2013 Is Gold Overpriced?
This paper builds a model of the gold price including a number of macroeconomics factors to test whether gold has been over priced between 1968 and 2012. They do this using a quantile regression framework and compare their results to simple OLS estimates using the same variables.
They included US unemployment and GDP to proxy the strength of the American economy. They use US inflation to reflect its inflation hedge characteristics, the US dollar’s trade weighted value, the US T-Bill yield and the price of oil. The advantage of the quantile regression in this area is that it allows an investigation of the effects of these factors on the gold price when they are at different levels so that we do not get their average effect on gold but also their effect when they are below and above average as well.
They find based on this analyses that gold was not overpriced until January 2009 and up to the end of their sample in March 2013.
Opinion: While the methodology is interesting some of its implementation is questionable. Some of the variables used in the regressions are usually found to be nonstationary and unless they cointegrate they regression results may be invalid. The large number of variables used are also subject to Baur and Glover’s (2012) criticism of over parameterisation.
Method: Quantile and OLS regressions
Full Citation: L. Ma and G. Patterson (2013) Is Gold Overpriced? The Journal of Investing 22(1):113-127.
Abstract: The price of gold has risen dramatically since 2001, and there have been intense debates regarding the current price of this asset and its long-term outlook. Here we study a long history of the price of gold and construct a quantile regression model to identify distributional relationships between the price of gold and macroeconomic indicators, financial market performance, and other relative factors. We find that while the traditional mean methodology indicates that gold is overpriced, the distributional results from quantile regression imply that gold is not overpriced in the current economic and financial environment. We provide gold price forecasts based on two economic scenarios that cover both the long-run economic environment and the recent economic environment experienced in 2009 and 2010.