Is the Efficient Frontier Hogwash?

[hedgehog]

To quote AgAuMoney: http://gyroscopicinvesting.com/forum/ht ... 7#msg38807

I had to start a topic for this because I haven't found any discussion regarding the efficient frontier but it's important.

I don't that much like the 40/60 allocation proposed by Modern Portfolio proponents but I kinda like the Efficient Frontier graph:

http://www.ifa.com/portfolios/#ChartFlashID2

Is it hogwash or relevant?

More: http://gyroscopicinvesting.com/forum/ht ... 6#msg69276

[craigr]

The efficient frontier is driving by looking in the rearview mirror. It only can tell you what worked best in the past, it can't do anything about the future. That's why when you move the analysis around by even a year or so the allocations recommended in an efficient frontier analysis will shift all over the map.

I say this often, but the only thing an investor can do is widely diversify with asset classes that have few overlapping risks. There is no way to know what was most efficient until decades after the fact. And then of course what will be most efficient from that point forward will be different. It's chasing pots of gold.

[hedgehog]

1st, all people on this forum use all kinds of backtesting, including you. Driving by looking in the rearview mirror, as you say. So why they still do it?

2nd, I just kinda like the idea to show a graphical correlation between risk and reward: Higher risk implies higher possible rewards. Does that make sense? So how would you show this correlation without looking in the rearview mirror?

[Xan]

I believe the point is made in Craig and Tex's book that backtesting can DISPROVE an investing theory or strategy, but not prove one.  The extensive backtesting on the PP primarily shows that it has gone through many scenarios and not yet been disproven, as opposed to most other options.

[craigr]

Yes to build on what Xan said, I primarily use backtesting to disprove ideas. Even then, there is an element of intuition that I think needs to be applied because the data set everyone draws conclusions from in terms of the U.S. market is very small. It's only around 80 or so years of kind of/sort of reliable data. And that assumes no survivorship bias. And even with that, only about 40 or so years most people commonly use as new asset classes came available.

So think about that for a second. The data everyone uses to prove that some asset allocation will be great going forward based on the efficient frontier is barely covering a single person's average lifetime. Data prior to that was spotty and the typical person probably couldn't have invested in the markets so easily. This is why I chuckle when I see people posting 200 years of stock data in various charts. If you think about all the history of just the U.S. over the last 200 years it is mind boggling the country still exists under one government. It could have blown apart in 1812, Civil War, WWII, etc. So taking this past set of data and trying to project it forward is a really bad idea in many ways.

Also I suggest reading the IFA disclosures (emphasis added):

http://www.ifa.com/btp/

The performance information presented in certain charts or tables represent backtested performance based on combined simulated index data and live (or actual) mutual fund results from January 1, 1928 to the period ending date shown, using the strategy of buy and hold and on the first of each year annually rebalancing the globally diversified portfolios of index funds. Backtested performance is hypothetical (it does not reflect trading in actual accounts) and is provided for informational purposes only to indicate historical performance had the index portfolios been available over the relevant time period.

Backtested performance does not represent actual performance and should not be interpreted as an indication of such performance. Actual performance for client accounts may be materially lower than that of the index portfolios. Backtested performance results have certain inherent limitations. Such results do not represent the impact that material economic and market factors might have on an investment adviser’s decision-making process if the adviser were actually managing client money. Backtested performance also differs from actual performance because it is achieved through the retroactive application of model portfolios (in this case, IFA’s Index Portfolios) designed with the benefit of hindsight. As a result, the models theoretically may be changed from time to time and the effect on performance results could be either favorable or unfavorable.

IFA has also changed their portfolios around over time depending on new asset classes coming available:

http://www.ifa.com/btp/historyofchange.html

I stand by my original point about the efficient frontier. It's chasing rainbows.

[rickb]

To show the relationship between risk and reward I think most people plot the return vs. standard deviation for a selection of increasingly risky portfolios (I would image William Bernstein has such a chart in one or more of his books).  Is your overall point here that if you do this for (say) a stock/bond mix, then as the mix goes from "less risky" (100% bonds) to "more risky" (100% stocks) the line won't always have a positive slope (i.e. by adding stock to a 100% bond portfolio, for a while you'll increase return and decrease risk)?  Although this might seem counter-intuitive (and it relates to the Permanent Portfolio as well), as you add risky (uncorrelated) assets you actually decrease overall risk and increase return - apparently violating the "more risk = more reward" dictum.  Once the overall return is dominated by the riskier asset, this line shows the expected "more risk = more reward" relationship.  The less risk/more reward anomalous portion of the curve follows from the math behind MPT - but violates our intuition.

[hedgehog]

If you think about all the history of just the U.S. over the last 200 years it is mind boggling the country still exists under one government. It could have blown apart in [...] WWII, etc.

Emphasis from me.

But thanks for your post, otherwise. (I am not American, but compared to the European situation in WWII )