Gold, Currencies And Market Efficiency

Ladislav Kristoufeka,

Miloslav Vosvrda,

Institute of Information Theory and Automation, Academy of Sciences of the Czech Republic, Pod Vodarenskou Vezi 4, 182 08, Prague, Czech Republic, EU


Gold and currency markets form a unique pair with specific interactions and dynamics. We focus on the efficiency ranking of gold markets with respect to the currency of purchase. By utilizing the Efficiency Index (EI) based on fractal dimension, approximate entropy and long-term memory on a wide portfolio of 142 gold price series for different currencies, we construct the efficiency ranking based on the extended EI methodology we provide. Rather unexpected results are uncovered as the gold prices in major currencies lay among the least efficient ones whereas very minor currencies are among the most efficient ones. We argue that such counterintuitive results can be partly attributed to a unique period of examination (2011-2014) characteristic by quantitative easing and rather unorthodox monetary policies together with the investigated illegal collusion of major foreign exchange market participants, as well as some other factors discussed in some detail.

Gold, Currencies And Market Efficiency – Introduction

For decades, the efficient market hypothesis (EMH) has been a building block of financial economics. In his fundamental paper, Fama (1970) summarizes the then-current empirical findings following the theoretical papers of Fama (1965) and Samuelson (1965). Fama (1991) then recalls various issues of the hypothesis and reviews the newer literature on the topic. The capital market efficiency is standardly parallelized with the informational efficiency so that the markets are efficient as long as all the available information is fully reflected into market prices (Fama, 1970). Depending on the level of information availability, the EMH is usually separated into three forms { weak (historical prices), semi-strong (public information), and strong (all information, even private) (Fama, 1991). The theory has been challenged on both theoretical (Malkiel, 2003) and empirical (Cont, 2001) grounds regularly, yet still it remains a popular and fruitful topic of financial research.

The empirical testing of capital markets efficiency has a long history across various assets. The already-mentioned review study of Fama (1970) focuses mainly on stock markets. In commodity markets, Roll (1972) and Danthine (1977) are among the first ones to study their efficiency arriving at contradicting results. In the same timeline, foreign exchange rates are investigated as well (Frenkel, 1976; Cornell and Dietrich, 1978). The termination of the Bretton Woods system in 1971 made the detachment of gold and currency prices interesting for research of the separate phenomena (Booth and Kaen, 1979). Nonetheless, the two still remain tightly connected. Koutsoyiannis (1983) focuses on the efficiency of gold prices and argues that the market efficiency cannot be refuted. Nevertheless, the author finds a tight connection between gold prices and the strength of the US dollar as well as the inflation, interest rates and a general state of the US economy. The gold prices and foreign exchange rates are thus found to be firmly interconnected, which is supported by another early study of Ho (1985). Frank and Stengos (1989) further suggest that simple linear testing of the gold (and silver) market efficiency need not be sufficient.

The efficiency studies of foreign exchange rates are quite unique compared to the mentioned stocks and commodities as the foreign exchange rates pricing has solid macroeconomic foundations such as the balance of payment theory, the purchasing power parity, the interest rate parity, the Fisher effect and others (Dunn Jr. and Mutti, 2004; Levi, 2005; Feenstra, 2008). These theories lead to different ways of efficiency treatment and testing.

Charles et al. (2012) examine the return predictability of major foreign exchange rates between 1975 and 2009. Using various tests, the authors show that the exchange rates are unpredictable most of the time. Short-term inefficiencies are attributed to major events such as coordinated central bank interventions and financial crises. The crises perspective is further studied by Ahmad et al. (2012) who focus on the Asia-Pacific region. They argue that the 1997-1998 Asian crisis was more disturbing compared to the 2008-2009 global financial crisis. In addition, the floating currency markets are found to be more resilient than the countries with managed currencies. Al-Khazali et al. (2012) further examine the Asia-Pacific region using the random walk and martingale definitions of the market efficiency. Out of 8 studied currencies, only three (Australian dollar, Korean won and Malaysian ringgit) are found to be efficient while the other exchange rates offer profitable trading opportunities.

Olmo and Pilbeam (2011) review the literature on the foreign exchange rate efficiency testing based on the uncovered interest rate parity. They suggest that the rejection of efficiency in this area of research may be due to significant differences in volatilities of the logarithmic changes of exchange rates and the forward premium, in addition to conditional heteroskedasticity of the data. The authors introduce a set of profitability-based tests of market efficiency based on the uncovered interest rate parity and they show that the foreign exchange rates are much closer to market efficiency than usually claimed. Chen and Tsang (2013) inspect whether interest rates structure (yield curve) can be used for foreign exchange rate forecasting. They show that it is the case on time horizons between one month and two years. They also argue that these results can help explaining the uncovered interest rate parity puzzle by relating currency risk premium to inflation and business cycle risks. Bianco et al. (2012) further discuss the potential of using economic fundaments for foreign exchange rates forecasting. Their fundamentals-based econometric model for weekly euro-dollar rates is shown to beat the random walk model for time horizons between one week and one month. Engel et al. (2015) construct factors from exchange rates and they use their idiosyncratic deviations for forecasting. Combining these with the Taylor rule, and monetary and purchasing power parity models, they improve the forecasting power of the model compared to the random walk benchmark for the periods between 1999 and 2007 but not for earlier periods down to 1987.


See full PDF below.