| 摘要: | 研究期間：10108~10207;The risk-neutral valuation relationship (hereafter RNVR) is the core of derivative pricing, especially for valuing options. In financial literature, it can be divided into two research groups for employing RNVR techniques to value options. One is to use no-arbitrage as a starting point and directly implement RNVR techniques to build up for the valuation of options. The representative ones are Cox and Ross (1976) and Kreps (1981). The other is based upon the representative agent’s preference type and employs general equilibrium approach to derive a RNVR option valuation model. The representative ones are Brennan (1979) and Stapleton and Subrahmanyam (1984). It has been made a great progress for later research group recently. For example, Câmara (2003, 2005) used transformed normal distribution to derive pricing formulae for options while Vitiello and Poon (2010) employed transformed GAMMA distribution to construct models for option valuations. The advantage in using preference-free option valuation models is that these models can value options with non-tradable assets as underlying. For example, these models can value weather derivatives. In addition, these models can be applied to study the corporate finance issues as well. For example, Câmara, Chung and Wang (2010) use this approach to estimate the implied cost of firm’s equity. Other examples, Chang,Christoffersen,Jacobs, and Vainberg (2011 ), and Lin,Paxson,Wang and Kuo(2011)) employed these models to estimate the option-implied measures of equity beta. The implied cost (beta) of equity has the property of forward-looking, and they will contain more future information compared with the values estimated by traditional approaches. The studies of this project contain three dimensions. In the first-year study, we will firstly extend the Vitiello and Poon (2010) model (under a multivariate transformed GAMMA distribution), and then combine our derived model with the ideas of Siegel(1995) and Lin,Paxson,Wang and Kuo(2011) to build up a general framework for estimating option-implied measure of equity beta. In a further step, we will prove that Lin, Paxson,Wang and Kuo’s (2011) model is a special case of ours. In the second-year study, we will collect data of S&P 500 index options and individual stock options from OptionMetrics. The data period will cover the period 2005 to 2009 and will include nearly three hundreds of individual stock options. We then compare three categories of estimated betas including traditional beta, implied beta estimated by Lin, Paxson,Wang and Kuo’s model, and implied beta estimated by our model, among them which better can have best future information contents and predictability. For carrying out detail studies, in the third-year research, we will divide the sample as two periods, before and after financial tsunami. Furthermore, we will divide the sample according to firm’s characteristics, such as firm’s size, industry, etc.. We then make comprehensive comparisons for the future predictability among these three estimated betas. |
