Price Discovery In The Chinese Gold Market
Queen’s University Belfast – Queen’s Management School
Queen’s University Belfast – School of Management
A few years ago, crypto hedge funds were all the rage. As cryptocurrencies rose in value, hundreds of hedge funds specializing in digital assets launched to try and capitalize on investor demand. Some of these funds recorded double-digit gains in 2020 and 2021 as cryptocurrencies surged in value. However, this year, cryptocurrencies have been under Read More
University of Technology Sydney
Yung Chiang Yang
University College Dublin (UCD) College of Business; Queen’s University Belfast
May 6, 2016
This study conducts price discovery analysis in the Chinese gold market. Our result indicates that the price discovery in Chinese gold market occurs predominantly in the futures market. The result is robust to the different measures of price discovery, namely information share, component share, and information leadership share. Partitioning the daily trades into three trading sessions, we find that the dominance of the futures market occurs in all trading sessions. We further investigate the sequential price discovery within the spot market or futures market. We find that the price discovery of gold spot market and gold futures market occur in the night trading session.
Price Discovery In The Chinese Gold Market – Introduction
This study investigates the price discovery for gold in China on two informationally linked markets – the spot market and the futures market. The gold spot market in China started in 2002 when Shanghai Gold Exchange was established. The gold futures market was established in 2008 at the Shanghai Futures Exchange. According to Lucey et al. (2013), the trading volume of gold contracts and its derivatives in China is ranked third largest in the world. The consumption of gold as well as the demands for investing and hedging in gold related products is growing rapidly in China due to the pace of underlying economic development. Moreover, according to the U.S. Geological Survey (2012), China is the world dominant gold producer with proven reserves that are ranked third in the world. The gold market in China is becoming more important and is of interest for investors and researchers. Lucey et al. (2014) find that the gold market in China is very disconnected from the other markets with negligible effect on or from other markets. This special characteristic makes the price discovery of gold within China an important topic.
Price discovery refers to efficient and timely incorporation of information into market prices through trading. If the price discovery process is timely and effective, then the market would be efficient (Fama, 1970). In an efficient market, prices reflect new information quickly and adequately (Lehman, 2002). In case of similar or related products traded at different markets, new information could affect these markets simultaneously. For instance, when gold contracts are traded in spot and futures markets in parallel, the price discovery can be defined as which price series is the first to fully reflect new information about the true underlying asset value. In short, price discovery studies attempt to answer the following questions: “Which gold market moves first?” and “Which gold product moves closer to the intrinsic value?”
We use three measures to study the parallel price discovery between the gold spot market and the gold futures market. The first measure is information share derived by Hasbrouck (1995). He uses the variance of the common factors innovation uncovered from a Vector Error Correction Model (VECM) to define price discovery. It measures the price variation contributed by different markets, with the proportion contributed by each market being defined as information share. The second measure is the component share measure proposed by Gonzalo and Granger (1995). Component share measures the contribution to the common factor by each market, where contribution is defined as a function of market error correction coefficients. The market error coefficients are obtained from the vector error correction model capturing only permanent shocks in asset price. Lucey et al. (2013) use both information share and component share in the study. The third measure, information leadership share, is proposed by Putni?š (2013) as an adaptation to the measures outlined in Yan and Zivot (2010). He finds that information share responds to both permanent and transitory shocks, while component share capture the transitory shocks. He suggests a new measure, information leadership share, by combining information share and component share measures and the new measure captures only the permanent shocks on the asset price. Hauptfleisch et al. (2015) use Putni?š (2013) information leadership share to confirm that New York leads the other financial centers in terms of gold price discovery. This exhibits the contrasting inferences drawn from using the unmodified Gonzalo and Granger (1995) and Hasbrouck (1995) that led to Lucey et al. (2013) concluding that in fact London was the dominant center in terms of gold price discovery.
Besides conducting parallel price discovery on gold spot and futures markets, we also carry out price discovery analysis across morning, afternoon, and night sessions of a trading day; sequential price discovery within gold spot and gold futures market. We employ three measures to compare the price discovery across trading sessions. The first sequential price discovery measure is variance ratio between two-scale realized variance and realized variance (TSRV/RV) proposed by Wang and Yang (2011). Intuitively, TSRV is a variance that is induced by pure information while RV captures both variances caused by information and microstructure noise. Therefore, the ratio TSRV/RV provides a measure for the price efficiency of a trading session. The second measure is a modified information share measure also proposed by Wang and Yang (2011). The information share of a particular trading session is its share of the total variance of the efficient price for the full trading day. The third measure for the sequential price discovery is weighted price contribution (WPC). WPC is a simple and convenient measure that uses the share of price change in different trading sessions to measure the level of efficient information, see, for example, Cao et al. (2000).
See full PDF below.