文稿分析成果lecture4market efficiency

1、REMINDER:MID-TERM EXAMWednesday 5th November2.15 pmIn room CB1.10Multiple choice questions: 30 questions in 45 minutesSyllabus is from material from weeks 1-4Lecture notes, readings in textbooks, class exercisesConstitutes 20% of unit assessmentStudents with special arrangements Contact me or Assess

2、ment OfficeMN50322 Investment ManagementSTOCK MARKET EFFICIENCYPortfolio selectionMarket efficiencyTypes of market efficiency Anomalies and behavioural financeReading: Brown/Reilly Ch 6; Lofthouse Ch 5/6MN50322 Investment ManagementPortfolio SelectionFour Principles of portfolio selection:Investors

3、want high expected returns and low standard deviationchoose from set of efficient portfoliosRisk of a stock depends on its contribution to risk of the portfolioCAPM or other asset pricing modelBeta measures marginal contribution of stock to the risk of the market portfolio Beta is a measure of the s

4、tocks sensitivity to movements in the market portfolioTwo fund separation: investors holds a portfolio of just two fundsthe market portfolio and the risk-free asset MN50322 Investment ManagementTwo-fund separationConservative investorMarket PortfolioCashrfAggressive investorStocksBondsEs MN50322 Inv

5、estment ManagementAll investors hold wealth in two funds: risk-free rate and market portfolioPortfolio SelectionProxy for market portfolio is to replicate an index portfolioCreate own portfolio that mimicks portfolio weights of stock market indexAlternatively invest in tracker/index fundsFunds that

6、mimick the return on a market indexPassive investment managementRelies on concept/belief in stock market efficiencyNo possibility of earning abnormal returnsContrast active vs passive managementKay Report (2012) comments on the rise of passive investment, but it is difficult to estimate the proporti

7、onMN50322 Investment ManagementStock Market EfficiencyCan a fund manager earn abnormal returns on the stock market?The informational efficiency of the stock market provides an answer to this question.If stock markets are informationally efficient then answer is:NO fund managers cannot earn ARsOn the

8、 other hand if markets are not efficient then fund managers might be able to make ARs?MN50322 Investment ManagementValuation of Assets Under null hypothesis of stock market efficiency, stock returns are unpredictable, because stocks are valued correctlyAsset pricing modelIf assets are incorrectly va

9、lued (mis-valued) then trading rule of the form: buy undervalued asset and/or sell overvalued asset, will make moneyMN50322 Investment ManagementAsset Pricing Assets are valued such that the required rate of return on an asset compensates the holder for the opportunity cost of capitalRequired Real R

10、ate of Return is the extra return over and above the return on treasury bills that the asset is expected to yield:ERi = rf + risk premiumRequired Nominal Rate of Return on an asset isEconomys real risk-free rate of returnExpected rate of inflation during holding periodRisk premium determined by the

11、uncertainty of returns How do we estimate the risk premium?Models usually derive an expected return for a stock, E(R), depending on the stocks risk CAPM APT MN50322 Investment ManagementStock Market EfficiencyMarket Efficiency is a conceptEfficient Markets Hypothesis (EMH) states that stock prices r

12、eflect information.If markets are efficient then new information is reflected quickly into market prices.Conversely, if markets are inefficient information is reflected only slowly into market prices, if at all.MN50322 Investment ManagementExamples: QuestionsIn 1998 Charlton played Sunderland in the

13、 first division play-offs, and the winner was promoted to the Premier League. On the Monday after the Sunday game Charltons share price rose, and Sunderlands fell. Who won? In 1990 Polaroid Corp. received a $924 million settlement from Eastman Kodak Co. in following a long-running patent-infringemen

14、t lawsuit: Polaroids stock price fell, and Kodaks rose. Why?MN50322 Investment ManagementAnswersCharlton beat Sunderland and won promotion to the Premier LeagueGenerating increased profits in gate receipts, TV revenues and sponsorship Dispute had been running for 20 years, and anticipated settlement

15、 had been incorporated into stock prices of both companiesSince Polaroid stock price fell, market must have been expecting a higher settlement value to be paid to PolaroidFinal settlement was better than expected for Kodak, and worse than expected for PolaroidMN50322 Investment ManagementForms of EM

16、H: Information structureIn order to provide a practical definition of market efficiency it is necessary to define the information structureThree forms of the EMH:1)Weak form2)Semi-strong form and3)Strong form.Fama (1991) redefines these forms as predictability, event studies and inside informationMN

17、50322 Investment ManagementWeak Form EfficiencyIf stock prices are weak form efficient, past prices contain no information about future changesprice changes are random.Kendall (1953) found that stock and commodity prices follow a random walkAutocorrelation tests (Lo & MacKinlay, 1988)rt = rt-1 + tRu

18、n tests (Defusco et al, 2004)Trading rule tests (Brock, Lakonishok, LeBaron, 1992)MN50322 Investment ManagementMN50322 Investment ManagementMN50322 Investment ManagementIntuition for Random WalkA random walk implies zero correlation between price change at t and price change at t+1, which is what we

19、 observe.If price cycles were predictable competition between investors would eliminate them.Arbitrage/Speculation will force prices to their efficient valuesSimple trading rule: BUY undervalued assetsSell overvalued assetsIf efficient, prices will only change on the basis of new information which b

20、y definition is randomhence price changes are random.MN50322 Investment ManagementSemi-Strong Form EfficiencyIf prices are semi-strong form efficient then prices reflect all public information .MN50322 Investment ManagementTest using: Event Study MethodologySample of firms that have experienced type

21、 of announcement, For each firm in sample, compute the abnormal returnARjt = Rjt - E(Rjt)where E(Rjt) is the expected return is estimated from an appropriate asset pricing model: CAPM, APT, Market Model etc. Asset pricing model is normally estimated using data in a pre-event estimation period.For an

22、y date around the announcement, compute average abnormal return across companies in the sample (AAR), and the cumulative average abnormal return over time (CAAR)AARt =j ARjt and CAARt = t AARt4. Test whether this average abnormal return is statistically significantly different from zeroMN50322 Inves

23、tment ManagementEmpirical EvidenceExamples:Stock splits Fama, Fisher, Jensen and Roll (1969) Capital restructures Masulis (1980)Dividend/Earning announcements Ball and Brown (1968), Rendleman, Jones and Latane (1982)Merger announcements Jensen and Ruback (1983)New Issues/Secondary Issues Smith (1986

24、)Empirical finding is that prices do react to information consistent with semi-strong form market efficiency MN50322 Investment ManagementPrice Reactions to Earnings AnnouncementsEvidence of post-earnings announcement drift, andPre-announcement drift (insider trading or news management?)MN50322 Inve

25、stment ManagementStrong Form EfficiencyIf prices are strong form efficient, all private information is reflected in pricesTwo contexts:When corporate insiders trade do they earn ARsProfessional money managersIn fact insider trading by directors is profitable Gregory, Matatko, and Tonks (1997) in lon

26、g-run; Friederich et al (2002) in short-run; Gregory, Tharyan and Tonks (2011) value/growthsuggesting markets are not strong form efficient, butBut performance of mutual funds do not generate abnormal returnsJensen (1968), Ippolito (1989), Kosowski et al (2006), Cuthbertson et al (2008)consistent wi

27、th strong form efficiency. Hence conflicting evidence on whether stock prices are strong form efficient MN50322 Investment ManagementPartial Conclusions?Unlikely that markets are strong form efficientBut early evidence (pre-mid 80s) prices seem to be weak and semi-strong efficientBUTAnomalies to wea

28、k form EMH (post-mid 80s)Behavioural financeRobert Shiller and Eugene Fama both awarded Economics Nobel Prize in 2013Fama supports EMHShiller rejects EMHMN50322 Investment ManagementSmall firm/January effect/PE ratiosBasu (1977,1983) identified P-E ratios as predictors of subsequent performance. In

29、particular high PE firms (glamour/growth stocks) have lower returns than low PE firms (value stocks)Also Banz (1981) and Reinganum (1981) suggested that this PE effect was related to firm size, that small firms tend to outperform large firms even after an allowance is made for the likely riskier cha

30、racteristics of small firms. Also relation between January effect and small firm effect.MN50322 Investment ManagementSmall firm/January effect/P-E ratiosMN50322 Investment ManagementAggregate levelExcess Volatility/Predictability/Mean ReversionShiller (1981) found that stock prices are excessively v

31、olatile compared to those given by the PV modelAlso Bulkley and Tonks (1989, 1992) for UKFama and French (1988) found that long run returns (5-year horizon) are predictable negative correlation: mean reversion Poterba and Summer (1988) find that eight year (long horizon) returns are only four times

32、more variable than one year returns, implying negative serial correlation.But one month (short horizon) returns are only 80% as variable as twelve month returns, implying positive serial correlation.Campbell & Shiller (1998, 2005) find evidence of mean reversion in PE ratios and DYs: when P-E ratio

33、hits bounds there is an adjustment in pricesMN50322 Investment ManagementMean Reversion in P-E ratiosUS Real S&P IndexUS S&P P-E ratioMN50322 Investment ManagementContrarian/Value strategiesDe Bondt and Thaler (1985) found that portfolios of loser portfolios outperform and winner portfolios underper

34、form subsequentlyMN50322 Investment ManagementLakonishok, Shleifer and Vishney (1994) examine “contrarian” strategy in more detail.“Value Strategies” buying stocks that have low prices relative to some measure of value (i.e. earnings, dividends, historical prices, or book assets). Vs “Glamour strate

35、gies” selling stocks that have high prices relative to valueFinds that value stocks significantly outperform glamour stocks over next five years (by 11% per year in Year +5)In a further test, La Porta, Lakonishok, Shleifer and Vishney (1997) examine whether glamour stocks have negative three-day ret

36、urns around subsequent earnings announcements, and value stocks have positive returns. This would be consistent with the market having the wrong expectations initiallyMN50322 Investment ManagementPanel C shows annual returns, and Panel A shows 3-day returns around earnings announcement dates MN50322

37、 Investment ManagementFama & French (JF 1998): value effect is an international phenomenonMN50322 Investment ManagementValue-Growth anomaliesIs it due to:Behavioural explanations: For example, naive strategies such as extrapolating past earnings growth, or over-reacting to news so that “glamour stoc

38、ks” are overpriced, Barberis, Schleifer and Vishney (1998, JFE) model of investor sentiment: 情緒innate conservatism先天保守: investors react slowly to new evidence“representativeness heuristic”: prone to identify patterns in random sequencesAre value stocks fundamentally riskier?Typically we think of val

39、ue stocks as boring and safeBut Zhang (2004, JF) argues that investment irreversibilities make value stocks riskier than glamour stocksMN50322 Investment ManagementMomentum strategies動(dòng)量策略 (Relative Strength)Form portfolios on basis of past performance, on basis of short-run positive autocorrelationJegadeesh