Small changes in the assumptions, as we all know/learned, of these models cause havoc. And given ten years from now, the actual results will probably not be near the median of the distribution or even anywhere on the expected curve. I see greenie’s point. But i still think the math is useful in framing the discussion. I.e. maybe adding a little x can increase return and lower risk. Just introducing the idea that adding treasuries could improve the sharp ratio makes the models useful. Thinking you’ve found a holy grail, i’m with greenie. I keep the math mental. The software vendors have learned to look elsewhere.
Remember that “considering something in the context of the portfolio” does not necessarily mean “use Mean-Variance Optimization.” What it means is “consider what this asset combination does to the risk and the return of the entire portfolio - however you calculate that.” MVO is just one way, and it has a number of problems. Equal weighting is another way. You can do equal weighting, but you’d be crazy not to at least look at what your new asset does to the portfolio’s total risk (as well as the return).
Even if you reject MVO for whatever reason (sensitivity to inputs + inability to generate reliable inputs is a decent reason), you still have to consider your investments in terms of what they do to your risk. A portfolio that is equally allocated between EEM and SPX futures is very different than a portfolio that is equally allocated between EEM and SPY (or SPY and SPX Futures).
Well, I think this goes without saying.
I live in a town where the as the price of WTI goes, so goes the economy. And I go crazy every time I see somebody who buys into another oil and gas asset. Stock in Chesapeake Energy, MLP’s from Legacy Reserves, etc. It seems like they just WANT their entire fortune to go to pot when (if?) the boom turns to bust. To sell one oil-and-gas asset to buy another one is kinda ludicrous.
I agree–if a person wants to appropriately diversify their portfolio, then they shouldn’t sell United Airlines and buy American Airlines. Maybe they should sell United and buy a railroad (Berkshire).
See? No math needed. No covariance matrices, no advanced topology and linear algebra and elementary analysis multivariate calculus. Don’t need to know CML’s and SML’s and efficient frontiers and CAPM’s. Don’t need to derive the expected StDev for a three-asset portfolio.
And I certainly didn’t need to study for three years to pass a test to know that.
The test is evidence that you were exposed to the concept and demonstrates that you probably have at least an elementary understanding of mvo, etc. That gives me comfort if i were to do business with you and pay for your opinion.
Right, but “considering an asset in the context of the portfolio” is not equivalent to saying “consider the asset in the context of a portfolio optimized using MVO”.
As a reminder, this is what you said:
There are decent reasons not to use MVO (such as the sensitivity to estimation errors, also discussed in L3), but there really are no reasons to say “consider the entire portfolio… …is more or less a bunch of baloney.”
The key insight of MVO, is that an uncorrelated asset - assuming the correlation is somewhat persistent - can add value to a portfolio even if its expected return seems a bit lower than you’d normally bet on (the argument for having bonds in general); it also suggests that there is little value in adding a highly correlated asset (which many people have an easier time grasping intuitively, as you did, but the people you advise don’t always get).
A lot of people misinterpret the correlation argument as saying “if it’s uncorrelated, you should always add it.” What they forget is that “if it’s uncorrelated AND you trust the correlation to be somewhat stable AND it generates positive excess returns AFTER transaction costs, then it’s useful to add IF YOU CAN LEVER UP AND DOWN APPROPRIATELY to readjust your risk tolerance.”
That’s a more complicated decision, but it definitely applies for some assets and some investors, so it’s not useless.
Many people expect fixed income (or at least interest rate risk) to have negative excess returns right now, and there’s a good-but-not-watertight argument for that. Even if you accept MVO and that bonds will be low-correlation with stocks going forward, MVO and pretty much any portfolio construction method will tell you to quit bonds altogether if the excess returns are expected to be negative.
OK–I’ll give you this. It’s certainly better than “find the expected return of A and expected return of B and stdev of A and stdev of B and the correlation between the two and plot it on an efficient frontier and the less correlation the better” crap they spew in the CFAI curriculum.
Diversification is a good thing–at least within asset classes. But diversying your asset classes starts to muddy the water. Does a person need exposure to stocks and bonds? Most probably do. Do they also need exposure to real estate? Do they need commodities? Do they need small-cap emerging markets? Do they need floating-rate high-yield international municipal bonds? Moreover, do they need to pay their advisor huge fees to tell them all this?
I think that once you get past the initial “% to stocks, % to bonds”, it becomes extremely hard to justify higher fees for what amounts to very little, if any, excess return. And I think that historical correlation models and CAPM’s and EMH etc. becomes more or less useless in real life.
Probably some cute salesperson at an investment bank gave him a handy and told him she liked his silver hair and he only heard part of it.
He must have some mojo. Didn’t his first wife give his mistress her blessing? That’s well done.
^Yeah, Snowball goes into a lot of detail about it. About how like he’s flying to Washington to meet with the woman who ran the Washington Post like every other week and stuff and his wife didn’t care at all.
Snowball was funny in that it sort of glossed over it while hinting heavily. How do you just blatantly call out the grandfather of capitalism in print for sleeping around? Can’t be done. I did feel sorry for his wife though.
If you’ve ever read Janet Lowe’s book about Benjamin Graham, she goes into great lenght about his womanizing. Doesn’t try to hide it at all.
Billionaires sleeping around? Get out!
^You must not be familiar with Mr. Bruce Wayne, CEO of Wayne Enterprises. He gets lots of hotties.
^ He is an underachiever. He can do way better.