Nassim Taleb has become a prominent talking head with the advice to invest mainly in Tbills, but then a modest portion in something with large and unquantifiable upside: the potential next Google or Harry Potter. I think this is foolhardy, because uncertainty is correlated with volatility and factor loadings, and historically, earnings estimate variability, high volatility, or high beta stocks, all underperform, statistically (ie, on average). So too for the biggest longshot horses, or highest payoff lotteries, call options, movies, etc. Things that can generate lottery-like payoffs, have lottery-like returns.
The annual per capita lottery expenditure in the US is about $150, and the rate of return is about -47% per dollar played. These investments clearly cater to what is commonly called those seeking risk, or positive skew, in particular. A study of the popularity (sales) of lotteries found that average payout (expected return) did not matter, but the size of the top prize was highly significant in motivating popularity. In other words, the $100 million super lotto has a lot of sales even though the probability of winning is so small it basically is outside the realm of intuition (1 in 150 million). A St.Louis Fed article finds that those with household income of less than $25,000 spent twice as much on lotteries on a per capita basis, than those with a household income over $100,000. Such bad decision making is probably not orthogonal to their poverty status. Investing in lottery tickets, however packaged, is a great way to stay poor.
So I found it interesting that the otherwise sober Arnold Kling fell prey to the bias of finding these highly volatile--and uncertain--investments:
What if, instead of borrowing, students could arrange for investors to pay their college bills in exchange for a fixed percentage of their future income... If I were an investor in this market, I would be inclined to make Black Swan style bets. Philosophy majors, for example. Sure, a lot of them will end up as pathetic adjunct professors. But some of them will eventually apply their intellect and creativity to business.
to reiterate, in a different post he noted:
I'd invest in philosophy majors! (But I still think they have the best option value, and that a few of them will be very rich in a decade or two)
Basically, Kling hypothesizes that those wild, free thinking philosophy majors will have a higher average income because of a few 'whales' in their sample. A couple Steve Jobs or some other dreamer. Kling has fallen victim to the allure of hope, in that for things we have no data on, the classes of investments with conspicuous winners seem attractive. Thats why every millionaire from Ebay or Amway, or home forex trader prominently shills a 'system' and says you too can be rich 'just like me'. It's a nice thought, usually packaged with lots of motivational speaking themes, stuff from What the Bleep do You Know, etc. But its all untrue. These are scams that make money for those selling hope . There are a couple success stories to be sure, but they are exceptions, and not sufficient to make the mean look good, any more the success of famous athletes and rappers implies it is a financial advantage to be a poor black kid.
Of course, I can't prove this (I bet it's in a paper somewhere), but the bias suggests Kling is wrong because he thinks too much of conspicuous successes, and using the dreaded 'availability hueristic', does not adequately represent the mass of philosophers who end up wearing a name tag at work. Gambling, Buffet would note, is investing based on hope as opposed to an edge, and hope is anything but audacious, it is all too common.
11 comments:
while skew preference might be irrational from an expected value viewpoint, would you go without car/house insurance? i mean the odds are with the seller here too, right?
immunizing against bankruptcy probability is part of maximizing one's geometric return. It's idiosyncratic risk that insurance companies can mitigate via the magic of diversification, and so its win-win. Plus, lotteries suggest people like positive skew, while insurance is paying to eliminate negative skew.
"I think this is foolhardy, because uncertainty is correlated with volatility and factor loadings, and historically, earnings estimate variability, high volatility, or high beta stocks, all underperform, statistically"
If you've read Taleb, then you know you've committed a ludic fallacy.
I love the way the Taleban thinks merely referencing The Oracle's neologisms is somehow compelling.
The bottom line is that people have been studying uncertainty for a long time, and practical examples as to an uncertainty metric invariably are positively correlated with traditional measures of variance or factor loadings. For example, the disagreement of analysts was considered a good proxy, or trading volume as a function of shares outstanding, or parameter uncertainty, or kurtosis, or jumps--all strongly positively correlated with simple variance, and all give the simple suckers relation between risk and return. If you have a metric of uncertainty that is uncorrelated, or negatively correlated, with variance, I'd love to hear it.
Of course, you can define relevant uncertainty as only that which is unmeasurable, but then you are saying something grammatically correct but meaningless.
My main reason for investing in philosophy majors is that they tend to be very intelligent, and intelligence correlates with income.
Yeah, engineers and scientists are schills anyway. It's all smoke and mirrors. Partial differential equations? Calculating eigen values? Fourier transforms? Any idiot can do that. Philosophy on the other hand....well perhaps philosophy majors are intelligent? They chose an easy major to inflate their GPA and party during college whereas the rest of the suckers were stuck in the lab or library till all hours of the night.
I saw a statistics floating around that the average starting salary of a UNC pyschology undergrad was $100K+. Some guy named Michael Jordan went there?
Arnold: I guess it depends how much logic is in the core curriculum of philosophy. It can be more like sociology, or math, depending.
Intelligence would correlate better to income if you left the Philosophy majors out of it. Talk about a group that has no business sense whatsoever.
There would be a case in here somewhere... However, you defeat your own purposes by the abuse of rethoric.
In this post, rather than lamenting that "of course, [you] can't prove" a point and berating the commenters ars "Taleban", it would have been more suitable to actually try to prove the (lost) point that was made.
j: I can't prove philosophy majors don't have the highest earnings. There is a lot of data, and I have mentioned much of it, noting that volatility and correlates with risk or uncertainty either have zero or negative correlation to returns, which suggests investing based on hoping for big ones (black swans) is a bad strategy, unless you are merely trying to invest on hope because there's a chance you can win huge. It's a sucker's game.
As per inferring 'Taleban' is somehow unsavory, I am referring to the commonality with the distinguishing characteristic of the Taliban, their true-believer thinking, and also playing off Taleb's name, of course. If you take away this is an ethnic or racist slur, it shows how lame Taleb acolytes are, hoping the 'racist' label will overcome their absence of evidence. The Taliban is mainly Arab, but partially Persian and Pakistani; they are muslims. Taleb is a Greek raised in Lebanon and not muslim, so the only connection might be that Taleb understands arabic too, but that would be a strange slur that makes no sense ('he hates Arab speakers!'--ORLY?). It's like me criticizing someone for calling me 'Frankenstein' by saying that is anti-semitic because 'stein' is often in Jewish surnames, a painfully obtuse objection
I'm still waiting for an example of uncertainty, risk, or negative skew, that is generally positively correlated with returns (ie, on average).
Perhaps buying cat bonds (so far, at least)? Not saying I disagree with you, but I think one could find contradicting examples
Post a Comment