An Almon Approximation of the Day of the Month Effect in Currency in Circulation

In recent years, increasing importance of active day-to-day asset-liability management by banks and liquidity management by central banks renders that short-term seasonalities in monetary series should be studied minutely. Currency in circulation is a typical example of such a series that displays strong intra-month seasonality. Earlier, the intra-month seasonality in currency in circulation in India had been modeled by dummy variables pertaining to specific days of months (Bhattacharya and Joshi, 2001). The paper reports a parsimonious representation of the Bhattacharya and Joshi (2001) model by approximating the coefficients pertaining to the thirty-one 'day of the month' dummies by a polynomial. The extension is interesting because it provides an example of polynomial approximation of coefficients pertaining to a group of indexed dummy variables, rather than lagged dependent or independent variables as in a standard Almon (1965) scheme. For a degree four approximation, an empirical comparison of the two models reveals marginal difference in explanatory powers and forecasting abilities.