Dynamics of firm-level export diversification in Botswana
Doctoral Thesis
2018
Permanent link to this Item
Authors
Supervisors
Journal Title
Link to Journal
Journal ISSN
Volume Title
Publisher
Publisher
Department
Faculty
License
Series
Abstract
This thesis investigates the firm-level dynamics of export diversification in Botswana. Botswana is a country characterised by a high level of export concentration, with diamonds dominating its export bundle. With the stock of diamonds expected to be depleted in the near future, Botswana faces the urgent challenge of diversifying its export bundle. While much analysis has focused on the product composition of exports, little focus has been placed on the role that firms play in driving the composition of exports in Botswana over time. This thesis fills this gap in the literature. The analysis draws on various unique and unexplored databases. Firstly, it uses a panel of transaction level data for the period 2003 to 2012 obtained from Statistics Botswana. Second, the transaction data are merged with a panel of manufacturing firm data for the period 2003 to 2012 obtained from Department of Industrial Affairs. Finally, tariff data at the product level (HS8) for the period 2003 to 2012 are obtained from the World Integrated Trade Solution (WITS). Drawing on these databases allows for a detailed firm level analysis of export diversification not previously possible for Botswana. The thesis is comprised of three main chapters in addition to the general introduction and concluding chapters. The first main chapter (chapter 2) uses the transaction data to document the stylized facts associated with Botswana’s firm-level export diversification. Their consistency is assessed with empirical evidence in other countries. The background analysis reveals that a majority of exporting firms (over 70%) export to a single export destination and a small fraction of firms (less than 25%) export to multiple export destinations. However, as found in the international literature, export values are highly concentrated amongst the multi-destination exporters. The analysis also reveals that diamonds dominate Botswana’s export bundle, which are exported to one major destination, being the United Kingdom. An additional focus of the chapter is the relationship between firm size and changes in export diversification, defined in terms of product and destination margins. To study the dynamics between firm export size and diversification, a Multinomial logit regression approach is adopted. This technique is used to estimate the predicted probabilities of moving between different product-destination categories as a firm grows in export value. The results reveal nonlinearities in the evolution of a firm’s diversification path. At low values of exports, firms concentrate on selling a single product to a single destination. As firms grow in export value, they expand the number of products to the destination rather than the number of destinations of that product. This is a striking contrast to results found in other countries whose diversification path has been found to be driven mainly by the expansion of the number of destinations per product (Stirbat et al., 2011; Cadot et al., 2013). Only at higher export values do the multi-product firms transition into exporting to multiple destinations. The contrasting diversification path for Botswana suggests that diversification into new export markets is a key constraint to growth and diversification of Botswana’s export bundle. Therefore, the remaining chapters of the thesis explore firm level factors determining export destination diversification. Chapter 3 looks at the role of firm productivity in driving the diversification of firm exports across destinations. An important component of this analysis is the productivity relationship associated with manufacturing firm’s exporting out of the Southern African Customs Union (SACU). To assess the relationship, the transaction data are merged with the manufacturing database. Given the high number of zero trade flows for many firms, the Zeroinflated Poisson regression model is used to estimate the link between firm productivity and export destination diversification. The results strongly support the prediction that more productive firms enter the export markets. The results also show that upon entering the export market, only the relatively productive firms become multi-destination exporters. The results also confirm the presence of a productivity premium for firms exporting out of the SACU region. Chapter 4 tests the complementary input hypothesis where access to imported intermediate inputs enhance productivity thus enabling firms to access more export destinations. It further assesses whether the impact varies across differentiated inputs or homogenous inputs. Three measures are used to proxy input complementarity, namely: number of product-source country pairs, number of source countries and total import value. Using a poisson model with fixed effects, the results provide strong evidence of a positive association between variety of imported inputs used by a firm and the range of destinations it exports to. The results are robust across all the measures of input complementarity. Given concerns regarding endogeneity of imported input use, the reductions in tariffs under the South Africa – European Union Trade, Development and Cooperation Agreement (TDCA) is used to instrument firm use of imported intermediate inputs. Using the Two-stage Residual Inclusion approach, the results confirm the productivity-enhancing effects of the input complementarity hypothesis on firm export destination diversification. These results, hence, suggest that firms stand to benefit from the productivity-enhancing effects of imported intermediate inputs which can boost their export destination diversification efforts.
Description
Reference:
Kebakile, P.G. 2018. Dynamics of firm-level export diversification in Botswana. . ,Faculty of Commerce ,School of Economics. http://hdl.handle.net/11427/30165