Small and medium-sized enterprises (SMEs) are the main drivers for the continuous development of a national economy.Financing has always been a challenge to SMEs. SMEs often resort to forming the Board of Directors to help improve financial capability. Current research on SME Boards of Directors mostly focuses on market structure, price difference, market liquidity,volatility, and yield relationships. However, much less attention has been paid to the extreme risk measurement of SME Boards. The paper applies the extreme POT model and VaR model to investigate extreme risks in SME Boards. We collected the data of 992-sized stock daily closing prices from Tsinghua Financial Data from December 1, 2005 to December 28, 2009.First, we adopted an exponential regression model to choose the threshold values of the POT model. The model simulated the tail distributions in SME stock markets. We then combined our VaR model with our POT model to appraise the extreme risk in SME stock markets, and compared these two models with respect to their normal distribution. In the first section, we employed POT model to simulate the tail distribution in SME Board. Threshold selection is the key for estimating parameters in our POT model. Due to the uncertainty of threshold selection in Hill figure, the exponential regression model is an effective way to select threshold values of POT model.Excess Distribution and Tail of Underlying Distribution are diagnostic plots to describe the POT model fit. POT appears to fit the tail distribution of data fairly well. The second section discussed the VaR model to estimate the extreme risks of SME Boards. Our VaR model measured the size of extreme risks and verified that combining VaR model with POT model is more efficient than a normallydistributed model. We also found that traditional risk measures based on the normal distribution sometimes overestimates risks.However, it is difficult to test LR statistics because normal distribution assumptions do not reflect pinnacles and fat-tails in the distribution of asset returns.In summary, our POT model with exponential regression method can effectively measure the extreme risk in SME Board. The SME stock market is a key for constructing the multilayer capital market system. Therefore, accurate depiction of SME stock market risks has significant reference values.%中小企业板有着区别于沪深主板市场的风险特征.本文针对传统基于正态分布假设的VaR模型低估风险问题,运用极值POT模型研究中小企业板的极端风险,利用指数回归模型实现了POT模型中阈值的定量选取,避免了样本Hill图在选取阈值时存在的局限性.实证结果表明,POT模型对中小企业板极端风险的估计比正态分布更有效.研究还发现,传统基于正态分布的风险度量没有体现金融资产收益率的尖峰厚尾特征,同时存在低估风险和高估风险,但都很难通过LR统计量检验.因此,基于指数回归法的POT模型更能有效刻画中小企业板的极端风险.
展开▼
