Wednesday, July 22, 2009

Betting on Black Swans

Nassim Taleb has become popular, though I don't think he has anything really profound to say (see my book review here). Indeed, a reviewer of my book Finding Alpha lamented I did not mention Taleb, but I did not see the point because I was making a serious point about asset pricing theory and Taleb is a pedestrian populizer, who like all popular populizers, is successful at convincing a lot of people he is saying something new and true. This is not the same as actually saying something new and true. His book does not add any new data, or theoretical insight, to this corpus of knowledge (e.g., the Rietz-Barro Peso problem). But as funds affiliated with Taleb are becoming popular I think it would be helpful to address those specific strategies. The basic premise is that financial economists, and investors, systematically neglect improbable events. The result is that out-of-the-money options, especially qualitative analogues like unconventional investment ideas picked up at cocktail parties, offer the best reward-to-risk ratio. Let's consider these.

You can invest in a Black Swan fund that buys out-of-the-money options such as Universa Investments run by his former partner Mark Spitznagel. Taleb, and Spitznagel, would argue their implementation is much more sophisticated than merely buying out-of-the-money options in that it also takes advantage of 'behavioral biases'. Now, as 'prospect theory' implies people ignore, or overweight, improbable events, such 'behavioral biases' allow a strategy a great deal of latitude. In practice simple ideas, such as the underpricing of out-of-the-money options is too simple to sell, so the vendor feels compelled to confabulate a pretentious but useless tweak. In this case, that just means one sells in-the-money options to lighten the expense (more gamma, less vega). Do the math, and this strategy simply shifts your payoff distribution so it has a funky nonlinearity, shifting returns from the [-10%,0] space to the [-∞,-10%] and [0,∞] space, but the same expected return. That is, say you buy $2 worth of out of the money put options, and offset this a little by selling $1 worth of in-the-money put options. This means you still hit a home run in the extreme event like 2008 (which was, statistically, improbable); you make less in the more probable adverse event; you lose less if the market rises. As empirical studies have found option returns to decrease as one goes out of the money (see Ni, or Bondarenko, or Coval and Shumway), the in-the-money vol sold is relatively underpriced, so the only behavioral bias this is leveraging is a marketing one. Considering that out-of-the-money options are most overpriced, this is a bad strategy.

Clearly in big moves such as 2008 this strategy outperforms. Yet on average, I doubt it. That is, much was made of the "65% to 115%" return reported in October of some Black Swan Funds (see WSJ here), but the VIX peaked then at 80, and is now at 23, while the market has rebounded. It would be interesting to know the subsequent returns, but everyone merely quotes the hearsay (always, 'a person close to the fund reported...') reported on their top tic. As Taleb is insistent that too many investors naively chase last year's winners, it would only be consistent to not make too much of one year's returns. Indeed, Empirica Kurtosis, his earlier hedge fund, also started out with a widely reported 60% return in 2000, but then folded quietly in 2004. I earlier said I would donate $10k USD to his favorite charity if he sends me audited financials showing the cumulative Sharpe ratio of Empirica Kurtosis LLP over its lifetime was greater than 0.5 (which is a poor hedge fund return, consistent with exit), and the offer still stands. Thus, over the long run, not a sequence of 2008s, this is a crappy strategy.

The other problem with out-of-the-money strategies is that option market makers hate being naked low-delta options. You don't have to buy too many to move the price, implying market impact is high. This makes these strategies even more expensive than any simulation when done in size. That doesn't bode well for these now-popular strategies.

Another claim by Taleb is that 'wild' risks from cocktail party chatter is especially useful. At some level this is like saying things that can't be quantified are really important, which has the nice property of being tautologically untestable, but in practice we understand what this means: funky investments outside of stocks, bonds, bank deposits, etc. Llama farms, gold medallions, and no-money-down real estate, are all unconventional, and all scams. Sure, the people pitching them show someone who got rich doing this, the same way Amway lines up its multilevel marketing operation, it is an appealing idea—why not me? Airport Holiday Inn conference rooms are always full of conferences on how to become rich following their 3-point plan, which invariably involves trading through them, or buying their $100 instruction manual. They have a horrible average return, because 'fraudulent scam' is not rare in this space, rather de rigueur. Staffcentrix offers advice and due diligence on home-based businesses offered on the internet. A principal there, Christine Durst, states that the ratios of scams to real home-based businesses is 54-1 on the internet, most following the plan of selling a modestly priced but useless information brochure, enough to make money, not enough to elicit a lawsuit. But the bottom line is that virtual out-of-the-money options are, if anything, more expensive than those traded on exchanges.

In my book Finding Alpha, I argue that people tend to pay for hope, and nothing offers hope better than the lottery-like returns available in potential Black Swans. Hope is a good thing, and motivates a lot of hard work and creativity. But it would be foolish to think that the more improbable, the more speculative, the more derided by economists, the better the risk-adjusted return merely because of this. This tendency to buy into lottery ticket leads to all sorts of really bad investments:
  • Higher volatility stocks have lower return than low volatility stocks
  • Higher beta stocks have lower returns than low beta stocks
  • out-of-the-money options have lower returns--and higher betas--than in-the-money-options
  • higher leveraged firms have lower returns than lower leveraged firms
  • firms is greater financial distress have lower returns than firms with low financial distress
  • Junk bond mutual fund returns are lower than investment grade mutual fund return over the past 22 years
  • sports longshots such as 50-1 odds horses, have lower returns than favorites
  • lotteries with the highest payouts have the lowest expected returns
  • IPOs have lower-than-average stock returns

[This is in my book, and also summarized on an SSRN paper I wrote here.] Notice a pattern? The more volatile, more uncertain, the lower the return. People pay a premium (accept a lower return) for 'Black Swans', because in one fell swoop, they can get rich and prove they were RIGHT! Instant satisfaction. Plus, if you really have the touch, why waste time choosing Coke over Pepsi, when you can choose between GM and Citi!

Basically one would be better off not chasing dreams via Black Swans, and stick to boring investments. Finding alpha—a risk-adjusted return premium—is very difficult, and it involves a niche specific to an individual's skills, which almost surely is not in investing anymore than the average person has alpha singing or writing romance novels. Black Swan investing is a sucker's game, endemic in markets, a perennial loser, and highlights asset classes to avoid, not pursue.


lance said...

So if beating the market is so easy -- buy low beta stocks with momentum, and hold on -- then why do you say that alpha is so hard to find? "See, beating the market is no problem, just follow these simple steps. And by the way, if you don't have any inside info, forget it, almost nobody beats the market."

Huh? So which is it?

Eric Falkenstein said...

making a couple percent extra by avoiding Black Swans is not a high profile strategy; it won't lead you to the top of a mutual fund winners list; it is not sufficient for a hedge fund strategy demanding a potential 1+ Sharpe. It's like index fund investing, which has Alpha relative to active management. A good idea, I think, but one most find unappealing (they are still only about 15% of equity funds).

It's alpha, but modest alpha. I think it would make an excellent scalable, long-only strategy, and would take any offer to manage such a strategy!

Anonymous said...

There was French cardinal by the name of de Retz who in the 17th century supposedly said/wrote:"Nothing sways the stupid more than arguments they can't understand". I think Taleb is the nihilistic Pied Piper of a bunch of folks who like the sound of what he is preaching but fundamentally have no idea what he is really saying.

Your collection of on average money losing lottery tickets is also a list of "trades". There may not be a systematic pay-off to high yield, but it is the trade de jour. There may not be a positive systematic pay-off to IPOs, but given this market there will probably be an IPO boom. People love to trade, they really don't love to invest.....

Anonymous said...

So then you would advise selling the 2by1 out of the money put spreads that you suggest are poor investments. ie take the other side of the black swan trades that underperform over time. After all if this black swan strategy is a loser than it would be simple to do the opposite and find alpha, no?

Eric Falkenstein said...

A -0.2 Sharpe strategy like this, with high transaction costs, probably has a -0.1 cost taking the opposite side.

Anonymous said...

Isn't your story just another behavioral bias that you keep criticizing?

Anonymous said...


on a diff note, do u have any thoughts on the certifications like "Financial Risk Manager" ( & CFA?

Anonymous said...

Another claim by Taleb is that 'wild' risks from cocktail party chatter is especially useful.

Where does Taleb say this? How can this be conflated with those scams? His example is that you might get an idea for a paper by talking to someone at a party, not unlike say Brian Eno's "Oblique Strategies".

It is BECAUSE your comments and research shows black swan strategies are sucker strategies that they work! That's the irony of the black swans.

Anonymous said...

To be an empiricist, I don't think we can know if black swan strategies "work" (whatever "work" means). They profit from tail events, and by definition there aren't many of those, so we don't have enough data to tell. My guess is that, like any other strategy, it depends.

Anonymous said... received $150,000 liberty reserve
The Mountain Mining Investment Limited is a unique, financial gold, nickel and uranium mining company whose business strategy is to acquire royalty interests in gold

production from its core assets in the Lake Victoria greenstone belt of Tanzania. Since this year we started nickel and uranium exploration.
Why To Invest?
- Registered Mining Company
- High own capital and reserve fund
- DDOS and Hacker protected
- Available worldwide
- Direct offshore investments
- Real Phone and Email Support
- Deposits are secured by reserve fund
20%-50% hourly for 120 hours (min $600-$29,999) (principal returned)
(Total Profit 2400% - 6000% in 5 days)500% per 6 hours for 48 hours (min $30,000-$9,9999)(principal returned)
(Total Profit 4000% in 2 days)150% hourly for 24 hours
(min $100,000-$1Million) (principal returned) (Total Roi: 3600% in 24 hours)
Date: 08/19/2009 16:57
Batch: 13237xx
From Account: U6525495 (Mountain Mining Investment)
Amount: $150,000.00

keep reinvest everyday

Mail to me if you made a deposit use my referral link.I will return 20% of your spend instant.100% guaranteed money back.

Make $50k libertyreserve money daily on forex trading
Read news about hyips.make money fast and easy

More news about paying hyips

Post Your AD on 1000+ Forums
The Best HYIP Rating. The Fairest High Yield Investment Programs Monitoring Service