Macroeconomic Volatility in Emerging Countries : A Mean-Variance Asset-Liability Management Model

This paper studies a novel approach for managing macroeconomic volatility in commodity exporting countries. As part of this study, we develop a sovereign risk management model in the context of an Asset-Liability Management (ALM) framework. Our first contribution is an extension of the parsimonious Nijman and Swinkels (2008) pension fund model to a much broader macroeconomic setting using new data while developing unconditional investment allocation and optimal sovereign hedging policies for exporting countries. As the strategies are designed in the static mean-variance setting, the optimal solutions are contingent upon discrete-time realisations of deterministic return moments. However, preliminary investigation of the distributional properties of the data lend credence to the consensus view that the time-invariance assumption may be notoriously incorrect, warranting a relaxation of the restrictive assumptions and the application of advanced statistical procedures. Consequently, as a second contribution, we account for partial autocorrelation, significant heteroskedasticity, cointegration and non-linear dependence in the sample data by adopting GARCH and Error Correction models to enhance the forecasting accuracy of optimal hedge ratios. Thirdly, we propose overlay hedging strategies that, when used in conjunction with the portfolio results, guarantee Pareto efficient allocation. Furthermore, we document substantial increases in Sharpe ratios and significant reduction in Cornish-Fisher value-at-risk and maximum drawdowns. Our portfolio allocation and hedging results suggest buying commodity-linked instruments where the current sovereign market exposure in an endowment is lower than the model implied optimal quantity and selling the instruments where the actual quantity exported exceeds the optimal threshold