Investments

Lakonishok and Smidt 1988

- tests seasonal ret anomalies over a long period of time using uniform database and methodology -skeptical of past lit regarding seasonal anomalies b/c it considers selection bias, noise, and data snooping -examines monthly, semimonthly, weekend, holiday, end of dec, and turn of month seasonalities -DJIA 1897-2986 which concentrates on large actively traded firms; useful index for representing short term market movements -rate of ret on monday is substantially neg -price increase around turn of month exceeds total monthly price increase -rate of ret before holidays is more than 20 times normal rate of ret (especially b/n christmas and new years) -no consistent monthly pattern in rets, or any consistent tendency for rets in first part of month to be higher -monday, holiday, and end of month rets deviate from normal average daily rets by less than bid ask spread of average stock

Markowitz 1991

-3 ways portfolio differs from theory of firm to theory of consumer 1) concerned w/ investors rather than manufacturing firms or consumers, 2) concerned w/ economic agents who act under uncertainty, 3) thoeyr which can be used to direct practice, at least by large investors w/ sufficient computer and database resources -assumes investors face known probability distr -investors concerned abt risk and return measured as a portfolio whole -discusses other papers dealing with pt -not all expected maximizers are equally served by mean-var approximations -some utility functions mean-var approximation is so good virtually no room for improvement -still need a lot done

Fama French 2015

-5-factor model directed at capturing size, value, profitability, and investment patterns in average stock rets performs better than 3-factor -main problem is failure to capture low average rets on small stocks whose rets behave like those of firms that invest a lot despite low profitability -RMW: diff b/n rets on diversified ports of stocks w/ robust and weak profitability -CMA: diff b/n rets on diversified ports of stocks of low and high investment firms (conservative and aggressive -no size effect in highest quintile of inv -in various mehtods used for factor construction, seem to be expected profitability and investment premiums for small stock; weak evidence for profitability and strong for investment -use GRS stat (gibbons, ross, shanken 1989) to test hypothesis that intercept is diff from 0 -hml is redundant in 5-factor in describing average rets b/c average hml ret is captured by exposures of it to other factors -1963-2013 - HMLO: sum of intercept and residual from regression of hml on other factors, then substitute for hml -other 4 have slopes that reflect they capture hml info -like ff93, ports of small extreme growth stocks produce neg 3-factor intercepts and prots of large extreme growth produce po ones -RMW and CMA slopes say port is dominated by microcaps whose returns behave like those of unprofitable firms that grow rapidly -3-factor will have issues when ports have strong tilts to high or low profit or inv

Options

-Black and Scholes 1973 -Merton 1973 -Conrad 1989 -Mayhew and Milhov 2004 -Goyal and Saretto 2009 -Roll, Schwartz, Subrahmanyam 2010 -Hu 2014

Index/ETFs

-Boehmer and Boehmer 2003 -Chen, Noronha, Singal 2004 -Agapova 2011 -Green and Jame 2011 -Evans and Fahlenbrach 2012

Professional investors

-Carhart 1997 -Cremers and Petajisto 2009 -Fama and French 2010 -Barras, Scaillet, Wermers 2010 -Kacperczyk, Van Nieuwerburgh, Veldkamp 2014 -Gennaioli, Shleifer, Vishny 2015 -Cremers, Ferreira, Matos, Starks 2016 -Dyck, Lins, Roth, Wagner 2019

Momentum

-DeBondt and Thaler 1985 -DeBondt and Thaler 1987 -Jegadeesh and Titman 1993 -Lakonishock, Shleifer, Vishny 1994 -Chan, Jegadeesh, Lakonishok 1996 -Daniel, Hirshleifer, Subrahmanyam 2998 -Jegadeesh and Titman 2001 -Grinblatt and Han 2005 -Hirshleifer, Lim, Teo 2009 -Moskowitz, Ooi, Pedersen 2012 -Novy-Marx 2012 -Asness, Moskowitz, Pedersen 2013 -Daniel and Moskowitz 2016

AFA presidential addresses

-Duffie 2010 -Titman 2013 -Stambaugh 2014 -Zingales 2015

Market efficiency and return predictability

-Fama 1970 -Fama 1991 -Keim 1983 -Lakonishock and Smidt 1988 -Fama and French 1989 -Campbell and Yogo 2006 -Avramov and Chordia 2006 -Ang and Bekeart 2007 -Stambaugh, Yu, Yuan 2012 -Fama and French 2002 -Stein 2009 -Welch and Goyal 2008 -McLean and Pontiff 2016

Investor awareness

-Grinblatt and Keloharju 2001 -Tetlock 2007 -Barber and Odean 2008 -Guiso, Sapienza, Zingales 2008 -Fang and Peress 2009 -Da, Engelberg, Gao 2011 -Engelberg and Parsons 2011 -Engelberg, Sasseville, and Williams 2012 -Tetlock 2011 -Yermack WP

Bonds

-Harris and Piwowar 2006 -Edwards, Harris, Piwowar 2007 -Starks, Yong, Zheng 2006 -Butler2008 -Ronen and Zhou 2013

Other papers

-Hong and Kacperczyk 2009 -Van Rooij, Lusardi, Alessie 2011 -Cohen, Diether, Malloy 2013 -Riedl and Smeets 2017

Behavioral

-Odean 1999 -Huberman 2001 -Hirshleifer and Shumway 2003 -Baker and Wurgler 2006 -Frazzini 2006 -Barberis and Xiong 2009 -Grinblatt, Keloharju, Linainmaa 2012

Asset Pricing

-Roll 1977 -Markowitz 1991 -Fama and French 1992 -Fama and French 1993 -Fama and French 1996 -La Porta 1996 -Ang, Hodrick, Xhing, and Zhang 2009 -Fu 2009 -Lewellen, Nagel, Shanken 2010 -Frazzini and Pederson 2014 -Fama and French 2015 -Hou, Xue, Zhang 2015 -Havey, Liu, Zhu 2016 -Campbell, Giglio, Polk, Turley 2018

Engelberg, Sasseville, Williams 2012

-Use show Mad Money to test theories of attention and limits to arbitrage -recommendations lead to large overnight rets that subsequently reverse over next few months -spike reversal pattern is strongest among small, illiquid stocks that are hard to arbitrage -uses daily Nielsen ratings as a direct direct measure of attetion finds that overnight ret is strongest when high income viewership is high -weak price effects among sell recommendations -evidence supports retail attention hyp of barber odean 2008 -2005-2009 data -match most recs to size and b/m quintiles based on nyse stocks -recs are disproportionately in low b/m quintile and are often recent winners -proxy limits to arbitrage: illiquidity, size, and idio vol

Stambaugh 2014

-addresses current trends in equity market by developing theoretical model that explains the decline in equity market owned directly by individuals and predicts future trend -objective: determine whether trends that have been present since the 1980s in investment management are consistent w/ downward trend in individual equity ownership and associated decline in noise trading -hyp: less noise trading implies a lower capacity for profitable active management -model implies that active management corrects most of the noise trader-induced mispricing that would ow exist -fraction of mispricing remains uncorrected b/c active managers impeded by both trading costs and idio risk -remaining mispricing allows active management to earn pos alpha at expense of noise traders -w/ less mispricing, active management becomes less necessary -1980-2012, noise traders or those who would be better off not investing have declines -b/c of this active management has also declined and passive investing or indexing has increased

Ronen and Zhou 2013

-aims to address some of the difficulties inherent in comparison of info efficiency across markets -use VAR to revisit the lead-lag relation b/n stocks and bonds; examine overnight liquidity; analyze effect of trade size on sensitivity to info and impact of timing of earnings announces -when institutional trade dominance and other bond features are accounted for, stock leads in earlier studies disappears -bivariate var approach may not always be optimal in addressing info eff of corp bond market -info incorporation patterns differ systematically across diff traders and bonds -top bond: bond that has highest institutional trade volume after firm earnings announcement - identity of top bonds changes over time and share common characteristics, such as age, maturity, credit quality, and complexity -large trades most informative around earnings announces -trace 2003-2006 -use analysts forecasts (ibes) and earnings announcement time stamps (factiva) -show importance of accting for several methodological issues: 1) cant exclude overnight bond trades cuz it biases towards ignoring relative efficiency, 2) pooling retail and inst trades can confound inferences, 3) var around corp earnings announces when info effects should be heightened -use logistic model to predict which bond will be "top" -4 predictor dummies: 1) on the run long term debt, 2) has at least 1 special feature, 3) one of issuers 3 most liquid bonds, 4) issuers highest rated bond before earnings announce -use weighted price contribution measure to examine contribution of top bonds overnight trades to bonds price discovery

Barras, Scaillet, Wermers 2010

-aims to estimate prevalence of skilled fund managers, how many managers possess truck stock picking skills, where these funds are located in the cross section (estimated) alpha distr, and impact of luck on long term and short term performance -data from monthly rets of over 2000 actively managed us open end equity mutual funds b/n 1975-2006 -use false discovery rate to estimate frequencies of false discoveries: funds w/ sig estimated alphas -separates funds into 1) unskilled, 2) 0 alpha, 3) skilled -more precisely estimates proportions of unskilled/ skilled funds in population and their location in the cross sectional estimated alpha distr -only .6% are skilled; 75.4% zero alpha -proportion of skilled fund managers has diminished rapidly over past 20 years -actively managed mutual fund performance is due to long term survival of a minority of truly underperofrming funds -rather than arbitrarily impose a prior assumption as in past studies, approach estimates w/ straight forward computation that uses p-values of individual fund estimated alphas -uses monte carlo experiment, show that approach provides a much more accurate partition of universe of mutual funds into zero alpha unskilled and skilled -baseline moedle 4-factor

DeBondt and Thaler 1985

-aims to present empirical tests of the overreaction hyp and test whether whether it is predictive -present empirical tests are the first attempt to use behavioral principle to predict a new market anomaly -2 hyp imply a violation of weak form market eff: 1) extreme movements in stock prices will be followed by subsequent price movements in the opposite direction; 2) ports of losers experience exceptionally large jan rets as late as 5 years after port formation -price reversals measured by diff in cumulative average residual rets b/n winner and lower ports -consistent w/ prediction of overreaction hyp, ports of prior losers are found to outperform prior winners -ports of losers experience exceptionally larger jan rets -overreaction effect is asymmetric and much larger for losers than winners -the phenomenon is qualitatively diff from jan effect and from seasonality in stock prices -1926-1982

Yermack wp

-analyze changes in apparel company stock price when M Obama wears designer outfits at public events -prices rise sig when she wears their clothing, increasing more for more closely watched appearances and less so by routine events -internet users search the names of individual designers w/ higher frequency after she wears their clothes at major events, associated w/ surges in online sales -could be a form of overreaction of small investors caught up in event; but data don't support b/c price increases appear to persist for weeks -use database of postings on internet blogs that track her schedule of appearances and outfits -event study with indicator of 1 for days in an event window; estimate market model parameters and event study effects simultaneously in one equation avoids overlap of estimation periods and event periods b/c many firms had multiple events -use google for search info on designers -evidence of informed trading by ab rets before event -gains stronger for us companies -none of her fave designers donated to obama's opponent in election -clothing manufacturers donated more to democrats than republicans, reversing long pattern

Grinblatt, Keloharju, Linainmaa 2012

-analyze whether IQ influences trading behavior, performance, and transaction costs -comines equity ret, trade, and limit order book data w/ 2 decades of scores from an iq test administered to nearly every Finnish male of draft age -after controlling for many factors, find that high iq investors are less subject to disposition effect, more aggressive about tax loss trading, and more likely to supply liquidity when stocks experience 1 month high -also show superior market timing, stock picking skill, and trade execution -Use application of FM reg to units of observation consisting of each pairing of an investor and trade in a stock -approach facilitates study of iq's marginal effect on performance while controlling for both investor and stock attributes -participation increases with iq -losee interaction coefficients indicate that lowiq investors are less likely to realize cap losses, particularly large ones, than high iq -examin how one iq groups trades influence those of other groups, regress group trading behavior in a stock on a given day (sell v buy v hold ratio) against average trading by all other groups together and against each of the other groups current and lagged excess trading behavior -all investors tend to herd with current and lagged trades of all investors in market -when controlling for wealth quintile, high iq ports outperform lower iq counterparts -for transaction costs use HEX microstructure data -at market close on day of trade, high iq executed limit orders outperform on buy and sell side; indicate limit order of high iq face lower adverse selection costs

Duffie 2010

-analyzes asset pricing dynamics w/ slow-moving capital and develops a model for the dynamics of asset prices in settings w/ inattentive investors -premise is as investors become aware of ways to trade around shocks, avialable capital increases and impact of shock should decline -finds that a higher number of attentive investors reduces the initial price impact and allows the reversal to occur quicker -the more inattentive investors there are, the slower capital will move, and the more incentive attentive investors need to buy it -identifying the time signatures of price impacts and reversals to supply shocks help identify the degree and form of capital immobility

Mayhew and Milhov 2004

-analyzes listing choices made by option exchanges in us, identifies characteristics that made stocks more likely to be selected for option listing, examines how these changed over time as markets matured, and test whether option intro impacts volatility of underlying stock -using listing decisions by us option exchanges 1973-1996, uses a logit framework to measure extent to which the prob of option listing is related to things such as volume, volatility, and market cap -tend to list options on stocks w/ high trading volume, volatility, and market cap -relative effect of these factors has changed over time -shift from volume to volatility after moratorium on new listings ended in 1980 -no evidence that volatility declines w/ option intro, in contrast to previous studies that don't use control samples

Boehmer and Boehmer 2003

-answer question whether increased competition among market centers has beneficial effects on trading costs and price discovery or detrimental effects b/c order flow is dispersed across several locations w/ much interaction -look to april 6 2001 when nyse announced it seeks regulatory approval for trading 3 most active etfs (QQQ, SPY, DIA) -assess market competition on april 15 2002 when nyse began trading an additional 27 etfs -hyp: nyse entry to etf market has no sig effect on trading costs -there are substantially lower trading costs across market centers after nyse enters market: overall liquidity improves b/c of reduction in diff spread measures and quoted depth increases; estimated price impact of trades declines -only infor shares for AMEX drop slightly after nyse enters -smaller proportion of informed trading may explain lower spread on amex but not for the overall etf market -trading costs dont revert back to preutp levels in any of the centers -3 measures of trading costs: quoted spread, effective spread (approximation for total price impact of a trade) and realized spread (app for temp price impact of a trade: market maker revenues net of costs of adverse selection -halve the diff b/n effective and realized for app permanent price impact, used to approximate info content of trades

Barber and Odean 2008

-argues decision to buy or sell stocks does not apply w/ equal force to individual investors who are net buyers of attention grabbing stocks; doesn't apply to institutional investors, since attention isnt as scarce -hyp: buying behavior of ind. investors is more heavily influenced by attention than their selling behavior; ind investors more influenced than prof investors -attention measures: 1) news, 2) unusual trading vol, 3) extreme rets -4 samples of investors: 1) accts at large disct brokerage, 2) smaller disct brokerage firm that advertises its trade execution quality, 3) accts at large retail brokerage, 4) professional money managers -data from 1) large disct brokerage, 2) small disct brokerage, 3) large full-service brokerage, 4) plexus group -high volume day: ratio of day trading vol to average vol over previous year -calculates buy-sell imbalance that day -sorts on rets and news -individual investors solve search problem by buying on high vol days, following extreme one day rets, and when stocks are in news -prof investors are less prone to this -attention-driven buying patterns dont generate superior rets -investors more likely to buy rather than sell attention grabbing stocks regardless of size

Roll 1977

-argues no correct test of 2-parameter asset pricing theory w/ no possibility of being done in future -previous models: Black, Jensen, Scholes 1972; Fama MacBeth 1973; Blume Friend 1973 -Finds it is difficult to test apt b/c it is difficult to include all compositions (individual assets) of the true market portfolio -only single testable hypothesis association w/ generalized 2-parameter apm is hypothesis that the market portfolio is mean-variance efficient -use proxy for market portfolio is hard b/c it could be mean-var efficient when market isn't -testing it by using linear relation b/n return and beta is hard b/c widely used portfolio grouping procedure can support theory even when it is false -from 2 different market portfolios, same security could have different values of beta

Frazzini 2006

-argues that presence of large subset of investors who display disposition effect can generate stock price underreaction to news and in turn ret predictability and post announcement price drift -test by using a measure of unrealized cap gains, constructed by using time series of net purchases by mutual fund managers and their cost basis in a stock to compute a weighted average reference price -analyzed transmission of info when firm specific info is released in form of public news -stock rets and accting data b/n 1980-2002 obtained from crsp and compustat; mutual fund data from Thomson Financial -disposition effect can induce underrreaction to news, leading to ret predict and post announce price drift -stocks w/ large unrealized cap gains have higher subsequent rets and underreact only to positive news -stocks w/ large unrealized cap loss underreact only to neg news -both consistent w/ disposition effect b/c investors reluctant to realize loss -loser funds show signs of disposition effect w/ magnitudes comparable to retail investors -bulk of profitability of the PEAD is concentrated in high overhang stocks: those w/ large unrealized cap gains tend to underreact to and only to pos earnings surprises, while neg overhang underreact only to neg news

Evans and Fahlenbrach 2012

-argues that whether or not redeemable claims safegaurd investors depends on whether investors use the correct criteria to evaluate funds -objective: determine whether retail investors can benefit from ability and willingness of inst investors to exercise market governance -examines performance of a subset of retail mutual funds that offer a version of the fund for inst investors -propensity score matching -treatment group is retail group that created a inst twin -examine change before and after creation on 3 channels: 1) direct expenses (expense ratio), 2) indirect expenses (return gap: diff b/n funds actual ret and ret on hypothetical buy and hold port of funds most recently disclosed holdings net of expenses) 3) manager effort (active share) -inst investor flows are more sensitive to high fees and poor risk adjusted performance than retail flows -consistent w/ greater monitoring on part of inst twin investors, retail fund performance increases w/ inst twin -morningstar database 1996-2009

Fama and French 1989

-assess current debate on whether predictability of stock and bond rets implies that market is inefficient or whether this is a result of rational variation in expected rets -main questions: 1) if expected rets on bonds and stocks move together and if same variable forecast both rets; 2) is variation in expected stock and bond rets related to business conditions -use 3 explanatory vars w/ objective to measure and interpret variation in expected excess rets for return horizons of 1 month, 1 quarter, and 1 to 4 years -excess rets measured as rets net of one month tbill rate -variables used to predict variation in stocks and bonds: 1) dividend yields (d/p), 2) default premium, 3) maturity premium -excess rets for both stocks and bonds are inversely related w/ business conditions throughout time -rets are high during periods of uncertainty or vol such as recessions and lower during periods of prosperity -variation of expected rets has rich mix of components that relate to both long and short term aspects of bus conditions -d/p and def imply that they track variation in expected rets that is largest for stocks and smallest for high grade bonds -spread and expected rets high in weak economy -term has similar slopes for stock and bond -term captures a premium in expected rets that is largely a function of maturity and so is similar for all long term securities -if variation in expected rets is common to different securities, then it probably rational result of variation in tastes for current vs future consumption or investment opps of firms

Dyck, Lins, Roth, Wagner 2019

-assesses whether shareholders drive e and s performance of firms worldwide -41 countries, institutional ownership is pos associated w/ e and s performance (causal evidence). -institutions are motivated by both financial and social rets -investors increase firms e and s performance following shocks that reveal financial benefits to e and s improvements -in the cross section, investors increase firms e and s performance when they come from countries w; a strong community belief in importance of those issues, but not otherwise -construct firm level e and s performance measure using line items from several diff data providers; combine measures of firm e and s performance w/ institutional ownership data and financial data 2004-2013 -find greater institutional is associated w/ increase in score for performance -use quasi-natural experiment of 2010 bp oil spill -if institutional ownership drives changes in firms environmental policies, then firms w/ greater inst ownership at time of shock are expected to be more reactive in improving environ performance afterwards; find this -to capture scope of e and s improvement, split into low and high (below and above median); expect greater effects in firms w/ initial low -mechanisms to push for e and s changes: 1) exit and selection (no), 2) voice (no), 3) private engagement most likely

Fama 1991

-attempt at comprehensive review of empirical studies on market eff hyp from Fama 1970 -replaces the 3 relevant info subsets with 3 new categories: 1) return predictability, 2) event studies, 3) tests for private info -postulates that costless info and costless trading are extreme assumptions -chooses power of a joint test b/n market eff and equilibrium apm as benchmark for efficacy of a model -finds that empirical studies since 1970 show that expected rets are large, non-mean reverting, and follow slowly decaying patterns -variation in expected rets is common across securities and markets, and is related in plausible ways to business conditions; leans toward conclusion that if it is real it is rational -extends predictability of past rets to var and autocorrelations of past stock rets -over long horizons, predictability of rets from past rets is more stable w/ high neg autocorrelation -ubiquitious problem in time series tests of market eff is that irrational bubbles in stock prices are indistinguishable from rational time varying expected rets -event studies show that market eff responds to new info; thus provide the most convincing proof of market eff; mostly due to efficient ways that prices adjust to firm specific events (div announcements, stock issues, tender offers, etc) -contends that possession of private info can lead to ab rets -tests on possession of private info by mutual funds and pension fund managers involve long run ab rets test; there is evidence of strong form

Harvey, Liu, Zhu 2016

-broad analysis of factors that influence capm since first appeared -assess appropriateness of stat sig standards as they apply to factors that have predictive power of expected retursn -assess factors proposed in 313 papers published regarding cross sectional ret patterns -uses multiple testing framework from fequentist perspective to re-evaluate past research and provide new benchmarks -focus on family-wise errors and false discover errors by using p-value adjustments: bonferroni, holm, and benjamini, hochberg, and yekutieli -a tstat of 2 is too low for current research -after taking multiplicity into account, minimum threshold tstat for 5% sig is about 3

Fama 1970

-builds theoretical and empirical basis for testing eff market hyp -efficient market: price of a security fully incorporates "all available info" -empirical basis of this paper divided into 3 relevant info subsets: weak, semi strong, and strong forms -Weak form: historical rets are reliable predictors of current rets -semi strong: ab ret can be earned from publicly available info -stron form: some insiders or investors have access to private info at detriment of other investors -empirical analysis of market eff is also a fair game -random walk and submartingale models are special cases of fair game -random walk hyp: idea that degree of dependence in the series of stock price changes isnt sufficient to allow past prices to predict future price in a way that can earn abnormal profit -empirical studies support weak form tests; evidence against doesnt hold if transaction fees considered -event studies of stock splits, annual earnings, new issues all show that info in these event studies concerning firms future rets are reflected in stock price; semi strong supported -strong form test not supported; 1 example: specialists on NYSE use their access to info to generate excess profits -expected return model: expected ret on a security is a function of its risk but differ in how risk id defined -fair game model: conditions of market can be stated in terms of expected rets (investors will not invest unless at least a fair game) -submartingale model: expected value of next period's price, as projected based on info, is equal to or greater than current price -random walk: current price of a security fully reflects all available info and assumes successive price changes are iid

Fama and French 1993

-captures common variation in stocks and bond returns in terms of risk factors - 3 for stock and 2 for bonds -excess market portfolio returns, size, b/m; maturity risk (term) and default risk (def) -want to answer question about average returns of cross section variations in risk factors -use time series regression of Black, Jensen, Scholes; give direct evidence on whether if assets are priced rationally, variables that are related to average returns must proxy for sensitivity to common risk factors in rets -use excess rets as DV and either excess rets or rets on 0-investment ports as explanatory vars -construct dv by dividing size and b/m vars in quintile cross-tabulation (25 dv) -excess market ret portfolio can't alone adequately explain cross section variations in stocks and bond rets -variations in size and b/m risk factors strongly explain variation -3-factor model is designed to capture relation b/n average ret and size and relation b/n average ret and price ratios

Tetlock 2007

-characterizes relation b/n content of media reports and daily stock market activities by focusing on immediate influence of WSJ "abreast of the Market" colum on us stock rets -use simple measure of media pessimism from content of wsj, provide evidence that news media content can predict movements in broad indicators of stock market activity -main hyp: high media pessimism associated w/ low investor sentiment, resulting in downward pressure on prices -informational theory posits that media pessimism is proxy for neg info abt fundamental values of equities that arent currently reflected in prices -uses general inquirer (gi) , a quantitative content analysis program, to analyze daily variation in wsj column from 1984-1999 -each day in sample, gathers newspaper data by counting the words in 77 predetermined gi categories from Harvard pyschosocial dictionary -high levels of pessimism robustly predict downward pressure on prices followed by reversion to fundamentals -pessimism measures sig predict neg rets to smb factor -neg sentiment seems to be longer lasting and have larger impact on small stocks

Engelberg and Parsons 2011

-compares behaviors of investors w/ access to diff media coverage of same info event to disentangle causal impact of media reporting from impact of event being reported -use zip codes to id 19 mutually exclusive trading regions corresponding w/ major cities -for all earnings announces of s&p 500 index firms, find local media coverage strongly predicts local trading after controlling for earnings, investor, and newspaper characteristics -local trading is strongly related to timing of local reporting -1991-1996 - days of extreme weather sever link b/n locally reported content and local trading -local trading is predicted on exact day of local coverage, controlling for all fixed effects -data: 1) earnings announcements, 2) local media coverage, 3) trading of retail investors -compustat and ibes cross checked -use SUE based on random walk model w/ price as deflator -media coverage from ProQuest's newspaper database -local bias: pool accts of all investors in a city and compute number of local stocks held by this set; then scale by number of local stocks available; compute similar stat for nonlocal firms

Agapova 2011

-compares investment of conventional mutual funds and etfs and study substitutability of their coexistence in the market -2000-2004 and matches etfs w/ index funds tracking same indexes (9 indexes) -conventional funds and etfs are subs, but not perfect -coexistence explained by clientele effect that segregtes the 2 vehicles into diff market niches -etfs may be preferred by tax-sensitive invesotrs while funds may be preferred by tax-exempt investors -in fund industry, demand can be measured by fund flows and price measured by feeds or fee adjusted ret -effectiveness measured as diff b/n fund ret and tracked index ret -tracking error is absolute value of effectiveness var -etfs generally track their underlying indexes more closely than fund counterparts on gross of fees basis -all investors in economy are examined in aggregate and considered as a single representative investor; aggregate flows to conventional funds and etfs are used as indicator of representative investors choice -substitution effect tested w/ 2 models used flowcf and flowetf as dvs -look at clientele effect: consider tax code change in may 2003 to see if sub levels changed

Welch and Goyal 2008

-comprehensively assess performance of vars that have been found to be good predictors of eq prem -objective: reexamine empirical evidence as of early 2006 and to evaluate each var using same methods, time periods, and estimation freqs -equity premium: total rate of ret on stock market minus prevailing short term interest rate -2 variable models: kitchen sink regressions and model selection which considers 2^k models for k vars -use 2 models that are rolling in their independent var construction: consumption, wealth, income ratio -on an annual prediction basis, there is no single var that meets all suggested investment criteria -evidence suggests that most models are unstable or spurious -OOS performance is useful model diagnostic for in sample regressions and is useful in itself for an investor who used the models tested -most of the models have performed poorly for over the past 30 years in sample and any early stat power was ofted based on years up to and on oil shock 1973-1975 -stock rets: CRSP and Shiller's website -risk free rate: tbill and NBER commercial paper

Lewellen, Nagel, Shanken 2010

-critique of why new apm don't add to what we already know about determinants of average and excess return of an asset if we rely on Rsq as marjor test of goodness of fit -size-b/m 25 ports have very strong factor structure that explains 90% and 80% of variations respectively -if factors of proposed models are highly correlated sith smb and hml but are uncorrelated w/ residuals of size-b/m port, using rsq as gof would be misleading -simulate artificial factors which are highly correlated w/ rets but not residuals -quarterly excess rets are calculated from 1963-2004 -obtaining a high cross sectional rsq is not difficult under arbitrage pricing models when returns have strong factor structure -improve new apm: 1) include other ports in tests (ex: sort by industry); 2) impose restrictions on risk premia; 3) use Rsq from gls; 4) use confidence interval -Lettau and Ludvisgon 2001 conditional consumption capm; conditioning var aggregate consumption to wealth ratio CAY -lustig and van nieuwerburgh 2004; conditioning var housing collateral ratio MYMO -Santos and Veronesi 2006; conitioning var labor income to consumption ratio

Hu 2014

-decomposes total stock order imbalance into an imbalance induced by option transactions and an imbalance independent of options -option induced imbalance sig predicts future stock rets in cross section after controlling for past stock and options rets -the imbalance independent of options has only transitory price impact -options order flow contains important info abt underlying stock value -compute option induced stock order imbalance: assumes market makers full delta hedge and customers do not hedge; computes imbalance by aggregating the signed option transactions weighted by respective delta exposure of each option contract -2008-2010 Trade Alert LLC, which matches each option transaction record w/ underlying stock price and computed iv; uses taq -predictive power comes from delta exposures of at the money and in the money options -measurement of info asymmetry are probability of informed trading (easley and ohara 1987), number of stock analysts following, bid-ask spread in stock market, adverse selection component of bid ask spread (glisten and harris 1988), and firm size -doesnt matter which is used, option induced imbalance always shows largest predictive power in group of informationally opaque firms -suggests informed trading in options market drives ret predictability from options order flow -short sale constraints measured by institutional holdings -firms w/ low inst ownership have largest predictive coefficient and highest stat sig for option imbalance -option imbalance predicts better when option trading is active and total options volume is high -event study around earnings announcements; option imbalance predicts cars days before

Merton 1973

-develops a theory of rational option pricing as an intermediate step toward a unified theory to answer questions abt pricing of a firm's liabilities, the term and risk structure of int rates, and theory of speculative markets -BS formulation is examined and an alternative derivation of their formula built to show that alt derivation is value under weaker assumptions -it is possible to develop a theory of term structure of int rates and the approach would also have application in theory of speculative markets -develops 10 theorems to state restriction on rational option pricing -shows that even if an option could be exercised early it wouldnt be

Avramov and Chordia 2006

-develops and applies framework to examine whether predictability of size, b/m, turnover, and past returns is explained by apm (8 diff models) -none of examined models capture impact of liquidity or momentum on cross section of individual stock rets -allosing beta to vary w/ size, b/m, and business cycle vars can improve pricing abilities of models -momentum profits are consistent w/ ap misspecification that varies w/ business cycle -time varying beta versions of multifactor models can capture size and b/m effect -runs regressions of excess stock ret on as factors w/ loadings that may vary cross sectionally and over time with stock level size and b/m ratio as well as w/ macroeconomic vars -cross sectional regs of risk adjusted rets, rather than gross rets, as dv on equity characteristics -null of exact pricing, such attributes should be stat insig in cross section -monthly rets, size, b/m, turnover, and lagged rets for sample of common stocks of NYSE, AMEX, and nasdaq listed companies 1964-2001

Zingales 2015

-devoted to addressing shortcomings of finance as a discipline in last 4 decades -attributes lack of effective answers to these shortcomings to bloated egos of financial economists who often fail to address certain failures of their discipline -identifies some serious problems facing finance discipline, and suggests solutions -shortcomings: anti-finance sentiment among the populace, the ease of engaging in rent-seeking activities in finance, etc -recommendation to these 1) financial economists need to understand limits of their theories and empirical studies 2) academics should act as watchdogs of financial industry 3) academics should incorporate ethics in their curriculum 4) academics should keep their research simple

Fama and French 1996

-disputes previous paper by Kothari, Shanken, Sloan 1995 who disagree with FF 1992 -argue that survivor bias doesn't explain relation that the annual and monthly betas produce the same inferences abt the beta premium and test whether size adds to explanation of average return provided by beta -finds monthly betas produce same inference -for portfolios formed on size and beta, average beta premiums from univariate regs of ret on beta underestimate pos relation b/n beta and average ret produce by size sort and overestimate relation be/n beta and average ret produced by beta sort -beta premium is more basic: it can't save the cAPM and beta alone can't explain expected return

Jagadeesh and Titman 1993

-documents strats which buy stocks that have performed well in past and sell stocks that have performed poorly in past generate sig pos rets of 3-12 month holding periods -contrarian strats are where investors buy past losers and sell past winners -analyze 16 strats by looking at past rets from 1-4 quarters and holding periods of 1-4 quarters -also look at a second set of 16 that skips a week b/n formation and holding period to avoid biases (bid ask spread, price pressure, lagged reaction effects) -1965-1989 from CRSP daily file -all stocks w/ available rets data in J months preceding port formation date are included in sample -winners minus losers port has pos rets for first 12 months after formation and half of the excess rets are gone after year 2 -stocks in loser prots experiences sig higher rets than stocks in winners port 8-20 months after formation - 2 interpretations: 1) investors who buy good and sell poor influence prices to overreact, 2) market unerreacts to info about short term prospects and overreacts to info about long term -conclude the hyp of market efficiency can be rejected at even the most conservative levels of sig

Lakonishok, Shleifer, Vishny 1994

-empirical evidence in support of why value strat, which is contrarian, is a profitable bet against popular or naive stock investment strats -most investors overreact to both pos and neg news, so they tend to project trends in stock prices for too much into future by buying overpriced glamour stocks and expense of underpriced value stocks -1963-1990 and classify firms into deciles before rets in each year -higher ab rets using value strats over popular strats -2-dimensional value strats produce rets of 10-11 percent per year more than the 2-dimensial glamour strats -fundamental to price ratios have high predictive powers, but cf to price ratio is most powerful -hypotheses: 1) stock w/ higher b/m, cf to price, and earnings to price ratios (value stocks) outperform corresponding stocks w/ lower fundamental to price ratios; 2) contrarian model forecasts that a value stock must have had low growth in past and be expected by market to continue to grow slowly -C/P (E/P) as proxy for low expected growth rate -sort stocks into 30%, 40%, 30% by GS and C/P then take intersections resulting from those classifications -extreme glamour (high gs, low c/p); value (low gs, high c/p) -value strats based jointly on past performance and expected future performance produce higher rets than more ad hoc strats such as that based exclusively on b/m ratio

Frazzini and Pederson 2014

-examine assumption of capm that all agents invest in highest expected ret per unit of risk; if agents overweight risky securities b/c of constraints to earn higher ret -this suggests that risky high beta ssets require lower risk adjusted rets than low beta assets -considers dynamic model of leverage constraints by presenting consistent empirical evidence from 20 international stock, treasure bond, credit, and futures marekts -betting against beta (bab) factor: rank securities by beta and divide them into high and low beta groups; then create simple ports that are long low beta and short high beta -findings imply that agents faced w/ fewer constraints could be applying leverage to safe assets and be compensated by investors who have to borrow w/ constraints -1926-2012 for stocks; 1952-2012 for bonds -5 propositions with support for all: 1) slope of sml depends on tightness of funding constraints (TED spread); 2) bab factors have pos average ret which increases in ex ante constraints and in beta spread of high and low beta securities; 3) bab has neg retrs when constraints tighten while expected future rets increase; 4) betas compressed to 1 when liquidity risk is high; 5) high constrained agents overweight high beta; low constrained agents overweight low beta

Keim 1983

-examines 1) anomalous neg relation b/n firm size and ab risk adjusted rets, 2) month by month empirical relation bn ab rets and mv of NYSE and AMEX common stocks, 3) possible explanations for January effect -CRSP daily stock files 1963-1979 -ranks all sample firms on mv and divides yearly distribution of mv equally into 10 ports on basis of size (200 firms) -nearly 50% of average magnitude of risk adjusted premium of small firms relative to large over this period is due to January abnormal rets -more than 26% of size premium attributed to large ab rets during first week of trading in the year at 11% to first trading day -data dont reveal sig seasonal behavior in any other month -hypotheses advanced are not able to explain jan effect: tax loss selling, info

Butler 2008

-examines ability of investment banks to evaluate soft info and to document that investment bank location matters -soft info may allow local investment banks to be more competitive underwriters of local bond matters, charging lower fees and securing better prices on the bonds they sell -uses SCD muni bond offerings from 41 states and DC 1997-2001 -measure investment banks reputation by annual market share -hand collect info on companys principal locations of business -focuses on issuers choice of investment bankers -investment banks w/ local connections have substantial comparative and absolute advantages over nonlocal -inv banks w/ an instate presence charge lower fees and sell bonds at lower yields than non local -strong for underwriting of high-risk and nonrated bonds -b/c choice of underwriter could be endogenous, use 2-sls; results consistent with ols -use probit framework to model choice of local v nonlocal in 1st stage, using an indicator var that takes value of 1 if state has more than 100 isues in sample period

Edwards, Harris, Piwowar 2007

-examines cost of trading corp bonds after national association of security dealers required dealers report all otc bond transactions through trace in 2002 -Trace data 2003-2005; NYSE ABS and dark trades to determine what effect transparency has on trans costs -secondary trans costs in corp bond market decrease w/ trade size -follow harris and piwowar 2006 to measure customer trans costs -more transparent bonds have lower trans costs -trans costs are lower for bonds that are: newly issued, have fewer years to maturity, lower coupon, higher credit ratings, and floating rates -transparency reduces customer costs by roughly 5 bps

Hirshleifer and Shumway 2003

-examines relation b/n morning sunshine in the city of a country's leading stock exchange and daily marketindex rets -when individuals are in a good mood they tend to make more optimistic choices and more pos evaluations -use that lit as motivation to see whether there is a relation b/n positive emotion and investor behavior -develop hyp using evidence that suggest people feel better when exposed to more sunshine -Hyp: sunshine is pos correlated w/ stock rets -occurrence of sunshine will caue prices to move -26 stock exchanges tests 1982-1997 -weather data from international surface weather observations: contain detailed weather descriptions for 3000 locations -use total sky cover as measure of sunshine -on a city by city basis the relation b/n cloudiness and market index for ret reported on city major exchange is neg -average rets rae almost montonically decreasing in cloudiness and effect on rets is negative in all 12 months -estimate city specific fixed effects model w/ panel corrected standard errors -in absence of trading costs, improve sharpe ratios by trading on weather

Fang and Peress 2009

-examines cross sectional relation b/n mass media coverage and stock returns -hyp: 1) impediments to trade: no-media premium reflects mispricing that arbitrageurs cant trade away, 2) investor recognition: investors arent aware of all securities in informationally incomplete markets -all companies listed on nyse and 500 randomly selected from nasdaq 1993-2002 -use number of newspaper articles abt a stock to proxy for stocks overall media exposure which comes from LexisNexis database for articles published in major us papers -stocks not covered in media earn sig higher future rets -larger for small stocks, stocks w/ high individual ownership, low analyst following, high bid ask spread, and high idio vol -media effect consistent with Chan 2003 that shows loser stocks w/ contemporaneous news experience neg ret drift and lowers w/out tend to reverse subsequently -no such drift or reversal for winners -0-investment ports formed going long stocks w/ no media coverage and short stocks with high coverage -4 diff models: 1) market model, 2) ff 93 3 factor; 3) carhart 97 4 factor; 4) 5-factor including pastor and stambaugh 2003 liquidity factor -barber odean 2008 show that individuals exert buying pressure on attention-grabbing stocks and subsequently underperform -if thats the case, expect alpha from long-short strat to come from short leg, but not the case -robustness: bid-ask bound; postearnings announcement drift; delisting bias; ipo underperformance; sector bias -media effect more concentrated among smaller stocks -3 possible causes: 1) continuations and reversals in rets, 2) lack of liquidity, 3) information frictions

Chen, Noronha, Singal 2004

-examines impact of rebalancing s&p 500 index on prices of newly added and deleted stocks -covers 1962-2000 and 3 distinct periods associated w/ changed in the popularity of indexing and the subsequent announcements -there is a permanent increase in price of firms added but not a permanent effect for those deleted -those deleted recover from temporary negative price effect after 3 months -rebalancing s&p has asymmetrical effects -investor awareness is higher for sotcks added -use volume turnover instead of trading volume so that unnaturally high volume in a few large stocks doesn't disproportionately affect market volume -very different turnover ratios: post 1976, median turnover around announce day is less than normal; pre 1976, turnover is higher than normal -existing hyp on price effect of index change: 1) long term downward sloping demand curve (no support); 2) price pressure (limited evidence) 3) certification (none); 4) liquidity w/out info production (none)

Starks, Yong, Zheng 2006

-examines jan effect anomaly through examinations of individual investors trades and investment in muni bond closed end funds -2 leading hyps why jan effect persists: tax loss selling and window dressing -tax loss selling: before end of year, investors sell securities in which they incurred losses to lower tax obligation for the year -window dressing: port managers sell poorly performing stocks to conceal holding them -monthly rets and index monthly ret from muni bond closed end fund prices -covers 168 muni funds 1999-2000 -empirical support for tax loss selling -jan ab ret are pos related to previous year end volume -buy-sell ratios confirms existence of jan effect -closed-end fund associated w/ brokerage firms display more tax loss selling -show sig jan effect in muni funds but not underlying bonds -year end volume measures: turnover (funds average monthly trading volume in nov and dec divided by number of shares outstanding at beginning of year) and volume ratio (funds average monthly trading volume in nov and dec divided by funds average monthly volume from feb to oct) -if tax loss selling is dominant explanation for ab dec and jan trading volumes, should observe diff in buy-sell ratio around year end which should be related to both current jan ret and previous years ret -uses taq and find that before year end, ratio less than 1 indicating selling pressure and over 1 in jan indication buying pressure

Fama and French 2010

-examines luck vs skill in cross section of mutual fund rets by studying if there are active mutual funds with pos true alpha, then they are balanced by active mutual funds that have neg alpha -data from 1984-2006 and construct vw ports -when alpha is estimated on net rets to investors, the cross section of precision adjusted alpha estimates suggets that few active fund produce benchmark adjusted rets to cover costs -there are some managers who have sufficient skill to cover costs -measuring alpha is a zero sum game; if someone earns pos alpha, someone else got neg one -active mutual funds win at expense of active investments outside mutual funds -use long histories of ind fund rets and boostratp simulations of return histories to infer existence of superior/ inferior funds

Ang and Bekeart 2007

-examines predictive power of dividend yields for forecasting excess rets, cfs, and interest rates -answer question of whether predictability exists since conventional wisdom attributes the variation of dividend yields to changing forecasts of expected rets, especially at long horizons -2 data sets: long data set for US, UK, and Germany at quarterly frequency from 1935-2001; shorter dataset for US, UK, Germany, and France at monthly freq 1975-2001 -build nonlinear pv model w/ stochastic discount rates, short rates, and dividend growth -model also includes standard errors developed by Hodrick 1992 -dividend yields predict excess rets only at short horizons together w/ short rate and dont have any long horizon predictive power -at short horizons, short rate strongly neg predicts rets -results robust in international data and arent due to lack of power -high div yields predict high future int rates

Conrad 1989

-examines price effect of option introduction 1974-1980 -introduction of individual options causes a permanent price increase in underlying security, beginning app. 3 days before introduction -uses data from wsj with 96 stock options -variance of individual underlying security declines following option intro -BS priced options by assuming they are redundant and can be valued w/ a no arbitrage relation; in their model, option intro can have no price effect on underlying asset -important to test b/c another idea is that option could allow investors to take positions in underlying security which werent possible before -having the price effect at intro and not announcement could indicate that it only helps those who will act as a dealer and use the new option to hedge; support found

Stein 2009

-goal is to explore consequences of individual investors being largely supplanted by institutional investors for market eff -argues that complications arise from basic economic logi when arbitrageurs inflict neg externalities on one another -2 distinct mechanisms this paper models abt externalities are crowded trade effect and fire sales effect -crowded trade: when arbitrageurs dont know in real time how many others are using same model and taking same position for broad class of quantitative trading strats; create a coordination problem and push prices further away -fire sale effect: arbitrageur takes a highly leverages position in a given trade, and a reduction in capital will force him to liquidate some of the holdings, further pushing down prices -downward sprial in prices -no clear theoretical presumtion that risk of sophisticated investors should be beneficial to market efficiency

Van Rooij, Lusardi, Alessie 2011

-examines role of financial sophistication in quality of financial decision made by economic agents -link intensity stock market participation to financial literacy -measure levels of financial sophistication: 2 indices of financial knowledge and literacy: basic literacy index and advanced literacy index -perform factor analysis on 2 indices and crease an index -create a stock market participation var from subgroup of respondents who invest in stocks -data from (dutch) DeNederlandsche Banks household survey for 2005; annual survey covering info about demographics and economic info characteristics -financial lit is important determinant of stock market participation -respondents w/ low fin lit are less likely to participate in fin market -majority of respondents display basic fin knowledge, but most dont understand advanced fin market concepts like diff b/n stocks and bonds

Green and Jame 2011

-examines trades of index funds and other institutions around s&p 500 index additions; specifically the strategic response of index funds to anticipated price pressure associated w/ additions -identifies index funds by searching for portfolio managers that trade over 450 of the s&p 500 and use data on 45 index additions b/n 1999-2005 -index funds purchase stocks beginning w/ announcement date and don't fully establish until weeks after effective date -strategic trading is more evident among large funds and for large and illiquid stocks -small and micap funds provide natural liquidity for index funds and stocks held by these funds experience smaller inclusion price effects -evidence suggests optimal strat for managers interested in maximizing AUM is to trade strategically around effective date -funds that trade index may be able to improve performance by trading strategically around composition changes -institutional data from Abel Noser Corp; captures all trades for port manager in a given month, which allow them to infer which funds are index -Thomson Financial data to examin changes in quarterly holdings of funds around changes

Chan, Jagadeesh, Lakonishok 1996

-examines whether predictability of future rets from past rets is due to market's underreaction to info (past earnings news) -3 diff measures of earnings news: 1) standardized unexpected earnings, 2) ab stock ret, 3) changes in analysts forecasts -relates evidence on momentum in stock prices to evidence on market's underreaction to earnings related info -all domestic primary stocks listed on nyse, amex, and nasdaq 1977-1993 -rank stocks on basis of either past rets or a measure of earnings news -drifts in future rets are predictable from prior rets and prior earnings news -each of the momentum strats provides diff info and no one is sufficient alone -bulk of evidence suggests that drift in future returns are not reversed, so momentum doesnt appear to be entirely driven by pos feedback trading

Fu 2009

-examines whether under-diversified investors are compensated for bearing idio risk -AHXZ 2009 found monthly stock rets are neg related to 1month lagged idio vol and this relation exists in other countries -3 questions asked: 1) do findings imply that the relation is neg; 2) if not, what is true empirical relation; 3) if true relation not neg, how AHXZ's findings explained -find idio vol are time varying and that 1month lagged value is not a good prxy for expected value -EGARCH model can capture time varying property of idio risk -by using this model, finds rets are pos related to EGARCH-estimated idio vol -AHXZ findings driven by a subset of small firms w/ high idio vol, where in the month these firms have high rets but next month they reverse resulting in neg abnormal rets -NYSE, AMEX, nasdaq 1963-2009 -amihud and mendelson 1986 first to propose role of transaction costs in ap; measure liquidity by bid-ask spread and find stocks w/ larger spreads are expected to have higher rets

Fama and French 1992

-explains main prediction of SLB model and provides detailed contradictions to model such as leverage and size -finds that average stock returns are not positively related to market beta -size and b/m predict cross section of average returns and help evaluate portfolio performance and measure expected returns -returns increase with b/m -returns decrease with size -tests conducted fail to support SLB -use all nonfin firms in NYSE, AMEX, and NASDAQ -finds betas by estimating them for portfolio then assign each stock that beta -then divide in size deciles and then again into 10 portfolios based on betas for stocks

Titman 2013

-explores how shocks to financial markets interact w/ externalities and why shocks to debt markets can affect investment choices very differently than shocks to equity markets -assumes there is asymmetric info about fundamentals, the implication of the participation shocks are not readily apparent, and focuses on participation shocks that have much longer term effects on overall market -address raises possibility that shocks to investor participation in debt and equity markets affect no only amount invests and how the investments are financed, but also types of investments that are taken -shock to supply of debt generates a tilt towards types of investments that have greater debt capacities -shock to equity supply generates a tilt towards investments that are better suited for equity financing

Stambaugh, Yu, Yuan 2012

-explores role of investor sentiment in a broad set of anomalies in cross sectional stock rets -authors consider setting in which presence of market wide sentiment is combined w/ argument that overpricing should be more prevalent than underpricing, due to short sale impediments -long-short strategies exploit the anomalies show profits -each anomaly is stronger following high levels of sentiment -short leg of each strat is more profitable following high sentiment -sentiment exhibits no relation to returns on long legs -looks at 11 apm and develops 3 hyp: 1)anomalies should be stronger following periods of high investor sent, 2) short leg of each long-short strat should have lower rets following hgih sent than low sent, 3) long legs of strats should have similar rets following high and low investor sent -investor sent measured using monthly market-based sentiment series in baker and wurgler 2006 -confirms hyps and finds that pricing anomalies stronger following higher sent -combines lit of short selling impediments keeps investors from trading away overpricing w/ sent -11 anomalies: 1) high failure probability have lower subsequent rets, 2) o-score as distress measure has lower rets, 3 and 4) post-issue years equity issuers under perform, 5) high accruals earn ab lower rets than firms w/ low accruals, 6) net operating assets is a strong neg predictor of long run stock rets, 7) mementum, 8) sort on gross profit to assets; more profitable firms have higher rets, 9) grow total asset more earn lower rets, 10) profitable firms have higher expected rets, 11) higher past investment predicts lower future rets

DeBondt and Thaler 1987

-in previous paper, found systematic price reversals for stocks that experience extreme long term gains or losses: past losers sig outperform past winners -reevaluates overreaction hyp and discusses new empirical findings relevant to the winner-lower, size, and jan effects, and the broader issues of time varying risk premia and market efficiency -excess rets for losers in test period are neg related to both long term and short term formation period performance -for winners, jan excess rets are neg related to excess rets for prior dec, possible bap gains tax effect -winner-loser effect cant be attributed to changes in risk measured by capm betas -test: construct arbitrage ports that finance purchase of losers by sellings winners short, the regress annual test period rets on market risk premium -winner-loser effect is not primarily size effect -earnings of winning and losing firms show reversal patterns that are consistent with overreation -compustat 2965-1984; use m/b measure of under/over valuation

Cremers and Petajisto 2009

-introduces new measure of active port management, active share: share of port holdings that differ from the benchmark index holdings -along with tracking error, it is used to characterize active management for all equity mutual funds -fund holding differ from benchmark in 2 ways: stock selection or factor timing -active share proxy for stock selection, puts equal weights on all active bets -tracking error proxy for factor bets, incorporates covariance of rets -data is domestic equity mut funds 1980-2003 -using active share, active management predicts fund performance relative to benchmark -high active share funds outperform benchmark before and after expenses; low active share funds underperform -funds w/ assets greater than 1 billion have size effect -small funds more active and largest funds are closet indexers

Moskowitz, Ooi, Pedersen 2012

-investigate 5 parts: 1) time series predictability of future rets across diff time horizons; 2) relation b/n time series and cross sectional mom strats; 3) what might be driving ts mom; 4) how much overlap and diff exist b/n ts mom strats and cs mom strats; 5) who trades on trends -58 futures contracts prices and scales these rets by their volatilities -sig ts mom in equity index, currency, commodity, and bond futures for each of the 58 -persistency in rets for 1-12 months that partially reverses over longer horizon -tsmom effect is distinct from cs mom, though theyre related -tsmom captures the rets associated w/ individual stock mom -speculators profit from tsmom at expense of hedgers -pos autocovariance in futures contracts rets drives most of the ts and cs mom effects and serial cross correlations and variation in mean rets is small -ts captures cs individual stock mom -examines trading activity of speculators and hedgers around ret patterns using weekly position data from commodity futures trading commision -speculators trade w/ it, being positioned to take advantage of pos trend for first 12 months then reducing positions as trends reverse -decompose tsmom coming from spot price predictability vs the roll yield stemming from shape of futures curve -both effects contribute to tsmom but only spot price changes w/ lont term reversals, consistent w/ overreaction idea in spot but not hedging

Cremers, Ferreira, Matos, Starks 2016

-investigates consequences of indexing in the equity mutual fund industry across 32 countries -distinguish b/n truly active funds and closet indexing, uses active share -cut-off of total net assets proxy for market share is 60% -hyp: since indexed funds have had recent growth, truly actively managed funds will increase their active share and lower fees -there will be fee reduction and higher alpha in mut funds industry due to competition -datasets from Lipper and FactSet/LionShares 2002-2010 -markets w/ more competition from explicitly indexed funds display active funds promote competition, thereby increasing excess rets and lower fees for active management -actively managed funds are more active and charge lower fees when face competition from low-cost explicit funds -active fund mangers will increase active share to keep up w/ competition

Baker and Wurgler 2006

-investigates impact of investors sentiment on cross section of stock rets -starts w/ simple theoretical predictions which identified 2 channels through which sentiment can have sig impacts on rets and other firms characteristic vars -2 channels are uniformed demand shocks and a limit on arbitrage -uniformed demand shocks vary across stocks cross section, while arbitrage opps are constrained -constraints on arbitrage vary across stocks -6 proxies to construct investors sentiment index using principal component analysis: 1) close-end fund discount; 2) nyse share turnover, 3) number of ipos, 4) average first day ret on ipos, 5) equity shares in new issues, 6) dividend premium -data includes all common stocks 1962-2001 -when beginning of period sent is low, small stocks, newly listed stocks, highly volatile stocks, unprofitable stocks, and nondividend paying stocks earn high rets -when sent is low, rets on extreme growth and distressed frims are high relative to average -stocks in middle decile are less affects by sent -address that pca cant distinguish b/n common sent component and common business cycle: regress proxies on growth in data such as growth in industrial production index and growth in consumer durables, nondurables, and services, and dummy for NBER recessions -orthogonalized proxies are qualitatively similar to before

Goyal and Saretto 2009

-investigates nature of rets and market characteristics of firms if temporary divergence implied volatility and historical volatility occurs -studies cross section of stock option rets by sorting stocks on diff b/n hv and iv -measurement and forecast volatilities are also mean reverting and thus long run diffs b/n iv and hv should be zero -estimates iv by averaging inverted iv from BS -this iv is averaged for cal and put contracts that are closest to in the money -compute hv from sd of daily realized ret over 12 months -use 0-cost trading strats and is repeated under straddle and delta hedged calls -use 2 ports: decile port (log diffs b/n hv and if) and pos and neg ports -data from Option metrics Ivy DB database 1996-2006 -zero cost strats involves taking long (short) postion in an option port w/ a large pos (neg) diff b/n hv and iv -monthly ab ret are very high and close to raw rets -investors overreact to current events by increasing their estimates of future volatilities after large neg stock rets -trading frictions reduce profitability of option port strat; before cost profits higher for illiquid options -margin requirements for short positions are roughly equal to one and a half times cost of written options, so liquidity considerations reduce but not eliminate economical importance -extreme rets to underlyings stocks are related to extreme changes in iv; investors overreact to high realizations of stock rets leading them to form an estimate of fv that is not justified

Barberis and Xiong 2009

-investigates what is driving disposition effect; looks to see if prospect theory preferences can predict a disposition effect -prospect theory posits that people evaluate gambles by thinking about gains and losses, not final wealth levels and that they process these gains and losses using a value function that is concave for gains and convex for losses -first implementation applies prospect theory to annual stock level trading profits and second implementation applies it to realized gains and losses -in both cases, look to see how results change as variation is allowed in expected stock ret and number of trading periods -when applying the firm implementation, does predict disposition effect in some cases, annual gain/loss of it often predicts the opposite -investors have a greater propensity to sell a stock trading at a paper loss than one trading at a paper gain -model applies to realized gains and losses predicts disposition effect more readily

Daniel and Moskowitz 2016

-mom strats can experience infrequent and persistent strings of neg returns despite strong pos average rets across numerous asset classes; these are partly forecastable -mom crashes occur in panic states, following market declines and when market volatility is high, and are contemporaneous w/ market rebounds -low ex ante expected rets in panic states are consistent w/ a conditionally high premium attached to the option like payoffs of past losers -implementable dynamic mom strat based on forecasts of its mean and var approximately doubles alpha and sharpe ratio of a static strat -1927-2013 us equity sample -in bear market, a mom portfolios up-market beta is more than double its down-market beta -if market rebounds quickly, mom strats crash b/c they have conditionally large neg beta -crsp; construct monthly and daily mom decile ports that are rebalanced at end of each month -while past winners generally outperform past losers, relatively long periods when mom has severe losses -14 of 15 worst rets occur when lagged 2 year market ret is neg and all occur inmonths where market rose contemporaneously -clustering shows they have relatively long duration -extreme losses are clusters -crash performance mostly attributable to short side, or performance of losers

Riedl and Smeets 2017

-objective: shed light on why investors hold socially responsible mutual funds, which are financially costly and dont necessarily perform better -combines administrative investor data, behavior in incentivized experiments, and survey data -compares realized port performance of socially responsible and conventional investors, examines investors expectations regarding ret and risk of sri funds -results indicate that financial motive are unlikely to be main driver -investors intrinsic social preferences and social signaling are major factors determining likelihood of holding sri equity funds -on average investors w/ strong social motivation are willing to forgo financial returns in order to invest in accordance w/ their social preferences -most socially responsible investors expect sri funds to earn lower rets than conventional funds, to achieve worse sharpe ratios, and to pay higher fees

Novy-Marx 2012

-objective: show that intermediate horizon past performance, not recent, drives mom -Unlike JT 1993, considers rets to mom strats while varying length of test period and time b/n test period and prot formation -int horizon includes 12-7 months prior formation -recent past 6-2 months -int horizon past performance drives mom not just in cross section but also in industries, investments, styles, etc -recent past performance commoves w/ inter horizon -recent winners that were inter losers underperform recent losers that were inter winners -recent rets are irrelevant after controlling at inter -all crsp stocks 1926-2010 -mom strats constructed each month by buying winners and selling losers (defined as upper and lower decile by NYSE breaks of cumulative rets over a period) -spanning tests to see if test strat is in span of explanatory strat -inter mom helps price recent past performance -results hold in industries, investments styles, foreign markets, commodities, and foreign market currencies -disposition effects associated w/ unrealized gains and losses dont explain performance, contrary to grinblatt and han 2005

Gennaioli, Shleifer, Vishny 2015

-present new model of investors delegating port management to professionals based on trust -trust in a manager reduces investors perception of riskiness of a given investment and allows managers to charge fees -money managers compete for investor funds by setting fees but b/c of trust they don't fall to costs - in equilibrium, fees are higher for assets w/ higher expected ret, managers on average underperform the market net of fees, but investors still prefer to hire managers to invest for them -model trusting as reducing the utility cost for investor of taking risk -net of fees, investors consistently underperform market, but experience less anxiety and earn higher expected returns than they would doing it on their own -in model, attitudes toward risk are shaped by 4 parameters: 1) baseline risk aversion; 2) anxiety, 3) reduction in anxiety from having trusted manager, 4) dispersion of trust in diff managers -each investor finds balance b/n rf instrument and risky equity thats managed by someone -optimal port is riskier if investor hires more trusted manager -in model, fees are not compensation for ab ret but a way to share risk premium b/n investor and manager -show model w/ multiple financial products -managers end up w/ strong incentives to pander to investors believes even if theyre misguided -implications: 1) arbitrage could be weaker than thought to be, 2) catering to hot products also destabilizes prices and could make bubbles

Black and Sholes 1973

-presents theoretical valuation formula for options derived from idea that if options are correctly priced in the market, is should not be possible to make profits by creating a port of long and short positions in options and their underlying stocks -valuation model interprets options as securities that are redundant and can be replicated in a continuous manner by investments in stocks and bonds -can also be applicable to corp liabilities such as common stock, corp bonds, and warrants -value of option will only depend on price of the stock, time to maturity, and variables that are constant -theoretically: relative volatility of the option is not constant and it depends on both stock price and maturity of option; value of option is independent of expected ret on stock -expected ret on option will depend on expeted ret of stock -Empirically: actual prices at which options are bought and sold deviate in certain systematic ways: option buyers pay prices consistently higher than those predicted -large transaction costs in option market, all of which are effectively paid by option buyers

Kacperczyk, Van Nieuwerburgh, Veldkamp 2014

-propose new def of skill as general cognitive ability to pick stocks or time market -find evidence for stock picking in booms and market timing in recessions -same fund managers that pick stocks well in expansions also time market in recessions -these fund managers sig outperform other funds and passive benchmarks -measure displays more persistence than either market timing or stock picking alone and predicts fund performance -to measure skill: construct estimates of stock picking (product of a fund's port weights in deviation from market weights and firm specific component of stock rets) and market timing (product of port weights in deviation from market weights and aggregate component of stock rets) -show skilled managers exist, select top 25% of funds in terms of on and show they good at the other -superior funds tend to be smaller and more active -crsp suvivorship bias free mutual fund database -thomson reuters sotck holdings database 1980-2001 -use pooled panel reg model w/ se clustered along fund and time dimensions -in recessions, skilled increase cash holdings, reduce high beta stock holdings, and tilt ports away from cyclical and towards more defensive sectors -define skill index as a weighted average of timing and picking where weights depend on business cycle -find performance measures increase monotonically w/ skill index

Da, Engelberg, Gao 2011

-proposes new and direct measure of investor attention using search frequency in Google (search value index SVI) -aims to 1) investigate relation b/n svi and existing attention measures 2) examine whose attention svi is capturing, 3) test attention theory of barber odean 2008, which argue that individual investors are net buyers of attention grabbing stocks and results in temp pos price pressure -data from google trends and focuses on Russel 3000 index from 2004-2008 -svi is correlated with but diff from existing proxies -svi captures investor attention in a more timely fashion -svi captures attention of retail investors (data from dash-5) -results consistent w/ barber odean 2008 -use IPOs b/c they follow pattern predicted by attention-induced price pressure hyp -retail attention measure by search volume is related to IPO first day returns and subsequent return reversal -searches related to ipo stocks increase by almost 20% during ipo week -more asvi outperform those w/out on first day

Hou, Xue, Xhang 2015

-proposes new factor model that 1) accounts for many asset pricing anomalies and 2) looks from the production side for direct link b/n atock rets and firm characteristics -use q-factor model to show that high investment stocks and low expected profitability stocks should earn lower expected rets -an empirical q-factor model consisting of the market factor, a size factor, and investment factor (investment to assets) and a profitability factor (ROE) largely summarizes cross section of average stock rets -a comprehensive examination of nearly 80 anomalies reveals that about 1/2 of them are insig in broad cross section -q-factor model's performance is at lease comparable to ff93 3factor and carhart 97 4 factor in capturing remaining sig anomalies -many seemingly unrelated anomalies seem to be diff versions of investment and profitability effects

Daniel, Hirshleifer, Subrahmanyam 1998

-proposes theory of securities market under and over reactions based on overconfidence and self attibution -builds on debondt and thaler 1985; theory assumes that investors view themselves as more able to value securities than they actually are so they overweight their own forecast and underestimate the forecast error variance -through self-attribution, confidence of investor grows when public info is in agreement w/ their info but doesnt fall when it is against -from 6 propositions proposed, theory finds that investors overreact to private info signals and underreact to public info signals -long run reversals w/ short term momentum explain event based ret predictability -wealth shifts from foolish to rational traders until price setting is dominated by rational traders, so overconfident informed traders lose money on average -event based predictability arise from underreaction only if event is chosen in response to market mispricing -overconfidence effects are more severe in less liquid securities

Grinblatt and Han 2005

-prospect theory(main element is an S-shaped value function that is concave (risk averse) in the domain of gains and convex (risk loving) in the domain of losses) -mental accounting (decision makers tend to segregate diff types of gambles into separate accts, then apply prospect theory to each acct by ignoring possible interactions -these can lead some investors to hold on to their losing stocks; tendency creates a spread b/n a stocks fundamental value and its equilibrium price, as well as price underreaction to info -spread convergence arises from random evolution of fundamental values and updating of reference prices -generates predictable equilibrium prices interpretable as possessing momentum -empirically, a var proxying for aggregate unrealized cap gains appears to be the key var that generates the profitability of mom strat -controlling for this, past ret have no predictability for cross section of rets -implication of model is to see mom in stock rets -model is diff from others in explaining mom by predicting (1-period) lagged cap gains are sufficient stats for forecasting the cross section of rets -weekly rets, turnover, and market cap from MiniCRSP 1962-1996

Hong and Kacperczyk 2009

-provide new evidence on market effects of social norms by looking at who owns sin stocks and examining the ret performance of the sin stocks -sin stocks: publicly traded companies involved in producing alcohol, tobacco, and gaming -time series and cross sectional regressions -social norms can have important consequences for markets; sin stocks are less held by norm-constrained institutions such as pension plans as compared to mutual or hedge funds that are natural arbitrageurs -they have higher expected rets than ow comparable stocks, consistent w/ them being neglected by norm-constrained and facing greater litigation risk -sin stocks receive less coverage from analysts

Jagadeesh and titman 2001

-provide out of sample tests for their 1993 paper w/ additional 9 years of data -momentum strats predict excess ret from buying winners and selling loser stocks over 3-12 months after port construction, followed by a reversal -momentum strats are due to underreaction, but a reversal then turns losers stocks to winner stocks after 12 months -authors construct winner and loser ports for 60 months following formation dates w/ sample covering 1965-1998 -construct overlapping ports; a moment decile portfolio in any particular month holds stocks ranked in that decile in any of the previous 6 - conrad and kaul 1998 is competing hyp considered which suggests that cross sectional expected ret is all that matters -confirm oos test for JT1993 in momentum portfolio yields post rets followed by reversal

Campbell and Yogo 2006

-provides evidence against long tradition of random walk models of stock prices -stock rets, especially at long horizon, and validity of inferences of 4 predictors of excess stock rets are challenged -propose bonferroni ttsest, a uniformly most powerful test, as an alternative to conventional ttest -3 desirable properties for empirical work: 1) can be implemented w/ standard regression methods and inference can be made through intuitive graphical output; 2) test is asymptotically valid under fairly general assumptions on the dynamics of predictor var and on dist of the innovations; 3) more eff than previously proposed tests in sense of Pitman eff (requires fewer obs for inference at same level of power -favor the local-to-unity asymptotic method of inference construction ahead of exact finite-sample theory method -find theoretically the persistence of stock rets is established and theoretical monte carlo simulations shows that q-test (ump) is better gof than ttest -empirically, high persistency of dividend yield, earnings price ratio, and short term interest rates predictor vars -null hyp of high persistency of long short yield is rejected -1952-2000, short term interest and long short yield spread are not good predictors of ret at annual frequency

Asness, Moskowitz, Pedersen 2013

-provides simple 3-factor model that captures global rets across asset classes, ff us stock ports, and a set of hedge fund indices -use liquid set of stocks for US, UK, Europe, and Japan -use 18 developed equity markets for country equity index -generates 48 test ports using 3 ports of low, middle, and high for each of the characteristics value and mom in each of the 8 asset classes (3-2-2) -create 0-cost long short ports that use entire cross section of securities w/in an asset class -high ret premium and sharpe ratio of a global across asset class diversified value and mom port -value effect: relation b/n an assets ret and ratio of its longrun (book) value relative to its current mv -mom effect: relation b/n an assets ret and its recent relative performance history -value performs poorly when funding liquidity rises; mom performs well during those times

Ang, Hodrick, Xing, Zhang 2009

-question: what is relation b/n idiosyncratic vol and returns -contributions: 1) anomalous relation b/n lagged isio vol and future average rets in US data exist in other markets; 2) neg spread in rets b/n stocks w/ high and low idio vol in international markets strongly comove w/ difference in rets b/n US stocks w/ high and low idio vol -idio vol can predict future aggregate market rets, but finds a strong neg relation b/n lagged idio vol and realized rets (contrary to prev lit) -FF 3-factor model used to measure idio vol -finds results in neg relation b/n idio vol and expected rets: implies there is underlying economic source behind phenomenon

Cohen, Diether, Malloy 2013

-show that legislation has simple yet previously undetected impact on stock prices -exploiting voting record of legislators whose constituents are the affected industries, show that votes of these interested legislators capture important info seemingly ignored by market -long-short port based on these vews earns ab rets per month following passage of legislation -seems most ret predictability from short side: bills in which interested senators seem to be especially neg relative to uninterested seem to result in large sig neg future rets -industries classified as beneficiaries of legislation experience sig more pos earnings surprised and pos analyst revisions in months after bill and sig higher future sales and profitability -show the more complex the legislation, the more difficulty market has in assessing impact -more concentrated the legislator's interest in the industry, more info the votes future rets -voting record to sign impact of a bill, pos or neg, for the set of firms it affects -explores whether economic interest measure capturing true info important for firm fundamentals; measures: industry level vw average profitability, industry level vw average sales growth -both timing and magnitude of ret predictability found are reasonable given subsequent news about industry fundamentals that appears in the few months following bill passage -real effects on industry profitability and sales growth that happen over the following year

Guiso, Sapienza, Zingales 2008

-studies effect that a general lack of trust can have on stock market participation -use a sample of almost 2000 Dutch households in 2003; participants asked questions on trust, attitudes towards risk, ambiguity aversion, and optimism ; from Dutch National Bank -Proposition 1: only investors w/ high enough trust will invest in stock market -proposition 2: the higher the level of investor trust, the higher his optimal port share invested in stocks conditional on participation -proposition 4: for any probability of being cheated, there exists a wealth threshold that triggers participation which increases in p -higher levels of trust lead to higher investment in market largely regardless of wealth -trust is not a proxy for risk aversion -trust has most effect on those w/ low education; education on stock market would increase investor participation

McLean and Pontiff 2016

-studies out of sample and post publication ret predictability of past uncovered vars that previously had shown to predict cross-sectional stock rets -answer question of whether a specific cross sectional relation continues out of sample and after relation has been published compared w/ in sample findings -79 different studies from peer-reviewed finance, acct, and econ journals and synthesize 97 vars where null hyp was rej -form long-short ports based on extreme quintiles of characteristic, compare each predictors ret over 3 distinct periods: 1) original sample period, 2) after publication, 3) b/n the 2 -find that long-short ret declines by 26% oos and declines more post pub -reject hyp that ret predictability disappears entirely and instead determine that there is a decay effect in port rets -stocks w/ predictor ports observe post pub increases in trading vol and is consistent w/ idea that academic research draws attention to predictors -suggest that academic pubs transmit info to sophisticated investors

Campbell, Giglio, Polk, Turley 2018

-studies pricing vol risk using first-order conditions of a long-term equity investor who is content to hold aggregate equity market instead of over-weighting value stocks and other equity ports that are attractive to short-term investors -conservative long-term investor will avoid such overweights to hedge against 2 types of deterioration in investment opportunities: declining expected stock rets and increasing vol -low-frequency movements in equity vol tied to default spread are priced in cross section of stock rets -merton's 1973 intertemporal capm shows that assets should deliver lower average ret in equilibrium if they are priced from conservative long-term investors' first-order conditions -econ model of stock rets that captures time variation in expected rets and vol and permits tractable analysis of lt port choice (VAR) -augmented w/ scalar affine stochastic vol model in which a single state var governs the vol of all shocks to teh var -derive 3 priced risk factors corresponding to 3 impotrant attributes of aggregate market parameter, so dont worry about lewellen, nagel, shanken 2010 critique -GMM to estimte 1921-2011 -news abt future var has sig vol and isnt correlated w/ contempraneous market rets -contrasting behavior of def and defo during episodes such as tech boom illustrate that taken in isolation, the relatively stable default spread throughout most of the late 1990s would predict little change in future market vol, but take into acct the declining eq prem over that period and high pe-adjusted default spread forecasted much higher vol -construct ports sorted by value, size and risk and b/m and ivol -growth stocks hedge news abt future real stock rets (known) and abt var of market ret (novel) -capm did good pre-1963, fails to price set of assets after -3-beta icapm does best -vol risk exposure is helpful in explaining many of equity anomalies in asset pricing lit

Harris and Piwowar 2006

-studies secondary trading costs in use municipal bond market -propose and implement new method for measuring transaction costs given unique market structure of muni market and very low trade freqs observed in most bonds -3 problems currently with bond data: 1) no quote data, so can't do transaction costs (like spreads based on benchmarks) for individual stocks, 2) scarcity of data, 3) timestamps are often off -uses econometric time series model in which every trade is a unit of observation to study how bond complexity and credit quality are related to transaction costs -price of a trade is the true value of it plus or minus a price concession depending on if the initiator was a buyer or a seller; transaction cost model separately estimates for customer trades and interdealer trades -absolute customer trans costs, measured as a fraction of the price, depends on size of trade, measured by dollar value; measure is the effective half spread -decompose true value into linear sum of time drift, short and long term municipal bond factor returns, and a bond specific valuation ter -factors together define a response curve defining average trade costs -muni trades 1999-2000 from MSRB surveillance database; data about characteristics of bonds for 3 dates from Kenny database -muni trades are sig more expensive than equivalent sized equity trades -muni market small trades are substantially more expensive than large trades -above 2 points because lack of transparency in bond market; large institutional investors probably have a good sense of value of bonds where small traders don't -actively traded bonds arent cheaper to trade than infreq traded bonds -bond trading costs increase credit risk, instrument complexity, time since issuance, and maturity -traders incur sig higher trans costs when trading complex bonds: those w/ attached calls, sinking funds, credit enhancement, egt. than when trading simpler bonds

Roll, Schwartz, Subrahmanyam 2010

-studies time series properties and determinants of option/stock trading volume ratio using a comprehensive cross section and time series of data on equities and their listed options -o/s is diff in option trading vol divided by stock trading vol -OptionMetrics 1996-2007 -option spread is strongly neg indicating that liquidity of option market is associated w/ greater trading -iv is pos showing that more volatile stocks attract more option trading -o/s is variable over time and is related in the cross section to size, trading costs, iv, leverage, and institutional holdings -o/s increases in 5 days leading up an earnings announcement and seem to affect prices -high o/s in conjunction w/ high cars before announcement imply smaller cars following

Huberman 2001

-test familiarity hyp; considers geographic distr of shareholders of 7 us regional bell operating companies, assembles and discusses evidence on how familiarity affects investment choices -also documents geographic distrib of rboc shareholders and relates the amt of money that individuals invest in rbocs to the typical us households net worth and stock holdings -7 rbocs used in sample that provide number of acct holders and number of shares being helod in 1996 for 48 states -the geographic distrib of shareholders demonstrates the propensity to invest in the familiar -investment in familiar conflicts w/ portfolio theory's advice to diversify

La Porta 1996

-tests for systematic errors in expectations that may explain high ret earned by value stock (low prices relative to bv, cf, earnings, dividends) using survey data on forecasts by stock market analysts -show that investment strats that seek to exploit errors in analyst forecasts earn superior rets b/c expectations abt future growth in earnings are too extreme -construct annual ports by compounding monthly rets on basis of expected growth in earnings and find that 1-yr postformation raw ret for stocks w/ low expected growth rates is much higher on average -analysts revise expectations in direction of errors -find neg event rets for high expected growth stocks around announcement dates makes risk hypothesis less likely -also no evidence that low expected growth stocks carry more risk -sort stocks into size decile ports on basis of June mv of equity, then compute ew rets CRSP of NYSE/AMEX, compustat, IBES (shows growth expectations too extreme) -expect average return on value stock should be higher than of temp winner -temp losers do better than glamour stocks -mixed evidence on extrapolation hyp: glamour stocks more likely to be overpriced than temp losers; inconsistent that value stocks will be more underpriced than temp winners -stand devs say risk in high E(g) more than low, not predicted by risk hypothesis -high have larger betas

Odean 1999

-tests whether individuals are overconfident by examining if they trade too much -if investors are overconfident in ability to interpret info, they could experience losses beyond transaction costs -tests whether trading profits of disct brokerage customers are sufficient to cover tradings costs b/n 1987-1993 -studies whether securities bought outperform those sold by enough to cover trading costs -study ret horizons of 4 months, 1 year, and 2 years following a transaction -securities overconfident investors buy dont outperform those they sell by enough to cover costs; on average, securities buy underperform those they sell - these investors buy securities that have experienced greater absolute price changes over previous 2 years than the ones they sell -buy similar numbers of winners and losers but sell more winners -sell securities which has risen sharply in weeks prior to sell

Tetlock 2011

-tests whether stock market investors appropriately distinguish b/n new and old info abt firms -staleness: textual similarity to previous ten stories abt same firm -firms stock rets respond less to stale news -firms ret on day of stale news neg predicts its ret in the following week -individual investors trade more aggressively on news when its stale -subsequent ret reversal is sig larger in stocks w/ above average individual investor trading activity -consistent w/ idea that individual investors overreact to stale info, leading to temporary movements in price -dow jones newswire 1996-2007 -ind investors increase tendencies to trade aggressively (order imbalance) in direction of news on stale days and inst investors dont -firms stock ret volatility decreases w/ staleness, especially for small firms -ret reversal increases as newswire messages decreases, earnings related words decrease, and firm size increases; day-t turnover decreases, prior week volatility increases, and prior week illiquidity increases (fm regressions) -use calendar time ports to estimate magnitude of stale news reversal -all stale mom port rets are neg

Carhart 1997

-uses a large and survivor bias free mutual fund database to explain persistence in eq mutual funds mean and risk adjusted rets and whether this is driven by 1 year mom -controls for survivor bias and documents common factor and cost based explanations for mutual fund persistence -database covers diversified equity funds monthly rets 1962-1993 -4-factor model combines ff 93 3-facotr w/ mom anomaly -4-factor model eliminates almost all of the patterns in pricing errors, which is an improvement from both capm and 3 factor -persistence in mutual fund performance doesnt reflect superior stock picking skill, but common factors in stock rets and persistent diffs in mutual fund expenses and transaction costs -only strong, persistent underperformance by worst return funds remains anomalous -JT 93 one year mom in stock rets accts for hot hands effect in mut fund performance -Hot hands funds infrequently repeat their abnormal performance -funds in top decile have rets strongly pos correlated with mom, while rets in bottom decile strongly neg correlated -expenses and turnover are related to performance -as turnover increases, ab ret drops

Fama and French 2002

-uses dividends and earnings to estimate expected stock return and help determine whether realized average ret is high or low relative to expected value -unlike previous studies, uses a longer period, unconditional expected ret, and focuses on comparing alt estimates of unconditional expected stock ret -1951-2000, dividend and earnings growth are much lower than equity premium produced by average stock ret -evidence suggests high average ret is due to decline in discount rate that produces a large unexpected cap gain -main conclusion: average stock ret of last half-century is a lot higher than expected -logic behind paper is that average stock ret is average div yeidl plus average rate of cap gain

Grinblatt and Keloharju 2001

-uses logistic reg to examine determinants of buying, selling, and holding on stock transactions over 2 year period -analyze how market participants behaving in equilibrium to characterize similarities and heterogeneity of investors -consider 6 categories of investors (nonfinancial, financial, government, nonprofit, household, and foreign) and conduct 5 tests to predict the determinants of decisions -1994-1997 from Finnish Central Securities Depository; analyze disposition effect and tax loss selling anomalies -disposition effect: tendency of investors to sell shares whose prices have increased, while keeping those that dropped in value -tax loss selling: selling an asset w/ cap loss to lower or eliminate cap gain realized -propensity to sell a stock is pos related to recent rets -stocks w/ large pos ret in recent past and w/ prices at their monthly high are more likely to be sold -effect of past ret on trading activity more important for pos past rets than for neg -disposition effect and tax loss effect 2 major determinants of propensity to sell a stock -tendency to hold onto losers is worsened for losses exceeding 30% -life-cycle hyp: rational economic agents should smooth their consumption by appropriately investing and borrowing based on expectations about lifetime income -youngest investors buy more and those past retirement age sell more

Hirshleifer, Lim, Teo 2009

-uses pyschological studies and those in finance that address investors distractions as motivations behind mom anomaly -tests limited investor attention around earnings announcements to determine whether the number of firm announcements creates a distraction -if attention is limited, earn announce by other firms can call investor attention away from task of valuing firm -investor distraction hyp: implies that number of competing news announcements has opposite effects on immediate sensitivity of firms stock to its earnings surprise, vs post event sensitivity -hyp 3 parts: 1) sensitivity of announcement ab ret to earnings news decreases w/ number of competing announces; 2) ab trading vol on day of announce decreases w/ number of competing announces; 3) sensitivity of post announce ab ret to earnings news increases w/ number of competing announces -use quarterly earnings announcement data to produce decile ranks based number of announces -main objective: assess diff b/n high news days and low news days -investors announe date reactions to earnings are sig less sensitive to new on high news days -robustness checks: confirm news dates; results could be from leakage, not weaker reactions; type of firms announce on difference days; measure ab rets differently