| Summary: | 碩士 === 淡江大學 === 土木工程學系碩士班 === 103 === The main goal of business is to maximize the enterprise value. To achieve this goal, it is important to build reasonable models to assess the enterprise value. This paper proposes an earnings-based and asset-based hybrid approach to supplement the existing valuation models, presents using quantile regression analysis and quantile neural networks to estimate the distribution of the enterprise value, and finally, explores the difference between the optimum valuation models for construction enterprises and other enterprises. Since the publicly traded company''s share price multiplied by the total number of shares was the reasonable estimation of the enterprise value, the stock market historical database was employed as the source of empirical data, a total of 14,985 data available. The following conclusions were obtained: (1) the proposed earnings-based and asset-based hybrid approach can more accurately predict the market value of the enterprise than the traditional approach. (2) The distribution of enterprise value can be estimated with quantile regression analysis and quantile neural networks. (3) At a specific return on equity, the company''s price-to-book value ratio presents a logarithmic normal distribution. The enterprise value can be obtained by multiplying the book value with the ratio, so the enterprise value also shows a logarithmic normal distribution. (4) The business valuation model of property development and construction industries is quite different from those of other industries. The enterprise values of the property development and construction industries can be significantly overestimated if the business valuation model for total industry is applied. (5) The patterns of Price-to-Book value ratio (PBR) curves are quite different in cases of the ROEs being larger or smaller than zero. The PBR becomes larger if the ROE is larger than zero and increasing. On the other hand, the PBR remains almost unchanged if the ROE is smaller than zero and decreasing. The Growth Value Model is highly useful and effective in various industries only if the ROE is larger than zero.
|
|---|
