Welcome to free paper download website

International Experience and Lessons of executive compensation, research methods

Author: LiSu From: www.yourpaper.net Posted: 2009-04-28 06:27:24 Read:
[Abstract] executive compensation, has long been a hot research topic. In order to improve the level of research and the quality of research, it is necessary to learn from foreign research in the field. Japan in the study of executive compensation, data acquisition, and the choice of methods of measurement and model specification and hypothesis testing practice were analyzed, and found some of the skills and experience that we can learn.
[1] remuneration; research methods; data; model; hypothesis testing

Introduction
The company's executive compensation issues in foreign countries has long been a hot research topic. Empirical research on this issue began in Weygand (2001) and Zengquan Li (2001), two articles. Since then, new research results are emerging. But the researchers are too concerned about whether a meaningful research results or findings against expectations, often neglected research methods possible impact on the results. Some researchers in the country, "It is noteworthy draw down is not a large sample of statistical regularities should pay system into success in the enterprise, inquire about its pay system was successful key factors to see whether can be transplanted grafted to Chinese enterprises "(Cui Wang, 2006). With this view, the author argues that research methods than the remuneration system more general applicability, but also contribute to the understanding of the essence of the problem. Continuous improvement and refinement of research methods and techniques to improve the the domestic executive compensation, an important aspect of the level and quality of research.
Chose Japan as a learning object, the following reasons: First, the disclosure of information on corporate executive pay has many similarities with Japan, how to overcome the problem of data sources and data quality is such study an important problem to be solved. Second, different from the Anglo-Saxon countries, in Japanese literature for executive compensation research focuses on the cash remuneration. As China's equity incentive pilot began in the late 1990s (Li Chunqi, Huang Qun-hui, 2002), has not been widely adopted. Therefore, the cash remuneration has become the main target of domestic research. Japanese academia on remuneration issues in the data selection, model specification and interpretation of the results there are many places that we can learn. While some aspects can not be directly borrowed, but for our pioneering research ideas or innovation of great help.

Second, the research data acquisition
The pay data sources and quality of the research has a very important role. In Japan, the source of remuneration data is also subject to considerable limitations, data quality, there are many defects. Japanese listed companies only be required to disclose the total of all executive compensation, but did not like the U.S. companies that require disclosure of individual compensation information. Researchers often is this total remuneration divided by the number of executives, to be converted to the so-called personal emoluments data. This approach can make data exists bias, and often is underestimated. Therefore, how to overcome the partial data misuse is a primary problem to be solved.
Kato and Kubo (2006) 51 Japanese companies pay data provided by the company of a compensation consultant. 18 listed companies, 33 of which are non-listed companies during 1986-1995. The main advantage of this data: First of all, the salary data of the remuneration of company executives personal information, so you can avoid some of the previously mentioned "bias". Second, the data structure of the panel data, omitted variables problem can be solved to some extent. Third, this data includes both listed companies, also includes a non-listed companies, so you can compare the two. However, this data is also present some drawbacks. Nakazato et al (2006) pointed out that although these data are very rich, but the selection is not random. In addition, Kato (1994) also pointed out that professional survey data often does not show the name of the survey data can not be merged with the information in the financial statements. Due to the lack of some of the variables reflect the characteristics of the enterprise, the data obtained by the professional company survey "is not suitable for multiple regression analysis. Some of the literature in the United States are often professional survey data to senior management compensation issues.
Japan required by law to pay taxes in excess of 10 million yen taxpayers should their identity and tax bills to be announced. Therefore, Kato and Rockel (1992) and Nakazato (2006) are the use of the tax reporting calculate the amount of personal emoluments high management. Kato and Rockel (1992) observed in 1985 to 599 the tax situation of the president of the company; Nakazato et al (2006) to 578 in 2004, President tax situation observed. Because Japanese law, the taxpayer's income tax rate is 37% more than 1.8 million yen. According to the tax rate can roughly calculate the taxable income of the taxpayer. The important feature of this data source can only get the pay data revenues in excess of a certain standard. Less than the amount of remuneration in this standard can only be considered to be below a certain value.
This type of data there are also some limitations. First of all, many executives report taxable income are lower than their actual income. And other local businesses, there are also a large number of company executives in Japan not part of a tax allowance. However, it is important that this downward bias will not be due to the different systemic change. In addition, many executives as well as income from other sources. Therefore, the remuneration paid in taxable income and would produce an upward bias is not. For example, some of the more wealthy corporate executives can be a larger amount of investment activities, so as to get more investment income, which is part of income does not belong to the category of remuneration.
For researchers, data sources and quality in a way that can not be completely controlled. Accordingly, in order to overcome the defects of the data, we also require a corresponding adaptation of the method of measurement for the specific data, thereby enhancing the effectiveness of the estimation result. The following discussion will focus on this problem.

Third, the choice of the type of data and measurement methods
Remuneration data collection is just the first step in problem analysis, in order to draw a reliable and valid conclusions, different data types must be compatible with the appropriate measurement methods. As mentioned earlier, data from different sources there will be some bias, therefore, by the choice of measurement method to try to weaken the data defects influence on the results. Because many studies did not bias the data the possible impact on the results of in-depth and detailed discussion so often ignored the choice of the econometric model, simply using the OLS regression method.
The data structure Kato and Kubo (2006) panel data, so you can use panel data fixed effects model to omitted variables that do not change over time. These variables include corporate executives of some personal characteristic variables. Fixed effects panel data regression model on the one hand can be streamlined; the other hand, can solve the omitted variables on the efficiency of parameter estimation. For example, a model is used in Kato and Kubo (2006): the ¡÷ ln (APAY,) i, t = the alpha ¦ÂdDROAi, t ¦Ìi, t wherein the natural logarithm of the dependent variable for the corporate executives annual cash pay; The only argument is the return on assets.
If only one year of data, you can not use the fixed effects regression. Kato (1994) stressed that "do not want to, but because they can not use the fixed effects model regression. In this case, it is necessary to add some variables to control the CEO personal characteristics change. Accordingly, in this case will not be able to use "simple" model. Nakazato (2006), only to use the data for the year 2004, and therefore can not use panel data fixed effects regression. To this end, the authors added from the manager's personal characteristics, the proportion of executive ownership, shareholding structure and characteristics of the Board control variables.
We have inspired the author in the regression using the Tobit model. Disclosure of taxpayer information, the minimum amount of tax due to the relevant laws and regulations of Japan. Therefore, the amount of tax is lower than the standard corporate executives pay off "review" (censor). Although this data can not know the specific amount of these corporate executives pay. But you can know that is below a certain amount. In order to explore the corporate executive pay issues from a wider range of data, it is necessary to use the tobit model. The authors believe that "This is the standard technique for dealing with the review of the data the one hand, this technology can be avoided sample selection bias, on the other hand, increased the number of observations in the regression analysis sample.
Be interpreted using different types of variables, different measurement methods should be used. Nakazato et al (2006) in its initial conclusion robustness test, the use of different forms of regression. For example, the High Income TP dummy variable, (if executives in 2004, taxpayers in excess of 10 million yen or more, is set to 1; 0 otherwise) as the explanatory variables can be used probit model regression; high tube in the tax on the list of the number of times as the explanatory variables can be used possion regression model.
Thus, a variety of measurement techniques can be used in the study of corporate executive pay issues, but the problem lies in unity with data types and features. The existence of a variety of measurement techniques can broaden our research ideas, and the choice of appropriate measurement techniques will help to improve the effectiveness and reliability of the conclusions of the study.
 1/2    1 2 Next Last
Please consciously abide by Internet-related policies and regulations.
Tips: Log in to comment, the user name to enter comments directly from your personal space, so that more friends to meet you.

Sponsored Links

Sponsored Links

Top