Week 5 - Market efficiency: empirical tests & evidence

Question 1

3 empirical implications of the Efficient Market Hypothesis (EMH)

Answer 1

1. Prices fully reflect all available information
2. the market responds only to new information
3. it is not possible to forecast price changes

Question 2

Weak form of market efficiency

Answer 2

Cannot make profits given observed history of PAST stock prices, trading volume and other trading data

Question 3

Semi-strong form of market efficiency

Answer 3

Cannot make profits given all PUBLICLY available information (quality of management, balance sheet composition, accounting practices etc.)

Question 4

Strong form of market efficiency

Answer 4

Cannot make profits given publicly and PRIVATELY available information

• test: PROFITABILITY of trading on inside information

*insider trading is illegal
- strong form efficiency does NOT hold b/c corporate insiders can profit on superior information not available to outside investors

Question 5

2 tests for Weak form efficiency

Answer 5

1. Technical analysis
   eg, support level, resistance level, FILTER RULES (which have a caveat of needing to sell AND buy -> TRANSACTION COSTS -> LIMITS TO ARBITRAGE)
2. Serial correlation of returns OVER TIME (autocorrelation)
   - based on regression
   Observations:
   - MOMENTUM STRATEGY: in medium horizons, stocks exhibit POSITIVE CORRELATION
   - CONTRARIAN STRATEGY (REVERSAL): in long horizons, stocks exhibit NEGATIVE CORRELATION

Question 6

Weak form efficiency - Implications + Evidence

Answer 6

Implications
1. Technical rules are not profitable
eg. if stock price has just increased/decreased, it does not follow that it will increase/decrease in the future
- if they had worked, everyone would have used them. Therefore, don’t work anymore.

Evidence
2. Research on the performance of technical rules
3. Small serial correlation in stock returns at short horizons

Question 7

2 tests for Semi-strong form efficiency

Answer 7

1. Fundamental analysis
   - using acct. and financial data to forecast future stock returns
2. EVENT STUDY of reaction to news
   - use ABNORMAL RETURNS
   - compute cumulative abnormal return
   - FLATNESS BEFORE & AFTER the event & INSTANTANEOUS EFFECT in the EVENT WINDOW convince us that there is no alpha and market is efficient

> 4 possible scenarios: jump at event date (market efficiency) -> info is REFLECTED immediately in price jump
- underreaction, overreaction after event date, information leakage prior to event date

Question 8

Semi-strong form efficiency - Implications + Evidence

incl. point on Mutual funds

Answer 8

Implications
1. Fundamental analysis is valueless (unless incorporates private info)
2. Prices respond to news quickly

Evidence
3. Rapid adjustment to new info
4. Generally, evidence favours market efficiency, eg. takeover & dividend announcements

5. on average, mutual funds do NOT outperform stock market indices (Carhart, 1997)
   - BENCHMARK is the market (unbeatable)

Question 9

Joint hypothesis

Answer 9

Every test of market efficiency is a joint test of:
1. the EFFICIENCY of the market
2. the ASSET PRICING MODEL used to calculate expected returns

ie. if we can make an alpha, is our model wrong or is the market inefficient?

Question 10

What is Momentum?

[PS10]

Answer 10

• The momentum effect is the empirical ANOMALY whereby PAST WINNERS (stocks
  that have performed relatively well during the past three months to one year) OUTPERFORM RECENT PAST LOSERS (stocks that have performed relatively poorly during the
  past three months to one year).
• Forming portfolios that buy winners and sell losers yields ABNORMAL PROFITS, if these portfolios are held for periods of 3 months to 1 year.

The abnormal return (measured by the INTERCEPT ALPHA estimated from CAPM or
Fama-French models) is usually about 1% per month.
- Fama and French (1996) CANNOT EXPLAIN the momentum effect using their THREE-FACTOR MODEL.

Question 11

Describe in detail how momentum strategies are implemented.

[PS10]

Answer 11

1. each month, STOCKS are RANKED according to their past returns (3-12 months).
2. The TOP GROUP of stocks (decile,
   quintile, etc.) constitutes the WINNER PORTFOLIO, and the BOTTOM GROUP constitutes the
   LOSER PORTFOLIO.
3. The strategy buys the winner portfolio and sells the loser portfolio.
4. These portfolios are held for a period of 3-12 months.
5. If we REPEAT this EACH MONTH we obtain a time-series of momentum returns, whose average represents the average momentum profit.
6. We can then use CAPM or Fama-French regressions to calculate ABNORMAL RETURNS (alphas).

Question 12

Why can momentum portfolios be called hedge portfolios?

[PS10]

Answer 12

Momentum portfolios are hedge portfolios to the extent that, by shorting and buying stocks,
1. we can ELIMINATE the portfolio’s EXPOSURE to RISK FACTORS.
2. Empirically we observe that winners and losers have the SAME BETA, so that the risk related to the market is hedged away (see eg. the table from Jegadeesh and Titman, 2001, reported in the lecture notes).
3. If this were not the case, momentum PROFITS could be due to differences in EXPECTED RETURNS justified by DIFF. RISK EXPOSURES.

Question 13

Discuss one Rational explanation and one Behavioural explanation for the profitability of momentum strategies.

[PS10]

Answer 13

Rational explanation
- differences in expected returns
- If winners were riskier than losers, we would obtain a systematic higher expected return for winners than for losers. Momentum profits would be due to RISK.
- This hypothesis is NOT supported by empirical evidence.
> The systematic differences between winners and losers should exist at any point in time, whereas empirically we observe that momentum profits accrue during the first year after portfolio formation,
> and later start decreasing and become NEGATIVE in the LONGER RUN (recall the paper by Jegadeesh and Titman, 2001, seen in the lectures).

Behavioural explanation
- investors underreact to news,
therefore prices do not immediately adjust to fully incorporate information
- The adjustment is gradual and this induces POSITIVE SERIAL CORRELATION in stock returns.
- Stocks with positive news continue to be winners for some time, and stocks with
negative news continue to be losers (recall the discussion on the post-earnings
announcement drift seen in the lectures).

Question 14

What is Reversal?
What is the test for its presence in data?

[2018]

Answer 14

1. Evidence that returns are negatively autocorrelated at long horizons (3-5 years)
2. DeBondt and Thaler (1995) find that “trends” in stock returns reverse over long periods

Test
- Stocks are allocated to portfolios based on their market-adjusted cumulative abnormal returns (CAR) during the previous 3-5 years
- Assign the top 35 stocks (or the top 50 or top decile) to the winner portfolio and the bottom 35 (or the bottom 50 or bottom decile) to the loser portfolio
- Measure the performance of the extreme portfolios over the following 1-5 year periods.

• Over the last 50 years, loser portfolios of 35 stocks OUTPERFORM winner portfolios of 35 stocks by 25% in the 3-year period following portfolio formation.

Question 15

1. Post earnings announcement drift
2. Which form of market efficiency does this finding violate?

[2018]
One potential explanation is that investors UNDERREACT to the information in earnings surprises (limited attention).

Answer 15

1. Earnings announcements generate long predictable trends.
   - Stocks with positive surprise keep trending upward
   - firms with negative surprise are trending downwards.
2. Violates SEMI-STRONG form of market efficiency.
   - Note that the past information in prices does NOT allow us to predict the effect of earning announcement and hence weak form is NOT violated.

Question 16

2 evidences in support of semi-strong form efficiency {my own Q}

Answer 16

1. Event studies
2. On average, mutual funds fail to outperform stock market indices (Carhart, 1997)
   » mutual funds are fund managers using publicly available information to try to beat the market (unbeatable)

Question 17

What is the Size effect anomaly?

[Lecture 6]

Answer 17

Banz (JFE, 1981) and Reinganum (JFE, 1981) showed that small capitalization firms on the NYSE earned higher average returns than is predicted by the CAPM from 1936-75.

Question 18

Firms with the highest B/M ratio (value stocks) have HIGHER returns than firms with low B/M (growth/ glamour stocks).
1. What are 2 explanations?
2. What is the Behavioural explanation? (LSV 1994; LLSV 1997)

[Lecture 6]

Answer 18

1. high B/M firms are underpriced, or
2. B/M is a risk factor affecting equilibrium stock returns.

Behavioural explanation
- Some investors tend to get overly excited about stocks that have done very well in the past and buy them up, so that these “glamour” stocks become overpriced
- Similarly, they overreact to stocks that have done very badly, oversell them, and these out-of-favor “value” stocks become underpriced.
- Value stocks provide superior returns because the market slowly realizes that earnings growth rates for value stocks are higher than it initially expected and conversely for glamour stocks.

Question 19

Event studies - What is the sign for Information leakage?

Answer 19

ANY CHANGE in CAR prior to the event date