Are Gold Bugs Coherent?
Trinity Business School, Trinity College Dublin; University of Ljubljana – Faculty of Economics
University of Dublin – Business School and Institute for International Integration Studies; York St. Johns Business School
January 22, 2016
We use wavelet models to surface the relationship between gold miners stock prices and the price of gold. We find that there is little relationship in the short run but some significant and long standing long run relationships. Gold prices appear to lead gold miner stock prices.
Are Gold Bugs Coherent? – Introduction
Significant research suggests that gold, in financial (futures) or physical form, can be a useful diversifier for portfolios. From early work by Sherman (1982) through Jaffe (1989) and to more recent work by Lucey et al. (2006) , Hillier et al. (2006) and Conover and Jensen (2009) the literature suggests a small, typically less than 5% allocation to gold is beneficial. Gold can be, in the medium to long term, a useful hedge. Gold has also been examined from the perspective of its potential as a safe haven (see inter alia Baur and McDermott (2010), ?, Joy (2011), Ciner et al. (2013), with the general conclusion being that it provides protection to a portfolio from extreme market events. Gold therefore can play a useful role in reducing a portfolio’s risk.
Gold is traded in a variety of ways and around the world. The two major centers for gold trading, the London over-the-counter (LOTC) spot market and the New York Mercantile Exchange Futures Market (COMEX), account for approximately 78.0% and 7.7% of the total gold turnover. Murray (2011) The London OTC market for bullion is the largest pool of gold assets, although the most important, from the perspective of setting a gold price, appears to be the New York futures market, COMEX in particular. See Haupt eich et al. (2016).
In 2011, estimated daily turnover in the international gold market was 4,000 metric tons, a money volume approximately the same as the daily dollar volume of trade on all of the worlds stock exchanges combined. If we were to consider gold as a currency it would be the fth largest currency.
Investors wishing to get exposure to gold in other forms can look to gold mining stocks (as well as variety of other approaches – see Batten et al. (2015). A complicating factor is that first, there are limited numbers of gold miners, and second in many cases it is impossible to obtain a pure mine substitute, as gold is frequently extracted in conjunction with other minerals. MacDonald and Taylor (1988) provides an early examination, of 20 South African miners, and finds that there is a statistically significant positive relationship between mine share prices and gold, with asymmetry evident in a greater relationship for higher cost miners. This is also the case in Blose and Shieh (1995), for 23 US miners and for australian miners in Faff and Chan (1998) and for US miners using more recent data, as in Borenstein and Farrell (2007). OConnor et al. (2015) however finds that that the gold price leads production costs suggesting that an examination of gold mining companies as alternatives is not a useful path. This finding is in line with the results of Areal et al. (2013) who find little benefit from a safe haven perspective of investing in gold mining companies.
Here we examine the relationship between the gold price and the NYSE ARCA Gold Bugs index of gold miner share prices over a 17 year period using wavelet analysis.
2.1. Continuous Wavelet Transformation
Wavelet multi-scale analysis is a technique appropriate for the estimation of spectral characteristics of a time series. In this paper, we measure the degree of local variability and o-variability between gold (returns) and changes in the Gold Bugs Index using the wavelet power, cross-wavelet coherence and the phase difference. Each of these are displayed in a three dimensional diagram that demonstrates time series information at different frequencies (low and high) and points in time simultaneously. The computational framework adopted for this study is fundamentally based on Torrence and Compo (1998) and Grinsted et al. (2004).
The wavelet transform approach is particularly applicable to financial and economic time series and has been widely documented in previous studies (In and Kim, 2006; Gencay et al., 2001; Percival and Walden, 2000). The pioneering work on wavelet multi-scale analysis in finance is documented in Ramsey et al. (1995), where they examine the contribution of the wavelet approach in detecting self-similarity in US stock prices, whilst Ramsey and Lampart (1998) study the money, income and expenditure link using the wavelet-based scaling method.
See full PDF below.