Hello everyone,
I am studying the effect of the market on the return series of precious metal assets. Following Hillier et al. (2006), Pullen et al. (2014), I will make use of their GARCH setup. The regression model assumes that precious metal commodities itself (bullion), stocks, mutual funds and ETFs are dependent on changes in the stock market and that this relationship is conditional on specific, extreme, market conditions. I will examine the safe haven features of these products. Following prior research as mentioned above, assumed is that the price of the precious metal asset is dependent on changes in the stock market. Furthermore, the relationship between the precious metal asset and the stock market is not constant and affected by extreme market conditions. As specified by Pullen et al. (2014) and Hillier, D., Draper, P., & Faff, R. (2006), a GARCH will be adopted, in which the first equation is the mean equation and the second the (conditional) volatility equation. as follows:
Equation 1 models the relation of the precious metal asset and stock market returns. The parameters are alpha1, beta 0/1/2 and the error term. rPMA is the return on the precious metal investment asset and rm is the return on the U.S. equity or the market. Two dummy variables are added to capture stock market volatility. They take a value of 1 if the market exceeds a certain volatility threshold given by the 5% (i.e., VOLq5), and 1% (i.e., VOLq1) quantiles of the return distribution. These dummies capture the safe haven abilities of precious metal assets during extreme market conditions.
I tried the following commands but stata omits the created dummies due to collinearity. So, what am I doing wrong
arch return_spcomp, arch(1) garch(1)
predict market_var, variance
sum market_var
tabstat market_var, stat(n mean min max sd p95 p99)
xtile quant_20 = market_var, nq(20)
gen mrkt_q5 = 1 if quant_20 == 20
sum quant_20
sum mrkt_q5
drop quant_20
xtile quant_100 = market_var, nq(100)
gen mrkt_q1 = 1 if quant_100 == 100
sum quant_100
sum mrkt_q1
drop quant_100
reg return_gold_bullion return_spcomp mrkt_q5 mrkt_q1
Many thanks for all your kind support. Any comments will be highly appreciated.
Kind regards,
Stijn
Hillier, D., Draper, P., & Faff, R. (2006). Do precious metals shine? An investment perspective. Financial Analysts Journal, 62(2), 98-106.
Pullen, T., Benson, K., & Faff, R. (2014). A comparative analysis of the investment characteristics of alternative gold assets. Abacus, 50(1), 76-92.
I am studying the effect of the market on the return series of precious metal assets. Following Hillier et al. (2006), Pullen et al. (2014), I will make use of their GARCH setup. The regression model assumes that precious metal commodities itself (bullion), stocks, mutual funds and ETFs are dependent on changes in the stock market and that this relationship is conditional on specific, extreme, market conditions. I will examine the safe haven features of these products. Following prior research as mentioned above, assumed is that the price of the precious metal asset is dependent on changes in the stock market. Furthermore, the relationship between the precious metal asset and the stock market is not constant and affected by extreme market conditions. As specified by Pullen et al. (2014) and Hillier, D., Draper, P., & Faff, R. (2006), a GARCH will be adopted, in which the first equation is the mean equation and the second the (conditional) volatility equation. as follows:
Equation 1 models the relation of the precious metal asset and stock market returns. The parameters are alpha1, beta 0/1/2 and the error term. rPMA is the return on the precious metal investment asset and rm is the return on the U.S. equity or the market. Two dummy variables are added to capture stock market volatility. They take a value of 1 if the market exceeds a certain volatility threshold given by the 5% (i.e., VOLq5), and 1% (i.e., VOLq1) quantiles of the return distribution. These dummies capture the safe haven abilities of precious metal assets during extreme market conditions.
I tried the following commands but stata omits the created dummies due to collinearity. So, what am I doing wrong
arch return_spcomp, arch(1) garch(1)
predict market_var, variance
sum market_var
tabstat market_var, stat(n mean min max sd p95 p99)
xtile quant_20 = market_var, nq(20)
gen mrkt_q5 = 1 if quant_20 == 20
sum quant_20
sum mrkt_q5
drop quant_20
xtile quant_100 = market_var, nq(100)
gen mrkt_q1 = 1 if quant_100 == 100
sum quant_100
sum mrkt_q1
drop quant_100
reg return_gold_bullion return_spcomp mrkt_q5 mrkt_q1
Many thanks for all your kind support. Any comments will be highly appreciated.
Kind regards,
Stijn
Hillier, D., Draper, P., & Faff, R. (2006). Do precious metals shine? An investment perspective. Financial Analysts Journal, 62(2), 98-106.
Pullen, T., Benson, K., & Faff, R. (2014). A comparative analysis of the investment characteristics of alternative gold assets. Abacus, 50(1), 76-92.
