Cross section pooling as against time series pooling in market analysis

Abstract

Market analysis essential for strategising market activity involves determination of optimal market response coefficients to marketing mix variables. Market response understanding is usually obtained from market research through surveys and panels. Pooling is a method of combining the advantages of surveys and panels, by combining several cross sections of time series variables (PCSTS). In the event of homogeneity hypothesis to be rejected and PCSTS not being considered appropriate, it is useful to check if the data of pooled time sections of cross sectional variables (or PTSCS) would permit pooling to be conducted and if so the procedures for pooling of data. Longitudinal data set of cigarette market in 46 U.S. states over the period 1980-1992, has been taken from the panel data source and subjected to the procedures for cross-section pooling. The pooling can be done by natural dummy variables such as the consumer price index (CPI). In this case the natural dummies act as fixed intercepts and output the sensitivities of the market to this natural dummy. It is inferred that PTSCS or cross-section pooling could be a valuable method to uncover market response estimates when PCSTS data estimation get invalidated due to failure of the homogeneity assumption.