Growth and Development [SP2017]
This site contains relevant guidelines for Growth and Development, Spring 2017. Students are asked to regularly check this course website for updates and most recent information. This site will grow with the semester and it will contain a section with a brief description of what we have learned in previous classes and what we will cover in the next one.
We meet Wednesdays 15:00-18:00 in the seminar room. The GnD workshop runs Fridays at 9am (specific dates below).
This graduate course covers advanced topics in growth and development economics. We will integrate micro data and tools and quantitative macro theory to study aggregate economic development and evaluate policy.
- First, we will discuss development facts from cross-sectional and panel household-level data from developing countries. Our unit of observation will be individuals and households. We will look at the behavior of consumption and
income, and whenever possible of wealth and land. We will also review some macro development facts from standard aggregate data (e.g. PWT, WDI, etc.) and discuss their relation to our micro facts.
- Second, we will introduce consumption theory (intertemporal choice and PILCH models with borrowing and savings constraints) and discuss its applications to development. Then, we will use theory to study consumption risk and insurance in village economies formalizing tests of market (in)completeness; this will include state-of-the-art estimations of the degree of partial insurance. Further, we will study models where incomplete markets arise endogenously (e.g. imperfect information and imperfect enforceability) and for which we can derive additional insurance tests that we will also take to the data. Here, we will pay particular attention to Sub-Saharan African countries and China. Finally, we will move beyond villages and discuss potential routes to link individual risk and insurance to aggregate development. This course is very much engineered to get this aggregation right. To do so, we will present quantitative macro models with heterogeneous agents and discuss their solution under individual/aggregate risk in non-stationary development processes. Specifically, we will analyze how to embed heterogeneity into aggregate models of structural transformation.
- Third, we will study a set of advanced topics in development: (a) health behavior and risk to understand the effects of disease (e.g. HIV, malaria) and assess the role of prevention strategies and palliative policies; (b) data and theory behind the relationship between fertility and income/land in developing countries; (c) missallocation issues arising from regulations on firms, land policies and gender-biased social norms; (d) models that explain low rates of new technology adoption (e.g. use of fertilizers) in some developing countries; and (e) measurement and use of subjective expectations on returns to schooling (and aspirations) in developing countries.
- Finally, we will discuss how to conduct field experiments. Here, our goal is to find useful ways in which we can integrate field experiments and structural modeling to exploit their complementarities when conducting policy evaluations.
Syllabus here [.pdf].
It helps if you have previous programming experience but it is not a pre-requisite. The programming language you use is at your discretion. STATA is a good application that allows you to upload and manipulate many large data sets at once very useful to 'fish for facts' or conduct your own research. If you are planning to do serious computational work in your research, I encourage you to learn FORTRAN (good alternatives are C or C++). This requires an initial fixed cost but it usually pays off. MATLAB is more user-friendly than Fortran and very popular in Economics. It is particularly useful if you are used to think in vector-matrix operations. FORTRAN and MATLAB are both good alternatives to solve Aiyagari-Bewley-Hugget economies. You have all these packages in the graduate lab.
Grades and Requirements:
The grade will be mainly based on one project to be delivered the last week of this course. Occasionally, you may have additional homework exercises. You will have to present advances of your projects regularly in class.
- Data: Cross-Sectional and Panel Facts from Household-Level Surveys in Developing Countries.
Consumption: Intertemporal Choice
- Measurement Issues.
- Tracking Lifecycle Behavior.
- Estimating income processes for developing countries.
Risk and Insurance in Village Economies
- PILCH models with credit and savings constraints.
- Aiyagari-Bewley-Hugget Economies.
Beyond Villages: Consumption Insurance and Economic Growth: Heterogenous Agents Structural Transformation (HAST) Economies.
- Complete markets tests and the measurement of the degree of partial insurance.
- Endogenously incomplete markets models: private information and limited commitment.
- Other aspects of risk and insurance: Formal vs. informal insurance; migration risk; off-village marriage; the trade-off between private vs. public insurance.
Integrating Field Experiments and Quantitative Macro Models to Evaluate Policy
- Health Behavior and Risk
- Fertility and Investment in Children
- Misallocation: Firms, Land and Gender-Biased Social Norms
- Politico-Economic Equilibria, Vested Interests, and Technology Adoption
- Wed Jan 11: We went over the course content and the rules of the game. Our helicopter view of the course got us into discussing some mesurement issues and why inequality, which is potentially the result of ex-ante and ex-post redistribution mechanisms, can matter for economic growth. We briefly introduced the concepts of consumption insurance, misallocation, and structural change.
- Wed Jan 18: (1h30min) We started to look into micro data. We introduced several tools that help track lifecycle behavior. The identification issue relating time, cohort and age affects was discussed. Welfare is individual, and we discussed how to compute adult equivalent measures of consumption. We showed the role of household structure in explaing lifecycle humps of consumption expenditure. We discussed empirical differences in the ability to smooth consumption over the lifecycle in rural and urban areas of poor countries. Last, we showed how inequality measures that use standard adult-equivalent consumption expenditure differ from inequality measures computed from actual individual data using the China Health and Nutrition Survey. [Slides: Tracking Lifecycle Behavior]
- Wed Jan 25: We went over LSMS data and some measurement issues related to consumption, income and wealth [Slides: Household Survey Data & Measurement Issues]. Then we discussed the estimation of income processes that separate transitory shocks from permanent shocks and contextualize it for poor countries. [Slides: Estimating Income Processes]
- Fri Jan 27 (9-10.30am): GnD Workshop. We discussed HWK1. Hadar, Aldo and Lluis presented work on the electrification of rural areas in Kenya. Anu presented the neighborhood peer effects in secondary school enrolment decisions by Bobonis and Finan. Viviana presented the misallocation in manufacturing in China and India paper by Hsieh and Klenow (2009).
- Wed Feb 1: (1h30min) We completed the discussion on the estimation of income processes in poor countries. We discussed the case of superior information, and also selection issues.
- Wed Feb 8: We went over the complete markets theory. We derived some testable implications of this theory a la Townsend. We discussed how to empirically conduct full-risk sharing tests using panel data. This included mean-based tests and cross-sectional variance-based tests. We also introduced a synthetic-cohorts tool to test for complete markets in the case where panel data is absent and panels of cross-sectional data are available (a la Attanasio and Davis). [Slides: Consumption Insurance: Some Complete Markets Tests]
- Fri Feb 10 (9-10.30am): GnD Workshop. We finished the paper presentations.
- Wed Feb 15: We covered a wide range of computational methods useful to solve economic models.
[Slides: Function Approximation]
[Slides: Numerical Integration and Differentiation]
[Slides: Nonlinear Sytems]
[Slides: Numerical Optimization]
[Slides: Value Function Methods]
- Fri Feb 17 (9-10.30am): GnD Workshop. Bruno is after the effects of the 2004 tsunamis on consumption insurance. Pablo discussed the CAP and "how not to" spend EU resources. Ana Karen introduced Mexican data to discuss the impact of welfare programs on income mobility.
- Wed Feb 22: We related aggregate productivity (TFP) to resource allocation. We started with the work of Hsieh and Klenow 2009 in manufacturing China and India. We then moved to land misallocation in poor Africa. We showed empirical evidence of land misallocation using the relation between land and individual farm productivity. We then quantitatively assessed the degree of land misallocation using a Lucas span of control model. Land reallocation raises aggregate agricultural productivity (and output) by 260 per cent. We are able to directly link those gains to access to land markets. We also discussed the effects of such gain in aggregate agricultural productivity on structural change. Then we discuss the potential broader implications for income inequality. We discussed some insights on selection and dynamic misallocation. [Slides: Misallocation and Productivity]
- Fri Feb 24 (9-10.30am): GnD Workshop. We continued discussing your research projects. We covered a wide range of research questions and topics. There is some misconception on how to assess policy in your projects. We will fix this next Wed.
- Wed March 1: Sven presented the resource misallocation HWK. Tomasso and Thomas went over the HWK related to the computation of a structural change economy with nonhomothetic preferences (i.e., an economy in transition). Then we introduced economies in which agents differ in some form of idyosincratic risk and the only instrument available to hedge against risk is individual savings, i.e., self-insurance. These are standard incomplete markets (SIM) economies, aka PILCH economies, or Ayagari-Bewley-Hugget economies. We discussed in detail how consumption insurance and aggregate capital are related in this economy through the notion of precautionary savings. Then we went over useful computational algorithms that solve this economy in general equilibrium. The main issue is that to forecast future rental prices of capital we need to keep track of the distribution of wealth (i.e., the aggregate state of the economy is a large object). [Slides: PILCH Models in General Equilibrium] Then we discussed how to assess policy change in this framework. This potentially includes changes in the tax code, social security, introduction of fertility policies, credit constraints, investment loans, etc. This assessment implies that we need to do welfare comparisons that require the computation of the entire transition of the economy from one state of the world (policy 0) to a new one (policy 1). Finally we discussed consumption equivalent variation to assess welfare implications of policy changes. [Slides: Policy Evaluation]
- Fri March 3 (9-10.30am): The outcome of socio-political processes can be growth-enhancing policies or growth-preventing processes. To study this we introduced the notion of dynamic politico-economic equilibria to formalize a model of technology adoption. In this setting, vested interests can generate stagnation and growth. For example, old agents with acquired skills on vintages of technology currently in production use might want to block innovation (new technology vintages) that might favors young entrepeneurial agents or unskilled workers (Krusell-Rios-Rull 1996). The complexity of these equilibria is that it is necessary for an agent to contemplate different current policy options to think not only about their respective effect on current prices and transfers, but on future prices, transfers, and politically determined policies as well. The politico-economic EQ requires two steps. First, find a a competitive EQ given a law of motion for policies. Second, find a political EQ that makes the law of motion of policies consistent with that comint out of the (endogenous) political process (Krusell-Quadrini-Rios-Rull-1997). We discussed how this relates to the literature on optimal taxation policy and time consistent policy. [Slides: Politico-Economic Equilibrium, Vested Interests, and Economic Growth]
- Fri March 10 (9-10.30am): The GnD shop opened. Guillermo and Niccolo want to add fertility to a structural transformation model to better replicate the data. Hadar, Mariam and Lluis discussed some ingredients for a model of sex selection that can potentially help explain the sex ratio imbalance in India. Pablo showed us some new aggregation on the CAP transfers. Vivian clarified her misallocation in clusters exercise using Mexican data.
- Wed March 15: Make-Up class. We will cover ex-post redistribution mechanisms.
- [Homework 1] Due Fri Jan 27 at 9am. Lifecycle behavior of consumption and income in poor countries.
- [Homework 2] Due Wed Feb 8 at 1pm. Estimating income processes in rural and urban areas.
- [Homework 3] Due Wed Feb 15 at 1pm. Testing Consumption insurance.
- [Homework 4] Due Wed Feb 22 at 1pm. Transitional dynamics and structural transformation. The Gollin-Parente-Rogerson economy.
- [Homework 5] Due Wed March 1 at 1pm. Explaining low TFP: A story based on resource misallocation.
- [Homework 6] Due Wed March 15 at 1pm. A Roy-Harrod-Todaro migration economy of 2 periods and a fully-fledged HAST economy.
- [Homework 7] Due Wed March 15 at 1pm. Politico-economic equilibria in a HAST economy.