an “average” return presupposes some set that you’re averaging.
Okay. The set I’m imagining is the set of investments that meet some criteria, like “US stocks listed on the NYSE” or “US large cap stocks” or even “home robotics stocks.” It seems meaningful to me to state that indexing across that set will provide “average” returns in the sense that it will provide low-variance returns centered around the set mean, relative to a picking a portfolio based on personal taste from stocks in that set. For investors as a whole, the effects of their personal taste must average out to the mean (once weighted by participation in that set), and so unless you have special knowledge you should expect your returns to be centered around the set mean, but with higher variance. For risk-averse investors, higher variance is a bad thing. Justifying the expectation that your returns will be higher than the set mean requires a statement like “I can pick home robotics stocks better than the market as a whole,” or the less questionable “I can pick industries better than the market as a whole”, but the safest is probably “I can’t pick better than the market as a whole,” which suggests indexing across all industries (and, barring legal barriers, across all countries).
People who have been successful through luck have just as compelling narratives how some feature of their {method/mind/character/upbringing/etc.} made their success inevitable.
I understand the outside view that most successful people are likely to be highly lucky, and are unlikely to have public correct causal models of their success. (It’s very unlikely that a boy band star or male actor would admit to having sex with a gay talent agent in order to land a position, for example, even if that’s the primary differentiator between them and their competition that did not go on to become famous.)
That said, as someone who formally studies good decision-making, I think I have a strong inside view of what works about Buffett’s decision-making, and it seems likely to me that it actually is a causal factor in his success. I have no doubt that luck played a role in Buffett making his way to the top, rather than just doing well for himself. (There are other residents of Graham-Doddsville, many of whom have also gotten rich, but none of whom have gotten as rich as Buffett.)
Well, let’s try to sort this out a little better. There’s a general superset of all investment opportunities. That’s too much to chew on, so let’s limit ourselves to a semi-arbitrary subset. As an example let’s take the standard large-cap US equity and define it as members of the S&P500 index. There is a variety of tradeable instruments (mutual funds, ETFs, futures, etc.) which will allow you to buy the S&P500.
indexing across that set will provide “average” returns in the sense that it will provide low-variance returns centered around the set mean
That is not correct. Owning an index security will provide returns similar to the set mean (defined appropriately, here the mean is cap-weighted), but these returns will not be low-variance. Provided the security that you chose is run competently, your tracking error with respect to the the index will be small and so low-variance, but the returns themselves can be as high-variance as they like.
If you pick a portfolio from the same set based on personal taste, the tracking error will be much greater, but the variance of the portfolio returns themselves might be greater or might be smaller. For example, if your taste runs to low-beta stocks, the variance of your personal-taste portfolio is likely to be less than the variance of the S&P500 index.
so unless you have special knowledge you should expect your returns to be centered around the set mean, but with higher variance
Again, this is not true. Indices are not homogenous, my personal taste will affect the variance of my portfolio. It might end up higher than the index or it might end up lower. Similarly, my personal taste will affect the expected returns as well.
higher variance is a bad thing
Given everything else equal, which it never is. If investors were only interested in expected variance, why even invest? Holding cash has zero variance in nominal terms. And, of course, different investors have different risk tolerances.
Let me also point out again that investing is a much broader field than just picking stocks and choosing a subset to invest in often has more significant consequences than deciding which securities from that subset to pick. The advice to “index” tells me nothing whether I should have municipal bonds in my portfolio, for example. Or any bonds. Or any equities. Or, actually, anything in particular.
Re Buffet, I fully agree that he possesses superior decision-making abilities. He’s probably better than 99.9% of the population. However his wealth is far beyond the average wealth of the top 0.1% best decision makers. More importantly, I am not sure how the superiority of Buffet’s personal abilities translates to believing that value investment can “beat the market”.
Provided the security that you chose is run competently, your tracking error with respect to the the index will be small and so low-variance, but the returns themselves can be as high-variance as they like.
This is what what I meant by “low variance centered on the set mean,” as I assumed the set mean was a stochastic object (with the variance that implies). Similarly, you can read “high variance” about the personal taste portfolio as “high tracking error variance,” which as you point out may be the result of low absolute variance in returns. Holding cash has a high-variance tracking error relative to the S&P 500, even though cash is zero nominal variance.
Given everything else equal, which it never is.
What does it add to the conversation to point this out?
I don’t get the impression that this conversation is productive. I think if someone doesn’t want to become proficient at finance, their best way to invest money is to set aside part of their salary* and buy into VTSMX every month. I don’t think they need to know what “dollar cost averaging” is, or why index funds are better than actively managed funds, and I think that crude heuristic arguments (“on average, money managers subtract value, you should expect to be average”) are the right way to convince them to invest this way.
As of yet, I don’t think I’ve seen you make a positive recommendation, which strikes me as worse than useless for a PSA thread. If finance is hard, then what?
*I feel the need to point out that I understand not everyone has a salary.
I’ve said it before—my point is that there are no good general-purpose universally-applicable recommendations for personal investing. They do not exist.
Certainly there are not-horrible recommendations. Investing in VTSMX is one of them. Putting your money into T-bills is another one. Putting it into TIPS is yet another one. Putting it into a 50-50 mix of high-grade corporate bonds and Russel 3000 is yet another one. Etc., etc.
All of these are not horrible. But none of them is particularly good or a good fit for everyone.
It’s like asking “what kind of food should I eat?” There are many not-horrible answers to it, starting with “follow the US government’s food pyramid”. But it’s not a particularly good answer and there is no single universal answer. People are too different for that. Same with investing—no generic answer is good enough.
However I agree that this particular chain of exchanges got too long. Your arguments seem incoherent to me and, no doubt, mine seem obtuse to you. If I get to writing a post on introduction to thinking about investing this conversation might continue...
Okay. The set I’m imagining is the set of investments that meet some criteria, like “US stocks listed on the NYSE” or “US large cap stocks” or even “home robotics stocks.” It seems meaningful to me to state that indexing across that set will provide “average” returns in the sense that it will provide low-variance returns centered around the set mean, relative to a picking a portfolio based on personal taste from stocks in that set. For investors as a whole, the effects of their personal taste must average out to the mean (once weighted by participation in that set), and so unless you have special knowledge you should expect your returns to be centered around the set mean, but with higher variance. For risk-averse investors, higher variance is a bad thing. Justifying the expectation that your returns will be higher than the set mean requires a statement like “I can pick home robotics stocks better than the market as a whole,” or the less questionable “I can pick industries better than the market as a whole”, but the safest is probably “I can’t pick better than the market as a whole,” which suggests indexing across all industries (and, barring legal barriers, across all countries).
I understand the outside view that most successful people are likely to be highly lucky, and are unlikely to have public correct causal models of their success. (It’s very unlikely that a boy band star or male actor would admit to having sex with a gay talent agent in order to land a position, for example, even if that’s the primary differentiator between them and their competition that did not go on to become famous.)
That said, as someone who formally studies good decision-making, I think I have a strong inside view of what works about Buffett’s decision-making, and it seems likely to me that it actually is a causal factor in his success. I have no doubt that luck played a role in Buffett making his way to the top, rather than just doing well for himself. (There are other residents of Graham-Doddsville, many of whom have also gotten rich, but none of whom have gotten as rich as Buffett.)
Well, let’s try to sort this out a little better. There’s a general superset of all investment opportunities. That’s too much to chew on, so let’s limit ourselves to a semi-arbitrary subset. As an example let’s take the standard large-cap US equity and define it as members of the S&P500 index. There is a variety of tradeable instruments (mutual funds, ETFs, futures, etc.) which will allow you to buy the S&P500.
That is not correct. Owning an index security will provide returns similar to the set mean (defined appropriately, here the mean is cap-weighted), but these returns will not be low-variance. Provided the security that you chose is run competently, your tracking error with respect to the the index will be small and so low-variance, but the returns themselves can be as high-variance as they like.
If you pick a portfolio from the same set based on personal taste, the tracking error will be much greater, but the variance of the portfolio returns themselves might be greater or might be smaller. For example, if your taste runs to low-beta stocks, the variance of your personal-taste portfolio is likely to be less than the variance of the S&P500 index.
Again, this is not true. Indices are not homogenous, my personal taste will affect the variance of my portfolio. It might end up higher than the index or it might end up lower. Similarly, my personal taste will affect the expected returns as well.
Given everything else equal, which it never is. If investors were only interested in expected variance, why even invest? Holding cash has zero variance in nominal terms. And, of course, different investors have different risk tolerances.
Let me also point out again that investing is a much broader field than just picking stocks and choosing a subset to invest in often has more significant consequences than deciding which securities from that subset to pick. The advice to “index” tells me nothing whether I should have municipal bonds in my portfolio, for example. Or any bonds. Or any equities. Or, actually, anything in particular.
Re Buffet, I fully agree that he possesses superior decision-making abilities. He’s probably better than 99.9% of the population. However his wealth is far beyond the average wealth of the top 0.1% best decision makers. More importantly, I am not sure how the superiority of Buffet’s personal abilities translates to believing that value investment can “beat the market”.
This is what what I meant by “low variance centered on the set mean,” as I assumed the set mean was a stochastic object (with the variance that implies). Similarly, you can read “high variance” about the personal taste portfolio as “high tracking error variance,” which as you point out may be the result of low absolute variance in returns. Holding cash has a high-variance tracking error relative to the S&P 500, even though cash is zero nominal variance.
What does it add to the conversation to point this out?
I don’t get the impression that this conversation is productive. I think if someone doesn’t want to become proficient at finance, their best way to invest money is to set aside part of their salary* and buy into VTSMX every month. I don’t think they need to know what “dollar cost averaging” is, or why index funds are better than actively managed funds, and I think that crude heuristic arguments (“on average, money managers subtract value, you should expect to be average”) are the right way to convince them to invest this way.
As of yet, I don’t think I’ve seen you make a positive recommendation, which strikes me as worse than useless for a PSA thread. If finance is hard, then what?
*I feel the need to point out that I understand not everyone has a salary.
I’ve said it before—my point is that there are no good general-purpose universally-applicable recommendations for personal investing. They do not exist.
Certainly there are not-horrible recommendations. Investing in VTSMX is one of them. Putting your money into T-bills is another one. Putting it into TIPS is yet another one. Putting it into a 50-50 mix of high-grade corporate bonds and Russel 3000 is yet another one. Etc., etc.
All of these are not horrible. But none of them is particularly good or a good fit for everyone.
It’s like asking “what kind of food should I eat?” There are many not-horrible answers to it, starting with “follow the US government’s food pyramid”. But it’s not a particularly good answer and there is no single universal answer. People are too different for that. Same with investing—no generic answer is good enough.
However I agree that this particular chain of exchanges got too long. Your arguments seem incoherent to me and, no doubt, mine seem obtuse to you. If I get to writing a post on introduction to thinking about investing this conversation might continue...