BIS国际清算银行-What-share-for-gold-On-the-interaction-of-gold-and-foreign-exchange-reserve-returns_30页_1mb

Summary of "What share for gold? On the interaction of gold and foreign exchange reserve returns" by Omar Zulaica

Core Content

This paper investigates the role of gold in foreign exchange (FX) reserve portfolios, focusing on whether gold's inclusion is justified from a risk-return standpoint. It explores the interaction between gold and fixed income instruments in the context of FX reserves, considering various portfolio durations and numeraires. The analysis goes beyond traditional portfolio optimization techniques by incorporating a range of risk-return metrics, including volatility, Value-at-Risk (VaR), expected shortfall, probability of loss, and risk factor concentration.

Main Findings

• Market Risk and Volatility: Gold exhibits substantial market risk, particularly in relation to portfolio volatility. For low-duration, reserve currency fixed income portfolios, small gold allocations (between 0% and 5%) can marginally reduce volatility. However, as gold allocations increase beyond 5%, the portfolio's volatility tends to rise sharply.

• Duration and Hedge Effectiveness: For higher-duration portfolios, gold can act as an effective hedge against yield changes. Gold allocations above 10% may be beneficial for portfolios with durations beyond 2 years.

• Diversification and FX Risk: Gold's diversification benefits are enhanced in the presence of foreign exchange risk, especially for portfolios with numeraires that are volatile against the portfolio's unit of account. For such cases, gold allocations above 20% may be optimal.

• Tail Risk and Extreme Events: In the context of tail risk and extreme events, gold allocations between 20% and 50% may be appropriate, particularly for portfolios with elevated yield sensitivity or exchange rate risks. This is especially relevant for portfolios with durations above 5 years or when performance is measured in a non-reserve currency numeraire.

• Risk Parity: The concept of risk parity, where each asset class contributes equally to portfolio volatility, is reached at a gold allocation of 10% in the case of a 2-year SDR fixed income portfolio. This indicates a balance of risk sources rather than a minimization of risk.

• Portfolio Return: The expected return of gold is close to that of the 2-year SDR fixed income benchmark. As a result, increasing gold exposure leads to minimal changes in the expected return of the FX reserve portfolio.

• Uncertainty and Predictability: The confidence bands around expected return estimates are wide, indicating the difficulty of predicting financial asset returns. Increasing gold exposure increases the uncertainty of the portfolio's expected return.

Key Metrics Analyzed

• Volatility: The impact of gold on portfolio volatility is examined across different gold allocations.
• Value-at-Risk (VaR): The risk of loss at a given confidence level is analyzed.
• Expected Shortfall: The average loss in the event of a VaR breach is considered.
• Probability of Loss: The likelihood of experiencing a negative return is evaluated.
• Single Factor Concentration (SFC): A measure of how much of the total variance is explained by a single risk factor, calculated using principal component analysis (PCA).

Empirical Methodology

• The study uses historical data from January 1981 to July 2020.
• Gold returns are modeled using gold futures prices.
• A 2-year SDR government bond portfolio is used to represent the fixed income component.
• The analysis includes 5,000 simulated scenarios to capture the full distribution of possible outcomes.
• A copula model is used to simulate multivariate return distributions, preserving individual empirical distributions and capturing non-linear dependencies in the tails.

Observations on Gold Holdings

• The global share of gold in FX reserves is approximately 14.6%.
• Advanced economies tend to hold a higher proportion of gold compared to emerging markets.
• Gold holdings are generally stable at the country level, with a standard deviation of 28 tonnes on average.
• The inclusion of gold in FX reserves is influenced by both quantitative and qualitative factors, including policy objectives, risk tolerance, and the numeraire used for performance measurement.

Conclusion

The paper concludes that determining the appropriate share of gold in FX reserve portfolios is non-trivial and depends on the specific characteristics of the portfolio, such as duration, numeraire, and risk tolerance. While gold may offer diversification benefits and hedge properties under certain conditions, its inclusion is not universally optimal and requires careful consideration of the portfolio's risk-return profile.