Elsevier

World Development

Volume 109, September 2018, Pages 497-510
World Development

Rural–Urban Linkages, Public Investment and Transport Costs: The Case of Tanzania

https://doi.org/10.1016/j.worlddev.2016.08.013Get rights and content

Highlights

  • We examine the impact of public investment on household welfare in Tanzania.

  • The impacts of different policies do not necessarily rest in the targeted sectors.

  • Rural households benefit most from investment in non-agricultural sectors.

  • The distributional consequences of government financing determine outcomes.

Summary

The economy of Tanzania, like those of many other sub-Saharan African countries, displays strong geographic and locational disparities. We develop a three-location spatial applied general equilibrium model calibrated to the 2001 Tanzanian Social Accounting Matrix to examine the impact of various public investment programs on household welfare across this diverse geography in which production and consumption are locationally specific and transport costs support equilibrium price differences across locations. We examine how different public investment packages combined with reforms in the transport sector alter the equilibrium structure and location of economic activity. The choice of financing arrangement matters for welfare, since tax incidence, relative price, and real exchange rate movements are non-neutral. We show that the distributional consequences of alternative investment programs may matter more in terms of household welfare than the direct consequences of targeting investment to particular sectors or locations. For instance, under some financing scenarios, interventions that aid agriculture may lead to decreases in the welfare of the rural unskilled labor force, because the financing mechanisms create distortions that effectively skew the terms of trade sufficiently powerfully against the rural unskilled as to outweigh the direct welfare-enhancing effects of the public investment. We also note that welfare gains are generated by the movement of rural workers out of quasi-subsistence agriculture into higher productivity jobs in other sectors and locations.

Introduction

Tanzania’s economy, like those of many other sub-Saharan African countries, displays strong geographic and locational disparities. In a stylized sense, it can be thought of as consisting of several distinct components, spatially separate and imperfectly connected. The capital, Dar es Salaam, is a vibrant urban economy with a strong government presence; given its location as a port on the Indian Ocean, it is reasonably well integrated into the global economy. At the opposite extreme, many interior rural regions are heavily agricultural and are poorly linked to national or global markets. Households in these rural areas consume much of what they produce, and they produce much of what they consume; they sell modest fractions of their agricultural output and purchase manufactured goods and services. In between the urban economy of Dar es Salaam and these quasi-subsistence rural regions lies a mixture of secondary cities, market towns, and well-connected commercially oriented rural areas. They have better access to markets in Dar es Salaam and the rest of the world than do the more remote rural areas, although they still face consequential transport and transaction costs with respect to those markets.

These spatial patterns are associated with corresponding differences in the patterns of production and consumption. They also are associated with substantial differences in prices, wages, and living standards. The spatial disparities create a variety of seeming paradoxes: an economy that is simultaneously open (in Dar) and closed (in rural areas), with respect to world trade; an economy that is subsistence oriented coexisting with one that is highly commercial; and an economy in which people are free to migrate but in which there are substantial differences in standards of living across locations.

Within an economy like this, public investments targeted to different sectors or locations can have different effects on national income and on welfare. This paper uses an applied general equilibrium model to explore the impact of a set of government policies on household welfare, taking seriously the spatial and sectoral differentiation of the Tanzanian economy. In particular, we focus on public infrastructure investments that may differentially affect the agricultural sector and on reforms that directly reduce transport costs. The model, which is calibrated to Tanzania’s 2001 Social Accounting Matrix, pays particular attention to inter-regional transaction costs and rural–urban linkages. We view the resulting model as occupying a middle ground between a highly stylized but equally transparent model (as in Gollin & Rogerson, 2014) and the highly detailed CGE models sometimes used for policy analysis, such as the 58-sector model deployed in Pauw and Karl (2010). The former are perhaps too stylized to be useful for realistic policy analysis, whereas the latter by necessity build in a high degree of structure that limit their usefulness in thinking about policies that might alter the underlying organization of the economy.

Our model generates a set of stylized but important insights. We show how different programs of investment generate substantially different impacts on the economy and also have different effects on the well-being of unskilled workers. In general, we find that the benefits of public investment are often felt in sectors other than those that are the primary target of the interventions. For example, although a large fraction of the unskilled labor force resides in rural areas and works in agriculture, we find that increasing public investment in the agricultural sector generally does not improve the well-being of those unskilled workers who remain in agriculture. It does, however, result in an outflow of workers from agriculture to other sectors. Similarly, we show that increased public investment in urban areas, particularly in secondary cities, may lead to welfare benefits for rural households.

The second key message from the paper is that the poverty impacts of different interventions are highly sensitive to the ways in which these are financed. As in Adam and Bevan, 2006, Adam and Bevan, 2014, outcomes are sensitive to government choices with respect to financing mechanisms. In our analysis, interventions can be financed through taxes or other forms of domestic revenue collection (e.g., tariffs). Alternatively, they can be financed through aid inflows—essentially gifts from abroad—or “deficit financed” which entails the direct crowding-out of private investment. All of these sources of finance create impacts on the economy and on the distribution of well-being. Taxes create obvious effects on the absolute and relative well-being of different households and locations through differential incidence; the same is true for tariffs. Foreign assistance appears at first glance to be relatively neutral as a source of financing, but the capital inflows associated with foreign aid lead to exchange rate effects and relative price effects that can impact the poor. We argue that the choice of financing arrangement, no less than the location or sector in which the public sector invests, will affect income distribution, and by extension the poverty impacts of public policies.

A third finding of the paper is that we estimate the potential welfare gains associated with interventions that directly reduce the transport costs that sustain the disintermediation of the economy. While it is impossible to accurately cost these interventions in the context of our stylized model, we show that reduction in the transport cost wedge directly accelerates processes of structural transformation. This process may be complicated, however, if transport costs represent rents to monopoly providers. In this case, policy reforms have potentially large effects on the pattern of domestic demand; the “pure” gains from transport cost reductions are mediated by the loss in rents to the transport sector.

Finally, we note that an important channel for these effects is through the sectoral and regional reallocation of labor. This finding is consistent with numerous recent papers suggesting that structural transformation and migration are important channels for welfare improvement—as opposed to increases in income for workers within sectors; e.g., Beegle, Christiaensen. Different interventions may lead low-skill workers to move across sectors or regions. Where there are important sectoral differences in income or productivity (e.g., Gollin, Young, 2013), sectoral movements have the potential to increase output and income considerably. Such welfare gains can easily be overlooked in household surveys that maintain constant sampling frames and do not track individuals or households that migrate.

Section snippets

Background

Tanzania is a country with strong spatial patterns of economic activity. Almost 80% of the population lives in rural areas and approximately the same proportion of the labor force works primarily in agriculture. Most households depend for their livelihoods on farming small plots of land, where they primarily produce food for home consumption. Small amounts of food (and non-food agricultural goods) are sold to market. Productivity in the agricultural sector is generally very low. Agriculture’s

Previous literature

The paper contributes to a recent and growing literature that has revisited the relationship between transportation costs and development patterns.

The relationship between transport costs and development is difficult to examine for several reasons. One is that general equilibrium effects may dominate partial equilibrium effects, at least in the long run. For example, a new road may induce many changes in economic activity patterns, including changes in prices, wages, movements of labor and

Model

Our model embeds the basic insights from Gollin and Rogerson (2014) within the framework developed by Adam and Bevan (2006). Thus both production and consumption are spatially distributed across three stylized geographic locations. One is a rural region, which produces only staple food, which is non-tradable internationally, and an exportable cash crop. A second region is the commercial capital, Dar es Salaam, which is assumed to do no agricultural production, but which serves as a production

Calibration and baseline parameterization

The model is calibrated to a stylized version of the 2001 Tanzania Social Accounting Matrix (SAM) produced by the International Food Policy Research Institute (IFPRI), as described in Thurlow and Wobst (2003).

Experiment design

We restrict our attention in this paper to two basic experiments. The first consists of an exogenous permanent increase in the rate of public infrastructure investment, in the context where public infrastructure investment has a positive rate of return. The greater abundance of public capital increases the productivity of private factors but may do so differentially across sectors and locations (see Adam & Bevan, 2006). For example, a rural feeder road-building program might disproportionately

Public infrastructure investment

Table 4(a)–(c) provides a summary of the public investment experiments. Panel (a) reports the change in broad macroeconomic balances; panel (b) the change in the (spatial) economic structure; and panel (c) focuses on the welfare effects of these changes alongside equilibrium changes in relative prices than underpin these welfare changes. The left hand panel corresponds to the case where public infrastructure is biased towards the production of agricultural output (public irrigation systems, for

Conclusions and next steps

This analysis offers some sobering conclusions with respect to efforts to study the impacts of investments in agriculture and non-agriculture. A recurring question in policy circles is whether governments should target agriculture or other sectors if they wish to improve the well-being of the poor. We find that general equilibrium effects make this a complicated question to answer: because sectors are linked through factor markets and through a complex input–output structure, the benefits and

Acknowledgments

We thank the International Growth Centre for research support, Radhika Goyal for excellent research assistance and we gratefully acknowledge the assistance of Beatrice Mkenda for sharing her valuable research on transportation costs in Tanzania.

References (38)

  • Paul Dorosh et al.

    Can cities or towns drive African development? Economywide analysis for Ethiopia and Uganda

    World Development

    (2014)
  • D. Gollin et al.

    Productivity, transport costs and subsistence agriculture

    Journal of Development Economics

    (2014)
  • H.G. Jacoby et al.

    On measuring the benefits of lower transport costs

    Journal of Development Economics

    (2009)
  • Adam, C. & Bevan, D. (2014). Public investment, public finance and growth: The impact of distortionary taxation,...
  • C. Adam et al.

    Aid and the supply side: Public investment, export performance and dutch disease in low-income countries

    World Bank Economic Review

    (2006)
  • Africa Infrastructure Country Diagnostic (AICD)

    Tanzania’s infrastructure: A continental perspectiveCountry Report

    (2010)
  • S. Asher et al.

    Market access and structural transformation: Evidence from rural roads in India

    (2016)
  • D. Atkin et al.

    Who’s getting globalized? Intra-national trade costs and world price pass-throughNBER Working Papers. 21439

    (2015)
  • A. Banerjee et al.

    On the road: Access to transportation infrastructure and economic growth in ChinaNBER working papers 17897

    (2012)
  • N. Baum-Snow et al.

    Roads, railroads and decentralization of Chinese cities

    (2015)
  • K. Beegle et al.

    Migration and economic mobility in Tanzania: Evidence from a tracking survey

    Review of Economics and Statistics

    (2011)
  • Buffie, E., Berg, A., Pattillo, C., Portillo, R., & Zanna, L.-F. (2012). Public investment, growth and debt...
  • Carleton, T. A. (2013). Spatial integration of staple food markets in Tanzania: Applying novel methods and data to...
  • L. Casaburi et al.

    Rural roads and intermediated trade: Regression discontinuity evidence from Sierra Leone

    (2013)
  • F. Caselli et al.

    Agriculture and structural transformation in an open economy: Evidence from Ghana

    (2014)
  • L. Christiaensen et al.

    Urbanization and poverty reduction: The role of rural diversification and secondary towns

    Agricultural Economics

    (2013)
  • Derksen-Schrock, K., Anderson, C. L., & Gugerty, M. K. (2011). Tanzania: Agricultural sector overview. EPAR brief no....
  • Donaldson, D. (2016). Railroads of the Raj: Estimating the impact of transportation infrastructure. American Economic...
  • P. Dorosh et al.

    Agglomeration, growth and regional equity: An analysis of agriculture- versus urban-led development in Uganda

    Journal of African Economies

    (2011)
  • Cited by (0)

    View full text