no panacea here.

David Merkel gives another: if you’re buying insurance, you need to be sure of your insurer’s ability to pay out in the event of a catastrophe. But in the markets, your insurer is precisely the sort of institution which is likely to be going bust as a result of some unexpected tail event. Writes Merkel:

In order for tail risk to be mitigated fairly, someone must keep a supply of slack high quality assets. Rather than the insurer doing that, why not have the investor do so? The insurer brings along his own cost structure. Why not self insure and bring down the risk level directly. Someone has to hold high-quality assets to mitigate risk; let the investor be that party; embracing simplicity and enjoying reduced risk without the possibility of counterparty failure.

This is pretty much the same conclusion I came to last year: the best way to hedge against tail risk is to put 90% of your assets in Treasury bills or other super-safe assets. This is not an extreme strategy at all; indeed, it makes perfect sense for anybody who’s both rich and conservative, including Suze Orman (“I have a million dollars in the stock market, because if I lose a million dollars, I don’t personally care.”)

The problem with tail-risk hedging is that everybody doing it wants to get healthy returns during good years and then have very little downside risk during bad years. It’s an impossible combination to achieve, and although Wall Street will happily sell tail-risk hedging products to those who want them, there’s absolutely zero evidence that they actually work in practice.

So if and when Universa comes out with a black swan ETF this month, my advice will be to stay well away. Indeed, insofar as it helps makes investors overconfident, on the grounds that they’re hedged, it’s likely to be positively dangerous — just as portfolio insurance was in 1987. Tail risk, it turns out, is just too amorphous to hedge. Sorry.

More From Felix Salmon
Post Felix
The Piketty pessimist
The most expensive lottery ticket in the world
The problems of HFT, Joe Stiglitz edition
Private equity math, Nuveen edition
Five explanations for Greece’s bond yield
Comments
19 comments so far

I would want to distinguish between an event like the 1987 crash (or LTCM blowing up) as compared to a prolonged bear market like after the tech bubble burst. In a CVaR optimization framework, depending on your preferences, it may make sense to pay a little extra to reduce the amount you could lose in a 1987-type situation (of course, depending on the volatility and shape of the tail, you may pay more or less in different periods). An even like a prolonged bear market, can be better thought of as a separate regime (see Ang & Bakaert in 2002). In this situation you might still want to have some tail risk hedging, but you also would want to significantly reduce equity exposure and increase bond and cash exposures as well.

Posted by jmh530 | Report as abusive

No such thing as a super safe asset. In times of high inflation you’ll be losing money hand over fist in T-Bills, not to mention rollover costs and assuming you can find ultra-short end government credits.

Posted by Danny_Black | Report as abusive

“Treasury bills”? My brother recently told me that Ron Paul is too status-quo oriented, and what we need is a currency credibly backed by ammunition.

But, yes, anything that does well during good times is suspect. Also, anything that does well during mediocre times. “More money has been lost reaching for yield than at the point of a gun.” — Raymond F. DeVoe, Jr.

Posted by dWj | Report as abusive

“the best way to hedge against tail risk is to put 90% of your assets in Treasury bills or other super-safe assets”

That’s not really a hedge, it’s more like putting it under a very secure mattress. Wouldn’t a hedge provide some upside during a down period to offset some of the losses that occur when the unexpected happens?

Posted by KenG_CA | Report as abusive

In property, they have sold “all risks” coverage so I suppose one need not identify any tail risk, just that there is risk. I wonder if the biggest question is whether that insurance will pay. It won’t if the tail risk is the sun blows up, but then you won’t be making claims either. The question, I think, is whether it’s sensible to buy the insurance. How do you price it? You as provider and you as buyer. If it’s sold as a claim on a stock of expected ultra-safe aspects, then it has to be expensive: you’re buying into a pool, aren’t you? That pool can’t do much because more return would put it at risk, so the expected yield should be low – unless you’re being sold fake ultra-safe stuff.

How do you as a buyer know a price? The next tail event might not affect you. It might wipe you out anyway. What if you’re hit in one area but not the area you’ve insured? Do people buy insurance for meteor strikes? How many buy trip insurance or flight coverage in case of a crash? We do all sorts of risky things everyday and yet most people don’t have much life insurance. If you have lots of money, maybe it makes more sense to buy art or jewels. Too many questions for this to be sensible.

Posted by jomiku | Report as abusive

The number of US-based institutional investors that investigated (and continue to investigate) tail-risk hedging is absolutely massive. Having very recently left a role as a derivatives specialists (listed index option-based strategies) at a $300B asset management group I can assure you that the number of accounts that we ran some form of tail-risk overlay for become substantial. Yes, the dedicated collateral to those overlays in tiny because of the efficiency of the margin requirements attached to listed index options, but the notional values they are structured to protect are substantial, and growing.

There is no such thing as a cheap put, true, and that is what many misguided clients came to us looking for. Still, the idea that the industry hasn’t evolved to the point where large asset allocators are now sophisticated enough to evaluate and implement option-based hedging approaches just isn’t correct. From targeted volatility funds to using options spreads like collars, put-spread collars, and condors to fund long put positions. These strategies have real and growing traction.

Posted by BRM_3 | Report as abusive

James Montier at GMO published an excellent and brief white paper titled “An Ode to the Joy of Cash” analyzing various tail risk strategies, including cash. Its a good read, with real numbers. The pdf can be found on the GMO website.

Posted by jose_invests | Report as abusive

You’re right, of course. It seems so obvious…

The fact that this is even being discussed now is proof that any effort to promote insurance against catastrophic risk is the triumph of greed over experience. There needs to be some corresponding catastrophic punishment for non-performance, like: if there is a catastrophic event and you can’t pay the insurance you have sold, you will be drawn-and-quartered. We’ll see how eager hedge fund managers are to pitch that.

Posted by jbernar | Report as abusive

Maybe they’re thinking that in a Very Large Crisis, the Govt bears the Risk. So the key is to invest in Risky Investments that the Govt has some reason for saving should things go bad.

Posted by DonthelibertDem | Report as abusive

I’m glad you highlighted the ridiculousness of the Black Swan ETF, since it’s a financial oxymoron.

I would love to hear what they’re buying to protect themselves from extreme events, and then provide a hypothetical event that would destroy their value.

Posted by djiddish98 | Report as abusive

It seems to me hedging with TBills is an alternative to investing as it is an expression of an increased marginal propensity to hoard: http://cobblehillbilly.blogspot.com/2011  /06/marginal-propensity-to-hoard.html

I suspect this is why Bill Gross: http://www.creditwritedowns.com/2011/06/ gross-education-debt-and-government-as-l ast-employer.html

and Barton Biggs: http://www.creditwritedowns.com/2011/07/ the-big-interview-with-barton-biggs.html

have both taken to talking about public investment and jobs: the private sector can not identify a realistic demand pull for future returns on investment and is thus increasingly hoarding wealth in bonds that create nothing but fiat returns.

Posted by jnewman | Report as abusive

You do not really want to do this for the rest of your life, do you?! Did they promise you something really juicy? A chalet on Candy Mountain? All the stinky cheese you can eat? What? Wouldn’t you really rather be a tech evangelist or blogger over at Boing Boing?

Posted by Uncle_Billy | Report as abusive

My biggest fear is that black swan derivatives will become popular without the appropriate safeguards. Some Black Swan derivatives, for example, bet on the collapse of China. Sino-Forest shows that China is still quite opaque for the level of FDI they are getting. So, what if they’re banks become extraordinarily insolvent due to the popping property bubble and China has to bail them out, causing either hyperinflation or much higher Chinese bond yields?

In that case, any sort of derivative betting on the event acts as a weapon of mass financial destruction, as Buffett put it. Banks have to be properly hedged for the event and I’m not sure Dodd-Frank can stop another Howie Hubler loading up on supposedly AAA CDS’s or put option writing on Chinese companies. In that case, I guess we’ll find out how well the living wills actually work.

Posted by mwwaters | Report as abusive

I agree that tail risk can’t be hedged.

I think a related topic of great interest the next few weeks will be what is truely the worlds most investable ultra low risk asset. Curently the market says it’s U.S. T-bills and German Euro Bonds.

I’m on the other side of that trade… put my money into ultra-high quality equities which generate meaningful %’s of their earnings in several currencies and who’s dividends (untaxable to me in IRA’s) are higher than the 10 year treasury. Like TFF I’m more worried about the dependability of the cashflows than the day to day volitility of the principal.

Posted by y2kurtus | Report as abusive

mwwaters, the irony of course of you mentioning Howie is that his trade was predicated on a crash in house prices. He was the one making a “Big Short”. Of course he didn’t want to be losing 100s of millions a year for what would have been 3 years so he funded it with writing insurance on AAA tranches and when correlations went to one he lost a bundle.

Just goes to show you can be right about the future and still lose money.

Posted by Danny_Black | Report as abusive

It looks to me as though there is a “tail risk hedge,” except that I don’t fully understand how to pull it off. If I did, I’d consider doing it myself and building near risk-free wealth.

It involves first becoming a well-known fund manager, but not necessarily one who can survive both up and down markets. Then somehow, after you blow up, like Meriwether or apparently like this Boaz Weinstein Felix mentions, you persuade people to set you back up in business anyway, and you still have a nice income.

Posted by SelenesMom | Report as abusive

It’s clear from looking at the comments, including my own, that we’re talking about two different forms of tail risk. One is the 3 standard deviations move from a price. That can be hedged because it can be modeled and priced. Another is what are now called black swans but which also be called that which the model doesn’t predict can happen; either the probability is too low, the risk is already hedged in the strategy being modeled or it simply can’t be anticipated. I imagine one can hedge one’s hedge one’s hedge but one can’t overcome the biases inherent in models that make one blind to unanticipated, even unimagined risk.

Posted by jomiku | Report as abusive

I think there are two parts to the question,

The first part, hedging tail-risk of a defined distribution, is what Wall Street claims to have lots of products to do. They accomplish this with varying success.

The second part, is the “black swan” type of event which is a completely different animal.

How do you hedge against the Weimar Republic hyper-inflation, the destruction of your country like most of Europe and Japan saw in WW II, the take-over of your country by a foreign power, such as the Iron Curtain countries, internal revolution (France late 1700s, Iran 1979), the collapse of a banking system (US 1929-33), civil war (US 1860s), etc.

These types of events require “products” much different from anything that Wall Street purveys. Jews in Germany in the 1930s needed Swiss bank accounts, diamonds, and an exit visa from the country, similarly for the ruling elite in Iran in 1979 and France in the 1700s.

Survival in the early years of the Great Depression required no debt and lots of cash under the mattress, not in a bank. Even owning farmland didn’t necessarily help due to the massive drought and topsoil loss, partially due to poor farming practices.

Ultimately, hedging “tail risk” can involve everything from having some shorts to owning a shotgun and shells under the bed along with cans of food and gold pieces.

Bill Bernstein at http://www.efficientfrontier.com did a good piece on tail risk in about 2000 when he estimated that the best that the various retirement calculators could do for actual confidence level was about 80% instead of the 95% that most touted because the potential of major events like war and societal collapse during a lifetime does not show up in any MPT statistics.

Posted by ErnieD | Report as abusive

many years ago, my dad read an amazing statistic in, I think the NYTimes, but it might have been some other, at the time well respected, source of news.
The statistic was that 30% of vietnam vets attempted suicide – which if you think about it for a few seconds, is an amazing number.
My dad snailmails (this is way, way before the web) the reporter, who sends him to a “source” who sends him to another source…eventually, my dad gets a Prof of Clin Psychiatry at the San Diego Veteran’s med center on the line, who, according to my dad, starts screaming, those bastards, they’ve been misquoting me for months, its 30% of Viet Vets hospitalized for acute psychosis who attempt suicide..which is roughly the same as the general population.
Or, never attribute to motive what can be explained by stupidity, or as Pascal is reputed to have said, it is easy to understand infinity, just contemplate human stupidity

Posted by joeenuf | Report as abusive
Post Your Comment

We welcome comments that advance the story through relevant opinion, anecdotes, links and data. If you see a comment that you believe is irrelevant or inappropriate, you can flag it to our editors by using the report abuse links. Views expressed in the comments do not represent those of Reuters. For more information on our comment policy, see http://blogs.reuters.com/fulldisclosure/2010/09/27/toward-a-more-thoughtful-conversation-on-stories/