Carryover Effects and Temporal Aggregation in a Partial Adjustment Model Framework

This paper discusses the impact of aggregating time series observations on the specification of an econometric model. In the framework of a familiar sales response model, temporally aggregated specifications are derived. It is illustrated that different aggregation schemes provide different models, each of them having appealing characteristics. Furthermore, the derived specifications illustrate that frequently calibrated models are misspecified and as such provide biased estimates which significantly affect the interpretation of the results. Some suggestions are provided for estimating the exact specifications in such a way that the obtained parameter estimates provide correct insights into the underlying response process.