5 ms·

> If you want to put your money in a fund, your odds are better if you pick one that has just had a bad year. Everyone - please don't make claims like this on

> If you want to put your money in a fund, your odds are better if you pick one that has just had a bad year. Everyone - please don't make claims like this on HN. It's not adding value to anyone's life when you speculate incorrectly on financial advice. I will pay you $500 if you can share serious evidence that the market can be outperformed by buying mutual funds that did poorly the prior year. (I am serious. Though of course, if it were true, you would stand to make billions beyond my $500.)

xioxox 9y ago
Not looked in this much, but here's an old paper from a simple search: http://onlinelibrary.wiley.com/doi/10.1111/j.1468-5957.1994.tb00350.x/abstract http://onlinelibrary.wiley.com/doi/10.1111/j.1468-5957.1994.... (I've only read the abstract)
- tedsanders 9y ago
  Great find. Though of course one study is unlikely to be definitive. (http://slatestarcodex.com/2014/12/12/beware-the-man-of-one-study/ http://slatestarcodex.com/2014/12/12/beware-the-man-of-one-s...) Here are 15 more articles on mutual fund persistence: http://www.altruistfa.com/readingroomarticles.htm#Persistence http://www.altruistfa.com/readingroomarticles.htm#Persistenc.... I have not read them all, but my impression of the literature is that: (1) There is slight positive serial correlation in mutual fund returns (2) Most of the slight positive serial correlation comes from bad performers staying bad performers (rather than good performers staying good performers) Theoretically, I find the notion of positive serial correlation far more likely than negative serial correlation. It seems more plausible to me that good fund managers would stay good than that managers would tend to oscillate from good to bad to good on an annual time scale. In general, negative serial time correlations are generally harder to explain because a negative number squared is positive. Any sort of "rebounding" effect will result in oscillatory behavior, which will only be detectable over a certain time scale. Measuring a positive linear relationship is easier than measuring a negative linear relationship, because the negative solution will have an associated time scale. (That is, solutions that drift/diffuse can be captured over many time scales & sampling rates, whereas solutions that oscillate can only be captured near the rate of oscillation. On time scales that are very zoomed in or very zoomed out, you won't notice that the system is oscillating.) Much of this is handwavy and applies only to linear correlation, but hopefully the point is clear nonetheless.
ScottBurson 9y ago
Well, I didn't say the odds were much better -- only that they were better :-)
- tedsanders 9y ago
  I am very bothered by your comments and your smiley face. Please do not give incorrect and speculative investment advice. You have supplied zero evidence that the odds are better. They are very very likely not better. I apologize for my lack of humor. I find the issue serious right now.

xioxox 9y ago

Not looked in this much, but here's an old paper from a simple search: http://onlinelibrary.wiley.com/doi/10.1111/j.1468-5957.1994.tb00350.x/abstract http://onlinelibrary.wiley.com/doi/10.1111/j.1468-5957.1994.... (I've only read the abstract)

tedsanders 9y ago
Great find. Though of course one study is unlikely to be definitive. (http://slatestarcodex.com/2014/12/12/beware-the-man-of-one-study/ http://slatestarcodex.com/2014/12/12/beware-the-man-of-one-s...) Here are 15 more articles on mutual fund persistence: http://www.altruistfa.com/readingroomarticles.htm#Persistence http://www.altruistfa.com/readingroomarticles.htm#Persistenc.... I have not read them all, but my impression of the literature is that: (1) There is slight positive serial correlation in mutual fund returns (2) Most of the slight positive serial correlation comes from bad performers staying bad performers (rather than good performers staying good performers) Theoretically, I find the notion of positive serial correlation far more likely than negative serial correlation. It seems more plausible to me that good fund managers would stay good than that managers would tend to oscillate from good to bad to good on an annual time scale. In general, negative serial time correlations are generally harder to explain because a negative number squared is positive. Any sort of "rebounding" effect will result in oscillatory behavior, which will only be detectable over a certain time scale. Measuring a positive linear relationship is easier than measuring a negative linear relationship, because the negative solution will have an associated time scale. (That is, solutions that drift/diffuse can be captured over many time scales & sampling rates, whereas solutions that oscillate can only be captured near the rate of oscillation. On time scales that are very zoomed in or very zoomed out, you won't notice that the system is oscillating.) Much of this is handwavy and applies only to linear correlation, but hopefully the point is clear nonetheless.

ScottBurson 9y ago

Well, I didn't say the odds were much better -- only that they were better :-)

tedsanders 9y ago
I am very bothered by your comments and your smiley face. Please do not give incorrect and speculative investment advice. You have supplied zero evidence that the odds are better. They are very very likely not better. I apologize for my lack of humor. I find the issue serious right now.