Noise and Expected Return in Chinese A-share Stock Market. By Chong QIAN Chien-Ting LIN

Transcription

1 Nose and Expeced Reurn n Chnese A-share Sock Marke By Chong QIAN Chen-Tng LIN 1

2 } Capal Asse Prcng Model (CAPM) by Sharpe (1964), Lnner (1965) and Mossn (1966) E ( R, ) R f, + [ E( Rm, ) R f, = β ] + ε E E ( R, R f, ( R m, ) ) = he expeced reurn of porfolo a me = he rsk free rae a me = he marke reurn a me } If he model holds, Bea β should be all ha maers n asse prcng n he world

3 } Frm sze effec Banz (1981) found ha small frms socks earn sgnfcanly hgher reurns han hose of large frms. } Book-o-marke equy effec Rosenberg e al. (1985) llusrae ha socks wh hgh book-o-marke equy rao have hgher reurns han socks wh low book-o-marke equy rao. } Momenum effec Tradng sraeges ha buy pas wnners and sell pas losers realze sgnfcan abnormal reurns over he 1965 o 1989 perod (Jegadeesh and Tman, 1993)

4 } Wh he movaon of mprovng Fama and French s (1993) 3-facor model, Carhar (1997) consruced a 4-facor model: R R = α + β ) + s SMB + h HML + wwml + ε f, ( Rm, R f, R Rf, R m, R f, HML = reurn of porfolo n excess of rsk free rae = excess reurn on marke proxy SMB = sze facor = book-o-marke equy facor WML = momenum facor } Sze, Book-o-Marke equy and momenum facors explan he varaon n sock reurns well.

5 } Nose Devaon from he sock value (Black, 1986). Arno, Hsu, Lu & Markowz (2007) argue ha he sock has negave nose n prce should ouperform he sock wh posve nose. } Nose and frm sze effec The large sze socks are he resul of a posve nose; The small sze socks are he resul of a negave nose (Arno, 2005). } Nose and book-o-marke effec Hgh (Low) book-o-marke equy rao mgh caused by negave (posve) nose. } Nose and momenum effec Posve feedback radng as a resul of exrapolang a pas good (poor) performance oo far no he fuure Lee and Swamnahan (2000). Nose generaon adds fuel o he connuaon of prce momenum.

6 } Research gap No emprcal nvesgaon has been done on he nose effec and he causal relaon of nose o sze, book-o-marke equy and momenum effecs. } Research quesons Does nose gve rse o sze, book-o-marke equy and momenum effecs? If no, does nose rsk premum exs afer conrollng for sze, book-omarke equy and momenum facors?

7 } Develop mehods for esmang he nose. } Invesgae he casual relaon of nose o frm sze effec, book-o-marke effec and momenum effec respecvely. } Provde emprcal evdence n supporng he exsence of nose rsk premum. } Improve he Carhar s (1997)4-facor model.

8 } Monhly daa on all A-shares raded on he SHSE and he SZSE n Chnese sock marke coverng he perod of January 2000 o January 2007 s obaned from Daasream. } To be ncluded n he sample, a frm mus have avalable nformaon on capalsaon and relevan accounng daa (P/E and P/B rao ec).

9 P/B mehod PB 1 =,( j n 1 1/ n = ( B j, / Pj, ) n= j ) P / B V = B * PB PB, P j, B j, P B V / = he benchmark P/B mulple (harmonc mean) for he arge frm a me = prce per share of equy for frm j a he end of perod = he book value per share for frm j a he end of perod = he fundamenal value per share of frm from P/B approach

10 P/E mehod PE, PE E j, P E V / 1 =,( j n 1 1/ n = ( E j, / Pj, ) n= j ) P / E V = E * = he benchmark P/E mulple (harmonc mean) for he arge frm a me PE = he earnngs per share for frm j a he end of perod = he value per share of frm esmaed from P/E approach Esmae he value of sock from P/E and P/B combned approach V P / E& P / B P/ E& P/ B V, = ( V P / E + V 2 P / B ) = he value per share of frm esmaed from P/E and P/B combned approach

11 } Measure he value of nose Accordng o Black (1986), he nose can be measured as he dfference beween he marke prce and fundamenal value. Nose: λ = P, V P / E& P / B } Porfolo Aggregaon Procedures The sample socks are sored by sze, book-o-marke equy, pas reurns and nose o consruc porfolos. The porfolos are rebalanced a he begnnng of each year. Nose Facor NMP = he excess average reurn of porfolo wh negave nose socks over porfolo wh posve nose frm socks a me.

12 } Reasons for usng panel regresson model Our sample daa se has boh cross-seconal and me seres dmensons, herefore, should be denfed as panel daa. For he economerc analyss of panel daa, we canno assume ha he observaons are ndependenly dsrbued across me (Wooldrdge, 2006). Hence,unobserved facors whch can be deeced by me seres approach may exs. } Panel regresson Model: R R f, = + β Rm, R f, ) α ( + s SMB + h HML + wwml + n NMP + a + ε a = he unobserved facor } Hypohess: H1: The nose effec s sgnfcan H : 1 n 0

13

14

15

16 } Deermne he nose effec and prove he exsence of nose rskpremum afer conrollng for Carhar s four facors. } Doubs on hose clams ha he nose no only creaes he hgher reurns for small sze and hgh book-o-marke equy socks o some exen, bu also plays a crucal role n undersandng he momenum effec. } Improve he Carhar s 4-facor model by ncludng he nose facor.

17 Thank You & Quesons