V.13:08 (321-328): The Utility Average Stock Market Indicator by Dennis Meyers, Ph.D.
Here's a new indicator for trading the Standard & Poor's 500 index, based on the relative performance of the Dow Jones Utilities Average to the S&P. Trading rules and supportive studies are also provided.
The Dow Jones Utility Average (DJUA) has long been acknowledged as a leading indicator for the stock market. The literature on the DJUA as a leading indicator has been mostly anecdotal. To date, no rigorously tested method using the utility average as a leading indicator has been published.
That's easy enough to remedy. Here, I will rigorously test two different methods using the utility average to predict the stock market direction. I will use weekly closes of the DJUA and weekly closes of the Standard & Poor's 500 index to represent the stock market. The test period I will use to derive a utility average indicator is data from January 4, 1980, to March 31, 1995; the out-of-sample† test period is from January 2, 1953, to December 26, 1975. My goal here is to develop an indicator using recent market data and to see if the indicator could work on substantially different market data. Testing and optimization is done using TradeStation 3.5 from Omega Research.
UD% PRICE CHANGE INDICATOR
One of the easiest indicators to test is the UD% indicator, which was described in my July 1995 STOCKS & COMMODITIES article. I optimized the UD% system on the utility average using weekly closes from January 4, 1980, to March 31, 1995. The logic of the UD% system is as follows: If the utility average rises u % from the weekly lowest low closing price that was recorded while we were short, then we would buy the S&P 500 average on its weekly close. If the utility average falls d % from the weekly highest high closing price that was recorded while we were long, then we would short the S&P 500. Figure 1 presents an abbreviated performance summary of the optimization results. The highest total net profit was the optimization objective. The best stable optimization parameters were u % =5 and d % = 14. "Stable" in this instance means that the net profit doesn't change substantially for small changes in parameters.