1、华中科技大学 硕士学位论文 财务报表分析与中国股市超额收益的概率预测 姓名：代金博 申请学位级别：硕士 专业：数量经济学 指导教师：王少平 20050512 Management Case Center of Peking University and the Economic Observer choosed twenty enterprises as the most honorific ones from Chinese enterprises in 2002, the 11 ones of them were listed in Shanghai or Shenzhen Securit
2、ies. This paper takes 11 A shares whose credit standing is excellent as researching sample, assuming their annual financial statements of 2002 are authentic and fair, we obtain 8 financial ratios from their statements, first the principal component analysis is applied to screen out some principal co
3、mponents with the 85% accumulated contribution ratio as a criterion. Second the single variable probit model is applied to screen out the significant principal components those will be used as independent variable, and the binary qualitative variable whose value depends on whether there is excess st
4、ock yield from 2003 April 30 to 2004 May 1 will be used as dependent variable, now we establish the forecasting model by applying the multi-variable probit model. Third, the MLE is applied to the model. Finally we take a sample test for the model. The following conclusion is drawn at last: On the co
5、ndition that financial statement is authentic and fair, the precision ratio of our probit model in forecasting whether there is excess stock returns is 82%, which outclasses the 50% of stochastic distribution, this shows that through some investing period, in the first, investors compute the 8 finan
6、cial ratios from financial statement of listed company which they want to invest in, in the second, the principal components are selected by applying principal analysis to these ratios, in the third , apply significant test to the selected principal ones and significant principal components are seek
7、ed out. Now apply the selected data to the probit model and the conclusion that whether there is excess yield can be drawed. If investors consult the conclusion when they select listed company, the probability that they get excess yield can be augmented greatly. Key words: financial ratio excess sto
8、ck returns principal component analysis probit regress 1 10 WTO QFII 1 1030 80% 16% 20 30 Benjamin Graham 2 3 30 2 45 Ball&Brown(1968), 6 Beaver(1968) Chung,Hay Y and Kim,Teong B(1994), 7 Holthausen,R.W.and D.F.Lacher(1992) 8 1998 Ball&Brown 1996 9 2000 1998 10 2003 11 Ou Penman 1989 1965 1983 68 Lo
9、gistic 0.6 0.4 12 Abarbanell Bushee 1988 Lev Thiagarajan 1993 9 1974 1988 133 14 1995 32 1993 5 3 5 14 9 9 2001 1999 Ou Penman 67%,Abarbanell Bushee 62%, Logistic 58% 2003 1998 2002 A Logistic 15 58% 8 A Logistic , 11 A Probit Model 4 11 2002 11 8 87% CAPM i a i a i Y 0 i a i Y 1 i Y Probit Probit M
10、odel Wald P 0.1 Probit Model P 0.1 Probit Model P 0.1 5 2.1.1 k k k k X X X , , , 2 1 L k k ) ( , , , 2 1 p k F F F k L i F k k X c X c X c F 1 2 21 1 11 1 + + + = L k k X c X c X c F 2 2 22 1 12 2 + + + = L k kk k k k X c X c X c F + + + = L 2 2 1 1 6 1 2 2 2 2 1 = + + + ki i i c c c L k j i j i F
11、F Cov j i , , 2 , 1 , , , 0 ) , ( L = = ) ( ) ( ) ( 2 1 k F Var F Var F Var L 2.1.2 , , , , , 3 2 1 p X X X X L 1 F 2 F ) ( , , , 3 2 1 k s F F F F s L L k s 1 X 2 X 1 X 2 X 1 X 2 X 1 X 2 X 1 X 2 X 1 X 2 X q 1 F 2 F 1 F 2 F 7 2.1 q q sin cos 2 1 1 X X F + = q q sin cos 2 1 2 X X F + = C X C X X F F
12、= - = 2 1 2 1 cos sin sin cos q q q q I C C C C = = - , 1 1 F 1 F 1 F 2 F 1 F 2 1 , F F 2 1 , X X 1 F 2 F 1 F 2 F 2 1 X X F 2.1.3 n k, 8 = nk n n k k X X X X X X X X X X L L L L L L L 2 1 2 22 21 1 12 11 ij X i j j j ij ij VarX EX X Y / ) ( - = Y j EX j j VarX j j i j i j i ij EY EY VarY VarY Y Y Co
13、v r = = ) , ( i j R , ) 1 /( - = n Y Y R T R l u R i l 0 3 2 1 k l l l l L i l 1 , , 1 ) , , , ( 2 1 = = = i T i ki i i i c c k i c c c c L L C : 1 = = C C CC T T C C R k = l l l L 2 1 9 - = k T T n YC Y C l l l L 2 1 ) 1 ( YC F = - = k T n F F l l l L 2 1 ) 1 ( k k k Yc F Yc F Yc F = = = L 2 2 1 1
14、) 1 ( k j Y i ) 1 ( k j F j ) , , 2 , 1 , ( 1 1 1 k l i ta a k m m l m m l k m m i i L = = = = = = l l l l l a i F l ta 1 F l F k l l ) ( , , , 2 1 k s F F F s L k 10 2.1.3 = nk n n k k X X X X X X X X X X L L L L L L L 2 1 2 22 21 1 12 11 j ij J p ip i i j i C F F c F c F c Cov F X Cov l = + + = ) , ( ) , ( 2 2 1 1 L i F i X ) , ( i i X F r 1 F 2 F k F 1 X 11 2 1
