Economics - Ph.D. / Sc.D.
https://hdl.handle.net/1721.1/7659
Sun, 31 May 2020 14:31:10 GMT2020-05-31T14:31:10ZGeometric methods in econometrics and statistics
https://hdl.handle.net/1721.1/124058
Geometric methods in econometrics and statistics
Mukhin, Yaroslav V.(Yaroslav Vadimovich)
Econometrics and statistics rely on asymptotic approximations to construct hypothesis tests and confidence regions. Asymptotic approximations can also be used more abstractly to study the quality (efficiency) of estimators and tests. These approximations are closely related to local (differential) properties of the functionals of the statistical model whose values are being estimated and tested. I consider statistical models and estimands motivated by economic theory and applications and study their local and also global properties: I study the local properties of functionals to characterize the efficiency bounds of their estimators and the directions of most rapid (gradient) change with respect to different metrics of distance on the model. I use gradient flows to describe global evolutions on the statistical model governed by changes in a scalar functional. These flows can be used to describe economic policy and to study structural estimators motivated by economic theory.
Thesis: Ph. D. in Economics and Statistics, Massachusetts Institute of Technology, Department of Economics, 2019; Cataloged from PDF version of thesis.; Includes bibliographical references (pages 143-150).
Tue, 01 Jan 2019 00:00:00 GMThttps://hdl.handle.net/1721.1/1240582019-01-01T00:00:00ZEssays in network economics
https://hdl.handle.net/1721.1/124057
Essays in network economics
Azar, Pablo Daniel.
This thesis is a collection of three chapters, each representing an individual paper. The first chapter studies how the formation of supply chains affects economic growth. It provides a new tractable model for supply chain formation. The main innovation in this model is that, firms can choose suppliers to maximize profits. Individual firms' actions determine the equilibrium input-output network, and affect macroeconomic variables such as GDP. We then apply this model to understand the effect of changing supply chains on American productivity during the 1987-2007 period. The second chapter studies how a monopolist may sell multiple goods to strategic bidders. The monopolist may face a series of combinatorial constraints. For example, it may be forced to allocate at most one good to each bidder, and it may have additional constraints on which bidders can be allocated which goods. Furthermore, the monopolist does not know bidders' demand distributions. Rather, it only knows one sample from the demand distribution corresponding to each bidder. Nevertheless, by developing new online optimization algorithms, we show how simple mechanisms can approximate the monopolist's optimal revenue. Finally, the third chapter, develops a new model of firm optimization to understand how shrinking electronics have contributed to increased productivity and welfare in the United States, during the 2002-2017 period. In this model, firms face constraints on the size of the products they can build. As intermediate inputs, such as electronics, shrink, the firms' production possibilities frontier expands, and GDP increases.
Thesis: Ph. D., Massachusetts Institute of Technology, Department of Economics, 2019; Cataloged from PDF version of thesis.; Includes bibliographical references.
Tue, 01 Jan 2019 00:00:00 GMThttps://hdl.handle.net/1721.1/1240572019-01-01T00:00:00ZEssays on the financial system of developing economies
https://hdl.handle.net/1721.1/122543
Essays on the financial system of developing economies
Sripakdeevong, Parit.
The thesis consists of three chapters, which explore the financial system of rural Thai Villages through the lens of the Townsend Thai Monthly Survey. The first two chapters examine savings behaviour, while the last chapter investigates borrowing. In the first chapter, I apply the estimation technique of Deaton and Paxson (2000) to the monthly version of the Townsend Thai Household Survey and find that individual savings rate decreases with individual age. This result contrasts the flat savings profile found when estimation is conducted at the household level. I further extend the Deaton-Paxson technique to deal with autocorrelation in panel data, and noisy estimates due to small sample size. Applying the appropriate corrections generates the inverse U shape savings rate profile predicted by the life-cycle model.; In the second chapter, I test the model of Amador, Werning, and Angeletos (2006) empirically against the data of the Townsend Thai Monthly Survey supplemented by Christopher Woodruff's behavioural survey. From the data, I estimate the model's inputs: the household's income distribution, level of risk aversion, and hyperbolic discount rate. The model identifies the subgroup of households where minimum savings policy is optimal and predicts the optimal minimum savings value. For this subgroup, I find that the predicted minimum saving values have a strong positive correlation with the actual amount saved by the household. The relationship is weaker for households outside the subgroup, as the minimum saving policy is no longer optimal. The third chapter, joint with Robert Townsend, documents the existence of active, high volume and relatively sophisticated money markets in villages in Thailand. As with traditional markets, loan repayment can be deferred through standard restructuring.; But there are also more complicated credit refinancing chains involving multiple parties and short/medium maturities. From risk sharing regressions, we find that borrowers surprisingly have higher income coefficient than non-borrowers. However, this is not because financial access is detrimental, but instead due to the self-selection into debt of more risk-tolerant individuals. Yet, those engaged in credit refinancing chains have the smoothest consumption against income shocks of all, as risk tolerance is dominated by low transaction and verification costs.
Thesis: Ph. D., Massachusetts Institute of Technology, Department of Economics, 2018; Cataloged from PDF version of thesis.; Includes bibliographical references (pages 94-97).
Mon, 01 Jan 2018 00:00:00 GMThttps://hdl.handle.net/1721.1/1225432018-01-01T00:00:00ZEssays in econometrics and machine learning
https://hdl.handle.net/1721.1/122542
Essays in econometrics and machine learning
Semenova, Vira.
Establishing the link between a cause and effect is a fundamental question in social science. Standard assumptions about human behavior (e.g., rationality) imply restrictions on the plausible values of the causal effect. In addition to this effect, these restrictions may depend on additional summaries of human behavior. Estimation of these additional parameters presents a trade-off between capturing the complexity of human's decision-making yet constraining it to deliver precise estimates. I resolve this tension by incorporating modern machine learning tools into the estimation of the additional parameters and deliver high-quality estimates of the causal effect and counterfactual outcomes. I estimate the causal effect in a two-stage procedure. At the first stage, I estimate the additional summaries of human behavior by modern machine learning tools. At the second stage, I plug the first-stage output into the sample analog of the restriction that identifies the causal effect. I modify the second-stage restriction to make it insensitive to any regularization biases present in the first-stage components. The second-stage estimate of the causal effect is of high-quality: it converges at fastest rate and can be used to test the hypotheses and build the confidence intervals for the values of the causal effect. I apply this idea in a wide class of economic models, including dynamic games of imperfect information, treatment effect in the presence of endogenous sample selection, and reduced-form demand estimation.
Thesis: Ph. D., Massachusetts Institute of Technology, Department of Economics, 2018; Cataloged from PDF version of thesis.; Includes bibliographical references (pages 209-213).
Mon, 01 Jan 2018 00:00:00 GMThttps://hdl.handle.net/1721.1/1225422018-01-01T00:00:00Z