BOX 2.2: Microeconomic Evidence on the Impact of Infrastructure in Sub-Saharan Africa

Two distinct approaches have been developed to understand the cross-country differences in aggregate productivity from a microeconomic perspective. The first approach is to evaluate the drivers of the total factor productivity (TFP) of an individual firm in a country relative to its counterparts in another country-say, cross-country differences in the ability to adopt more efficient technologies (Parente and Prescott 1994) or to operate technologies efficiently (Bloom, Schankerman, and Van Reenen 2013; Bloom, Draca, and Van Reenen 2016). The second approach is to examine resource misallocation as a key factor in accounting for cross-country differences in TFP.

The resource allocation process is characterized by: (a) the types of establishments operating in the economy (if newcomers are more productive than incumbents or exiting firms), and, (b) the allocation of labor and capital across establishments in operation. Distortions in either of these two features will lead to resource misallocation and, in turn, lower aggregate TFP. The resource misallocation approach aims not only at quantifying the extent of resource misallocation (indirect approach), but also understanding the underlying sources of misallocation (direct approach)-namely, product and labor market distortions, trade restrictions, financial frictions, informality, entry barriers, and credit market imperfections, among others (Restuccia and Rogerson 2013).

In this context, poor infrastructure networks can restrict factor mobility and hamper productivity. Shiferaw et al. (2015) examine the impact of having an improved road network on the entry decision and entry size of manufacturing firms in Ethiopia. The analysis is conducted using geographic information system-based panel data on the road accessibility of Ethiopian towns and census-based panel data for manufacturing firms over 1996-2008. The authors construct three measures of road infrastructure: (a) total distance that can be traveled during a 60-minute drive, (b) total area accessible during the 60-minute drive, and (c) total travel time from a particular locality to major economic destinations.^a The evidence shows that: first, the quality of local road infrastructure is positively associated with the number of firms present in the locality. However, the number of firms has no significant relationship with the connectivity of the road infrastructure. Second, the size of new entrants is more strongly associated with connectivity rather than with the quality of the local road infrastructure. In sum, local road infrastructure is important because it enables more firms to set up, but more extensive market connectivity may be important for the entry of large firms. In other words, poor infrastructure could be a source of resource misallocation through the selection channel.

Gollin and Rogerson (2014) build a general equilibrium model where the size of subsistence agriculture results from the interplay among sectoral productivities (in agriculture and manufacturing) and transportation productivity. Manufacturing goods are produced in urban areas. Agricultural activity takes place in near and remote rural areas. There are iceberg transportation costs of moving goods into and out of the remote region. Labor is mobile across regions, and people living in remote regions belong to the subsistence agriculture sector. After calibrating the model for Sub-Saharan Africa, it is found that improvements in agricultural productivity and lower costs of intermediate inputs free up labor from the agriculture sector-primarily from subsistence agriculture. Improving transport productivity (lower iceberg transportation costs) helps move individuals from subsistence agriculture into manufacturing, leaving the share of workers living in the near region unchanged. A boost in manufacturing productivity alone, by contrast, lowers the share of population in subsistence agriculture, but the magnitude of this effect is lower than that of greater agricultural productivity or lower transport costs. Therefore, structural transformation at low levels of development is mainly driven by productivity surges in agriculture and transportation. Economically speaking, a 10 percent increase in agricultural TFP combined with a 10 percent reduction in transport costs leads to a 14-percentage-point reduction in the labor share in subsistence agriculture. The welfare effects are significant-comparable to raising consumption per capita in the economy by 62 percent.

Other research papers provide evidence that deficient infrastructure hinders Africa's development. There is evidence that an unreliable supply of electricity deters firm-level investment in Uganda (Reinikka and Svensson 1999). Insufficient power-generating capacity limits growth in Ghana (Estache and Vagliasindi 2007). Poor transport infrastructure also imposes a high cost in the region. Diao and Yanoma (2003) show that growth in the agriculture sector is constrained by high marketing costs, which largely reflect poor transport (as well as other infrastructure) facilities. In general, it is found that higher costs of transportation alter the incentives for agricultural investment (Renkow, Hallstrom, and Karanja 2004; Stifel and Minten 2008). Overall, deficient infrastructure hinders the growth impact of private investment-especially foreign direct investment-in Africa (Lumbila 2005).

a. The measures in (a) and (b) capture primarily local improvements in road infrastructure, while that in (c) is a more comprehensive measure of how roads affect the connectivity of firms with local and distant markets.