Determinants of exchange rate in Ghana: Cointegration and wavelet analysis.
No Thumbnail Available
Date
2016-08
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
KNUST
Abstract
We empirically examine the determinants of the nominal exchange rate in Ghana using both time series econometric techniques and wavelet application for annual data spanning the period 1960-2014. The determinant variables under consideration included: inflation, supply of money(M1 and M2), GDP growth and trade openness. In particular, Engle-Granger method was used to test for the causal relationship, wavelets power and wavelets coherency
spectrum for direction of co-movements among the variables were applied. In addition, we determine the long and short run relationships between nominal exchange rate and its determinants using autoregressive distributive lag (ARDL) and other variants of ordinary least square cointegration techniques such as fully modified(FM), dynamics and integrated modified-OLS. As required in standard econometric analysis, unit root tests were initially performed using Augmented Dickey-Fuller, Philips-Perron and Kwiatkowski Philip Schmidt and Shin methods to assess the data generating process of the interested variables. The time series econometric analysis showed that all the variables exhibited bidirectional relationship with exchange rate with the exception of supply of money which showed unidirectional causation, from money supply to exchange rate. In the cointegration analysis, we found that all the variables showed statistically significant long term equilibrium relationship although money supply and GDP had ambiguous signs. However, inflation and trade openness coefficients were theoretically consistent with expectation. Also the lagged dependent variable and the dummy variable for exchange rate regime appears to influence exchange rate dynamics in the long run. In short, all the interested variables have positive coefficients, signifying depreciating effect on the exchange rate. The adjustment coefficient obtained was 60.78%, indicating that it takes approximately 20 months for exchange rate to fully revert to equilibrium from any deviations. The results from the wavelet application are also striking. Wavelets power spectrum clearly shows activities in the short to medium frequency band of the variables. The cross coherency test also indicated that exchange rate depreciation leads (or Granger cause) the other variables, with the exception of trade openness which drives exchange rate depreciation.
Description
A Thesis Submitted To The School Of Graduate Studies In Partial Fulfilment Of The Requirements For The Award Of Master Of Philosophy In Finance.