| The cost of equity capital is the investors’ required rate of returns,which plays a decisive role in the fields of micro-financial decision-making and macroeconomic management,but it can not be obtained directly and need to rely on some mathematical models to estimate.Based on the practice of western capital market,it has formed two types of mathematical models:one is the risk compensation models based on the historical returns,such as CAPM,APM,FFM,five factors model;the other is the implied cost of capital models based on the expected returns,such as GGM,GLS,CT,PEG,MPEG and OJ model.The domestic researches started lately and these researches mainly focused on the influential factors of cost of equity capital,which generally adopted foreign models directly and estimated the cost of equity capital of domestic listed companies,then applied these estimated results to the empirical researches on the influential factors.In the course of researches,the selection of estimating models was more casual without considering the validity of these models;when adopted the implied cost of capital models,earnings forecast data usually were obtained from analysts or generated by statistical models,some scholars also used the actual earnings,which without considering the reliability of earnings forecast.This two cases leads to estimated results of cost of equity capital are different and draws some paradoxical conclusions.Based on the basis of the existing researches,the paper focuses on the estimation,evaluatin and analysis of the cost of equity capital,that adopts the implied cost of equity capital models,including GGM,GLS,CT,PE,PEG,MPEG,OJ model,to estimate the cost of equity capital of domestic non-financial listed companies over the period 2004—2014 basing on earnings forecast obtained from financial analysts and generated by pooled cross-section regression model respectively,then tests the reliability of estimated results and analyses the characteristics of the cost of equity capital.This paper is helpful to the future researches on the selection of mathematics models and the origin of the earnings forecast,also to the role of decision of the cost of equity capital.Based on the reliability tests of the cost of equity capital,the paper finds that estimated results of PEG,MPEG and OJ model based on earnings forecast of pooled cross-section regression model are more reliable.Specifically including:(1)For the reliability of the earnings forecast,comparative analysis of forecast error and accuracy shows that analysts,earnings forecast are more accurate than earnings forecast of pooled cross-section regression model in the short-term,but this advantage would be weakened with the extension of the forecast term,earnings forecasts of pooled cross-section regression model are more accurate in the long-term;multivariate regression test of risk factors and univariate regression test of future-relized returns also show that earnings forecast of pooled cross-section regression model are more consistent with theoretical expectations;(2)For the reliability of the models,this two kinds of test show that PEG,MPEG and OJ model are more reliable wether based on analysts,earnings forecast or earnings forecast of pooled cross-section regression model,the mean of all models,estimated value also has some reliability,GGM,GLS,CT,PE models,reliability are very poor.Based on the test of reliability,the paper also analyses the characteristics of the cost of equity capital of domestic listed companies,and finds that there are significant differences among industries,which remains some stability in different years.There are significantly annual and regional differences of the cost of equity capital,which are consistent with fluctuations of macroeconomy and stock market,and influenced by the size of company,ownership and shareholding structure.In the sample session,there is a reasonable boundaries of the cost of equity capital and risk premium in the long-term,the average cost of equity capital is about 9.58% and the equity risk premium is about 3%.The innovation of the paper is mainly reflected in the following three aspects:First,the paper considers the validity of the estimating models when it estimates and analyses the cost of equity capital of the listed companies.The selection of the estimating models has an important impact on the estimated results and different estimating model always generates dramatically different result.The domestic researches mainly focused on the influential factors of cost of equity capital without considering the validity of these models.The paper adopts some estimating models and makes a judgement of the validity of these models by empirical tests.Second,the paper considers the reliability of the earnings forecast data when it estimates and analyses the cost of equity capital of the listed companies.The quality of earnings forecast data is an important influential factor of the estimating results,validity.The domestic researches rarely considered the reliability of earnings forecast.The paper estimates the cost of equity cost basing on earnings forecast obtained from financial analysts and generated by pooled cross-section regression model respectively,then makes a judgement of the reliability of earnings forecast by empirical tests.Third,the paper focuses on the effective measurement and analysis of the cost of equity capital.The domestic researches rarely considered the quantity and characteristics of the cost of equity capital.Based on the test of reliability,the paper analyses the characteristics of the cost of equity capital from different angles and discusses the reasonable boundaries of the cost of equity capital and risk premium. |
