The 65th BNUBS Master Forum Was Held Successfully
Time :2018-03-27

    On the afternoon of December 26, 2017, the 65th BNUBS Master Forum was successfully held in room 1620 of rear main building. Professor Qiao Duo from the Department of Economics of the School of Asian and African Studies at the University of London who was invited by the forum delivered a speech titled “Logically Misleading of Endogeneity Bias — A Methodological Reflection on the Application of Measurement Models”. The lecture was chaired by Prof. Li Shi from BNUBS.

    Being a professor of economics at the School of Asian and African Studies, University of London, Qin Duo is also a Ph.D. in Economics from the University of Oxford's Nuffield College. She was an economic expert in the Economic Research Department of the Asian Development Bank, a senior quantitative analysis expert at State Street Bank, and economic advisor and expert of Oxford Finance Group and Oxford Economic Forecasting. With the main research directions of macroeconomics and econometric development of emerging markets, empirical finance, and international economics, she has published several monographs and many articles in the Journal of Banking and Finance, Econometric Theory, and Journal of Development Economics. She won the 2014 Sun Yefang Financial Innovation Award. 

    At the beginning of the lecture, Professor Qin Duo introduced the reason of endogeneity, that is, to explain that variables are related to the random error term. The sources are mainly synchronic deviation (SB), measurement error (ME), omission-related variable (OVB) and self-selection bias (SSB) and other factors. She pointed out that currently, the mainstream approach to solve the endogeneity of the academic community is to find tool variables. But due to the great randomness of instrumental variables, the introduction of instrumental variables is likely to change the research problems, and the error between the estimated coefficient and the original variable is very large, leading to incorrect interpretation. Then Professor Qin Duo shared the processing method of micro data with everyone. For the problem of non-stationary data missing, it cannot be directly replaced with zero during processing. If the data can be fitted out, then it can be used, otherwise the missing value cannot be used. In the end, Professor Qin Duo also shared machine learning-related methods and techniques with everyone, and everyone benefited a lot.

    After the speech, teachers and students actively asked questions about the research methods and conclusions of the thesis, and expressed their opinions. Professor Qin Duo answered all questions and put forward her own views. The lecture ended in a warm applause.