Study On Investment Style Identification And Drift Risk Of Chinese Stock-based Open-end Funds

After the 1980s, security investment funds industry got the fast development,the open-end fund is entered the explosive rapid development era. Capital markets mature development and also promoted the prosperity of funds, As the voluminous issuance of fund products, especially increases in size of open-end stock fund, the space for fund industry's innovative development being opened already. In the end of 2010, there are sixty-one fund companies established, and they manage 708 open-end funds,702 funds of them belonging to stock fund. Indeed, the open-end funds now already become important institutional investors in China's capital market, at the same time, as an important tool, they also play a critical role in promoting the stability and health development of capital market in China.The development of U.S. fund market experiences a process from disorder to order, from order to efficiency. Regarding the design of fund products, there also experienced from a process of only providing a kind of financing services to a combination process of providing fund products and financing services; As to financing objective, fund buyers experience a process of pursuing excess returns to desiring returns, and even to huntering risk-return process.Regarding fund financing strategies, there also has a process of changing from dependence on individual fund manager's ability to investment style ideals, and then to investment style theory. Fund investing style and investment style recognition now already widely accepted in today's U.S and European capital market. Concerning about how the future development of our country's fund industry fund products'design and investing style ideals can be ahead of the fund investment community, this dissertation has systematically researched the phenomenon of investment style drifts in Chinese open-end fund community, therefore, providing a theoretical guidance for the future development of our country's fund industry.Studies about investment style focus may be traced back to the early 1990s, the investment style is still a hot topic of fund analysis nowadays. Fruitful previous researches showed that the investment style accounted for a large part of the fund's performance in practice and theory. This is because investment style drift is a double edge of the knife, it can produce short-term excess return, however, at the same time, reflects great risk behind it. Their fractal features of those newly rising fund markets lay foundations to study the feasibility of investment style drift phenomenon in their country. On the basis of this, this paper uses ARFIMA (the Autoregressive Fractionally Integrated Moving Average process) and HYGARCH (the "hyperbolic GARCH")models to study vestment style drift phenomenon in China's open-end fund community. Based on studying the fractal characteristics of long memory of fund style asset return series, we have studied systematically the non-linear features and risk measures of Chinese open-end stock funds. Mean time, on the basis of results researched done by empirical studies in this paper, a scientific measure standard of scaling investment style drift building up, so that it can be employed by financial institutions, such as fund companies, fund managers and regulatory departments for both internal and regulatory purposes.Due to newly-rising and track-changing characterisitcs of China's security market,and the numbers of Chinese stock-based open-end funds available still not large enough, in order to expand the sample volumes and data horizon, The sample period adopted in this paper starts from 1st January 2006 when the Chinese style indices became available and ends in 31th Decmber 2010,which includes five years(20 quarters). This provides us with a sample size of 72 observations of Chinese open-end stock funds, which consisting of 36 stock funds and 36 active allocation funs respectively. During studying period, each fund has 1214 daily net value,253 weekly net value, and 60 monthly net value respectively. Accordingly, we can get 1213 daily earnings ratio,252 weekly earnings ratio, and 59 monthly earnings ratio for each fund respectively.For ongoing study on the topic presented here, Innovations of research methodologies of this dissertation mainly involve several aspects as followings:First, Introducing the Autoregressive Fractionally Integrated Moving Average process, (or ARFIMA) and the "hyperbolic GARCH"(or HYGARCH)models, The two popular long memory models in advanced econometrics introduced aims to study Chinese open-end stock funds'style drifts; Meantime, we use R/S (rescaled range statistic) and modified R/S (modified rescaled range statistic) analysis to explain the t fractal characteristics of long memory of Chinese fund style drift and non-periodicity,such as "non-cyclicities" or "uncertain and time varying periodicities"Second, On the basis of Comparing two traditional approaches to fund's investment style recognition, in line with those fractal characteristics of Chinese fund markets, we come up with a new fund's investment style recognition based on box-fractional dimensions and a quality measure standard of fund's investment style drift degree--identical standard of drift(ISD); Furthermore, we bring up methodologies of fund's investment style analysis based on box-fractional dimensions, give the calculation formula of elastic fractional dimensions and self-exciting threshold of fund's investment style drifts. And so we expect we can provide a new method for studying fund's investment style drifts.Third, Based on qualifying fund's investment style drifts, we draw a formula for calculating fund's investment style drifts'earnings. We introduce skew-t distribution and skewed Student-t distribution to scale the peak, fat tail and skewed characteristics of investment style drift return series. When measuring the risks of fund's investment style drifts by using GARCH model, we find ARFIMA-GARCH model estimated more robust.So far, our research has drawn three main conclusions as followings:First, By using Classical R/S and modified R/S analysis, the empirical studies of the fractional autoregressive integrated moving average with hyperbolic generalized autoregressive conditional heteroscedasticity (ARFIMA-HYGARCH) model show that Chinese fund style asset return series has a genuine long-memory and non-periodicity features. Research done about Chinese fund's investment style recognition analysis based on box-fractal dimensions of fund's investment style recognition concludes that there has a common appearance of fund's investment style drift, and our conclusions are more likely to be suitable for realistic background of fractal markets.Second, furthermore, this research done based on methodologies of the analysis of elastic fractional dimensions shows that fund's investment style drifts are a common thing in Chinese fund community. And this demonstrates fund market in China is non-perfect efficient, and appears some certain fractal features. These tell us realistic feasibility why fund's investment style drift.Finally, Fund's investment style drift risk scaled by modeling fund's investment style drift risk VaR-GARCH framework, we concludes that Chinese open-end fund community experience widespread fund's investment style drift risks. Therefore, our conclusions possibly can provide theoretical guidance and supports for our country's securities'regulatory departments to regulate and control serious style drift phenomenon, standardize issuance of fund products and investment behavior.