Understanding the Political Economy of Agriculture in the Tropics

This paper offers a new explanation for why some governments consistently impose such high agricultural tax rates that their revenues decline, and consistently spend what tax revenue they have on things other than productive public investments such as agricultural R&D. We propose an intertemporal political economy model, in which taxes and R&D levels are set in a repeated game between government and producers. Panel data from Africa sustain the hypothesis that high sunk costs in production are potential targets of predatory taxation, making it timeinconsistent for governments to announce low-tax, high-investment policies. Thus there is persistent stagnation, with high tax rates and little investment, until new production techniques reduce fixed costs or new policymaking institutions allow credible commitments to a low-tax, high-investment regime. Acknowledgements Many thanks to Phil Pardey for sharing data on agricultural R&D, and to USAID’s collaborative agreement on Equity and Growth through Economic Research (EAGER) for financial support. JEL Classification F43, O41 \Growth-Africa\ASSA-Masters&McMillan.doc February 9, 2003 Understanding the Political Economy of Agriculture in the Tropics Empirical studies often find that policy differences account for a large fraction of crosscountry differences in the level and growth rate of per-capita income (e.g. Hall and Jones 1999, Sachs and Warner 1997). Such findings beg the question of why policies differ. A particularly urgent puzzle is why some governments impose confiscatory taxes that actually reduce tax revenues (McMillan 1998), and why such policy failures persist across successive governments in some regions, notably in Africa (Easterly and Levine 1998). To formulate an intuitively plausible, politically relevant and empirically testable model of policy formation, we focus on two specific instruments – output taxation and investment in public goods -and treat economic growth as the equilibrium outcome of an infinitely-repeated game between government officials and the private sector. The model generates persistent policy failure when private investment is irreversible, and producers expect (or fear) that policymakers will raise tax rates in the future. This approach builds on the time-consistency concept of Kydland and Prescott (1977), and the applications to intertemporal price policy of Besley (1997), McLaren (1996) and Gilbert and Newbery (1994). We add interaction with public investment in R&D so as to construct a mechanism whereby policy choice underpins changes in total factor productivity and economywide growth, and test the model empirically using cross-country data. Taxation and Investment in the Tropics 2