Investigating product market Integration in the southern African development community : a price-based approach
Doctoral Thesis
2015
Permanent link to this Item
Authors
Journal Title
Link to Journal
Journal ISSN
Volume Title
Publisher
Publisher
University of Cape Town
Department
Faculty
License
Series
Abstract
This thesis extends the price-based empirical literature on border effects and product market integration to the South ern African Development Community (SADC). The analysis draws on a unique integrated dataset of district-level monthly average retail product prices spanning a number of districts in Botswana, Malawi, South Africa and Zambia. The thesis is comprised of four main chapters. The first main chapter provides an empirical analysis of the extent to which product prices are integrated within and between the four SADC countries . The results reveal large and persistent absolute deviations from the law of one price both within and between each of the countries over the period from 200 6 to 2009 . Price deviations are found to be higher between SADC countries than within the individual countries, although there is considerable heterogeneity in the magnitude of these deviations across products. On average, absolute price deviations between-country pairs are smaller for countries adjacent to each other and for countries that share common membership in the Southern African Customs Union (SACU) . Simple econometric estimates based on the standard regression approach in the literature show that absolute price deviations between district pairs in the region increase the further apart the districts are from each other ; and are 11.8% higher, on average, between districts separated by a national border. Overall, there is no clear evidence that product prices in the SADC region became more integrated between 2006 and 2009 (although product prices between the Common Market for Eastern and Southern Africa countries did become more integrated over this period), despite the liberalization of tariffs under the SADC Protocol on Trade. The second main chapter critically evaluates the standard empirical methodology used to estimate border effects. The evaluation identifies several different sample selection effects that bias estimates of distance and border effects in the existing literature, and demonstrates the sensitivity of estimates of transaction costs and border effect s for the SADC region to these sample selection biases using quantile regressions. The results show that the standard pooled OLS estimate s reported in much of the existing literature (and in the first main chapter) suffer from a sample selection problem which biases the estimated distance and border coefficients downwards relative to the true cost of trade. The chapter also demonstrates the impact of two additional product and distance sample selection biases that are not dealt with in the Borraz et al. (2012) application of the quantile regression approach. Finally, it shows that not accounting for variation in within product quality across districts results in omitted variable bias which raises the estimated distance and border coefficients . The chapter proposes novel extensions to the quantile regression methodology to allow for the analysis of cross-country border effects; and to account for sample selection bias arising due to product and distance sample selection effects . The third main chapter applies the modified quantile regression methodology to precisely estimate average and individual border effects in the SADC region for the 2006 to 2009 period . The quantile regression results show the effects of borders in raising price dispersion in the region are generally comparatively lower between the SACU (23.1%) and Common Market for Eastern and Southern Africa ( 26.6%) countries compared to the remaining SADC country pairs (36.2%). There are also clear differences in the magnitude of border effects for contiguous (24.1%) and non-contiguous (41.1%) SADC countries, providing evidence of incremental border effects as products are traded across multiple borders. While the magnitude of the border effect estimates are sensitive to the estimation technique employed, the ranking of the border effects across different regional trade agreements and for contiguous versus non-contiguous countries is robust across different specifications. Finally, the results in the third main chapter reveal that, on average, there was little change in the magnitude of border effects in the SADC region between 2006 and 2009, despite the accelerated liberalization of tariffs on intra-SADC trade over this period. The final main chapter of the thesis unpacks the contribution of preferential tariffs to the South Africa-Zambia border effect over the period from 2002 to 2009. This addresses the lack of studies in the literature of the direct contribution of tariffs to border effects on international relative prices. The estimation results reveal an almost perfect pass-through of preferential tariffs onto domestic prices in Zambia. They also show that preferential tariffs account for a significant portion of the border effect . After accounting for the role of preferential tariffs, the additional increase in prices in the Zambian market caused by the border effect falls from 29.4% to 16.1%. This general result is qualitatively the same even after accounting for variation in the intensity with which products are traded between the two countries. Even so, a simple analysis of the trend in the South Africa-Zambia border effect indicates that the impact of crossing the border in raising Zambian retail prices actually increased between 2002 and 2009. The results do suggest, however, that the increase in the border effect over this period would have been more substantial in the absence of the phasing down of preferential tariffs. On balance, the evidence presented in this thesis indicates that markets in the SADC region remain fragmented, with little sign of greater product market integration either within or between countries. This is despite the explicit policy focus on trade reform that has accompanied the introduction of the SADC Protocol on Trade. Trade liberalization alone appears not to be sufficient in generating greater product market integration within the region.
Description
Keywords
Reference:
Balchin, N. 2015. Investigating product market Integration in the southern African development community : a price-based approach. University of Cape Town.