Detection and Forecasting of Islamic Calendar Effects in Time Series Data: Revisited

This paper is an attempt to revisit the pioneering work of Riazuddin and Khan (2002). A complete business cycle has been elapsed (2002-2010) since their study, so there is need to review the results with additional information. This revisited attempt, based on a theoretically specified framework, arrived at similar results and found significant impact of Islamic calendar. The Islamic months of Ramadan and Zilhaj have positive impact on currency holdings and negative impact on deposits. Although stylized facts indicate that consumer prices are significantly higher during Ramadan but econometric investigation rejects the upward exogenous shifts in prices during Ramadan. Therefore, structural relationship analyzed in co-integration framework has shown that inflation is not directly impacted by the Ramadan but indirectly through increase in its determinants. Inflationary tendencies during Ramadan are not due to exogenous increase by producers and retailers but possibly due to demand surge in the wake of redistribution of income. The months of June and December have positive effects on deposits and negative effects on currency in circulation indicating the presence of window dressing. Finally, as seasonal factors have important role in determining economic time series, therefore, ignoring those in monthly time series models will lead to omitted variable bias and inappropriate forecasts.