University of Wisconsin-Madison Department of Agricultural and Applied Economics Development Economics Preliminary Examination January 2003 Answer any three questions. All questions have equal weight. Please type your answers. Your answers must be submitted to Ian Coxhead, Development Prelim Chair, 413 Taylor Hall no later than 5 pm on Friday, January 10. 1. The three parts of this question invite you think about theoretical and empirical issues in the measurement and interpretation of poverty in economic growth. Consider an economy of n people for which we define a standard money-metric poverty line p. Let yi denote the income of the ith person in this economy. Let H denote the standard headcount ratio poverty measure, I denote the income gap ratio defined as: ( p yi ) I yi p ; pnH and, P2 denote the degree two Foster-Greer-Thorbecke poverty measure defined as: 2 ( p yi ) H [ I 2 (1 I ) 2 C p2 ] , p yi p 2 where C p is the squared coefficient of variation of income among the poor. 1 P2 n a. Explain why P2 is preferable on axiomatic grounds to H and I as a measure of poverty. b. Alesina and Rodrik (“Distributive Policies and Economic Growth,” Quarterly J of Econ (109)2: 465-490) report the following results for an endogenous growth regression: 2 Suppose that you are inspired by these results and want to econometrically explore the impact of poverty on growth. Which of the above poverty measures (P2, H or I) would you want to use for your empirical analysis? Do you think that your econometric results would be sensitive to which poverty measure you use in the regression? In answering this question, please develop your intuition about why initial poverty levels might shape subsequent growth. c. Finally, please sketch out a model that formalizes your intuition on the impact of poverty on growth. Discuss your theory’s implication for the expected relationship between growth and the different poverty measures. 2. Raising productivity through technological innovation is widely recognized as fundamental to economic development. Correspondingly, development economists have continued to work hard to improve their understanding of the factors that shape the dynamics of technology adoption, especially in agricultural areas of developing countries. Many papers (e.g., Feder and Slade, Besley and Case, Foster and Rosenzweig) have emphasized the potentially important role that social networks and/or neighborhood effects may play in the learning process associated with the adoption and other technology use decisions of farm households. This problem asks you to consider the role of social networks or neighborhood effects in shaping technology adoption processes in developing countries, and to deal explicitly with theoretical and empirical challenges in modeling these effects. [Note: Parts a and b could be relatively brief (at most 3-4 paragraphs each)]. a. Begin by identifying contexts (characteristics of technologies, households, markets, and/or institutions) where you think social networks or neighborhood effects are likely to be of importance to understanding technology adoption processes in developing countries. b. Consider next the appropriateness of the Foster and Rosenzweig theoretical approach (deploying an optimal input use model) given your answer to part a. and discuss an alternative one might use to model the technology adoption decisions of households. Be explicit about the basic theoretical structure of your alternative. c. Next, evaluate the theoretical and empirical treatment of neighborhood effects in the Foster and Rosenzweig article in light of the “reflection (or identification) problem” raised by Manski (Review of Economic Studies 1993, vol 60, no. 3: 531542) and others (e.g, O. Bandiera and I. Rasul, “Social Networks and Technology Adoption in Northern Mozambique”). d. Given your answers to a-c, propose a theoretical modeling framework that could guide an empirical investigation of the role of neighborhood effects in shaping technological advance in rural areas of developing countries? You can be as explicit as necessary about the context you have in mind for the investigation and how that shapes your approach. 3 3. A recent article in The Economist (December 19, 2002) highlights transport costs and inadequate infrastructure as factors inhibiting development, illustrating the African Development Bank’s finding of “a strong link between poverty and remoteness”. (The article cites the rising cost of a bottle of Coca-Cola with distance from the bottling plant in Yaoundé, Cameroon: 300 local currency units in Yaoundé; 315 in a town 125 km away, 350 in a village 100 km further yet, and so on). This question asks you to consider the spatial incidence of some common developing-country tax policies, their consequences and cures. Consider an agricultural economy in which all commercial farmers grow one crop, which is exported through the port. The farm-gate price of the crop, k, depends on distance from the port (x) and fuel costs (a* per kilometer); thus k(x) = p – ax defines the farm-gate price at any distance x from the port, given price p and fuel cost a. Now assume that exports are taxed at a rate t, such that the export price at the port is p = p* – t, where p* is the world price in domestic currency terms. Assume that production uses only family labor and a locationspecific resource, land, which is uniform in every other respect. a. Demonstrate that the incidence of the tax (as a percentage of the farm gate price) is increasing with respect to distance from the port, and quantify its implications for individual farm profits. b. Clearly there is a distance from the port beyond which profit-maximizing producers will not produce the export crop. Suppose that producers beyond this ‘boundary of cultivation’ instead produce non-traded goods (such as a locally-consumed grain crop), and that land used in production of either crop must be cleared from forest. Develop a concise analysis quantifying the demand for land (i.e., the pressure for deforestation) and its implied price as a function of trade policy and transport costs. c. If property rights in forest to be cleared for agriculture less than fully enforced, and if standing forest has positive social value, what implications do policy reform or exogenous price changes have for the expansion of agriculture, and for aggregate social welfare? d. Based on your answer to parts b and c, provide a policy-oriented commentary on the relationship between poverty, trade and infrastructure policies, and the use of natural resources (here, forests) in poor agrarian economies. Build your analysis, either formally or informally, on the supposition that poverty alleviation is an important goal of development policy. 4. In Charles Perrings's 1989 J. Dev. Econ. paper "An optimal path to extinction", the author presents an analysis from which he argues that in poor agrarian economies a failure of trade entitlements (to use Sen’s (1981) terminology) may lead to resource use decisions that result in accelerated environmental degradation and thus to a failure of production entitlements as well. Specifically, Perrings proposes that the opening of agrarian sub-Saharan African economies to trade will likely result in immerizing growth in agriculture, and further that " ...exchange rate and price fluctuations are, potentially at least, just as damaging as rainfall 4 fluctuations" in exposing agrarian economies to the exigencies of poverty (p. 19). Yet structural adjustment programs adopted (or at least attempted) by most countries of subSaharan Africa consistently include measures that increase the exposure of farmers to trade and to such economic fluctuations. Evaluate the robustness of Perrings' economic and empirical arguments, addressing the question: does economic liberalization cause environmental degradation? Feel free to add evidence, either economic or empirical, drawing on arguments that go beyond the Perrings' model, but use this additional evidence as a complement, not a substitute, in analyzing the impacts of liberalization on degradation. 5. Poverty rates are disproportionately high in rural areas of most world regions. This observation has led some to advocate policies that will improve the access of poor rural households to land. Others argue that such policies will be ineffective. This question asks you to consider the debate between these two positions both theoretically and empirically. a. Define total income for a rural family as: Y P.Q( LQ , A ) w( L f LQf ) wLQh r A , where the agricultural commodity, Q, is produced with a constant returns to scale technology using labor ( LQ ) and land (A). Farm labor is the sum of family and hired labor ( LQ LQf LQh ). To keep things simple, assume that the land to which the family has access is fixed at A . The family has a total labor stock of labor ( L f ) that can be allocated to farm and non-farm activities. Some of the labor allocated offfarm may be used up in searching for jobs, and the number of days employed as a function of time allocated to off-farm activities is ( L f LQF ) , where 0 1; 0; 0 . P is the price of the commodity Q, w is the wage rate, and r is the land rental rate. Assuming that families allocate resources in order to maximize income, under what circumstances will a policy that enhances land access (i.e., increases A ) for a poor family actually make them better off in income terms? When will improved land access have no effect on family income? In answering this question, you should define a poor family as one for whom LQH=0 and LQf< L f . You may find the analysis easier if you note that under constant returns to scale, the income of a poor family can be written as: y P Aq (lq ) w L f lq A r A , where the lower case letters denote values per-unit of land. 5 b. Suppose that you now want to use the model above to empirically determine whether or not policies designed to improve land access are good anti-poverty measures. Using the model, please explain what empirical proposition you would test. c. Finally, assume that you had data on a random sample of rural households (rich and poor, landed and landless). Econometrically, how would you undertake the task of estimating the test statistic you identified in part (b) above. 6