Thursday, February 13, 2014

A Core of Mutual Funds: Part 1.

Three is really enough.

Why is the portfolio limited to three funds?   There are reasons grounded in financial economics, strong behavioral economic foundations as well as important practical considerations.  Each will be addressed.  Part 2 of this posting will describe the considerations relevant to the selection of the funds.  Then there will be a description of the three funds, which will be followed by a discussion of how much to invest in each fund and how to adjust the weights over time.

Financial economics.

Most mutual fund families and most 401(k)s will offer at least three good mutual funds.  In fact, most offer a lot more.  For the long-term investor, the broader variety is based upon the assumption that diversification across asset types allows one to achieve a better risk-adjusted rate of return.  It is always possible, in hindsight, to show that that could be true.  The general problem with the approach is that extensive diversity of asset classes will work well under normal circumstances.  It draws heavily on applying Modern Portfolio Theory (MPT) to the mix of assets.  It assumes that the investor is applying a trading algorithm to try to maintain the risk-return level continuously.  That algorithm may be as simple as rebalancing once a year, or it may be as complicated as the most sophisticated computerized trading system.

Asset allocation can be made to sound as complicated as possible.  However, it is actually a lot simpler than advocates of Modern Portfolio Theory make it sound.  The problem with the more complicated applications of Modern Portfolio Theory is that it is based upon faulty assumptions.  To be successful, it depends upon asset correlations (i.e., technically, the covariance matrix between all asset classes).  It assumes those covariances are stable.  Nothing demonstrated the fact that the covariances are not stable better than the recent financial crisis.  So, the theory works well under normal circumstances, and it results in a theoretical improvement in returns for any given level of risk.  When put into practice, any improvement in return is very, very slight, and, as the saying goes, “It works as long as it works and then it doesn't work.”

One does not have to understand the role of asset correlations in order to see the limitations and prudent use of the idea behind Modern Portfolio Theory.  Another way to envision the role of asset allocation is by looking at the performance of different asset classes over various periods.  A common way to show that data is a quilt chart.  If one analyzes a quilt chart, it becomes apparent that the biggest difference in performance is between bonds (or cash) and all other assets.  When there is a major market dislocation (a market collapse) or a major economic decline, all assets except bonds (or bonds and cash) have negative returns.  That exception is important.  The threshold between positive returns and negative returns is extremely significant mathematically and psychologically.

Behavioral economics.

Modern portfolio theory has been around since the 50s.  It is a very good theory if you assume that individuals’ risk tolerances are consistent over time.  However, there is considerable evidence that individuals do not have stable risk tolerances.  Further, people are not good at accurately judging their own risk tolerance, much less forecasting it. In addition, behavioral economists have shown that an abundance of options does not necessarily lead to better decisions, and the number of options influences one's perception of risk.

So, the three fund restriction may theoretically involve taking a little more risk.  In actual practice, most non-professional investors do perfectly well with the more restricted diversification of asset classes.  It may be that selecting fewer options requires them to acknowledge the risk involved in any investment.  Thus, they may be less likely to misdiagnose their own risk tolerance.  It may also be that fewer options actually anchor the risk tolerance at a higher level of both risk and return.  

It also would seem logical to assume that the more mutual funds selected the more likely it is that the investor will pick at least one mutual fund that is poorly run.  Having that blooper in one’s portfolio undoubtedly has some effect on one's risk tolerance.  Behavioral economists have pretty much been able to confirm that witnessing poor performance lowers one's risk tolerance.

Practical considerations.

First, the idea that one needs to hold a large number of mutual funds in order to be diversified into multiple asset classes is nonsense.  There are individual funds that invest across multiple asset classes.  Some try to assess an individual's risk tolerance and then adjust the asset mix to maintain that risk profile.  They may do it by periodically adjusting the mix across asset classes, or they may do it continually.

An example of the first type that periodically adjusts the asset mix would be one of the many target date funds.  Their assessment of the risk tolerance is based upon the individual’s age as reflected in the target date the individual selects.  The notion that all individuals of a given age should be exposed to the same risk, regardless of health, wealth, or disposition, does not make sense.  Despite the fact that it is absolutely absurd to assume that one's age is the only determinant of the appropriate risk, target date funds are very popular.

Their popularity seems to originate from the fact that they are simple.  Further, the periodicity with which they rebalance assets to their target mix is largely arbitrary.  It may be that their benefit is that they relieve the individual of having to make any asset mix decisions.  However, there is absolutely no reason to think that they make the right asset mix decision for any particular individual.

At the other extreme are the funds that continually adjust to a target asset mix.  As explained above in this posting, they suffer from the assumption that any individual can measure a different individual’s risk tolerance. Further, any improvement in performance from such an approach is so slight that many funds that pursue this approach have to take on leverage in order to realize a large enough yield benefit to justify their management fees.  The best examples of such an approach are some of the hedge funds (e.g., Long-Term Capital Management) that have blown up spectacularly.

Even though funds that over-diversify and try to determine the appropriate amount of risk have shortcomings, that does not mean that funds cannot provide some benefit of asset diversification.  Rather, these funds illustrate that multiple funds are not required in order to get asset diversification.

Every mutual fund has a minimum amount that has to be invested.  Further, some have charges that are steeper in percentage terms on smaller holdings.  If the amount of capital to be invested can only justify holding one mutual fund, a target date fund will provide asset diversification.  That diversification will result in less volatility.  However, in many cases,  where the amount to invest only justifies one mutual fund, strong arguments can be made for the need for the investor to expose themselves to greater volatility in the form of equity holdings beyond what would be characteristic of a target date fund.

In general, when considering mutual funds, the trick is for the investor to select funds that provide the appropriate amount of asset diversification for that unique individual.  It is also important to ensure that, once that desired level of diversification is achieved, it remain stable unless the investor chooses to intentionally change it.  Further, as discussed above in this posting, the most important aspect of diversification is exposure to both equities (stocks) and debt (bonds).  However, bonds only represented a diversifier and a source for funds when equity markets turned down.  Over the long-run, equities are going to outperform bonds.

No comments:

Post a Comment