2019
On May 8 2019, several economics student presented their thesis at the Meeting of the Minds. Below is a summary of their work:
In his thesis, Eric Huang (Advisor: Onur Kesten) analyzed the mechanism for allocating housing to CMU undergraduate students. Eric used data from 2014 to 2018 collected by CMU’s housing services. The model constructed, allowed Eric to consider several alternative policy simulations. Of interest, was the removal of the retention phase (an initial time-period where students can elect to retain their current housing choice). In his work, Eric found that removing an initial retention phase, while it increases fairness, it reduces overall welfare.
In her research, Amelia Gilson (Advisor: Onur Kesten) focused on college acceptance decisions. In particular, she analyzed the role that changing demographics in the US have on college admission yields. Amelia constructed her own panel-data from multiple sources spanning a total of 225 universities. An interesting finding of her work is that, surprisingly, a higher fraction of female student, leads to lower acceptance yields. Several hypotheses where analyzed including the possibility that high-yield universities admit fewer females or that female students apply to a greater number of universities.
Jules Ross (Advisor: Nicholas Muller) studied the impact of state and local subsidies on solar panel installations. This is a key policy adopted by countries to meet greenhouse gas emissions reduction goals. Jules first compiled a dataset documenting the variation across states of policy subsidizing solar panel installation. As a following step, she then constructed several statistical models relating policy (both state and local) to solar panel installation rates. A key finding of this research is that policy interaction at state and local level can be significant to explain the observed pattern of solar panel adoption.
In his thesis Adam Tucker (Advisor: Onur Kesten) focused on forecasting. The goal of his research is to compare the performance of classical forecasting models with recent machine learning methodologies. The context for his analysis are business cycles for the US economy. Adam key finding is that, when trying to forecast a recession, a classical econometric tool such as a probit, outperforms more sophisticated methods. This result highlights the limited applicability of machine learning methods to time-series data where data is unavailable in large amount.
For this presentation Adam won the Statistics Oral Presentation Competition
Jiyoung Kim (Advisor: Ali Shourideh) centered her research on study of income inequality. Specifically, her research asked if geography matters in generating unequal labor-market outcomes. Jiyoung constructed a dynamic discrete-choice model to determine the cost that a worker has when moving across locations in the US. The idea is that if costs are high, a worker in a low-wage location cannot take advantage of high wages elsewhere in the country. The first, result is that moving cost are significant, especially for less educated workers. Second, Jiyoung found that moving costs have explanatory power in explaining wages and hence do indeed contribute to overall wage inequality.
For this work Jiyoung was awarded the Best Senior Honors Thesis from the Undergraduate Economics Program