Mechanical Trading Strategy Using The Hui/Gold Ratio

Using data from the past seven years, I've developed a mechanical trading strategy using the HUI/Gold ratio as a guide. The premise is simple: the value of the HUI index is fundamentally based on the price of gold, everything else is just noise. One should be able to determine a theoretical "correct value" for the HUI index, depending solely on a ratio of the HUI to the price of gold. Any advance in the HUI index far above an idealized HUI/gold ratio should be sold. Any decline of the HUI index far below the HUI/gold ratio should be bought.

The idealized HUI/gold ratio is not a constant. It varies with the price of gold because the leverage of mining companies varies with the price of gold. I've developed an algorithm which determines the "fair price" for the HUI based on the price of gold. The algorithm provides and upper and lower band for the idealized HUI/gold ratio based on a function of 100 day moving average of the price of gold. Anytime the HUI/gold ratio (the black line) rises above the upper limit (the red line), it means that the HUI index is overvalued and should be sold. Anytime the HUI/gold ratio falls below the lower limit (the green line), the HUI index is undervalued and should be bought. This formula has been remarkably successful for the past 7 years especially when the price of gold is above 280. This strategy has returned 170% since it was started in 1996. During the same time period, the HUI has fallen almost 40%! Below is a chart of the HUI/Gold Strategy. My data goes back to June of 1996.

A couple of observations:

• The strategy has been amazingly accurate at picking tops and bottoms of the HUI so long as the price of gold is between $290 and $400. The formula appears to break down at prices far below $290, witness the way-too-early buy point in June 1998 at 91.52. My hypothesis is that the model only works when gold is profitable to mine. At prices below $290, many mining companies become unprofitable. Profitable PM companies can easily be valued at a multiple of their profits but the stock price of unprofitable companies can be very unpredictable. Below $290, the HUI index appears to have fluctuated wildly based on emotion, reserves, political considerations, etc. rather than on profitability.

• I also doubt the accuracy of this model if the price of gold far exceeds $400. The formulas employ exponential functions which start to go parabolic at high gold prices. If I had more data, I would like to back test the model back to the early 90's when gold was higher. Unfortunately, the HUI didn't exist as an index until 1996.

Here is a close-up of the strategy during the last 400 days. I've added the price of gold to this chart so you can see how the HUI tends to predict the movement of gold.

The strategy has been particularly effective during this time period. Although it missed part of the blowoff top in theSpring of 2002, it caught most of the other buy and sell points almost perfectly. The strategy was a little early in picking the bottom of the most recent downswing in the HUI because the moving average of gold was still highly skewed by the recent run to $390. Nevertheless, a follower of this strategy would be up 200% since September 21st, 2001. During the same timeframe, the HUI advanced 70% and gold advanced 12%.

As of today, April 25^th, we are still in the buy zone on the HUI, but just barely. If past price patterns persist, the HUI/gold ratio will soon rise above the buy band and enter the "hold" region on its way to the next sell band. If the price of gold stays relatively stable in the $330 to $350 zone, I anticipate a sell signal at a HUI/gold ratio of .450, probably in two months or so. At a price of gold of $340, that would translate to a HUI valuation of 153. Obviously, if gold surges higher, the sell signal would occur at a higher HUI valuation.

A note of caution: this strategy was developed by backtesting data. With hindsight, I was able to fine tune the formulas somewhat to coincide with the peaks and valleys of the HUI index. As with most backtested strategies, it is likely that it will not be quite as successful going forward. Nevertheless, since the strategy is based on fundamental valuation, I feel comfortable with the general concept. I plan to scale into positions as the buy points approach, and scale out as the sell points approach. I am fully invested in PM's right now with no cash on the sidelines.

In my next essay, I will focus on some variations of this strategy. Specifically, what should you do with your cash after receiving a sell signal on the HUI?