Jump to content
Objectivism Online Forum
Sign in to follow this  
prosperity

Modern Portfolio Theory Debunked?

Rate this topic

Recommended Posts

I have never really been comfortable with the CAPM and modern portfolio theory, especially since it always seems like something is missing when trying to fit clients into nice, neat, little "risk boxes" on a piece of paper. Mainly because there is no asset allocation questionnaire that can accurately assess what a person should invest in for the rest of their lives. If life was static and unchanging, you might make a case for that.

I had read a few explanations as to why CAPM and MPT were not really all that great, but this article peaked my interest, and I wanted to share it:

http://www.travismorien.com/FAQ/portfolios/mptcriticism.htm

With the exception of his statement that he believes markets are not efficient, I would have to say the author makes some excellent points. What do you think?

Share this post


Link to post
Share on other sites

The core idea that modern portfolio theory propounds is this: that the covariance/correlation of returns between the various parts of a portfolio are critical in the determination of overall risk. At that high level of abstraction, it makes sense, and is just a little different from the old saying "don't keep all your eggs in one basket". It takes that one step further and says: "don't keep all your eggs in closely correlated baskets"... where one basket fails when the others do.

All the math behind the MPT is simply a formal way of proving the point.

However, the problems arise when one tries to actually quantify risk and then plug those values into an MPT-based model. That's where things start to break down. Different analysts have different views on the amount of risk in a particular investment, as well as on the degree of correlation between two investments. So, going in search of a measure that does not depend on opinion, practitioners use past data. This is the first mistake, since past data does not speak to risk, which is about the future. The second poor assumption that is made is to equate risk with volatility, particularly short-term volatility. These two poor assumptions render the actual calculations fairly meaningless for the purpose of investing.

Share this post


Link to post
Share on other sites

I think Mandelbrot (book) makes the very good point that the reliance on normal distributions in economic theory is probably wrong.

So your underlying process in the theory has the wrong behavior.

Share this post


Link to post
Share on other sites

The problem with this analysis is is looks at individual stocks to show that there is no correlation between "risk" and return. The results stated, that lower "risk" stocks have higher return and higher "risk" stocks have lower return, holds true in individual instances. It's from aggregating stocks that the higher returns from volatile stocks arise.

The basis of the theory (and I believe that it holds up to empirical analysis) is that stocks in general have a random volatility that has logarithmically even (log normal) distribution. That is, a stock is as likely to have a 10% rise as a 9.09% fall [ ln(1 + 0.10) = -ln(1 - 0.0909) ]. (If a stock rises 10% one time period and falls 9.09% the next, the result is that the price has returned to its original value.) If you own two stocks, to take the simplest case, and one rises 10% and the other falls 9.09%, you end up with a portfolio gain of 0.909%. Whereas if one rises 5% while the other falls 4.762%, you end up with a portfolio gain of only 0.23%.

I'm not a stats whiz, so I don't know exactly where or if the theory falls apart from there.

(edit) It should be noted that the log-normal distro theory does not hold hold true for stocks that lose all value (there would have to be a finite possibility of an infinite return to offset the instances of total loss of value), so is suspect. The apparent aggregate rise in stock values is possibly due more to the general tendency of the economy to increase in value over time.

Edited by agrippa1

Share this post


Link to post
Share on other sites
The core idea that modern portfolio theory propounds is this: that the covariance/correlation of returns between the various parts of a portfolio are critical in the determination of overall risk. At that high level of abstraction, it makes sense, and is just a little different from the old saying "don't keep all your eggs in one basket". It takes that one step further and says: "don't keep all your eggs in closely correlated baskets"... where one basket fails when the others do.

All the math behind the MPT is simply a formal way of proving the point.

However, the problems arise when one tries to actually quantify risk and then plug those values into an MPT-based model. That's where things start to break down. Different analysts have different views on the amount of risk in a particular investment, as well as on the degree of correlation between two investments. So, going in search of a measure that does not depend on opinion, practitioners use past data. This is the first mistake, since past data does not speak to risk, which is about the future. The second poor assumption that is made is to equate risk with volatility, particularly short-term volatility. These two poor assumptions render the actual calculations fairly meaningless for the purpose of investing.

Right, there is a serious problem when you go to measure risk (when defined as "standard deviation from the mean"), because how do you do it? Do you measure just the "downside"? and if so...how? in reference to what?

It sounds like you are saying "good in theory, bad in practice". I'm not so sure it's even a very good theory though. And the abstraction only makes sense under a definition like "the probability or likelihood that a desired event will not occur", which falls outside of the realm of investing since "officially" they have declared it to be "standard deviation from the mean".

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

Sign in to follow this  

  • Recently Browsing   0 members

    No registered users viewing this page.

×