Modelling heterogeneous speculation in Ghana’s foreign exchange market: Evidence from ARFIMA-FIGARCH and Semi-Parametric methods

In this paper, we explore the weak form efficiency of Ghana’s foreign exchange (FX) market and analyse the existence of speculative activity and correlated shocks in the market. We use high and low frequency data covering May 31, 1999 to November 30, 2017. For robustness, four rigorous methods are employed. Our findings are as follows: First, the efficiency of the FX market is non-homogenous. This gives very little room for speculative trading options, hence, we surmise that speculative activities cannot necessarily account for the self-driven shocks in Ghana’s FX market system. Second, the cedi/dollar market inefficiency is concealed in conditional returns, and toggles between persistence and anti-persistence for the high and low data frequencies respectively. Third, varying significant persistence is detected for the volatility returns for all market series, however, the evidence is more pronounced for daily-, absolute-, and conditional volatility returns. These data dynamics prove useful and should be considered when examining empirical behaviours of asset markets. In summing up, investors and policy makers could rely on the findings and their implications in making decisions on investment and exchange rate control system.