hi Dan Duchardt,
Loosely translated, "if you put your money in the right assets, and avoid the wrong ones, you come out way ahead." No question about it. But I, or anyone else can look backwards in time and identify the best and worst asset allocation. There should be no surprise that there is a large disparity between people who were in the right places and people who were in the wrong places, but right and wrong in this context might mean guessed right or wrong, not that they had a superior or inferior way of approaching the market
you touch on a very important point. it is a different piece of the puzzle from what i was talking about before (that asset allocation, as opposed to individual security selection, accounts for the vast majority of return variance), but it is the next logical topic.
let's back up a bit and look at the big picture:
an imaginary investor desires to maximize returns for a given level of assumed portfolio risk. the investor cannot know with certainty what asset classes will provide the best returns in the future, but can look at historical returns for different classes. he can also look at historical volatilities (standard deviations (SDs)--a proxy for risk) for these different classes, as well as the correlations of these SDs to each other. the investor can also look at available information on the outlooks for each asset class, although such information may already be priced into those classes. let's assume that's the starting point.
now, the investor knows that asset classes are the primary determinants of return variance. this should be obvious to tech investors--the Nasdaq almost doubled in six months, but is down some 60%+ since then. at the same time, value stocks, for example, have had an excellent run. so you could take the brightest mutual fund manager in Nasdaq stocks, and the dumbest manager in value stocks, and the dumb value manager would have outperformed the Peter Lynch of the Nasdaq since march 2000. the reverse would have been true for the preceding six months. that's asset allocation for ya.
so, our investor nods his head and says yeah, that makes sense, but what good does it do me to know this unless i have a crystal ball and can switch asset classes as appropriate? well, the answer is there is no crystal ball, but there is something intriguing out there. what is it?
this is what it is: if you create a portfolio that deliberately mixes asset classes that have different degrees of correlation with each other (covariances), your portolio's return will (obviously) be the weighted average return of its component asset classes, BUT (not so obviously, and quite interestingly), the standard deviation of the portfolio as a whole will be less than the average of the SDs of the component assets.
so what does this mean? what matters most for the investor is how the portfolio as a whole does, as opposed to how this or that 3% of the portfolio does. the two things the investor is most concerned about--assumed risk and return--are interrelated. by mixing asset classes, the investor is able to lower the standard deviation for the portfolio as a whole for a given level of expected return.
our investor can create a graph with return (%) on the y-axis and SD on the x-axis. whatever the expected level of return he wants to go for, he wants to minimize his risk. therefore, he seeks to combine disparate asset classes to achieve the best risk/reward ratio for his targeted return. of the infinite possible asset combinations, there is one combination for each expected return which has the lowest level of risk for that return. that is the most "efficient" portfolio for that target return. the entire set of efficient portfolios over the range of target returns (i.e., the ones with the lowest risks for their respective target returns) is called the efficient frontier.
obviously, our investor would like his portfolio to be on the efficient frontier.
but there is a problem: while it is possible with hindsight to recognize the best risk/reward combinations (i.e., the efficient frontier for the past), we cannot know for certainty what the efficient frontier is in the future. the reason we cannot know is that the inputs are uncertain. what are the inputs? they are, for each asset class, things like its return going forward, its volatility (SD) going forward, and its correlation to other asset classes.
which is to say, the investor is not certain that his chosen portfolio will be the most efficient going forward. in fact, it probably not be. all he can do is try to put together a combination of assets that should do well under most conditions, and should have a relatively high level of efficiency. that is certainly not perfect, but to my mind, it's a heckuva lot better than betting the whole wad on a single asset class (especially one that's just enjoyed the greatest bull market in 200 years!).
apropos, i wrote the following on another thread regarding a book by Roger Gibson (called "Asset Allocation"):
Gibson's book has some very interesting charts in it. one shows a chart of reward/risk (return% on y-axis, standard deviation on x-axis) of four asset classes from 1972-1997: S&P500, EAFE (foreign large caps, which during the survey period were largely Japanese and British firms), GSCI (commodities) and NAREIT (REIT index). each of these four classes is plotted by itself on the chart, and all possible equal combinations of these assets (e.g., 50/50 S&P and REITs; 33/33/33 GSCI, EAFE and S&P) are also plotted. the worst portfolios were the individual assets by themselves (e.g., 100% GSCI). the best portfolio was a combination of all four (25/25/25/25 S&P, GSCI, EAFE, S&P). "best" in this case means a portfolio that yielded a very high return versus its standard deviation. portfolios of this type are called "efficient".
i also discussed some other resources in that post #reply-16165958
OK, gotta run now, i'll address your other points later... |
