A Multi-Factor Model For 8 Gold Stocks' Returns

|
Includes: ABX, AU, EGO, GFI, GG, GOLD, KGC, NEM
by: Alberto Abaterusso

Summary

An attempt to measure the exposure of gold stocks' returns to six variables: Risk Free Rate, Gold (metal), NYMEX (Crude Oil), DXY (US dollar Index), DJIA, S&P 500 index.

Only NYMEX (Crude Oil price) and DXY (US dollar Index) factors provide significant explanatory power for all of the gold stocks over the period observed.

Considering a gold stock section of a portfolio, the multi-factor model has very well predicted returns on gold stocks, -16.17% ('predicted') against -16.45% (actual).

My attempt is to measure the exposure of gold stocks' returns to six variables:

  • Rf, the weekly rate of Tbill return as free risk return variable. Data are downloaded from here.
  • Return on gold (metal). GOLDAMGBD228NLBM, weekly prices downloaded from Ice Benchmark Administration, Gold Fixing Price 10:30 A.M. (London time) in London Bullion Market, based in U.S. Dollars [GOLDAMGBD228NLBM], retrieved from FRED, Federal Reserve Bank of St. Louis.
  • NYMEX (Crude Oil price). Crude Oil prices, weekly data, downloaded from here.
  • DXY (US dollar Index). Weekly data are downloaded from here.
  • DJIA. Weekly data are downloaded from here.
  • S&P 500 index. Weekly data are downloaded from Yahoo Finance.

Weekly prices of gold stocks are also downloaded from Yahoo Finance.

I run multiple regression with excel according to multifactor model.

Regression results for multifactor model applied to gold industry stock returns:

Rgt = Alfa + Betagold*Rgold + Betanymex*NYMEX + Betadxy*DXY + BetaDjia * DJIA + BetaS&P500 * S&P500 + BetaRfRf.

The table below shows the beta coefficients that are statistically significant to explain returns on gold stocks over the period January 17, 2006 - December 29, 2014 (weekly observations):

Click to enlarge

I found that only NYMEX (Crude Oil price) and DXY (US Dollar Index) factors provide significant explanatory power for all of the gold stocks over the period observed.

Return on the market (S&P 500) is statistically significant only for three gold stocks: Newmont Mining (NYSE:NEM), AngloGold Ashanti (NYSE:AU) and Eldorado (NYSE:EGO).

DJIA is statistically significant only for Eldorado Gold Corporation.

There is a negative relation between returns on the gold stocks and DXY index.

The following eight pictures show summary reports of the gold mining stocks during the observed period, 467 observations (weekly data):

Summary output for Goldcorp Inc. (NYSE:GG):

Click to enlarge

Click to enlarge

Summary output for Kinross Gold Corporation (NYSE:KGC):

Click to enlarge

Click to enlarge

Summary output for Barrick Gold Corp. (NYSE:ABX):

Click to enlarge

Click to enlarge

Summary report for Goldfield Ltd. (NYSE:GFI):

Click to enlarge

Click to enlarge

Summary Output for Newmont Mining Corporation:

Click to enlarge

Click to enlarge

Summary output for AngloGold Ashanti Ltd.:

Click to enlarge

Click to enlarge

Summary output for Eldorado Gold Corp.:

Click to enlarge

Click to enlarge

Summary output for Randgold Resources Limited (NASDAQ:GOLD):

Click to enlarge

Click to enlarge

Even though adjusted R squares are low, the F-test of overall significance determines that the relationships examined are statistically significant.

Furthermore the residual plots don't show any kind of trend.

The experiment below examines the prediction of gold stocks returns:

I have checked actual weekly and annual returns on gold stocks against 'predicted' weekly and annual returns on the gold stocks, to see how close the first with the second 'predicted' values by the multifactor model were.

The table summarizes actual and 'predicted' weekly and annual returns on the gold stocks as well as those on the variables:

Click to enlarge

Consider, for example, the predicted versus actual returns for Barrick Gold Corporation. The 'predicted' return on NYSE:ABX for year 2015 was -23.92%; the actual return was -26.52%.

The 'predicted' weekly return on NYSE:ABX for year 2015 was -0.46%; the actual return was -0.51%.

The table below shows weekly and annual return on the six variables used in the multifactor model:

Conclusion:

From end of 2012 onwards gold prices and gold stocks tumbled.

Over year 2015 gold stocks tumbled even further than what I 'predicted' through a multifactor model as used in this article.

Considering a gold stock section of a portfolio, the multifactor model has very well predicted returns on gold stocks, -16.17% ('predicted') against -16.45% (actual).

I found out that only NYMEX (Crude Oil price) and DXY (US dollar Index) factors provide significant explanatory power for all of the gold stocks over the period observed and that there is a negative relation between the latter and gold stocks' returns.

Gold stocks are less volatile than crude oil (NYMEX), but both move towards the same direction...

The fundamentals of over-supply remain and so there is no reason to believe that the oil price crash will turn at this point. (here).

Disclosure: I/we have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours.

I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.