The delisting risk warning model is the key one for defending the normal order of the stock market and the interests of investors. It plays an important role in optimizing the allocation of resources in the capital market and in taking the health of the market. The delisting risk of any companies has a significant impact on its stock price, and even leads it to be delisted. Once the delisting early warning model is flawed, wrong predictions are inevitable. When the "good" company is predicted as a "bad "one, it will damage the shareholder rights and interests of the "good" company, even lead the company to be delisted from the capital market. And when the "bad" company is predicted as the "good "one, it will directly mislead the investors and the public, even makes the wrong investment decisions and causes to huge losses.
The method ofradar map(RC) presents the advantages of intuitive, uncomplicated, and interpretable in the comprehensive evaluation of the existing research. The drawback is that the evaluation results may be completely different and even the opposite when the order of indicators changes in radar map. This may be one of the reasons why radar map has not been applied to delisting early warning research and application. We overcome the drawbacks above, in order to apply the radar char method to establish the delisting early warning model.
The principle of using the equal interceptradar map(EIRC), our proposed in this paper, to carry out the delisting risk is as follows: we present the delisting risk by using the two areas in radar map enclosed respectively by the single indicator and all of indicators. The larger the area is, the greater the delisting risk is. When any one of the two areas enclosed approaches or reaches its maximum area value in the radar map, the delisting risk of the company by *ST is greater. On the contrary, the smaller the area in radar map is, the less delisting risk of the company by *ST is.
In this paper, we improve the RC based on equal intercept transformation and establish the delisting risk early warning model based on the improved RC. We perform polygon by using EIRC which connect the vertices of two adjacent indicators through the points on the same equal intercept. The area of the polygon enclosed by EIRC is constant regardless of the order of the indicators. And to RC in existing research, the polygon is formed by connecting directly the vertexes of adjacent indicators. The area of the polygon enclosed by RC changes with the order of the indicators changing. This can cause the evaluation result of the same company to be completely different or opposite, if the order of the indicators changing.
In existing research on financial early warning, the main drawback has two folds. First, existing study do not pay attention to the non-financial indicators in delisting early warning research for Chinese listed company. In fact, non-financial indicator, such as “audit report opinions”, is the fundamental criterion of the delisting early warning. Taking the delisting risk earlywarning model in the Stock Exchange Listing Rules of the Shanghai Stock Exchange as an example, there are all 12 indicators, where including 4 financial indicators and 8 non-financial indicators. However, in the indicator system research of delisting risk early warning, existing researchers present either ignoring non-financial indicators or selecting many financial indicators that are not included in the current delisting rules. Therefore, the indicator system used by existing research does not match with actual *ST criterion of Chinese listed company.
Second, the predictive models in existing research, such as Probit regression, are essentially the ones for discriminant or determining the probability of the company by *ST. It requires a large number of samples to fit and estimate the parameters in model. It is easy to produce prediction errors when the sample size is limited and parameters cannot be provided.
This paper introduces all 12 delisting situations, including of 4 financial indicators and 8 non-financial ones, according to the Stock Listing Rules issued by Shanghai Stock Exchange. We establish the delisting earlywarning model in line with the real situations, and change the disadvantages of existing researches that are out of touch with the real delisting rules. We analyze three situations of risk by *ST for Chinese listed companies and verify the performance of EIRC model our proposed in this paper.
In the empirical research section, this paper predicts the risk by *ST of three Chinese listed company, including of the Anhui Heli Co., Ltd. (Anhui Heli), Kunming Machine Tool Co., Ltd. (Kunming Machine Tool), and Hongda Co., Ltd. (Hongda), by using their data in the past five years (2011-2015) respectively.
For Anhui Heli, empirical research shows that the risk by *ST of single indicator is very small because the area of each indicator is much smaller than the maximum area of single indicator in EIRC. The total risk by *ST is also small, because its total area value 0.00358 is much smaller than the maximum area of 1.55. From these two aspects, the risk by *ST for Anhui Heli in 2016 is very small, that is, the possibility by *ST is negligible.
For Kunming Machine Tool, although the total area value 0.1517 is much smaller than the maximum area value 1.55, the area of the indicator “net profit” reaches the maximum value 0.1294. It will be *ST in 2016, because it meets the situation of “The audited net profit of the last two fiscal years is continuously negative or continuously negative after being retrospectively re-stated” which is the *ST rule of Shanghai Stock Exchange.
For Hongda, the area of the indicator “audit report opinion” reaches 0.10352, which is close to the maximum value 0.1294 of single indicator. This presents that the risk of the indicator “audit report opinion” is larger. The *ST risk of every other indicator and total risk are both lower. Although these do not meet the situations of early warnings by *ST, its management should still take measures to improve the corresponding processing of “audit report opinion” to reduce the risk by *ST.