Government demand and firm performance: New evidence for sub-Saharan African Countries

This project seeks to analyse the role of public procurement as a determinant of firm performance in developing countries. Public procurement is increasingly being employed as an industrial policy tool by East African countries (and developing countries more generally) searching to upgrade domestic industrial capabilities and spur innovation while promoting sustainable development. The recently launched implementation plan for the ‘Buy Uganda, Build Uganda’ policy and amendments to the guidelines on public procurement in Uganda, for example, provide for explicit preferences for local companies in certain sectors. However, evidence on the effect of procurement on firm performance across developing countries is extremely scant (Evenett and Hoekman, 2005). Using firm-level data based on a survey administered by UNIDO, we plan to provide a timely contribution to this emerging policy debate in Uganda and the East African Community (as EAC member states including Rwanda, Kenya and Tanzania are also adopting similar procurement reservations).

To conduct the analysis, we will build on the 2010/11 UNIDO African Investor Survey (AIS). This survey generated information on 6,492 domestic and foreign-owned firms based in 19 African countries. Compared to existing firm-level data on low-income countries (e.g. the World Bank Enterprise Survey), this survey is unique in that it includes a specific question on the share of output that is sold to government entities. Among the 4,600 firms (about 68% of total sample) that answered this specific question, almost 29% reported a positive share of their sales going to the Government. Considering the full sample of firms and countries, Government absorbed on average about 8.1% of firms’ sales. This information allows for an analysis of the effect of public procurement from local firms, while differentiating between firms on the basis of their ownership, size, sector, location, and other heterogeneous characteristics. The methodologies used (e.g., OLS & quantile regressions) will centre on the estimation of a general functional relationship linking a firm’s performance to a vector of firm-specific factors and a coefficient representing the demand coming from the Government. We will also include a rich set of country and industry fixed effects and interaction effects to rule out some of the unobserved factors that can influence the relationship between a firm’s performance and sales to the government. Finally, we will conduct a battery of robustness checks to test the stability of our results to issues arising from sample selection and endogeneity.

The contribution of this project is expected to be fourfold. First, we will provide an overview of the existing evidence and economics literature on public procurement in developing countries. Second, we will introduce new evidence (descriptive statistics) on the role of government demand in firm sales across countries, sectors, and according to firm characteristics, as reflected in a representative sample of firms surveyed by UNIDO in 19 SSA countries. Third, we will undertake an econometric analysis of the effect of procurement on the productivity performance of firms, accounting for the potential heterogeneity across countries, sectors and firm-level characteristics. Fourth, we will distinguish between domestic and foreign suppliers (in terms of ownership/equity) and evaluate the potential impact of introducing preference margins for domestic firms on their performance.