Works in Progress:
- "Information - Theoretic Model Selection and Estimation for Interval Data" with Amos Golan and Aman Ullah [Job Market Paper]
We develop an efficient information-theoretic estimator for analyzing interval-valued, and symbolic data. Rather than applying the traditional least squares or likelihood methods to estimate some moments of the intervals (as often done), we use the complete information in the sample and identify the best model (parameters) that are consistent with the data generating process. It is an iterative approach. In addition, we impose minimal structure and statistical assumptions. We provide a large number of sampling experiments as well as a few empirical examples.
- "Transfers, Health and Income Shocks within Social Networks"
The 2004 Kagera Health and Development Survey(KHDS 2004) from Tanzania is used to investigate whether gifts and transfers serve as insurance against health and income shocks and whether they have a systemic redistributive component: transfers flow from wealthy to poor, or vice-versa. Since there are huge discrepancies between gifts and transfers reported by receiving and giving households, we use interval-valued data estimation method developed by Golan, Tuang and Ullah (2015) by incorporating intervals of all observed or reported information. We have not found evidence that suggests risk-sharing against health shocks, but there are suggestive findings of altruistic or social norms, helping among family networks of households. Geographical distance does not play an important role while social proximity measures such as kinships and sharing the same religion does play a role in determining gifts and transfers within family networks.
- "Interval Capital Asset Pricing Model"
We reexamine the well-known Fama and French's (FF) three-factor (2003) and five-factor (2013) Capital Asset Pricing Model (CAPM) as an interval-valued CAPM. We apply the IT-GME (Golan, Tuang and Ullah - 2015) on the monthly data of FF's original 25 portfolios from July 1963 to February 2014. From the symbolic data analysis literature, the Fama and French model of excess stock returns can be viewed as a "range" model of the interval-valued CAPM that suffers from the loss of "level" information. By employing the complete information in the interval-valued CAPM, we can capture the variability of the portfolio returns over time and produce more efficient parameter estimates. The approach highlights the benefits of applying interval estimation even when the main interest is in the "range" or "difference" modeling.