Nevertheless, it seems possible that the financial success of Paulson and others was a consequence of careful analysis and shrewdness, and that other people of sufficiently high intellectual caliber and rationality would have been able to predict it as well.
I have an economist friend who predicted the tech bubble, but shorted too early, and ended up losing quite a bit of money in the process. Knowing that chickens are coming home to roost is not enough; you need a good idea of when they’re going to arrive, and to have the solvency to be able to stick to your beliefs. (If Keynes’s comment—”The market can stay irrational longer than you can stay solvent”—fits you, then shorting bubbles seems unwise.)
As is always the case in finance, those who recognized the impending pop of the housing bubble kept their analysis secret
I do believe there were people predicting the bubble would burst before it burst, and some of them were even people who don’t predict a bubble bursting every year. It’s not clear to me that a handful more mathematicians pointing out impending disaster would significantly shift public opinion, and hoping that the government would act to burst a bubble early rather than keep it inflated seems like an antihistorical hope.
Often having left-wing political beliefs
One wonders if they would have seen the minority aspect of the meltdown coming, then. Qualitative reasoning can be as suspect as quantitative reasoning, and it’s not clear to me mathematicians are that much less prone to qualitative errors.
What I find most interesting about this is that the situation is not that Simons succeeded where other mathematicians of the same caliber had failed – rather, Simons is virtually the only pure mathematician of his caliber to have left academia.
A professor in his 40s once told me that the department where he got his PhD had run the numbers, and financial success was inversely correlated with grades / research skills / intellectual ability; the top people got professorships and were middle class, whereas the people who were at the bottom of the class had left academia, and several of them made millions on Wall Street.
I also hear that the supply of physicists as quants is much larger now than it was ~20 years ago, and so it’s a respectable living but not a massively profitable one; it’s not clear that you could start a competitor to Renaissance today and do nearly as well.
Much of what you say seems like valid counterpoint. I’m getting out of my depth, insofar as I know very little about finance :-). I hope that other more knowledgable people will step in.
A professor in his 40s once told me that the department where he got his PhD had run the numbers, and financial success was inversely correlated with grades / research skills / intellectual ability; the top people got professorships and were middle class, whereas the people who were at the bottom of the class had left academia, and several of them made millions on Wall Street.
I don’t find this at all convincing, e.g.:
It’s only one department
He may have been misremembering the situation
“Bottom of the class” could reflect motivation rather than just intellectual ability.
It’s well known that academia doesn’t pay so well.
It seems to me to be weak support for the claim “if top tier math/physics talent went into finance, they could plausibly make billions, since fourth tier talent make millions.” (Fourth tier meant in the sense of this comment, but third tier might be more appropriate.)
There are also famous high-profile failures, though, such as Scholes at LTCM. (He did only win the Nobel prize in Econ, though, which is not a particularly strong predictor of math ability.)
Madoff fooled Simons for 10 years. He told Stony Brook to invest in Madoff in 1991. In 2003, Renaissance decided that Madoff was a fraud and pulled out. Simons told Stony Brook to pull out, but they left in 50%.
Many people investing with Madoff thought he was a fraud, just that he was defrauding someone else (eg, frontrunning his brokerage clients). It’s not clear that they pulled out because they realized he was a fraud or because they worried that Spitzer would investigate him. Their suspicions triggered a failed investigation of Madoff not because they complained to the SEC, but because when the SEC audited RenTech, it found emails discussing Madoff.
Oddly, the wikipedia article seems to be based on the first WSJ article I linked, not the Bloomberg one it cites.
It’s mild evidence against the “top level math talent will be able to identify bubbles before they burst” hypothesis, but only mild because I don’t think Madoff was transparent enough to be obviously unstable in the way that the housing market was.
How would we go about making it such that someone who contemplated perpetrating that sort of fraud would be very likely to conclude, “Nah, there’s no way that would work; if I tried that, I’d just end up in prison like Madoff”?
Considering that the wikipedia article about Madoff didn’t exist until the scandal was unveiled … probably. (I checked the Wikipedia history when I first heard about it in Dec 08, check the history for yourself. )
Hmm. I thought I mentioned Renaissance in this post, but for some reason it says Medallion, which is one of their funds. Apparently I derped earlier, fixing it now.
You’re a PhD in math with no specialist knowledge of finance. Do you think that you personally could do better financial analysis than, say, a Business major with a CFA and a decade or two of industry experience? If no, why do you think mathematicians are the way forward for the field?
I would expect the mathematician to be more likely to be a good coder, and more likely to lean on data analysis that she understood from first principles.
The closest thing to finance from first principles is just to assume the EMH, use the CAPM theory to constuct a market portfolio, and practice a policy of pure commission/tax avoidance. Finance from first principles is incredibly boring—the only way to make it interesting is to start getting into second-order effects.
The comment was “lean on data analysis”, as in “lean on data analysis to do finance”. Yes, the analysis itself is not “finance” per se, but most of the data analysis done in finance is simple math that gets used in complex ways. It’s easy enough to calculate P/E, but figuring out whether that implies you wanting to buy or sell is a much harder question. About the only way to use first principles to actually make buy/sell decisions is, as I said, to use the CAPM and just embrace your ignorance.
That seems a little extreme; presumably there’s a difference between using statistical tests as a heuristic you don’t understand, and using statistical tests in a well-understood way, even if you’re not deriving finance from first principles.
Also, CAPM isn’t actually true (i.e. assumptions never hold in the real world), whereas statistics is.
Did your friend think about purchasing put options on a continuous basis? No short squeeze risk and you wouldn’t have to worry about timing. I assume the information from your friend’s purchases would have been arbitraged in to the price of the underlying asset somehow.
(I’m not a quant and there might be something wrong with this idea.)
The problem is that a put strategy bleeds money on option costs, whereas a short doesn’t. (Shorts bleed on dividends, but in a lot of industries that’s cheaper). Also, long-term puts deeply out of the money(which you want, to minimize costs and maximize leverage) are an incredibly thin market, which tends to make trading difficult and expensive.
The problem is that a put strategy bleeds money on option costs
That shouldn’t be an inherent problem… there are some strategies that tend to make money most of the time but occasionally go drastically wrong, and this is an example of a strategy that tends to lose money most of the time but occasionally goes drastically right. Just calculate the expected value as usual, right?
It’s not a crippling problem. Nassim Taleb, the Black Swan author, ran a hedge fund with precisely that strategy(though his options were omnidirectional—his thesis was that we underestimate all forms of tail risk), and he made a mint in 2008. But it’s something that needs to be kept in mind.
Nassim Taleb, the Black Swan author, ran a hedge fund with precisely that strategy...and he made a mint in 2008
My understanding was that his fund was a failure and shut down before 2008, and the only evidence for his claim to have made money in 2008 was his word (with nothing about how well his strategy has performed on net).
Taleb reportedly became financially independent after the crash of 1987[15] and made a multi-million dollar fortune during the financial crisis that began in 2007, a development which he attributed to the mismatch between statistical distributions used in finance and reality.[36] Universa is a fund which is based on the “black swan” idea and to which Taleb is a principal adviser. Separate funds belonging to Universa made returns of 65% to 115% in October 2008.[20][37]
That said, the fund he founded, one by the name of Empirica, was shut down in 2004, though it was actually producing positive returns at the time. Apparently he was seriously ill, and wanted to become an author full-time as well. However, the “ran” in my above post was incorrect.
There’s at least one leprechaun there: citation 36 says nothing at all about whether Taleb made any money during the 2007 crisis, much less whether he realized a net return since 1987 greater than indexing. I’ve replaced it with a citation-needed.
That said, the fund he founded, one by the name of Empirica, was shut down in 2004, though it was actually producing positive returns at the time.
Eh, maybe it wasn’t bleeding too badly, but 2004 wasn’t a year anyone should be posting negative returns, and at least one of their funds was shut down for failing so badly:
When the Internet bubble burst in 2000, the Empirica Kurtosis fund posted a 57 percent return that year, according to company documents summarizing its results. In contrast, the S&P 500 fell 10.1 percent in 2000. The fund went into a tailspin starting with the 2001 decline, followed by drops of 13 percent in 2002 and 3.9 percent during the first two months of 2003. That’s when the partners closed the fund.
(Yes, 2001 was a loss. As Tavakoli asks, how the deuce does a ‘black swan’ fund lose money after 9/11?)
Apparently he was seriously ill
Articles I read phrased it as he ‘feared’ a recurrence of throat cancer, which honestly sounds a bit like ‘our CEO is resigning to spend more time with his family’.
Separate funds belonging to Universa made returns of 65% to 115% in October 2008.[20][37]
Spitznagel is more of an Austrian than a black-swan guy; since he started in 2007 and apparently managed small amounts like <$100m, some wins are not very strong evidence… I tried to find any estimate of Universa’s net return to investors from 2007 to now after fees etc, but I only found a fluffy Wikipedia article claiming “Spitznagel’s investment performance ranks as one of the top returns on capital of the financial crisis, as well as over a career” and citing The Dao of Capital—written by Spitznagel. So...
My first quote on Taleb was from memory, and looking into it it seems that my memory was a lot more black-and-white than the facts are. I’d love to get some concrete data on the two fund families, but hedge funds are notoriously hard to pry data out of, so I’m not sure we’ll get any.
but hedge funds are notoriously hard to pry data out of, so I’m not sure we’ll get any.
And in this case, a career is being built partially out of claiming that one of the families was a great success and proves the worth of a notoriously egotistic man’s ideas, so the situation is even worse than usual.
If risk/reward profiles that bleed money most of the time but occasionally make it big look inherently less attractive to investors, should we expect those strategies to be underplayed relative to their expected value?
Probably, but it doesn’t need to be a pure strategy. A normal portfolio hedged with a bit of crash insurance in the form of deep-OTM puts can be a sensible play in ordinary times. I don’t know how many people actually do that, though—judging by the market size, not many.
I have an economist friend who predicted the tech bubble, but shorted too early, and ended up losing quite a bit of money in the process. Knowing that chickens are coming home to roost is not enough; you need a good idea of when they’re going to arrive, and to have the solvency to be able to stick to your beliefs. (If Keynes’s comment—”The market can stay irrational longer than you can stay solvent”—fits you, then shorting bubbles seems unwise.)
I do believe there were people predicting the bubble would burst before it burst, and some of them were even people who don’t predict a bubble bursting every year. It’s not clear to me that a handful more mathematicians pointing out impending disaster would significantly shift public opinion, and hoping that the government would act to burst a bubble early rather than keep it inflated seems like an antihistorical hope.
One wonders if they would have seen the minority aspect of the meltdown coming, then. Qualitative reasoning can be as suspect as quantitative reasoning, and it’s not clear to me mathematicians are that much less prone to qualitative errors.
A professor in his 40s once told me that the department where he got his PhD had run the numbers, and financial success was inversely correlated with grades / research skills / intellectual ability; the top people got professorships and were middle class, whereas the people who were at the bottom of the class had left academia, and several of them made millions on Wall Street.
I also hear that the supply of physicists as quants is much larger now than it was ~20 years ago, and so it’s a respectable living but not a massively profitable one; it’s not clear that you could start a competitor to Renaissance today and do nearly as well.
Much of what you say seems like valid counterpoint. I’m getting out of my depth, insofar as I know very little about finance :-). I hope that other more knowledgable people will step in.
I don’t find this at all convincing, e.g.:
It’s only one department
He may have been misremembering the situation
“Bottom of the class” could reflect motivation rather than just intellectual ability.
It’s well known that academia doesn’t pay so well.
It seems to me to be weak support for the claim “if top tier math/physics talent went into finance, they could plausibly make billions, since fourth tier talent make millions.” (Fourth tier meant in the sense of this comment, but third tier might be more appropriate.)
There are also famous high-profile failures, though, such as Scholes at LTCM. (He did only win the Nobel prize in Econ, though, which is not a particularly strong predictor of math ability.)
There’s Renaissance technologies making obscene returns consistently for decades.
According to Wikipedia, the founder advised people to invest with Bernie Madoff...
Madoff fooled Simons for 10 years. He told Stony Brook to invest in Madoff in 1991. In 2003, Renaissance decided that Madoff was a fraud and pulled out. Simons told Stony Brook to pull out, but they left in 50%.
Many people investing with Madoff thought he was a fraud, just that he was defrauding someone else (eg, frontrunning his brokerage clients). It’s not clear that they pulled out because they realized he was a fraud or because they worried that Spitzer would investigate him. Their suspicions triggered a failed investigation of Madoff not because they complained to the SEC, but because when the SEC audited RenTech, it found emails discussing Madoff.
Oddly, the wikipedia article seems to be based on the first WSJ article I linked, not the Bloomberg one it cites.
Hindsight bias?
It’s mild evidence against the “top level math talent will be able to identify bubbles before they burst” hypothesis, but only mild because I don’t think Madoff was transparent enough to be obviously unstable in the way that the housing market was.
Yeah, it’s a cheap shot. Madoff fooled a lot of people, and detecting fraud takes a different skill set.
How would we go about making it such that someone who contemplated perpetrating that sort of fraud would be very likely to conclude, “Nah, there’s no way that would work; if I tried that, I’d just end up in prison like Madoff”?
Trump up the trials, studiously avoid mentioning the obvious fact that there’s a bunch of people living large on top of a Ponzi scheme right now.
Considering that the wikipedia article about Madoff didn’t exist until the scandal was unveiled … probably. (I checked the Wikipedia history when I first heard about it in Dec 08, check the history for yourself. )
Hmm. I thought I mentioned Renaissance in this post, but for some reason it says Medallion, which is one of their funds. Apparently I derped earlier, fixing it now.
You’re a PhD in math with no specialist knowledge of finance. Do you think that you personally could do better financial analysis than, say, a Business major with a CFA and a decade or two of industry experience? If no, why do you think mathematicians are the way forward for the field?
I would expect the mathematician to be more likely to be a good coder, and more likely to lean on data analysis that she understood from first principles.
The closest thing to finance from first principles is just to assume the EMH, use the CAPM theory to constuct a market portfolio, and practice a policy of pure commission/tax avoidance. Finance from first principles is incredibly boring—the only way to make it interesting is to start getting into second-order effects.
Finance != data analysis.
The comment was “lean on data analysis”, as in “lean on data analysis to do finance”. Yes, the analysis itself is not “finance” per se, but most of the data analysis done in finance is simple math that gets used in complex ways. It’s easy enough to calculate P/E, but figuring out whether that implies you wanting to buy or sell is a much harder question. About the only way to use first principles to actually make buy/sell decisions is, as I said, to use the CAPM and just embrace your ignorance.
That seems a little extreme; presumably there’s a difference between using statistical tests as a heuristic you don’t understand, and using statistical tests in a well-understood way, even if you’re not deriving finance from first principles.
Also, CAPM isn’t actually true (i.e. assumptions never hold in the real world), whereas statistics is.
Did your friend think about purchasing put options on a continuous basis? No short squeeze risk and you wouldn’t have to worry about timing. I assume the information from your friend’s purchases would have been arbitraged in to the price of the underlying asset somehow.
(I’m not a quant and there might be something wrong with this idea.)
The problem is that a put strategy bleeds money on option costs, whereas a short doesn’t. (Shorts bleed on dividends, but in a lot of industries that’s cheaper). Also, long-term puts deeply out of the money(which you want, to minimize costs and maximize leverage) are an incredibly thin market, which tends to make trading difficult and expensive.
That shouldn’t be an inherent problem… there are some strategies that tend to make money most of the time but occasionally go drastically wrong, and this is an example of a strategy that tends to lose money most of the time but occasionally goes drastically right. Just calculate the expected value as usual, right?
It’s not a crippling problem. Nassim Taleb, the Black Swan author, ran a hedge fund with precisely that strategy(though his options were omnidirectional—his thesis was that we underestimate all forms of tail risk), and he made a mint in 2008. But it’s something that needs to be kept in mind.
My understanding was that his fund was a failure and shut down before 2008, and the only evidence for his claim to have made money in 2008 was his word (with nothing about how well his strategy has performed on net).
Per Wikipedia:
Taleb reportedly became financially independent after the crash of 1987[15] and made a multi-million dollar fortune during the financial crisis that began in 2007, a development which he attributed to the mismatch between statistical distributions used in finance and reality.[36] Universa is a fund which is based on the “black swan” idea and to which Taleb is a principal adviser. Separate funds belonging to Universa made returns of 65% to 115% in October 2008.[20][37]
That said, the fund he founded, one by the name of Empirica, was shut down in 2004, though it was actually producing positive returns at the time. Apparently he was seriously ill, and wanted to become an author full-time as well. However, the “ran” in my above post was incorrect.
There’s at least one leprechaun there: citation 36 says nothing at all about whether Taleb made any money during the 2007 crisis, much less whether he realized a net return since 1987 greater than indexing. I’ve replaced it with a citation-needed.
Eh, maybe it wasn’t bleeding too badly, but 2004 wasn’t a year anyone should be posting negative returns, and at least one of their funds was shut down for failing so badly:
(Yes, 2001 was a loss. As Tavakoli asks, how the deuce does a ‘black swan’ fund lose money after 9/11?)
Articles I read phrased it as he ‘feared’ a recurrence of throat cancer, which honestly sounds a bit like ‘our CEO is resigning to spend more time with his family’.
Spitznagel is more of an Austrian than a black-swan guy; since he started in 2007 and apparently managed small amounts like <$100m, some wins are not very strong evidence… I tried to find any estimate of Universa’s net return to investors from 2007 to now after fees etc, but I only found a fluffy Wikipedia article claiming “Spitznagel’s investment performance ranks as one of the top returns on capital of the financial crisis, as well as over a career” and citing The Dao of Capital—written by Spitznagel. So...
My first quote on Taleb was from memory, and looking into it it seems that my memory was a lot more black-and-white than the facts are. I’d love to get some concrete data on the two fund families, but hedge funds are notoriously hard to pry data out of, so I’m not sure we’ll get any.
And in this case, a career is being built partially out of claiming that one of the families was a great success and proves the worth of a notoriously egotistic man’s ideas, so the situation is even worse than usual.
If risk/reward profiles that bleed money most of the time but occasionally make it big look inherently less attractive to investors, should we expect those strategies to be underplayed relative to their expected value?
Probably, but it doesn’t need to be a pure strategy. A normal portfolio hedged with a bit of crash insurance in the form of deep-OTM puts can be a sensible play in ordinary times. I don’t know how many people actually do that, though—judging by the market size, not many.
I’m not sure; I can ask the next time I talk to him (which will be over a month from now).