Productivity growth, technical progress and efficiency change in Spanish retail trade (1995–2004): a disaggregated sectoral analysis

Abstract

This article analyses productivity growth in Spanish retail stores during the period 1995–2004. It is also interested in analysing the influence of regulation/deregulation processes on the efficiency and productivity of the firms. The analysis is carried out from a disaggregated sectoral perspective at the 4-digit NACE code level. The non-parametric Data Envelopment Analysis approach is used to compute Malmquist productivity indexes. These are decomposed into efficiency change and technical change. Big differences are found in the productivity growth for each sector. First, six retail sectors experienced positive productivity growth, while six saw productivity growth decrease. Second, most sectors experienced technical progress. Third, some sectors improved their efficiency, while others became less efficient. Fourth, the TFP improvements were almost entirely due to technical progress, and only four sectors improved their efficiency. The findings obtained from the analysis of the deregulation of opening hours show two important facts: (i) the differences in the productivity and the efficiency of the firms between sectors, and (ii) the losses /improvements of efficiency of the firms in relation to the effects of the shop opening hours or the lack of adaptation to the environment.

Keywords:

• retail stores
• productivity growth
• efficiency changes
• technical progress
• regulation
• retail sector
• data envelopment analysis

Acknowledgements

The author would like to thank the editor, and the two anonymous reviewers for their valuable comments.

Notes

1. The kernel estimator of the value of the density function of that variable at point x, f(x), is given by: , where K(·) is the so-called kernel function, and h is the bandwidth (or smoothing parameter) that controls the regularity of the estimated curve. Technical details for the smoothing parameter (h) can be found in the works of Sheather and Jones (1991) and Park and Marron (1990). Marron's website offers the Matlab routines for obtaining these parameters (see http:\\www.stat.unc.edu/faculty/marron.html).

2. Both differences exist statistically significantly intra- and inter-sector. Test Kruskal Wallis among cross-section same sector (5211: Statistical = 124,45 P-value = 0.0000) (5248: Statistical = 135,98 P-value = 0.0000), Intra-sector (for every year), for example, in 2004; 5211 vs 5248 Test Kolmogorov Smirnov Statistical asymptotic K-S = 5.13923 P-value 0.0000.