Time-Series Support for Balassa's Productivity-Bias Hypothesis: Evidence from Korea

In this paper we use the Johansen and Juselius cointegration technique and quarterly data over the period 1979–1993 to test the productivity-bias hypothesis between Korea and four of its major trading partners (Germany, Japan, the United Kingdom, and the United States). The results show that in all four cases the deviation of purchasing-power parity (PPP) from the equilibrium exchange rate has a long-run relationship with the productivity ratios, supporting the notion that as Korea becomes relatively more productive, the Korean won appreciates in real terms.     

Fixed-to-mobile call substitution and telephony market definition in Korea

An appropriate market definition is critical in most antitrust cases. In practice, antitrust authorities define economic markets in a deterministic manner with little concern about the risk involved in defining markets incorrectly. In contrast, this article proposes a probabilistic market definition method by which antitrust authorities can establish a statistical confidence level for their intended market-definition judgments. As an application, we examine the likelihood that the fixed-line and mobile telephony services in Korea can compete in the same economic market. Combining critical loss analysis with a hierarchical Bayes model for stated preference data, we find some evidence for the separation of the fixed-line and mobile telephony markets in present-day Korea. After discussing certain possible regulation biases for market definition, we predict that the two markets will converge in the near future as the mobile price premium continues to decrease.     

A micro,econometric investigation of choice of technology

This paper reports on a micro study of the determinants of technological choice utilizing a sample of observations at the man-machine level from operating plants. The large sample size permits econometric modelling - based on engineering and process analytic principles - of the production relationships involved. It also allows statistical testing of hypotheses. Among the determinants analyzed in the context of differentiated inputs and outputs are policies affecting prices on the factor and product markets. The particular case examined is the choice between imported and locally made looms for cotton textile weaving in Korea. However, methodological concerns are given equal weight in the discussion.     

Long-Run Elasticities of the Demand for Money in Korea:Evidence from Cointegration Analysis

Most recent studies have employed the cointegration technique to investigate the long-run stability of the demand for money. This study considers the case of Korea. It is shown that in the long-run while Ml monetary aggregate is cointegrated with income, interest rate, and the exchange rate, M2 is not. However, results from error correction models reveal that both Ml and M2 have short-run relationship with their determinants. [E41]     

Asymmetric Effects of Money on Inflation: Evidence from Korean Data

New Keynesian macroeconomic models predict that monetary shocks have asymmetric effects on inflation, i.e., inflation responds more to positive monetary shocks than to negative shocks (Asymmetry Hypothesis A), and that the asymmetry is positively associated with movements in average inflation (Asymmetry Hypothesis B). Korean data are shown to provide little support for Asymmetry Hypothesis A and strong support for Asymmetry Hypothesis B. [O53, E12]     

Does Reputation Contribute to Reducing Organizational Errors? A Learning Approach

In this study I examine the effect of a firm's reputation for product quality on its effort in learning to reduce its product defect rate. Theoretical ideas on the motivation of learning associated with social aspiration levels and the self-serving bias combined with social categorization suggest that poor quality reputation firms are more likely than their counterparts with a good reputation to attend to potential product defects and consequently reduce their defect rate. However, a stream of research on the motivation of learning stemming from historical aspiration levels and slack search leads to a different argument: a reputation for good quality is more likely to provide firms with a motivation to avoid product defects. I build upon these two competing arguments and hypothesize that stronger motives for learning exist in situations where firms have either a weak or strong reputation for product quality. My study of product recalls in the US automotive industry highlights an inverted U-shaped relationship, indicating the liability of an intermediate reputation in reducing product defects.     

基于分层模型的银行客户信用风险预测研究

本文基于商业银行客户信贷记录数据集,通过运用拉普拉斯分层模型对客户的信用风险进行预测研究.利用客户群体存在差异化的特点,采用XGBoost机器学习算法来选择分层特征以及结合多元特征的组合形式来预测客户的违约情况.在不同分层特征结构下依次对比拉普拉斯分层模型、单独模型、共同模型和随机森林四个模型的预测效果,并建立模拟数据集来对拉普拉斯分层模型的性能进行验证.研究发现:(1)拉普拉斯分层模型的预测精度是最高的,预测性能具有稳定性;(2)本文数据集所适用的最佳分层特征是贷款金额、年龄和婚姻;(3)分层特征的选择和数量会依据不同数据而产生相应变化,并非一成不变.结合本文的研究思路和结果,以期为商业银行在客户信用风险评估实践中提供新的思考和建议.     

对导入式智能审计模式的再思考

导入式智能审计模式是专门用于注册会计师财务报表审计的一种计算机审计模式,其核心在于会审软件、数据导八、全面重算、智能审计、社会对账、安全通信和日志记录七大技术要素,它们组合起来形成有别于其他任何审计模式不同的特点.导入式智能审计可以通过会审软件来实施.