Seeking Alpha
Long/short equity, long-term horizon, portfolio strategy, dividend investing
Profile| Send Message|
( followers)  


  • Should the small investor play Wall Street's game?
  • Continuous trading strips individual investors of one of their strongest advantages over professional traders.
  • There are a number of companies, including WiseBanyan, Betterment, and Wealthfront, that bill themselves as investment advisers. An algorithm is not an adviser.

Jason Zwieg's column, The Intelligent Investor, is always interesting. His book Your Money And Tour Brain is a delightful read, both interesting and informative. The column in the March 8 Wall Street Journal article entitled "The Incredible Shrinking Management Fee" was a good report on a new type of asset management company. However, it left much to be desired if it was intended as a presentation of the phenomena for potential investors.

First, the subtitle, "If a new company has its way, the cost of portfolio management could be zero," reflects the problem. The article discusses the number of companies, including WiseBanyan, Betterment, and Wealthfront, that bill themselves not as portfolio management, but as investment advisers. One thing to note is that Wealthfront does not seem to be a fiduciary. The article contains no reference to whether Betterment or WiseBanyan are fiduciaries. In other words, they do not have a legal obligation to act in the investor's interest.

They have a business obligation to do so, but not a legal obligation. That is true of most investment companies like mutual fund companies, but it is unusual for someone who is taking financial management responsibility not to be a fiduciary. Turning over asset allocation decisions to a financial manager who is not a fiduciary should make an investor uncomfortable. However, there are some mutual funds that retain some asset allocation control, and occasionally they are appropriate for an individual investor. However, such a mutual fund could, and should, be purchased from a well-established company with a track record. Further, the fund's prospectus will describe explicit limitations on the variability of the asset allocation.

The second thing to be aware of is that these companies are venture-capital financed. The important thing to remember about startups is that the majority of them will not survive. Some will go bust, some will just return investors' money, some will be ho-hum investments, but it is the return on the remaining ones that justifies the investments. If one in 10 is very successful that is good because they are owned directly by the investor. So, the one in 10 can return enough to offset all the others.

Investing through a startup as opposed to investing in a startup is a different story. There is no potential for that really big payoff if the company is successful because the investor does not have an ownership stake in the company. At the same time, if these startups do not succeed, that could have implications. These companies are entering a very competitive industry. So, it is important to be extremely confident about the success of the business model and the individual company which the investor chooses to have manage the money. The implications of the failure of one of these businesses would require a lot of research to understand. That, in-and-of-itself should be a powerful deterrent for some small investors.

Third, whenever one turns money over to someone else to manage, it is very important that the investor trust the money manager. Since the companies seem to be exclusively web-based, there is less of a basis for establishing a judgment about their honesty. Plus, because these are new companies, there are no track records on which to base judgments. Not surprisingly, the senior executives at some of the companies have venture-capital backgrounds. Perhaps, information about whether they made money for investors in their venture-capital firms would be relevant, but even that is only indirectly indicative and would have to be interpreted very carefully. It should be a concern to investors even though it is more than likely that the senior executives are honest, hard-working individuals.

The point is that trust is a judgment the investor has to make. In the case of these companies, there is very little information on which to base the judgment. One cannot count on anyone else making that judgment. For example, the Securities Exchange Commission completely missed Bernie Madoff's fraud. Similarly, one cannot rely on the Financial Industry Regulatory Authority, a self-regulatory body commonly called FINRA. A recent (March 5) Wall Street Journal article had the telling subtitle "Analysis Shows More Than 1,600 Stockbrokers Have Bankruptcies or Criminal Charges in Their Past That Weren't Reported."

Fourth, given the companies' descriptions of their investment approaches, the investment will probably work well under normal circumstances. However, how well it works for an individual investor depends upon the quality of the analysis of the investor' risk tolerance. The firms probably have a questionnaire to determine one's risk tolerance. What they are trying to do is apply Modern Portfolio Theory (MPT) to the asset allocation based upon the results of the questionnaire. They use trading algorithms to try to maintain the risk-return level continuously.

Modern portfolio theory has been around since the '50s. It is a very good theory if one assumes that individuals' risk tolerances are easily measured and consistent over time. However, there is considerable evidence that individuals do not have stable risk tolerances. Jason Zwieg's book and many of his articles, especially those reporting on the research of behavioral economists, have pointed out that risk tolerance is neither stable nor independent of investment performance. Consequently, it is a bit surprising that he does not point out this problem with the approach that the companies are using. Risk tolerance is one of those concepts that works nicely in theory, but it appears to be totally meaningless when applied to reality.

Interestingly, the same author, in the same feature, in the same newspaper, published an article the next week entitled "The Bull That Got Away." The point of that article was that recent market developments present risks because they influence investors' perceptions of risk. The research on which that conclusion is based is extremely fascinating. Unfortunately, titles like, "Path Dependence in Risky Choice: Affective and Deliberative Processes in Brain and Behavior," "Gambling with House Money in Trying to Break Even: The Effect of Prior Outcomes on Risky Choice," and "Neural Signature of Fictive Learning Signals in a Sequential Investment Task," makes the issue sound far more complicated than it is. Put simply, recent returns can distort an investor's behavior for a variety of reasons. A very important one of those reasons is that recent stock market gains and losses change how the brain assesses risk.

One could reasonably argue that the principal advantage of an algorithm over a human adviser is that the algorithm has no emotions. The problem is when the algorithm, which has no emotions, interfaces with the investor who does. There is no reason to believe that investors would not systematically move in and out of this investment with detrimental timing. Investors have shown that they do that with other investments. That is a problem that no investment vehicle can overcome. It is also worth remembering that an algorithm is a product of a human who is affected by the same emotions. Who is to say that the algorithm will not be changed (i.e., upgraded to Version 2.0) in response to the emotional reaction of the programmer to recent returns. However, ignoring the issues related to measuring risk tolerance is only the tip of the iceberg.

Fifth, there is a basis for suspecting that algorithmic trading encourages an incorrect assessment of risk. To understand the issue, consider an article written by The Numbers Guy in The Wall Street Journal. The article published on January 4, 2013 entitled "Don't Let Math Pull the Wool Over Your Eyes" makes the case that many people, including holders of graduate degrees, professional researchers and even editors of scientific journals, can be too easily impressed by math. Since that is the case, it is likely that the investor and the companies offering the service do not understand the limitations of their approach. Thus, it is unlikely they correctly assess the risk.

Sixth, the marketing pitch that is quoted in the Zwieg article: "Investment management can be more expertly done by an algorithm than by a human adviser," should be cause for concern. It could illustrate overconfidence and hubris on the part of those developing the services. It also might represent a cynical willingness to use the tendency for people to be overly impressed with math. Finally, as discussed below, it represents an incredible level of naiveté on the part of the developers.

Seventh, anyone who works with models is aware that they are models. They have limitations and are based on assumptions. The algorithm is simply a model of how the market functions. All algorithms that stress diversification across asset types are applications of Modern Portfolio Theory. Modern Portfolio Theory is an excellent theoretical framework for making asset allocation decisions. There is little doubt of the validity of the saying, "the only free lunch in investing is diversification." It is an excellent theory, but that is all it is.

Asset allocation can be made to sound complicated. However, it is actually a lot simpler than advocates of Modern Portfolio Theory make it sound. It is just diversification. The problem with algorithmic application of Modern Portfolio Theory is that it is based upon the correlations and covariances between all asset classes. It assumes they are stable. Nothing demonstrated the fact that they are not stable better than the recent financial crisis. (Technically one might argue that they do not have to be stable. They just have to be measurable and "forecastable." However, the combination of discontinuities and the tendency for nonlinear models to be unstable necessitates adding even iffier assumptions than stability in the correlation matrix).

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. Consequently, frequently professional managers try to take advantage of the slight improvement by increasing the leverage. That, of course, creates its own risk. However, the most common problem with the approach is that, as the saying goes, "it works as long as it works, and then it doesn't work."

There is an eighth point that was missed in the article. Algorithmic trading is much the rage with many on Wall Street. What is not clear is whether there is any return to continuous, algorithmic trading in-and-of itself. The largest firms involved in the activity also are accessing information not available to the general public or only available to the general public with a delay. Research that separates the returns due to the information advantage from the return due to their trading algorithm is sorely needed.

Further, and this should be a major concern, if these new firms serving the small investor are trading with an algorithm that is similar to those used by one of the larger firms but only receives the information on which the trades are based after it has been available to the larger firms, there may be no advantage regardless of how rational the algorithm is. In fact, the trading by these firms may only enhance the returns of the larger firms which have better information.

Ultimately, an individual investor will do better by just avoiding continuous trading as a method of reducing volatility and increasing return. Volatility between the beginning date and the ending date of a trade is actually quite irrelevant to the return from the trade. It is important in Modern Portfolio Theory, but only matters to the individual investor if he or she is willing to accept volatility as the definition of risk.

Continuous trading strips individual investors of one of their strongest advantages over professional traders. The individual knows how long he or she intends to hold the asset whereas the professional investor has to continuously worry about the mark-to-market value of their assets. For the individual investor giving up that advantage in order to conform to the tenants of Modern Portfolio Theory seems like a foolish choice.

Disclosure: I have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it. I have no business relationship with any company whose stock is mentioned in this article.

Source: Does Algorithmic Trading Make Sense For Small Investors?