基于LSV模型的公募基金羊群效應(yīng)及交易特征研究

目錄前言 1羊群效應(yīng)相關(guān)型原理 1易為型：LSV 1益相模型 3文采模型 4公募基金羊群應(yīng)分析 5本擇 5算輯點(diǎn)考 6股篩及易向 6參選擇 6調(diào)因分及響 7募金群應(yīng) 8公基羊效著 9公基羊效買入于出 10公基羊效上漲于跌 10公基羊效中等模金高 10公基的票買入率同市收負(fù)相關(guān) 公募基金交易征分析 12益因?qū)θ簯?yīng)影響 13值子羊效影響 15值位因?qū)θ盒в绊?17資收率子羊群應(yīng)響 19值子羊效影響 21手因?qū)θ簯?yīng)影響 23交額子羊效應(yīng)響 25總結(jié)和展望 27文結(jié) 27來望 27風(fēng)險(xiǎn)提示 27插圖目錄圖1LSV型法程圖 3圖2半報(bào)年基量 5圖3參交的金閾值羊效影響 6圖4不參下整分布 8圖5羊效與金關(guān)系 圖6各告股平入概與期證指益散圖 12圖7上收率羊應(yīng)影響 13圖8本收率羊應(yīng)影響 13圖9上收率買群效影響 13圖10本收率買羊群應(yīng)響 13圖期益對(duì)羊群應(yīng)響 14圖12本收率賣羊群應(yīng)響 14圖13上收對(duì)群應(yīng)整影響 14圖14本收對(duì)群應(yīng)整影響 14圖15上均EP羊效應(yīng)響 15圖16本均EP羊效應(yīng)響 15圖17上均EP買羊群應(yīng)響 15圖18本均EP買羊群應(yīng)響 15圖19上均EP賣羊群應(yīng)響 16圖20本均EP賣羊群應(yīng)響 16圖21上均EP羊效應(yīng)體響 16圖22本均EP羊效應(yīng)體響 16圖23上市率位對(duì)羊效影響 17圖24本市率位對(duì)羊效影響 17圖25上市率位對(duì)買羊效影響 17圖26本市率位對(duì)買羊效影響 17圖27上市率位對(duì)賣羊效影響 18圖28本市率位對(duì)賣羊效影響 18圖29上市率位對(duì)羊效整影響 18圖30本市率位對(duì)羊效整影響 18圖31上凈產(chǎn)益對(duì)羊效影響 19圖32本凈產(chǎn)益對(duì)羊效影響 19圖33上凈產(chǎn)益對(duì)買羊效影響 19圖34本凈產(chǎn)益對(duì)買羊效影響 19圖35上凈產(chǎn)益對(duì)賣羊效影響 20圖36本凈產(chǎn)益對(duì)賣羊效影響 20圖37上凈產(chǎn)益對(duì)羊效整影響 20圖38本凈產(chǎn)益對(duì)羊效整影響 20圖39上平流市對(duì)羊效影響 21圖40本平流市對(duì)羊效影響 21圖41上平流市對(duì)買羊效影響 21圖42本平流市對(duì)買羊效影響 21圖43上平流市對(duì)賣羊效影響 22圖44本平流市對(duì)賣羊效影響 22圖45上平流市對(duì)羊效整影響 22圖46本平流市對(duì)羊效整影響 22圖47上平換率羊群應(yīng)響 23圖48本平換率羊群應(yīng)響 23圖49上平換率買入群應(yīng)響 23圖50本平換率買入群應(yīng)響 23圖51上平換率賣出群應(yīng)響 24圖52本平換率賣出群應(yīng)響 24圖53上平換率羊群應(yīng)體響 24圖54本平換率羊群應(yīng)體響 24圖55上平成金對(duì)羊效影響 25圖56本平成金對(duì)羊效影響 25圖57上平成金對(duì)買羊效影響 25圖58本平成金對(duì)買羊效影響 25圖59上平成金對(duì)賣羊效影響 26圖60本平成金對(duì)賣羊效影響 26圖61上平成金對(duì)羊效整影響 26圖62本平成金對(duì)羊效整影響 26表格目錄表1羊效顯性與統(tǒng)量Hi,t大關(guān)系 9表2公基羊效小 9表3市上、跌群效差異 10表4公基半報(bào)報(bào)股平買概率 前言那么如何用定量的方法衡量這種從眾心理呢？本文基于早期國外學(xué)者提出的羊群效應(yīng)相關(guān)模型原理本章詳細(xì)介紹LSV模型的相關(guān)原理及算法邏輯，并簡(jiǎn)述其他幾種常見的衡量羊群效應(yīng)大小的方法。交易行為模型：LSV1992年，LakonishokJ、ShleiferARTheimpactofinstitutionaltradingonstockpricesLSV其中：

Hi,t|Pi,tPt|AFi,tHi,t為羊群效應(yīng)的度量值，其數(shù)值越大，則表示市場(chǎng)上的羊群效應(yīng)越大；Pi,t為在t報(bào)告期內(nèi)股票i的凈買入基金數(shù)量占比，即：P= Buyi,ti,t BuySelli,t i,tBuyi,t為在t報(bào)告期內(nèi)股票i的凈買入基金數(shù)量，Selli,t為在t報(bào)告期內(nèi)股票i的凈賣出基金數(shù)量。Pt為在t報(bào)告期內(nèi)，所有股票的凈買入基金數(shù)量占比的平均值，即：Nt Buyi,t= i1 Nt NtBuyi,tSelli,ti1 i1Nt為在t報(bào)告期內(nèi)所有基金交易的股票全集數(shù)量。AFi,t為調(diào)整因子，表示在t報(bào)告期內(nèi)，股票i的凈買入基金數(shù)量占比與同時(shí)期所有股票的凈買入基金數(shù)量占比均值之差的絕對(duì)值的期望，其計(jì)算公式為：AFi,tE(|Pi,tPt|)經(jīng)濟(jì)學(xué)上，調(diào)整因子AFi,t表示在不存在羊群效應(yīng)的市場(chǎng)中，|Pi,tPt|的期望。由于此處取了絕對(duì)值，因此該期望并不會(huì)等于0。假設(shè)不同基金經(jīng)理購買某只股票為獨(dú)立事件，且其購買該股票的概率為Pt，由此可

Buyi,t

Mi,t、Pt

Buyi,t~Bin(Mi,t,Pt)，其中Mi,tBuyi,tSelli,t，即參與交易該股票的基金數(shù)量。由二項(xiàng)分布的概率公式可知，某只股票被不同基金經(jīng)理購買m次的概率為：PBuy

Pm

)itmi,t Mi,t t t將上述公式帶入AFi,t的計(jì)算表達(dá)式中可得：AFi,t

Mi,t|m0

mMi,t

P|*Cmt i,tt

*Pm*(1P

t Mi,tmt i綜上所述，羊群效應(yīng)Hi,t的含義如下：ti（（的期望）Hi,t公募基金整體的羊群效應(yīng)則可以表示為報(bào)告期t各個(gè)股票羊群效應(yīng)的平均值，即：tNtHi,tHti1 NtRLSV1999Mutualfundherdingandtheimpactonstockpricesi,tHi,t|i,ttHSi,tHi,t|tHi,tHSi,t小于Hi,tLSV模型詳細(xì)算法流程如圖1所示：圖1LSV模型算法流程圖資料來源：首創(chuàng)證券收益率相關(guān)模型正如前文所述，LSV模型主要基于投資者對(duì)各個(gè)股票的交易行為（為CHFHCH模型該模型為學(xué)者Christie、Huang于1995年在文章《FollowingthePiedPiper:DoIndividualReturnsHerdaroundtheMarket i,t i,t tni1(RRn1tRi,tti收益t在上述基礎(chǔ)上，引入兩個(gè)啞變量RMup、RMdn，構(gòu)建下述二元回歸方程：t t+*RMup*RMdnt up t dn t tttRMupX%10；RMdnX%10。X15updntt小，各個(gè)股票分化程度較小，市場(chǎng)上投資者的交易行為趨于一致，羊群效應(yīng)顯著。綜上由CH模型的算法邏輯可知該方法有一定缺陷，僅能識(shí)別市場(chǎng)波動(dòng)較大時(shí)候的羊群效應(yīng)，對(duì)市場(chǎng)的小幅波動(dòng)無能為力。FH模型上述CH模型只能測(cè)算全市場(chǎng)的羊群效應(yīng)，而無法對(duì)某類投資者單獨(dú)衡量。學(xué)者JiangHM2018DoesHerdingBehaviorRevealSkill?AnAnalysisofMutualFundPerformanceFM文章首先引出下述兩個(gè)變量：Diff

j,i,t

Wj,i,tWj,i,t1WDiffFund

i,t

j,i,t1WFundi,tWFundi,t1WFundi,t1其中Diffj,i,t表示t時(shí)刻基金j中股票i相對(duì)于t-1時(shí)刻持倉數(shù)量或權(quán)重的變化比例；DiffFundi,ttit-1jtDiff

*

R*

Size*

PB*PB j,i,t j,t j,t i,t

j,t i,t

j,t i,t

j,t i,t

j,i,tRi,tt-1iSizei,tt-1iPBi,tt-1iDiffFundi,tj,t可以反映t-1tjtjtj本文擬采用模型LSV公募基金羊群效應(yīng)分析樣本選擇201420221、剔除被動(dòng)型指數(shù)基金及被動(dòng)型ETF基金；2、剔除指數(shù)增強(qiáng)型基金及指數(shù)增強(qiáng)型ETF基金；3、剔除滬深港主題基金；4、剔除投資海外市場(chǎng)的基金；5、剔除t-1時(shí)期半年報(bào)或年報(bào)缺失的基金；圖2半年報(bào)、年報(bào)基金數(shù)量資料來源：聚源數(shù)據(jù)220142022400201420183504502022170040%。測(cè)算邏輯幾點(diǎn)思考股票篩選及交易方向ttt-1t報(bào)t-1在實(shí)際建模過程中，本節(jié)對(duì)每只基金持有的所有股票做如下篩選：10.1%20.1%3t5參數(shù)選擇本節(jié)測(cè)算參與交易基金數(shù)量閾值對(duì)羊群效應(yīng)的影響，如圖3所示：圖3參與交易的基金數(shù)量閾值對(duì)羊群效應(yīng)影響資料來源：聚源數(shù)據(jù)伴隨閾值M增大，市場(chǎng)羊群效應(yīng)大小先揚(yáng)后抑，于M>=10時(shí)達(dá)到最大，隨后逐步

Mi,t|m0

mMi,t

P|*Cmt i,tt

*Pm*(1P

Mi,tmt 可知，伴隨基金數(shù)量M增大，調(diào)整因子快速下降，此階段Mt 盾，導(dǎo)致羊群效應(yīng)在M增大初期穩(wěn)步上升；當(dāng)MM||MM>=5調(diào)整因子分布及影響下4（）可知：1MM5001.5%2、買入概率P以50%為中心，調(diào)整因子呈對(duì)稱分布，且中間大、兩端低。當(dāng)M為5時(shí)：=518；=5（或57.7。M605%。圖4不同參數(shù)下調(diào)整因子分布51015202530354045505560657075510152025303540455055606570758590952030405060708090100110120130140150160170180190200210220230250260270280290300310320330340350360370380390400410420430440450460470480490500資料來源：聚源數(shù)據(jù)公募基金羊群效應(yīng)本節(jié)以基于交易行為的LSV模型為基礎(chǔ)，測(cè)算國內(nèi)公募基金的羊群效應(yīng)，為本文第四章分析公募基金交易特征做鋪墊。學(xué)術(shù)界普遍認(rèn)為，羊群效應(yīng)的顯著水平與統(tǒng)計(jì)量Hi,t數(shù)值大小的關(guān)系如表1所示：表1羊群效應(yīng)顯著性水平與統(tǒng)計(jì)量Hi,t大小關(guān)系顯著程度 不顯著略顯著顯著非常顯著Hi,t數(shù)值 0%至5%5%至10%10%至20%20%至50%資料來源：聚源數(shù)據(jù)Lakonishok、ShleiferARTheimpactofinstitutionaltradingonstockprices1%至3%20142022H表示H2100的二項(xiàng)分布假設(shè)下買入（或賣出）該股票的基金數(shù)量期望值多13.7只。表2公募基金羊群效應(yīng)大小羊群效應(yīng)羊群效應(yīng)2014H2羊群效應(yīng)羊群效應(yīng)2014H215.2%18.3%12.3%2015H113.9%15.9%11.4%2015H212.8%12.5%13.0%2016H112.5%13.6%11.4%2016H213.1%13.0%13.1%2017H113.4%10.8%15.1%2017H213.0%11.4%14.2%2018H113.9%14.3%13.5%2018H213.4%15.2%11.9%2019H111.1%11.2%11.1%2019H213.6%14.4%12.8%2020H113.7%15.0%12.3%2020H216.2%17.2%15.3%2021H114.5%16.7%12.0%2021H216.0%17.5%14.0%2022H112.7%13.4%12.1%2022H214.8%16.6%12.6%平均13.7%14.5%12.8%資料來源：聚源數(shù)據(jù)見表282%見表20422210%15%100%。表3市場(chǎng)上漲、下跌時(shí)羊群效應(yīng)差異漲跌閾值中證全指漲跌方向羊群效應(yīng)買入羊群效應(yīng)賣出羊群效應(yīng)上漲 14.02%15.07%12.86%0%下跌13.35%13.78%12.80%上漲5%下跌13.93%13.34%15.34%14.28%12.54%12.42%上漲10%下跌14.08%13.34%15.65%14.28%12.53%12.42%上漲15%下跌14.08%13.26%15.65%14.36%12.53%12.27%資料來源：聚源數(shù)據(jù)t-110（110整體上伴隨基金規(guī)模下降，羊群效應(yīng)先揚(yáng)后抑，中等規(guī)?；鹧蛉盒?yīng)相對(duì)較高。此外，買入羊群效應(yīng)明顯高于賣出羊群效應(yīng)。圖5羊群效應(yīng)與基金規(guī)模關(guān)系資料來源：聚源數(shù)據(jù)（表45%55%2015202155.5%。表42014H2

2015H1

2015H2

2016H1

2016H2

2017H1

2017H2

2018H1

2018H2

2019H1

2019H2

2020H1

2020H2

2021H1

2021H2

2022H1

2022H2股票平均 50.1%46.2%52.3%51.5%49.8%49.0%48.4%50.9%50.9%50.5%52.9%49.7%47.7%49.6%55.5%50.1%52.3%買入概率資料來源：聚源數(shù)據(jù)

見圖-50%圖6各報(bào)告期股票平均買入概率與同期中證全指收益散點(diǎn)圖資料來源：聚源數(shù)據(jù)公募基金交易特征分析具體而言，本章將當(dāng)前報(bào)告期（半年報(bào)或年報(bào)，下同）公募基金交易的所有股票，按照當(dāng)期或上期因子值從大到小的順序分組，分別計(jì)算每組股票的平均羊群效應(yīng)大小，分析各組之間的差異及規(guī)律。110因子值最小。收益率因子對(duì)羊群效應(yīng)影響圖7上期收益率對(duì)羊群效應(yīng)響 圖8本期收益率對(duì)羊群效應(yīng)響2014H22015H12015H22016H12016H22017H12017H22018H12018H22019H12019H22020H12020H22021H12021H22022H12022H2組

組2 組

組4 組5 組

組7 組

組9組10 資料來源：聚源數(shù)據(jù) 資料來源：聚源數(shù)據(jù)圖9上期收益率對(duì)買入羊群應(yīng)影響 圖10本期收益率對(duì)買入羊效應(yīng)影響2014H22015H12015H22016H12016H22017H12017H22018H12018H22019H12019H22020H12020H22021H12021H22022H12022H2組1

組2 組3 組

組5 組6 組

組8 組

組10 資料來源：聚源數(shù)據(jù) 資料來源：聚源數(shù)據(jù)從圖中可以看出：上期收益率影響：見圖9賣出羊群效應(yīng)：如圖11所示左側(cè)顏色較深，表明上期收益較高的組賣出羊群效應(yīng)較高，反之賣出羊群效應(yīng)較低。但各報(bào)告期稍有差異。當(dāng)期收益率影響：羊群效應(yīng)：（8整體呈U（1。買入羊群效應(yīng)：10賣出羊群效應(yīng)：12圖上期收益率對(duì)賣出羊效應(yīng)影響 圖12本期收益率對(duì)賣出羊效應(yīng)影響2014H22015H12015H22016H12016H22017H12017H22018H12018H22019H12019H22020H12020H22021H12021H22022H12022H2組1

組2 組3 組

組5 組6 組

組8 組

組10 資料來源：聚源數(shù)據(jù) 資料來源：聚源數(shù)據(jù)圖13上期收益對(duì)羊群效應(yīng)體影響 圖14本期收益對(duì)羊群效應(yīng)體影響資料來源：聚源數(shù)據(jù) 資料來源：聚源數(shù)據(jù)估值因子對(duì)羊群效應(yīng)影響本節(jié)以股票EP（市盈率倒數(shù)，TTM，下同）因子為例，分別測(cè)算上期平均EP、本期平均EP對(duì)羊群效應(yīng)的影響。2014H22015H12015H22016H12016H22017H12017H22018H12018H22019H12019H22020H12020H22021H12021H22022H12022H2圖15上期平均2014H22015H12015H22016H12016H22017H12017H22018H12018H22019H12019H22020H12020H22021H12021H22022H12022H22014H22015H12015H22016H12016H22017H12017H22018H12018H22019H12019H22020H12020H22021H12021H22022H12022H2組1 組2 組3 組4 組5 組6 組7 組8 組9組10

組1 組2 組3 組4 組5 組6 組7 組8 組9組10資料來源：聚源數(shù)據(jù) 資料來源：聚源數(shù)據(jù)2014H22015H12015H22016H12016H22017H12017H22018H12018H22019H12019H22020H12020H22021H12021H22022H12022H2圖17上期平均2014H22015H12015H22016H12016H22017H12017H22018H12018H22019H12019H22020H12020H22021H12021H22022H12022H22014H22015H12015H22016H12016H22017H12017H22018H12018H22019H12019H22020H12020H22021H12021H22022H12022H2組1 組2 組3 組4 組5 組6 組7 組8 組9組10

組1 組2 組3 組4 組5 組6 組7 組8 組9組10資料來源：聚源數(shù)據(jù) 資料來源：聚源數(shù)據(jù)EPEP15買入羊群效應(yīng)：與上述羊群效應(yīng)相似。賣出羊群效應(yīng)：與上述羊群效應(yīng)相似。當(dāng)期平均EP影響：與上期平均EP影響相似。2014H22015H12015H22016H12016H22017H12017H22018H12018H22019H12019H22020H12020H22021H12021H22022H12022H2圖19上期平均2014H22015H12015H22016H12016H22017H12017H22018H12018H22019H12019H22020H12020H22021H12021H22022H12022H2組1 組2 組3 組4 組5 組6 組7 組8 組9組10資料來源：聚源數(shù)據(jù) 資料來源：聚源數(shù)據(jù)（圖21上期平均EP對(duì)羊群應(yīng)整體影響 圖22本期平均EP對(duì)羊群應(yīng)整體影響資料來源：聚源數(shù)據(jù) 資料來源：聚源數(shù)據(jù)估值分位數(shù)因子對(duì)羊群效應(yīng)影響本節(jié)以股票近一年市盈率（TTM，下同）分位數(shù)因子為例，分別測(cè)算上期末市盈率分位數(shù)、本期末市盈率分位數(shù)對(duì)羊群效應(yīng)的影響。2014H22015H12015H22016H12016H22017H12017H22018H12018H22019H12019H22020H12020H22021H12021H22014H22015H12015H22016H12016H22017H12017H22018H12018H22019H12019H22020H12020H22021H12021H22022H12022H22014H22015H12015H22016H12016H22017H12017H22018H12018H22019H12019H22020H12020H22021H12021H22022H12022H2組1 組2 組3 組4 組5 組6 組7 組8 組9組10

組1 組2 組3 組4 組5 組6 組7 組8 組9組10資料來源：聚源數(shù)據(jù) 資料來源：聚源數(shù)據(jù)2014H22015H12015H22016H12016H22017H12017H22018H12018H22019H12019H22020H12020H22021H12021H22014H22015H12015H22016H12016H22017H12017H22018H12018H22019H12019H22020H12020H22021H12021H22022H12022H22014H22015H12015H22016H12016H22017H12017H22018H12018H22019H12019H22020H12020H22021H12021H22022H12022H2組1 組2 組3 組4 組5 組6 組7 組8 組9組10

組1 組2 組3 組4 組5 組6 組7 組8 組9組10資料來源：聚源數(shù)據(jù) 資料來源：聚源數(shù)據(jù)從圖中可以看出：上期末市盈率近一年分位數(shù)影響：羊群效應(yīng)：（3買入羊群效應(yīng)：如圖25所示右側(cè)顏色較深，表明上期末市盈率近一年分位數(shù)較低的組買入羊群效應(yīng)較高，反之買入羊群效應(yīng)較低。賣出羊群效應(yīng)：如圖27各組之間區(qū)分度不明顯。當(dāng)期末市盈率近一年分位數(shù)影響：U（0。買入羊群效應(yīng)：26賣出羊群效應(yīng)：282014H22015H12015H22016H12016H22017H12017H22018H12018H22019H12019H22020H12020H22021H12021H22014H22015H12015H22016H12016H22017H12017H22018H12018H22019H12019H22020H12020H22021H12021H22022H12022H22014H22015H12015H22016H12016H22017H12017H22018H12018H22019H12019H22020H12020H22021H12021H22022H12022H2組1 組2 組3 組4 組5 組6 組7 組8 組9組10

組1 組2 組3 組4 組5 組6 組7 組8 組9組10資料來源：聚源數(shù)據(jù) 資料來源：聚源數(shù)據(jù)圖29上期市盈率分位數(shù)對(duì)群效應(yīng)整體影響 圖30本期市盈率分位數(shù)對(duì)群效應(yīng)整體影響資料來源：聚源數(shù)據(jù) 資料來源：聚源數(shù)據(jù)凈資產(chǎn)收益率因子對(duì)羊群效應(yīng)影響2014H22015H12015H22016H12016H22017H12017H22018H12018H22019H12019H22020H12020H22021H12021H22014H22015H12015H22016H12016H22017H12017H22018H12018H22019H12019H22020H12020H22021H12021H22022H12022H22014H22015H12015H22016H12016H22017H12017H22018H12018H22019H12019H22020H12020H22021H12021H22022H12022H組1 組2 組3 組4 組5 組6 組7 組8 組9組10資料來源：聚源數(shù)據(jù) 資料來源：聚源數(shù)據(jù)2014H22015H12015H22016H12016H22017H12017H22018H12018H22019H12019H22020H12020H22021H12021H22014H22015H12015H22016H12016H22017H12017H22018H12018H22019H12019H22020H12020H22021H12021H22022H12022H22014H22015H12015H22016H12016H22017H12017H22018H12018H22019H12019H22020H12020H22021H12021H22022H12022H2組1 組2 組3 組4 組5 組6 組7 組8 組9組10

組1 組2 組3 組4 組5 組6 組7 組8 組9組10資料來源：聚源數(shù)據(jù) 資料來源：聚源數(shù)據(jù)從圖中可以看出：上期凈資產(chǎn)收益率影響：羊群效應(yīng)：31如圖33整體呈U賣出羊群效應(yīng)：如圖35所示上期凈資產(chǎn)收益率較低的組賣出羊群效應(yīng)較高。當(dāng)期凈資產(chǎn)收益率影響：與上期凈資產(chǎn)收益率影響相似。2014H22015H12015H22016H12016H22017H12017H22018H12018H22019H12019H22020H12020H22021H12021H22014H22015H12015H22016H12016H22017H12017H22018H12018H22019H12019H22020H12020H22021H12021H22022H12022H22014H22015H12015H22016H12016H22017H12017H22018H12018H22019H12019H22020H12020H22021H12021H22022H12022H2組1 組2 組3 組4 組5 組6 組7 組8 組9組10

組1 組2 組3 組4 組5 組6 組7 組8 組9組10資料來源：聚源數(shù)據(jù) 資料來源：聚源數(shù)據(jù)圖37上期凈資產(chǎn)收益率對(duì)群效應(yīng)整體影響 圖38本期凈資產(chǎn)收益率對(duì)群效應(yīng)整體影響資料來源：聚源數(shù)據(jù) 資料來源：聚源數(shù)據(jù)市值因子對(duì)羊群效應(yīng)影響2014H22015H12015H22016H12016H22017H12017H22018H12018H22019H12019H22020H12020H22021H12021H22014H22015H12015H22016H12016H22017H12017H22018H12018H22019H12019H22020H12020H22021H12021H22022H12022H22014H22015H12015H22016H12016H22017H12017H22018H12018H22019H12019H22020H12020H22021H12021H22022H12022H組1 組2 組3 組4 組5 組6 組7 組8 組9組10資料來源：聚源數(shù)據(jù) 資料來源：聚源數(shù)據(jù)2014H22015H12015H22016H12016H22017H12017H22018H12018H22019H12019H22020H12020H22021H12021H22014H22015H12015H22016H12016H22017H12017H22018H12018H22019H12019H22020H12020H22021H12021H22022H12022H22014H22015H12015H22016H12016H22017H12017H22018H12018H22019H12019H22020H12020H22021H12021H22022H12022H2組1 組2 組3 組4 組5 組6 組7 組8 組9組10

組1 組2 組3 組4 組5 組6 組7 組8 組9組10資料來源：聚源數(shù)據(jù) 資料來源：聚源數(shù)據(jù)從圖中可以看出：上期平均流通市值影響：羊群效應(yīng)：（9買入羊群效應(yīng)：與上述羊群效應(yīng)相似，且單調(diào)性更為明顯。賣出羊群效應(yīng)：與上述羊群效應(yīng)相似，但單調(diào)性略弱。當(dāng)期平均流通市值影響：與上期平均流通市值影響相似。2014H22015H12015H22016H12016H22017H12017H22018H12018H22019H12019H22020H12020H22021H12021H22014H22015H12015H22016H12016H22017H12017H22018H12018H22019H12019H22020H12020H22021H12021H22022H12022H22014H22015H12015H22016H12016H22017H12017H22018H12018H22019H12019H22020H12020H22021H12021H22022H12022H2組1 組2 組3 組4 組5 組6 組7 組8 組9組10 組1 組2 組3 組4 組5 組6 組7 組8 組9組10資料來源：聚源數(shù)據(jù) 資料來源：聚源數(shù)據(jù)圖45上期平均流通市值對(duì)群效應(yīng)整體影響 圖46本期平均流通市值對(duì)群效應(yīng)整體影響資料來源：聚源數(shù)據(jù) 資料來源：聚源數(shù)據(jù)換手率因子對(duì)羊群效應(yīng)影響圖47上期平均換手率對(duì)羊效應(yīng)影響 圖48本期平均換手率對(duì)羊效應(yīng)影響2014H22015H12015H22016H12016H22017H12017H22018H12018H22019H12019H22020H12020H22021H12021H22022H12022H

2014H22015H12015H22016H12016H22017H12017H22018H12018H22019H12019H22020H12020H22021H12021H22022H12022H資料來源：聚源數(shù)據(jù) 資料來源：聚源數(shù)據(jù)2014H22015H12015H22016H12016H22017H12017H22018H12018H22019H12019H22020H12014H22015H12015H22016H12016H22017H12017H22018H12018H22019H12019H22020H12020H22021H12021H22022H12022H22014H22015H12015H22016H12016H22017H12017H22018H12018H22019H12019H22020H12020H22021H12021H22022H12022H2組1 組2 組3 組4 組5 組6 組7 組8 組9組10

組1 組2 組3 組4 組5 組6 組7 組8 組9組10資料來源：聚源數(shù)據(jù) 資料來源：聚源數(shù)據(jù)從圖中可以看出：47買入羊群效應(yīng)：與上述羊群效應(yīng)相似，但單調(diào)性略弱。賣出羊群效應(yīng)：與上述羊群效應(yīng)相似，且單調(diào)性明顯。當(dāng)期平均換手率影響：與上期平均換手率影響相似。2014H22015H12015H22016H12016H22017H12017H22018H12018H22019H12019H22020H12014H22015H12015H22016H12016H22017H12017H22018H12018H22019H12019H22020H12020H22021H12021H22022H12022H22014H22015H12015H22016H12016H22017H12017H22018H12018H22019H12019H22020H12020H22021H12021H22022H12022H2組1 組2 組3 組4 組5 組6 組7 組8 組9組10

組1 組2 組3 組4 組5 組6 組7 組8 組9組10資料來源：聚源數(shù)據(jù) 資料來源：聚源數(shù)據(jù)圖53上期平均換手率對(duì)羊效應(yīng)整體影響 圖54本期平均換手率對(duì)羊效應(yīng)整體影響資料來源：聚源數(shù)據(jù) 資料來源：聚源數(shù)據(jù)成交金額因子對(duì)羊群效應(yīng)影響圖55上期平均成交金額對(duì)群效應(yīng)影響 圖56本期平均成交金額對(duì)群效應(yīng)影響2014H22015H12015H22016H