University at Albany
 

Ph.D. Job Market Candidates

Placement Director:
Professor Kajal Lahiri
E-mail: klahiri@albany.edu
Phone: (518) 442-4735

Job Market Candidates (2019-2020)

Jihye Kim        CV | E-mail | Homepage



Primary Fields: Health Economics
Secondary Fields: Health Policy, Applied Econometrics
Job Market Paper: Educational Attainment and Health-Adjusted Life Expectancy Among White and Black US Adults and the Elderly, 1998-2016


References: Kajal Lahiri (Advisor), Pinka Chatterji, Chun-Yu Ho



Fangning Li        CV | E-mail | Homepage



Primary Fields: Econometrics
Secondary Fields: Applied Econometrics, Macroeconomics
Job Market Paper: Consistent Model Averaging Using Elastic-Net


References: Zhongwen Liang, Daiqiang Zhang, Ulrich Hounyo



Yichuan Wang        CV | E-mail

Primary Fields: Industrial Organization
Secondary Fields: Health Economics, Applied Econometrics
Job Market Paper: Mergers and Acquisitions in Medicare Part D: Assessing Market Power and Consumer Reflect


References: Pinka Chatterji , Chun-Yu Ho



Cheng Yang        CV | E-mail | Homepage



Primary Fields: Econometrics, Forecasting
Secondary Fields: Applied Microeconomics
Job Market Paper: The Role of Model Averaging, MIDAS, Boosting, and Bootstrap in Forecasting State Tax Revenues and Their Uncertainty

References: Kajal Lahiri (Advisor), Ulrich Hounyo, Zhongwen Liang



Xiaoqi Zhu        CV | E-mail

Primary Fields: Applied Econometrics
Secondary Fields: Health Economics
Job Market Paper: Medicaid Asset Test Elimination and Household Savings


References: Pinka Chatterji (Advisor), Kajal Lahiri, Yue Li








Job Market Papers (2019-2020)

Jihye Kim

Title: Educational Attainment and Health-Adjusted Life Expectancy Among White and Black US Adults and the Elderly, 1998-2016
Abstract: Population aging in the United States underscores the importance of quality of life as well as longevity. The health-adjusted life expectancy (HALE) incorporates both mortality and morbidity information by combining disease prevalence with mortality data from life tables to represent general health status of a population as a single index. HALE is interpreted as the average number of healthy years an individual in a population could expect to live given the mortality and general health status of that population in each year. In this paper, I estimate the HALE of people over age 55 in the United States by gender, race, and education levels over four periods (1998-2000, 2002-2006, 2008-2012 and 2014-2016). To construct a comprehensive measure of morbidity status of each sub-population, I use 17 doctor-diagnosed diseases and 4 self-rated health conditions from the Health and Retirement Study panel with comorbidity adjustments. I show that higher educational attainment was strongly associated with longer expected healthy years of life in general. Women with middle- and high-educational attainments had higher HALE, with strongly diminishing trends. Noticeably, white men with high-education had especially high HALE with no diminishing trends. Black men had generally lower HALE than other race and gender groups, especially among those with low-education. The findings confirm that educational attainment is the major factor to HALE and disparities in HALE were observed in within and between race and gender groups.


Fangning Li

Title: Consistent Model Averaging Using Elastic-Net
Abstract: This paper proposes a new model averaging method in the linear model setup named consistent model averaging (CMA) based on Tikhonov regularization. The CMA estimator is consistent in the sense that it converges to the infeasible optimal model weight that minimizes conditional risk in finite sample. Given the number of regressors p, first we show that ideally model averaging over the maximal collection of 2p models is equivalent to averaging over a subcollection of only singleton and pairwise models in the sense of achieving the same minimum risk, which reduces computational burden substantially. Then we propose the CMA estimator based on Tikhonov penalty. The Tikhonov penalty turns out to be essential for the consistency of the CMA estimator. We derived the √n-consistency and asymptotic normality of the CMA estimator in fixed-p case, as well as its deterministic L2 error bound when p diverges with sample size n. Interestingly the CMA estimated model weight can be interpreted as probability amplitude. An additional elastic-net penalty is motivated in CMA estimation to stabilize solution and encourage sparsity. Further issues such as heteroscedasticity and sparse coefficients are addressed, so that CMA can handle heteroscedastic errors and cooperate nicely with variable selection procedures such as lasso and SIS. Simulation results show that CMA with elastic-net penalty performs better than the original elastic-net estimator and Mallows' model averaging estimator when population R2 is moderate. We also illustrate the better performance of CMA with an application of predicting wages.


Yichuan Wang

Title: Mergers and Acquisitions in Medicare Part D: Assessing Market Power and Consumer Reflect
Abstract: I examine horizontal mergers amongst Part D insurers with the aim of assessing how the mergers and acquisitions which causing falling numbers of Part D plan related to premium, market share and plan’s characteristics. I applied differences-in-differences identification strategy to panel data on plans offered between 2007 and 2019 to document the effects of mergers. The results reveal that stronger market power as mergers cause premiums and market share to rise in national market. But premiums fall for merging insurers that restructure plans and renegotiate contracts with drug suppliers by consolidating existing plans. The results also reveal significant market power raising plan’s aggregate market share.


Cheng Yang

Title: The Role of Model Averaging, MIDAS, Boosting, and Bootstrap in Forecasting State Tax Revenues and Their Uncertainty
Abstract: In recent years models with mixed frequency have been extensively used to forecast low-frequency variables such as GDP and inflation, but we are the first to use this framework in state government revenue forecasting. New York State has a notorious record of passing late budgets. In order to facilitate budget negotiations, which often center on forecasts, we develop a Mixed-Data Sampling (MIDAS) model for revenue forecasting using jagged edge data sets. Since recessions generate big errors in revenue forecasts, we pay special attention to incorporate leading indicators for recessions in tax revenue forecasts. Thus we forecast yearly tax revenues using monthly data on tax receipts and also two dynamic factors extracted from a set of selected monthly and quarterly indicators specific to the New York State and the U.S. economy separately. These three models are combined with optimal weights to generate monthly multi-period forecasts. The weights of the two dynamic factors are high at horizons more than 11 months, after which the monthly tax revenue variable picks up in its contribution as uncertainty is resolved. Additionally, by combining we gain forecast efficiency at all horizons. Our sample covers fiscal years 1986-2020; and data till 2007 is used in estimation to generate and evaluate out-of-sample forecasts over 2008-2018. To coincide with the budget process, our forecasts start 18 months, and are continuously updated monthly till the end of the fiscal year. Our model allows for identification of reasons for forecast revisions as new information arrives on a monthly basis in a transparent manner. We document significant gains in forecast accuracy. The relative gain in forecasting efficiency is particularly significant during the cyclical downturns. Counterfactual analysis is implemented to study the marginal contribution of data at each horizon. We estimate the variances of combined out-of-sample forecasts with blocking-based residual bootstrap methodology. With the variance estimates, we provide fan charts for each fixed target fiscal year showing the underlying forecast uncertainty. As an extension, we integrate boosting with factor-augmented MIDAS (Boosting-FA-MIDAS) and evaluate its forecast performance. The results show boosting also improves upon individual MIDAS models but only outperforms forecast combination at a few medium horizons.

Xiaoqi Zhu

Title: Medicaid Asset Test Elimination and Household Savings
Abstract: This paper assesses the effect of eliminating Medicaid asset tests for low-income families on household saving behavior. I use Difference-in-Difference method and data on low-educated household heads with children from the Survey of Income and Program Participation (SIPP) to estimate the elimination effects. The identification strategy exploits the exogenous variation in the timing of asset tests removal among states. The findings indicate a strong negative impact of eliminating asset tests on household holdings of liquid assets, and the effects are greater for states with stricter asset limit levels before the elimination. This implies that while Medicaid is indubitably essential for the poor, the role that Medicaid plays for those relatively rich may be crucial as well through alleviating the need for precautionary savings. This study finds no evidence of the elimination effect on non-liquid assets or household net worth (excluding the value of primary residence and first car).