A structural time series approach to modelling multiple and resurgent meat scares in Italy

Abstract

This paper exploits a structural time series approach to model the time pattern of multiple and resurgent food scares and their direct and cross-product impacts on consumer response. A structural time series Almost Ideal Demand System (STS-AIDS) is embedded in a vector error correction framework to allow for dynamic effects (VEC-STS-AIDS). Italian aggregate household data on meat demand is used to assess the time-varying impact of a resurgent BSE crisis (1996 and 2000) and the 1999 Dioxin crisis. The VEC-STS-AIDS model monitors the short-run impacts and performs satisfactorily in terms of residuals diagnostics, overcoming the major problems encountered by the customary vector error correction approach.

Acknowledgements

The authors would like to thank the Editors, the anonymous referees and Bhavani Shankar for their valuable comments on earlier versions of this article. This research was partially funded by the European Commission under programme QLK1-CT-2002-02343. The authors remain solely responsible for any error or omission.

Notes

¹ For a thorough discussion of the VEC-AIDS see Pesaran and Shin (2002), while Fanelli and Mazzocchi (2002) test its performance as compared to alternative estimation methods.

² A broader model, where all coefficients are allowed to vary smoothly through random walks as in Mazzocchi (2003) was rejected by empirical evidence and results are not reported here.

⁵ The VEC-AIDS on the whole sample period was estimated under the assumption of two cointegrating relationships.

³ The data set of this study slightly differs from the one employed in Fanelli and Mazzocchi (2002) as the Italian household budget survey has been restructured in 1997. To ensure continuity and comparability, we limit the analysis to beef and chicken, while all other meats are aggregated in the residual equation.

⁴ Maddala and Kim (1998) provide an extensive literary survey and examples of the issues of cointegration analysis in the contest of structural change. More recent work on the impact of structural breaks on the rank of cointegrating space is that by Johansen et al. (2000), Leybourne and Newbold (2003), while Hansen and Johansen (1999) and Hansen (2000) explore the issue of changes in the cointegration parameters.

⁶ Wang and Bessler (2002) show that VAR-type demand models return better forecasts when homogeneity is not rejected, while evidence is inconclusive in case of rejection.

⁷ A general-to-specific specification search was conducted starting from a model where all parameters were allowed to change over time and selection was made on the basis of goodness-of-fit criteria and other residual diagnostics. Only estimates of the selected STS-AIDS model are reported in this paper.

⁸ As in Masih and Masih (1996), coefficients for computing the short-run elasticities are obtained by adding up the prices and expenditure coefficients over the number of lags included in the VEC-STS-AIDS model. The same assumption has been maintained for testing symmetry and homogeneity.