Improving accuracy with forecast combination : the case of inflation and currency in circulation in Nigeria

This study shows how the application of forecast combination improves the accuracy of forecasts of economic variables. Using data from January 2009 to December 2014 on the Nigerian inflation rate, and forecasts of currency in circulation (see Ikoku, (2014)) as examples, we find that by combining forecasts of both variables using the regression-based method, the mean absolute percent errors of the combined forecasts were lower than the forecast errors from the individual models of the variables.