Stock market models are generally used for market timing, a subject which has a deservedly bad name. Market timing has led many investors to ruin because they missed big moves. For example, after stocks bottomed in 2009, many models didn’t give buy signals until stocks were already up 30% or more. The delay was based on the model’s design and buying after the signal would have been profitable for many investors. Unfortunately, investors are prone to override their models. In the summer of 2009, sentiment remained bearish and some investors were waiting for a pullback to get in. By the time the pullback came in 2010, the S&P 500 had gained almost 90%. Investors could have gained 60% even if they missed the bottom which demonstrates the type of gains possible by identifying the direction of the big trend, rather than trying to time the exact moment the market turns.
Models are not perfect and there will always be losing trades in any model. But the goal of a good model is to help investors identify profitable trading opportunities while avoiding big mistakes. A well-designed model will never be on the wrong side of the market for long.
To define a well-designed model, we turned to the work of Ned Davis Research, an institutional research firm that is widely respected among traders working at hedge funds and other large investment funds. According to Ned Davis, a good model should minimize risk by using a “weight-of-the-evidence” approach to making decisions. This means models should always include at least two indicators and at least one indicator in any model should be based on external indicators, which are not related to the price action in the stock market. External indicators can measure investor sentiment, actions of the Fed, economic conditions or other factors unrelated to prices.
Ned Davis Research (NDR) follows dozens, if not hundreds, of models. As individual investors we don’t have time to follow that many and some of us might only have time to follow one. To demonstrate the value of simple models, I will use one of NDR’s models that uses just two indicators and follows monthly data. It’s a model any investor should be able to update and follow in less than ten minutes a month.
This model is designed to benefit from the age old advice, “don’t fight the Fed.” Years ago, traders focused on the Federal Reserve, paying close attention to weekly reports on money supply and tracking short-term interest rates. This was before CNBC and the proliferation of economic news sources. In those simpler times traders noticed the Fed was a driving force behind the direction of the short-term trend. If the Fed was increasing the money supply or lowering short-term rates, stocks were generally going up. When the money supply was shrinking or rates were rising, stocks tended to fall. This led traders to act in ways where they “don’t fight the Fed” meaning they would always try to stay on the same side of the market as the Fed.
In our model, which NDR named the “Don’t Fight the Fed Model,” we’ll use the short-term interest rate on financial commercial paper. Commercial paper consists of short-term loans, usually for nine months or less, issued by financial companies in this case. Commercial paper is a liquid source of financing for banks. Traders will set interest rates in this market based on their perception of short-term risks for banks. If rates are rising, investors in commercial paper are most likely viewing banks as a risky investment and demanding a higher risk premium for lending money to banks. In this environment, with rates rising, we would expect to see stocks falling as concerns about risk spread to stock market investors.
Under the opposite conditions, when rates are falling, commercial paper investors are viewing banks as less risky. We would expect stocks to rise in this environment where concerns about risk are decreasing.
That’s our first indicator. We will use the interest rate on 3-month financial commercial paper. The Federal Reserve makes it easy to find this data and offers it for free at their web site. The indicator is formally called the “3-Month AA Financial Commercial Paper Rate.” We then add a 10-month moving average (MA) to the data. This can be done in Excel or any spreadsheet software or you could do it by hand since only 10 data points are needed. When rates are below the MA, the indicator is bullish. A sell signal is given when the rate is more than 0.10 percentage points above the MA. The delay in requiring rates to be at least 0.1% above the MA minimizes the risks of whipsaw trades. For example, if the rate is 0.12% and the MA is 0.11%, the rate is above the MA but no sell signal is given. The sell signal would require rates to be 0.21% or higher, at least 0.1% above the MA. This might be confusing at first, but the results are worth the extra thought required to understand the signal.
A long-term chart of this indicator created in Excel is shown below.
A shorter-term view shows the indicator is bearish right now.
This indicator has been bearish since August 2015.
The second indicator in our Don’t Fight the Fed Model is a simple moving average of the S&P 500. Here we use a 12-month MA and buy signals are given when the index is above the MA. This indicator is bearish when the end of month price of the index is below the MA. You could use SPDR S&P 500 ETF (NYSE: SPY) or the S&P 500 index for this calculation. Both are available for free at a number of web sites and moving averages can be added at a number of those sites.
This model will only take you a few minutes to update each month and bookmarking the Fed’s web page will help you find the data quickly. The web page’s address is shown below the chart of commercial paper rates.
The question we have to answer now is whether or not the work will be worth the effort.
When they published their study in 2014, NDR offered the following results based on a back test using data starting in 1968.
This model would have gotten you out of stocks in 2008, 2000, 1990 and 1974, allowing you to miss most of the major bear markets of the past fifty years. It also provided a signal to get back in after the bull market resumed.
Since the model was published by NDR, stocks have drifted higher but you would have moved to cash for brief periods and missed part of last summer’s selloff.
Based on the results, this strategy meets the objective of a well-designed model to minimize risk. Most importantly, you get few outright sell signals. This means you would be in cash only about 11% of the time and those are exactly the times when risk is highest. The model also signals those times when you could be most aggressive, which is when both indicators are bullish. Using leveraged ETFs you could possibly double the returns achieved in the back test.
Right now, the indicators are mixed with the S&P 500 being bullish and commercial paper being bearish. This tells us we should be in the market but it would be prudent to hold some conservative positions or even some cash.
At first glance, this model might seem a little challenging to implement. But after bookmarking the data sources and creating a simple spreadsheet, the Don’t Fight the Fed model will tell you when to avoid risk and when to add risky positions. In the long run, the time you spend updating this model could be the most valuable ten minutes of your month.