Saturday's edition of the FT) Previous evidence(Pope, Morris and Peel, Journal of Business Finance and Accounting, Summer 1990) found the market does react to information about dealing by insiders. That is, directors buy/sell decisions are on average taken to be signals of good/bad news about the company and the market responds by bidding up/down the share price, ensuring corporate insiders earn profits. Pope et al. also found evidence that an uninformed investor could have earned abnormal returns by simply trading based on publicly available information about the dealings of corporate insiders. This suggests share ices do not react instantaneously to this publicly available information and the stock market is not ly SSFe with respect to this type of information signal b. Inconsistent with WFE: it suggests changes in share price(or returns)are not independent There has been renewed interest in recent years in share price patterns. Early research tended to support the hypothesis of WFE. One line of recent research(initiated by Fama and French, Journal of Political Economy, 1988, Journal of Financial Economics, 1989, and Poterba and Summers, Journal of Financial Economics, 1989) has exploited advances in statistics to look at returns for holding periods exceeding a year. a popular interpretation of this research is that it shows returns over long holding periods(more than a year)exhibit 'negative autocorrelation. If returns are negatively autocorrelated, low returns tend to follow high returns and vice versa. We call this pattern of predictability in share prices, mean reversion Returns average out at some constant level so that a run of higher than average returns follows a run of lower than average returns and vice versa-average returns move back(revert )to the mean Some academics argue that mean reversion provides proof that share prices deviate from fundamental values for reasonably long periods. However, two of the original authors( Fama and French)are careful to point out that the evidence is also consistent with market efficiency and changes over time in the returns expected on shares as the economy moves between recession and boom. Recent research (Kim, Nelson and Startz, Review of Economic Studies, 1991)finds evidence that the pre-War period drives the mean reversion results with persistence in returns(mean aversion) characterising the post-War period There are two further strands of empirical literature looking at share price patterns. One of these suggests there is evidence of longer-term overreaction in share prices. Specifically, some evidence fo the UK and the US suggests the following as a successful investment strategy i. track share prices for a 3 to 5 year period ii.rank all stocks in order of their performance over this period; iii. identify the top 30, or 50, or 10% of best performing stocks (past winners) and the corresponding worst performing stocks(past losers) iv. buy the past losers and sell the past winners and maintain this strategy for a corresponding 3 to 5 year period The second strand suggests there is evidence of short-term momentum(underreaction) in the UK and the US. The strategy suggested here is to identify past winners and losers over a 3 to 12 month period to buy the winners and sell the losers and maintain this strategy for a corresponding 3 to 12 month period. For a recently published research study on the momentum effect in the UK, see Liu, W.,N. Strong, and X. Xu. 1999. The profitability of momentum investing Journal of Business Finance and Accounting, 26,(November/December), 1043-1091 Note that inconsistency with WFE must imply inconsistency with both SSFE and SFE c.This is a loose statement of a statistical model known as a random walk. It implies that you predict next period's price change based on previous prices. The statement is consistent weak-form efficiency, but says nothing about either SSFE or SFE This statement may be entirely consistent with SSFE First, in an efficient market, share prices respond only to surprises(news; new information). This means, for example, that share prices react only to unexpected earnings announcements(the difference C Gary Xu AcF2 14 Princip les of Finance© Gary Xu AcF214 Principles of Finance 3 Saturday's edition of the FT.) Previous evidence (Pope, Morris and Peel, Journal of Business Finance and Accounting, Summer 1990) found the market does react to information about dealing by `insiders'. That is, directors' buy/sell decisions are on average taken to be signals of good/bad news about the company and the market responds by bidding up/down the share price, ensuring corporate insiders earn profits. Pope et al. also found evidence that an uninformed investor could have earned abnormal returns by simply trading based on publicly available information about the dealings of corporate insiders. This suggests share prices do not react instantaneously to this publicly available information and the stock market is not fully SSFE with respect to this type of information signal. b. Inconsistent with WFE: it suggests changes in share price (or returns) are not independent. There has been renewed interest in recent years in share price patterns. Early research tended to support the hypothesis of WFE. One line of recent research (initiated by Fama and French, Journal of Political Economy, 1988, Journal of Financial Economics, 1989, and Poterba and Summers, Journal of Financial Economics, 1989) has exploited advances in statistics to look at returns for holding periods exceeding a year. A popular interpretation of this research is that it shows returns over long holding periods (more than a year) exhibit `negative autocorrelation'. If returns are negatively autocorrelated, low returns tend to follow high returns and vice versa. We call this pattern of predictability in share prices, `mean reversion'. Returns average out at some constant level so that a run of higher than average returns follows a run of lower than average returns and vice versa-average returns move back (revert) to the mean. Some academics argue that mean reversion provides proof that share prices deviate from fundamental values for reasonably long periods. However, two of the original authors (Fama and French) are careful to point out that the evidence is also consistent with market efficiency and changes over time in the returns expected on shares as the economy moves between recession and boom. Recent research (Kim, Nelson and Startz, Review of Economic Studies, 1991) finds evidence that the pre-War period drives the mean reversion results with persistence in returns (mean aversion) characterising the post-War period. There are two further strands of empirical literature looking at share price patterns. One of these suggests there is evidence of longer-term overreaction in share prices. Specifically, some evidence for the UK and the US suggests the following as a successful investment strategy: i. track share prices for a 3 to 5 year period; ii.rank all stocks in order of their performance over this period; iii. identify the top 30, or 50, or 10% of best performing stocks (past winners) and the corresponding worst performing stocks (past losers); iv. buy the past losers and sell the past winners and maintain this strategy for a corresponding 3 to 5 year period. The second strand suggests there is evidence of short-term momentum (underreaction) in the UK and the US. The strategy suggested here is to identify past winners and losers over a 3 to 12 month period, to buy the winners and sell the losers and maintain this strategy for a corresponding 3 to 12 month period. For a recently published research study on the momentum effect in the UK, see Liu, W., N. Strong, and X. Xu. 1999. The profitability of momentum investing, Journal of Business Finance and Accounting, 26, (November/December), 1043-1091. Note that inconsistency with WFE must imply inconsistency with both SSFE and SFE. c. This is a loose statement of a statistical model known as a random walk. It implies that you can't predict next period's price change based on previous prices. The statement is consistent with weak-form efficiency, but says nothing about either SSFE or SFE. d. This statement may be entirely consistent with SSFE. First, in an efficient market, share prices respond only to surprises (news; new information). This means, for example, that share prices react only to unexpected earnings announcements (the difference