Managing Variability in Thrift Savings Plans

A basic knowledge of variability is essential for managing your investments in the Thrift Savings Plan (TSP). I am always interested in the FedSmith.com TSP related articles regarding decisions investors are making in managing their funds and moving money between the various funds.
How do you avoid buying high and selling low? On the flip side, how do you take advantage of buying low and selling high? Would you get better returns if you did nothing and took advantage of averaging? By contributing every month, you buy at a range of highs and lows that, over time, may result in a positive gain, such as the case with the C Fund that invests in stocks. In reviewing the performance of the C Fund from 1988 to 2011, the total cumulative return has been 261.85% and the average annual return was 10.91%.
Figure 1 shows a control chart of the C Fund’s annual returns. This chart includes a center line (CL), an upper control limit (UCL) and a lower control limit (LCL). The UCL and LCL identify the expected or normal range of variability in the annual return.

ClarkTsp_CFundAnnualReturns.png


Investor Herding and Tampering
In December 2009, The Wall Street Journal published an article titled "Best Stock Fund of the Decade: CGM Focus". The results of its analysis indicated that the CGM Focus Fund “rose more than 18% annually and outpaced its closest rival by more than three percentage points. Despite the Fund’s success, the typical investor lost 11% annually. The gap between the CGM Focus Fund’s return and the investor return was considered “the worse of any fund tracked by Morningstar.” This situation was discussed in a blog by Robert Folsom, who referred to it as investor herding. Basically, investor herding is when individuals join the crowd in taking action that is unplanned and driven by emotion with a sense of urgency to get in or out of the market.
Dr. W. Edwards Deming would refer to this investor behavior as tampering. Tampering is performing an action that makes a stable system worse. In the case of the C Fund, annual results from 1988 to 2011 indicate a stable system. “Stable” means that the annual return fell between the UCL of 64.55 and the LCL of 42.73 and exhibited no statistically significant trends. Investors determine if the range and results are good or bad.
Investment experts have indicated that the interdependencies of the global market, instant communications, speculation, and more involvement by governments in the market can add to a more dynamic and unpredictable investor climate. This makes “plotting points” to help separate the “signals from the noise” even more important.
Any system contains two types of variation: variation that is common (e.g., normal, usual, expected) and variation that would be considered special (e.g., unusual, unexpected, an outlier). A stable system is more predictable and contains only common causes. An unstable system contains both common and special variation.
The aim for managing variability is to minimize the two types of mistakes that can be made:
Mistake 1. Reacting to a change in price as if it were from a special cause when it actually came from a common cause. An example of mistake 1 would be selling in 2008 when the share price was at its “record low” of -36.99. Other examples are headlines that contain keywords regarding market activity as soaring, plummeting, rising, falling, etc. The keywords imply special or unusual activity even though the respective prices fall within what can be considered a normal or expected range the majority of the time.
Mistake 2. Treating the change in price as if it were from a common cause when it actually came from a special cause. Examples of mistake 2 would include not knowing how to detect a special cause and not knowing how to correctly assess whether a special cause is a permanent or temporary event.
Never making mistake 1 or 2 is impossible. The aim is to regulate the frequency of the two types of mistakes to achieve minimum economic loss. Deming estimated that failure to understand how to assess a performance trend in order to separate the usual (common) from the unusual (special) results in situations where 95% of changes made typically make things worse. This was certainly the case regarding the investor behavior in the CFG Focus Fund.
Plotting and assessing key trends of home mortgage lending, borrowing and defaults would have provided early warning of the impending financial collapse in 2008. As members of the “herd” were being led on the path to losing some or part of their investments, others were making billions betting on the market realities of the high risk associated with making what historically has been considered bad loan practices. The repeal of portions of the Glass-Steagall Act that removed the checks and balances on banks and security firm affiliations is another variable that some members of Congress believe contributed to the crisis.
The foundation of Deming’s work was based on the contributions of Dr. Walter Shewhart, who developed the new paradigm for managing variability with the development of the control chart in 1924. Deming’s work with the Japanese after WWII, which then expanded worldwide, was considered by the editors of FORTUNE magazine as among the “greatest contributions in business history.” Deming remarked that if he was to reduce his message to just a few words, it all had to do with reducing variation. Reducing variation requires a basic knowledge of common and special causes of variation. In 1986, Deming estimated that it would be another 50 years (2036) before Shewhart’s contributions were more commonly understood and applied.

When to Change, When to Leave Things Alone
There are several trend interpretation standards to help identify common and special causes. One of the simpler standards is what I refer to as the rule of seven (7), which I typically recommend as a great start point in assessing trends. The steps include the following:


  • Create a trend chart and add the center line (the average).
  • Analyze the trend for statistically significant patterns. Unusual patterns or special causes include the following:
    • A run of 7 consecutive points going up or down.
    • A run of 7 consecutive points either above or below the center line.
    • Points that appear too far away from the average line. (These points estimate the upper and lower limits, which can also be determined using a control chart).
I have found that elementary school kids can immediately apply the standard. They also understand the link between their behavior (e.g., doing their homework, participating in class, working hard) and their results. For example, I sponsored a team of kids who presented their trends on the results of class math scores at an international quality conference sponsored by theAmerican Society for Quality. The kids had one significant drop (special cause) in scores that was questioned by the quality professionals in attendance. The kids admitted that they had a substitute teacher that day, so they were less disciplined in completing their assignments.

A Way Ahead
We are in year 88 of a 112-year paradigm shift to a situation where the concept of managing variability becomes more common. I would encourage you to review my previous articles that I have written to raise awareness of the paradigm and to help shorten the learning curve.
The techniques for “plotting points” to identify common and special causes of variation can be applied to not only help manage investments but also help improve the performance of any system or process. This includes plotting points to identify common and special causes of variation in organizational performance, daily commuting time, your weight, golf scores, crime statistics, the unemployment rate, and the national debt as a percent of gross domestic product, to name a few.
There are also numerous ways of analyzing the data. For example, you can identify trends by calculating a moving average or by analyzing the data by different time periods (e.g., hour, day, month, quarter, year, decade).
Regardless of the method used, it all boils down to the bottom line—did the method result in decisions that resulted in a positive gain? In the case of moving money in and out of TSP funds, are you making money, or are you following the herd or performing some other type of tampering?
 
Back
Top