P1-2 Consistent modelling of vertical price transmission processes at different levels of spatial disaggregation

PhD student: Said Tifaoui
Thesis Committee: Prof. Stephan v. Cramon-Taubadel, Prof. Helmut Herwartz, Prof. Bernhard Brümmer

Economists study vertical price transmission (VPT) processes to gain insights into the interactions between prices at different levels of a marketing chain. The nature of these VPT processes can cast light on the nature of competition in the marketing chain. Hence, analysts employ time series methods such as vector error correction models (VECMs) and non-linear variants of these models to depict complex, regime dependent VPT processes (e.g., asymmetric, threshold, smooth transition, and Markov-switching VECMs). However, disaggregated prices for individual food products in individual stores do not display unit root behaviour. Typical retail prices tend to persist and most deviations from the “regular” price are negative, often as a result of temporary discounts. Furthermore, prices generally reflect so-called psychological pricing, meaning that they end in “9” (e.g., 7,99). Hence, spatial aggregation has a significant impact on the time series characteristics of price data. The results of VPT analysis are also influenced by spatial aggregation; for example, analysis with aggregated data tends to point to slower and more asymmetric VPT processes than does analysis with disaggregated (i.e., individual retail store) price data.

The project “Consistent modelling of vertical price transmission (VPT) processes at different levels of spatial disaggregation” aims to investigate the impact of data aggregation on the estimation and testing for VPT in food products. This research focuses in developing appropriate, possibly panel estimation techniques, for estimating VPT processes with disaggregated retail price data that can be used in place of the standard cointegration methods that are used with aggregated prices.

back to overview of all projects