Macroeconomic Volatility and Sovereign Asset-Liability Management

Abstract

For most developing countries, the predominant source of sovereign wealth is commodity related export income. However, over-reliance on commodity related income exposes countries to significant terms of trade shocks due to excessive price volatility. The spillovers are pro-cyclical fiscal policies and macroeconomic volatility problems that if not adequately managed, could have catastrophic economic consequences including sovereign bankruptcy. The aim of this study is to explore new ways of solving the problem in an asset-liability management framework for an exporting country like Ghana. Firstly, I develop an unconditional commodity investment strategy in the tactical mean-variance setting for deterministic returns. Secondly, in continuous time, shocks to return moments induce additional hedging demands warranting an extension of the analysis to a dynamic stochastic setting whereby, the optimal commodity investment and fiscal consumption policies are conditioned on the stochastic realisations of commodity prices. Thirdly, I incorporate jumps and stochastic volatility in an incomplete market extension of the conditional model. Finally, I account for partial autocorrelation, significant heteroskedastic disturbances, cointegration and non-linear dependence in the sample data by adopting GARCH-Error Correction and dynamic Copula-GARCH models to enhance the forecasting accuracy of the optimal hedge ratios used for the state-contingent dynamic overlay hedging strategies that guarantee Pareto efficient allocation. The unconditional model increases the Sharpe ratio by a significant margin and noticeably improves the portfolio value-at-risk and maximum drawdown. Meanwhile, the optimal commodities investment decisions are superior in in-sample performance and robust to extreme interest rate changes by up to 10 times the current rate. In the dynamic setting, I show that momentum strategies are outperformed by contrarian policies, fiscal consumption must account for less than 40% of sovereign wealth, while risky investments must not exceed 50% of the residual wealth. Moreover, hedging costs are reduced by as much as 55% while numerically generating state-dependent dynamic futures hedging policies that reveal a predominant portfolio strategy analogous to the unconditional model. The results suggest buying commodity futures contracts when the country’s current exposure in a particular asset is less than the model implied optimal quantity and selling futures contracts when the actual quantity exported exceeds the benchmark.