Research
Job Market Paper
Panel VAR Model With Latent Group Structures
Univariate panel models with interactive fixed effects has been well discussed in previous studies. This paper studies the multivariate panel vector autoregression (PVAR) model with group-based factors. It is flexible as the number of groups, the group membership in each group and the number of group factors in each equation are not specified, and it can be extended to group-specific heterogeneous coefficient Panel VAR. Furthermore, it is a parsimonious structural model and easy to compute. We derive the asymptotic distribution and establish consistency of the estimator for N and T that tend to infinity.
Working Papers
Group Patterns in Income Inequality and Economic Growth
The relationship between income inequality and economic growth has been debated for a long time. This article seeks to take into account the latent group patterns across countries using the grouped fixed effects estimate, to examine the growth-inequality nexus. The results of the study indicate that there is a non-linear relationship between inequality and growth. Specifically, our results are consistent with the Kuznets curve which hypothesizes that inequality has a positive impact on growth at low levels of inequality, but a negative impact at high levels of inequality. We also reveal heterogeneity in the response of growth to inequality in the different groups, implying that the impact of inequality on growth varies across countries. The results are robust to the use of two different Gini indexes and to different specifications of the model.Â
Critical Review of Carbon-Emitting Energy as an EKC Regressor: New Evidence from US State-Level Data
The introduction of total fossil fuel or energy consumption as a regressor variable has become increasingly common practice practice in research on the carbon Kuznets curve. Given the way of calculating CO2 emissions, this empirical strategy implies a clear endogeneity problem. We purpose an alternative model to avoid the econometric problem derived from it, which is applied to a panel data set for U.S. state-level. Our empirical results reveals that bias on these parameters of interest may be important enough to avoid this mistaken empirical strategy.