DJIA-Gold Ratio as a Stock Market Indicator
Posted in Technical Trading
April 6, 2009
A reader requested: “Please test the following hypothesis [presented by Simon Maierhofer, co-founder of ETFguide.com] in the article ‘Gold’s Bluff – Is a 30 Percent Drop Next?’: Ironically, gold is more than just a hedge against market turmoil. Gold is actually one of the most accurate indicators of the stock market’s long-term direction. The Dow Jones measured in gold is a forward looking indicator.” For this analysis, we test relationships between the spot price of gold and the level of the Dow Jones Industrial Average (DJIA). We go back only as far as 1968, because: “On March 17, 1968, …the price of gold on the private market was allowed to fluctuate…[, and] in 1975…the price of gold was left to find its free-market level.” Using month-end data for the spot price of gold in dollars per ounce and the level of DJIA over the period 1968-2008 (492 months), we find that:
The following chart plots the ratio of the level of DJIA to the price of gold in dollars per ounce over the entire sample period. Visual inspection suggests that the ratio may be very slowly mean reverting and that the ratio is low (high) when the stock market is low (high). However, visual inspection is imprecise, and it is not obvious that this ratio reliably predicts the future level of DJIA at any particular forecast horizon. In other words, it is not obvious that mean reversion follows any particular schedule.
For a closer look at predictive power, we relate the month-end level of DJIA/gold to the next-month change in DJIA.
The following scatter plot depicts the relationship between monthly changes in DJIA to the value of DJIA/gold at the end of the prior month over the entire sample period. The Pearson correlation between the two series is -0.05 and the R-squared statistic is 0.00, indicating that DJIA/gold at the end of a month explains about 0% of the change in DJIA the following month. In other words, the ratio of DJIA level to spot gold price is not a useful signal at a one-month forecast horizon.
Might DJIA/gold work well at an annual forecast horizon?
The next scatter plot depicts the relationship between annual changes in DJIA to DJIA/gold at the end of the prior year over the entire sample period (41 years). The Pearson correlation between the two series is -0.21 and the R-squared statistic is 0.04, indicating that DJIA/gold at the end of a year explains about 4% of the change in DJIA the following year. In other words, a low (high) DJIA/gold is a little bit predictive of a strong (weak) stock market at a forecast horizon of one year.
Extending the forecast horizon strengthens the correlation, but reduces sample size. For example, using five-year intervals rather than one-year increases the R-squared statistic to 0.33 but reduces sample size to eight.
Might it be the price of gold, rather than the level of the stock market, that drives mean reversion of DJIA/gold?
The final scatter plot depicts the relationship between annual changes in spot gold price to DJIA/gold at the end of the prior year over the entire sample period. The Pearson correlation between the two series is -0.01 and the R-squared statistic is 0.00, indicating that DJIA/gold at the end of a year explains about 0% of the change in gold price the following year. In other words, it appears to be the stock market and not the gold market that drives fluctuation in the level of DJIA/gold.
In summary, evidence from the past 41 years suggest that the ratio of DJIA to gold price is not useful (may be marginally useful) for predicting stock market behavior at a forecast horizon of one month (one year). Lack of availability of long-run free market gold prices impedes measurement of its usefulness at multi-year forecast horizons.
Note also that the potential for “wild” return distributions undermines the meaning of commonly applied statistical metrics.
