Modeling Stock Trading Day Effects Under Flow Day-of-Week Effect Constraints
David F. Findley, Brian C. Monsell
By deriving an invertible linear relation between stock and flow trading day regression coefficients, we show how flow day-of-week effect constraints can be imposed upon the day-of-week-effect component of the stock trading day model of Bell used in X-12-ARIMA. As an application, a new one-coefficient stock trading day model is derived from the constraints that give rise to the one-coefficient weekday-weekend-contrast flow trading day model of TRAMO and X-12-ARIMA. We present summary results and some details of a quite successful application of the new model to the manufacturers’ inventory series of the U.S. Census Bureau’s M3 Survey. (JEL C87, C82).
Time series, RegARIMA models, seasonal adjustment, trading day adjustment, X-12-ARIMA, X-13A-S, M3 survey, inventory series