|
Faculty Affiliate Working Papers
Abstract: Recent available trade data with related- and non-related party breakdown show that US multinational enterprises (MNEs) use different trading strategies in the China region relative to other countries. US trade with the China region in 2002-2007 is characterized more by arm’s-length (or non-affiliated) transactions. The paper also provides a baseline understanding of US states’ engagement with the China region. Three patterns stand out. First, there is great variability in current engagement via trade: exports to the region range from 1% to 28% of states’ exports. Second, compared to exports to other countries, exports to the region are highly concentrated. At the extreme, for some states, 96-98% of their exports are computer and electronic products. Third, gravity regressions show that exports to Hong Kong are positively correlated with the relative size of the Hong Kong-born population in the states. Also, there is no evidence that stricter labor regimes necessarily leads to lower state exports. An investigation of its direction and some implications for outward investment, August 2009. Co-authored with Sumon Kumar Bhaumik Abstract: The Chinese state undertakes large scale investments in a number of countries under the auspices of economic cooperation related investment (ECI). While there are suggestions that it is an extension of China’s soft power aimed at facilitating Chinese FDI in those countries, often for access to natural resources, there is no systematic analysis of this in the literature. In this paper, we examine this investment of the Chinese state over time. Our results suggest that the pattern of investment is indeed explained well by factors that are used in the stylised literature to explain directional patterns of outward FDI. They also demonstrate that the (positive) relationship between Chinese ECI and the recipient countries’ natural resource richness is not economically meaningful. Finally, while there is some support for the popular wisdom that China’s willingness to do business with a country is not strongly affected by its level of corruption, there is much weaker support, if any, for the hypotheses that China favours doing business with countries where political rights are limited. Co-authored with Wei Shi and Carol Wang Abstract: Chinese commercial banks have experienced tremendous growth over the past decade but have received limited academic attention due to data collection difficulty. We’ve successfully compiled a hand-collected panel dataset of Chinese commercial banks governance characteristics from 1998 to 2007. We empirically examine the relation between board governance and the profitability of Chinese commercial banks. We find that board governance has significant impact on Chinese banks’ performance. Specifically, higher board ownership, lower percentage of insiders on board, and lower block ownership are associated with better bank performance. In addition, to improve bank performance, Chinese bank managers should also focus on effectively control of bank’s operating cost, increasing net interest margin, and closely monitoring loan productivity. This is the first study conducted on the efficacy of Chinese banks’ governance system and its relation with banks’ profitability. Empirical evidence from this study has important policy implications in reforming China’s banking system into a more transparent and more efficient market driven system. Co-authored with Guang Li Nie Abstract: As of 2008, 132 million credit cards have been issued in china. But many card holders are not active and are called “churn holders”. Research suggests that a bank can increase its profits tremendously even with a small improvement in its retention rate. In this paper, we apply two data mining algorithms to build churn prediction model using proprietary credit card data from a large Chinese bank. Four types of variables are examined which include customer information, credit card information, risk profiles, and transaction activities. Instead of putting the 135 variables into the model directly, we select the variables considering not only its correlation but also its economic sense. We design a misclassification cost measurement formula taking the two types error and the economic cost into account to evaluate the model. The selected variables suggest that the demographic information makes little contribution to the churn prediction. Both card information and transaction activities which reflect card holder behavior work very well in the model. Our result shows that regression performs better than decision tree. However, decision tree based models provide rules that can guide banks to make marketing strategies. Banks may adopt both methods for analysis to retain their credit card customers. |
Copyright 2011. College of Business Administration. Contact webmaster.