This is one I encounter extremely frequently in real life…
Say you believe yields are too low and treasuries are a poor investment. So you refuse to put any in your portfolio. Not 10%, not 5%, not even 1%. Even though they are not very correlated with your equities and can stabilize your portfolio and even boost your average return (never mind the sharpe ratio) if added at a relatively small weight (say 10%).
I mentally refer to this as the “failure to see assets in a portfolio context” bias. Some people simply refuse to have any bonds, real estate, commodities, or even equities (somewhat rare) because they simply don’t like the prospects of that asset class. By people, I don’t mean the average joe but investment professionals who probably are aware of modern portfolio theory.
I suppose there are some elements of mental accounting, status quo, and overconfidence here but they don’t seem to explain the root cause of this psychological bias.
Maybe they just don’t understand those asset classes well enough to want to put capital behind it. It could be a bias, though it could also be a simple matter of wanting to stay within their circle of competence.
familiarity bias, maybe? The thing is you have to believe in market efficiency to say someone else has a bias for not following it. Maybe they know something that others don’t know or are not interested in finding out.
Is there a bias for people who blindly follow EMH?
I’d call this ‘balls deep bias’ - if you’re a lay investor and you like equities and don’t like teasuries, why only go 95% shaft deep into equities? Go balls deep
but really, it’s just good old overconfidence bias. You’re 100% confident in your forecast so you don’t think you’ll need the diversification benefit
Yes, I’ve taken L3. Yes, I know that we’re supposed to “consider the entire portfolio” when we build portfolios. But I still think it’s more or less a bunch of baloney.
As a whole, commode-ities have only returned about 4.5% per year with a 19% std dev. Think–lower return than bonds, higher risk than stocks.
Bonds are a little better–they barely outpace inflation. (maybe)
And spare me the “benefits of diversification when assets are less than perfectly correlated” crap. Heard it. Learned it. Passed the test. Still don’t believe in it. Just like the one-liner said–diversification only works when one investment underperforms the other.
Strange that you would say that considering an asset in the context of the portfolio is just baloney. Would you consider an asset in the context of its tax consequences also baloney, after all, the company’s either a good company or it’s not?
I understand that sometimes transaction costs (whether money or mental energhy) make it impractical to rebalance the entire portfolio because you are adding a new asset to it, so sometimes you just decide to be lazy and throw some money at it because it looks good, but considering the portfolio is pretty important for figuring out how much of something to add. Sometimes you discover you ought to be having a lot more of something than you think; other times a lot less.
If you’re in the same asset class all the time, the correlations may all be pretty high anyway, but then you should be looking at the relative volatilities for sizing information.
Markowitz theory is exactly that–theory. Looks really good in a textbook, good test material, and virtually useless in real life (IMHO, of course). Sure, we can look at historical correlations, and come up with an optimized mean-variance portfolio, or a minimum-variance portfolio, based on future expectations. But after real life happens, where does your portfolio land? On the efficient frontier? Or somewhere else? Did Asset A have the expected return and standard deviation that you expected? Did it have the same correlation as asset B? Did you appropriately combine them with risk-free asset C to come up with the proper allocation? (Probably not. Real life happens.)
Tax consequences, on the other hand, exist in real life. They are exactly that. Tax. Consequences. It’s not just textbook theory. Lots of checks are written to Uncle Sugar on April 15. Moreover, dividends and interest aren’t random patterns. We know the yield on the mutual fund we expect to buy. We know the interest rates on municipal bonds, and we can use these expectations to our advantage, because we know (at least within a fairly narrow corridor) what our tax consequences will be of such decisions.
You can make the same arguments about an investment in an individual asset (did you get the return you expected, was the downside risk what you thought it would be, did the valuation mean-revert approprately, etc.). OK, so I grant you that tax liabilities are more certain than investment returns (assuming you know enough about the returns to calculate the liabilities), but is your conclusion that every investment is BS because you can’t predict its future exactly?
Markowitz isn’t theory (CAPM is theory; Markowitz is just math). What part of the math is BS? You can say that volatilites and correlations aren’t stable, that expected returns are expected, rather than guaranteed. But you can say that about any investment, even if you measure risk some other way than volatilities. If that’s your reason for saying it’s BS, then you should be saying that about every investment that isn’t riskless, which is pretty much all of them.
I think it’s legitimate to say something like “doing the full portfolio optimization is a lot of work and [in our experience] doesn’t convincingly do any better or worse than just having an equally weighted portfolio of my best picks, so we don’t bother to do it.” Or you might say “It’s not a bad idea, but it’s too sensitive to noise in the inputs [more true in Long/Short portfolios than Long-only]” Those are a legitimate statement that one can agree with or disagree with without sounding ignorant. But saying “it’s just BS” does sound like “I’ve decided not to pay attention to math.” And even if you don’t optimize for one of the above reasons, you still shouldn’t be adding highly correlated assets to a portfolio, or you are likely be liable for churning to generate fees.
Adding highly correlated assets doesn’t add much value to a portfolio, that much is true, if you add one, you should be removing the other (or sizing it down) because the only reason to add a highly correlated asset is 1) you think it’s going to grow faster (but in proportion) than the other, so you are substituting a higher growth asset for a lower growth one, or 2) you think there is some risk that is present in the thing you are removing that is not in the thing you are adding (and high correlation reflects that the risk hasn’t materialized yet).
A lot of people have a problem with CAPM, and there’s more justification for those concerns, though what one replaces CAPM with to generate expected returns for a given set of risk factors then becomes the key question. Personally, I think there are a lot of problems with CAPM, but the equation I use still looks like CAPM’s security market line because I choose a simpler model - the index model - which happens to have the same final form, even though the assumptions are different. So I sometimes say I use CAPM, but what I really do is use something that happens to look like CAPM.
No. Making investments isn’t BS. Even trying to allocate assets isn’t BS. But calculating ER and StDev for 84 assets and optimizing them for minimum variance accordng to Markowitz theory, then going back and recalculating it all to see what the “new” correlations and CAPM expectations are on a quarterly basis is, IMHO, an exercise in futility.
As noted in another thread, cat + monkey > CFA + Jim Cramer.
EDIT - you edited while I was posting. Now I have to re-read and possibly re-edit.