4 ms·

Seems to me there's an obvious flaw in the efficient market hypothesis. It states that, essentially, you can find no sustainable edge because the market rapidly

by Tycho 7y ago

Seems to me there's an obvious flaw in the efficient market hypothesis. It states that, essentially, you can find no sustainable edge because the market rapidly reacts to information. Meaning, if there is some information relevant to expected investment performance, investors will immediately act on it, extinguishing the information advantage. But what if the opportunity is some kind of abstract pattern that doesn't make any sense to humans? You can hardly say that people would efficiently react to patterns that are entirely uninterpretable to them. So it leaves the door open for these sort of statistical artifacts - abstract side-effects or unintuitive properties of the overall system - to provide a sustainable source of alpha.

empath75 7y ago
I think the idea is that once someone figures out how to exploit it, others will soon.
- Tycho 7y ago
  But I don't see why that would be true if we're talking about arcane patterns in data. Who's to say anyone's even looking at the same data, let alone the same pattern? If we were talking about intuitive patterns that map to logic and investor psychology, then sure, others will reliable stumble upon the same ideas. But if we're talking about unintuitive patterns in non-obvious datasets, the search space is far too big to assure that effect.
- thedudeabides5 7y ago
  Sure, but isn't this as much an argument for efficient markets as anything. Literally hundreds of thousands of folks have died on the very hill Medallion is holding. The fact that there is just one firm with this track record is evidence that markets are pretty hard and relatively efficient. Whether markets are so perfectly efficient that no firm can really beat the market, well that just seems like a pretty academic and unhelpful question. You don't need some grand perfect market hypothesis to say 'man those returns look a bit too good to be true, what's really going on in there?"
jerf 7y ago
Nobody believes in the 100% efficient market. It's not even possible. But if someone's got a bead on a 99.99% efficient market, they can make a lot of money, whereas if you say "Ah ha! The market is not perfectly efficient! Here, I can make money doing this!", you probably won't make much. (Then there are those who think the market is like maybe 10% efficient, in which case my response is, if it's so inefficient, go make money. That would imply all sorts of arbitrage opportunity. Why slave away in whatever job you may have now when there's such an inefficient market just waiting for someone smart & clever to pluck it? The answer is that where ever the market may be on the range of 0-100% efficient, to the extent that such a measurement is even particularly definable, while it may not be exactly 100.00000%, it's much closer to that than the 0% side.) (Remember, efficient doesn't mean good or moral, which I think is the most common mistake I see on HN. It means something much more like "there are no easy arbitrage possibilities". It hasn't got anything to do with "goodness" or "morality" any more than analyzing gravity or electromagnetism in terms of its goodness or morality makes any sense.)
- dmwallin 7y ago
  Your assessment assumes a perfectly rational, 100% transparent market. The market obviously has a distribution of rationality and transparency. The most efficient parts of the market are going to be the parts that are largely rational and/or highly transparent. Making money in the remaining more irrational/opaque parts of the market is not as simple as just identifying it as inefficient, it requires either finding a pattern in the irrationality or a way of unlocking hidden information(ideally legally).
  - throw0101a 7y ago
    > ... or a way of unlocking hidden information(ideally legally). One example of which is hedge funds using satellite imagery: * https://news.ycombinator.com/item?id=20243810 https://news.ycombinator.com/item?id=20243810 * https://newsroom.haas.berkeley.edu/how-hedge-funds-use-satellite-images-to-beat-wall-street-and-main-street/ https://newsroom.haas.berkeley.edu/how-hedge-funds-use-satel... * https://www.theatlantic.com/magazine/archive/2019/05/stock-value-satellite-images-investing/586009/ https://www.theatlantic.com/magazine/archive/2019/05/stock-v...
dcolkitt 7y ago
The efficient market hypothesis is a lot like Newtonian gravity. It's not a perfect theory, but it's a pretty close approximation that pretty much covers most any domain outside very exotic conditions. If somebody comes up to you, a random Joe Schmoe, and tells you they have an investment that consistently beats the market on a risk-adjusted basis. Well... You can pretty much guarantee that they're full of shit. Even just analyzing papers that passed the rigor of academic peer review, the sizable majority of market anomalies fail to replicate on an out-of-sample basis[1]. Like Newtonian gravity, to the extent that deviations exist it's either of very small magnitude or in very exotic conditions. The market anomalies that do exist, like the HML value effect or momentum effect, don't substantially improve the optimal portfolio and go through very long periods of deep drawdowns. Or in the realm of the very exotic, there are small teams, in aggregate making up much less than 0.1% of the market, that do consistently out-perform. There common characteristics of these strategies are that they have tightly limited capacity, involve a huge up-front investment in technology and expertise, almost never offer to manage outsider money, and are extremely hard to replicate even by other experts. [1] https://www.nber.org/papers/w23394 https://www.nber.org/papers/w23394
- throw0101a 7y ago
  I was under the impression that the Fama-French (3/5) factor model do pretty well at explaining why certain stocks/sectors do better than the market average.
- Tycho 7y ago
  If the value and momentum and size factors are legitimate, then it stands to reason that there could be other more obscure, less intuitive and more profitable anomalies. Are those limited to very small magnitudes? Maybe, but I don't see why that should necessarily be so. It seems like EMH is predicated on the idea that the only investment opportunities are ones that map to human reasoning.
- o_nate 7y ago
  Another analogy would be the theory of perfectly competitive markets. It makes sense for economists to assume this in many cases, even though that would imply that consistently profitable firms are impossible.
zneveu 7y ago
As long as there are market regulations, and as long as people have some kind of morals the market will stay a long ways from being perfectly efficient. Any profit-making trade that you can't make because of legal or ethical reasons shows a crack in the efficient market. That may not be a bad thing, but it's so important that we recognize these imperfections and keep them patched so that they're not exploited.

by Tycho 7y ago

empath75 7y ago

I think the idea is that once someone figures out how to exploit it, others will soon.

Tycho 7y ago
But I don't see why that would be true if we're talking about arcane patterns in data. Who's to say anyone's even looking at the same data, let alone the same pattern? If we were talking about intuitive patterns that map to logic and investor psychology, then sure, others will reliable stumble upon the same ideas. But if we're talking about unintuitive patterns in non-obvious datasets, the search space is far too big to assure that effect.
thedudeabides5 7y ago
Sure, but isn't this as much an argument for efficient markets as anything. Literally hundreds of thousands of folks have died on the very hill Medallion is holding. The fact that there is just one firm with this track record is evidence that markets are pretty hard and relatively efficient. Whether markets are so perfectly efficient that no firm can really beat the market, well that just seems like a pretty academic and unhelpful question. You don't need some grand perfect market hypothesis to say 'man those returns look a bit too good to be true, what's really going on in there?"

jerf 7y ago

Nobody believes in the 100% efficient market. It's not even possible. But if someone's got a bead on a 99.99% efficient market, they can make a lot of money, whereas if you say "Ah ha! The market is not perfectly efficient! Here, I can make money doing this!", you probably won't make much. (Then there are those who think the market is like maybe 10% efficient, in which case my response is, if it's so inefficient, go make money. That would imply all sorts of arbitrage opportunity. Why slave away in whatever job you may have now when there's such an inefficient market just waiting for someone smart & clever to pluck it? The answer is that where ever the market may be on the range of 0-100% efficient, to the extent that such a measurement is even particularly definable, while it may not be exactly 100.00000%, it's much closer to that than the 0% side.) (Remember, efficient doesn't mean good or moral, which I think is the most common mistake I see on HN. It means something much more like "there are no easy arbitrage possibilities". It hasn't got anything to do with "goodness" or "morality" any more than analyzing gravity or electromagnetism in terms of its goodness or morality makes any sense.)

dmwallin 7y ago
Your assessment assumes a perfectly rational, 100% transparent market. The market obviously has a distribution of rationality and transparency. The most efficient parts of the market are going to be the parts that are largely rational and/or highly transparent. Making money in the remaining more irrational/opaque parts of the market is not as simple as just identifying it as inefficient, it requires either finding a pattern in the irrationality or a way of unlocking hidden information(ideally legally).
- throw0101a 7y ago
  > ... or a way of unlocking hidden information(ideally legally). One example of which is hedge funds using satellite imagery: * https://news.ycombinator.com/item?id=20243810 https://news.ycombinator.com/item?id=20243810 * https://newsroom.haas.berkeley.edu/how-hedge-funds-use-satellite-images-to-beat-wall-street-and-main-street/ https://newsroom.haas.berkeley.edu/how-hedge-funds-use-satel... * https://www.theatlantic.com/magazine/archive/2019/05/stock-value-satellite-images-investing/586009/ https://www.theatlantic.com/magazine/archive/2019/05/stock-v...

dcolkitt 7y ago

The efficient market hypothesis is a lot like Newtonian gravity. It's not a perfect theory, but it's a pretty close approximation that pretty much covers most any domain outside very exotic conditions. If somebody comes up to you, a random Joe Schmoe, and tells you they have an investment that consistently beats the market on a risk-adjusted basis. Well... You can pretty much guarantee that they're full of shit. Even just analyzing papers that passed the rigor of academic peer review, the sizable majority of market anomalies fail to replicate on an out-of-sample basis[1]. Like Newtonian gravity, to the extent that deviations exist it's either of very small magnitude or in very exotic conditions. The market anomalies that do exist, like the HML value effect or momentum effect, don't substantially improve the optimal portfolio and go through very long periods of deep drawdowns. Or in the realm of the very exotic, there are small teams, in aggregate making up much less than 0.1% of the market, that do consistently out-perform. There common characteristics of these strategies are that they have tightly limited capacity, involve a huge up-front investment in technology and expertise, almost never offer to manage outsider money, and are extremely hard to replicate even by other experts. [1] https://www.nber.org/papers/w23394 https://www.nber.org/papers/w23394

throw0101a 7y ago
I was under the impression that the Fama-French (3/5) factor model do pretty well at explaining why certain stocks/sectors do better than the market average.
Tycho 7y ago
If the value and momentum and size factors are legitimate, then it stands to reason that there could be other more obscure, less intuitive and more profitable anomalies. Are those limited to very small magnitudes? Maybe, but I don't see why that should necessarily be so. It seems like EMH is predicated on the idea that the only investment opportunities are ones that map to human reasoning.
o_nate 7y ago
Another analogy would be the theory of perfectly competitive markets. It makes sense for economists to assume this in many cases, even though that would imply that consistently profitable firms are impossible.

zneveu 7y ago

As long as there are market regulations, and as long as people have some kind of morals the market will stay a long ways from being perfectly efficient. Any profit-making trade that you can't make because of legal or ethical reasons shows a crack in the efficient market. That may not be a bad thing, but it's so important that we recognize these imperfections and keep them patched so that they're not exploited.