ECON 437: Behavioral Finance

From BluWiki

[edit] 1/18/2007 First day

Didn't go to this class...

[edit] 1/23/2007 �Noise Traders and the Limits to Arbitrage,� Part I

These notes were deleted...

[edit] 1/25/2007 �Noise Traders and the Limits to Arbitrage,� Part II

Big thing: the difference between the arbitrage trader & the noise trader is the two curves on the graph. It's also Pt.

[edit] 1/30/2007 No class!

Make up in two weeks!

[edit] 2/1/2007 "Noise Traders and the Limits to Arbitrage," Part III

[edit] Finish up Shleifer

• FUNDIMENTAL*

Chapter 2:

• Overlapping generations structure
• Two assets (safe asset, and unsafe asset)
• In Shleifer's article, noise traders make all the money as they ride the bubble, but eventually they go bankrupt.
• Arbitrageurs then beging buying more of the unsafe asset to cover their shorts
• Noise traders do better because they take more risk.
• Weaknesses:
  • Unsafe asset's price doesn't haven any impact on its behavior (Hirschleifer deals w/ this)
    • It's as if the unsafe is not in any real business. There won't be any feedback back to the market.
  • Only 1 kind of noise trader, either a pessimist or an optimist.

[edit] Hirschleifer's article

• To kinds of investors
• Three periods 1,2,3
• Random true cash flow
• "Feedback" profits - because a random stock gets more money, they can do stuff with it, and become more successful.
  • Example: Amazon, got the cash, created a real business w/ warehouses. AOL, which got the cash & bought AOL Time Warner.

[edit] 2/13/2007 "Anomalies - Part I"

[edit] 4/12/2007 "Serial Correlation and Review for Second Mid-Term"

Topics since last exam:

• Closed end fund puzzle
  • Know what a closed end fund is, as opposed to an open end fund.
    • OPEN END FUND: You invest money into fund, and they make new investments
    • CLOSED END FUND: You buy stock, and the stock represents a portion of the assets in the fund.
    • Most of the time, closed event funds have a big discount vs. the assets they are holding?
• Market efficiency
  • Information determines prices
  • Volatility argument - market is way too volatile
• Serial correlation
  • Why, and how is it a violation of EMH?
  • What are the reported violations and how can the EMH theorists explain them?

[edit] Closed end fund puzzle

• Brought to market when "enthusiasm" is great - you can't sell a closed end fund to the public when things aren't going well. *Therefore, it's originally sold at a premium.
• However, within six months the fund falls to a discount of an average 10%.
• Amount of discount is subject to mean reversion - meaning when people are enthusiastic, the spread narrows.

The big question: why does the fund not trade at net asset value?

• Almost all theories have been tested...
• Only Steven Ross thinks he has the answer: management fees explain discount. However, the rest community disagrees.

[edit] Serial Correlation

[edit] Efficient Market Hypothesis (EMH)

Serial correlation is a discussion about whether market is efficient

• EMH means the market reflects all available information.
• Current price is the best estimate of true value

KNOW DEFINITION OF MARKET EFFICIENCY At least one question on exam will be designed to test your understanding of EMH.

[edit] What about CAPM?

CAPM is a very specific EMH theory. EMH could be true even though CAPM is found out to be false. Know how to describe EMH, if CAPM is false!

All CAPM says is that if you just buy a few stocks, that's too risky. You should diversify away risk, you can eliminate a lot of risk by simply diversifying away. Therefore, a stock's return should reflect the risk it contributes to the overall portfolio (BETA!!)

Note, based on data, CAPM has NOT been proved...

[edit] The Foundation Upon Which the Serial Correlation Rock is Founded

Two key articles, expected to know both:

• De Bont & Thaler - 1985
  • MEAN REVERSION!!! (in long term data)
• Fama French 1992 ??
  • Not required reading
• Jagdeesh and Titman
  • Earnings momentum: Markets "underreact" to a single item of good or bad news (in short term)
  • Price momemtum: stocks that have done well in the last 6 months continue to do well, same is true with stocks that have underperformed (in short term) < Attributed to either Debondt or Jagdeesh, not sure

[edit] Into the mix

• Fama and French, 1993

See slides

[edit] Calendar Effects

• See slides

• Cooper, McConnell
  • Identify two kinds of january effects:
    • January predicts the rest of the year
    • Stocks that do poorly in December do well in January
  • Concluded that the second january effect overshadows the first

[edit] Value - Growth Debate

• Formalized in Fama French, 1992
  • Argued value over growth stocks
  • Got immediate attention, everyone bought value stocks
  • Value got hammered for the next 10 years

• Petkova & Zhang, 2005
  • Time varying risk?
  • Periods when value beats growth?:
    • "Betas on value stocks got real big, betas on growth stocks got real small" right before growth beat value... not sure if this is right
  • See slides
  • Value beats growth, but only in a small part. Rest is explained by alphas, which are are estimated to be mostly positive and significant.
  • Burton finds this article unconvincing

[edit] Combining

• Hanna and Ready, Jul 2005 (Attack Haugen and Baker, 1996)
  • Combine price to book, and combine with earnings momentum. This produces 3% alpha (3% no-risk return).
    • Accounts for transaction costs
  • What about mean reversion? Growing consensus that it is captured by Price/book
• DEVASTATING to EMH

[edit] Search for additional "risk factors"

• Sadka, 2006
  • Illiquidity can explain earnings momentum
  • Burton finds this article unconvencing
• Chordia & Shivakumar, 2006
  • LIKES this article
  • Looks at price momentum & earnings momentum
  • Finds that price momentum is nothing but earnings momentum in disguise
  • PRO earnings momentum
• Kothari, Lewellen, Warner, 2006
  • In aggregate, earnings announcements are not good predictors for the return of the entire market
  • No predictive value for market as a whole

[edit] Finally, the risk of "data mining"

• No reading for this
• If you look long enough, you'll find what you're looking for.

How do you get around it?

1. Give theoretical model & math (explanation) as to why it works
2. Show that it can be predicted in an outside sample.

[edit] review for final

Will be similar to second exam, but there will be "at least one unanswerable question." Spend bulk of your time studying noise traders, limits to arbitrage, and serial correlation.

Final will be twice as long as what you're used to.

The best guide to the final will be the second mid term, twice as many question.

Burton will be here 3 days before the final (8th, 9th & 10th).

[edit] noise trader equilibrium models & limits to arbitrage

nose traders

• know definition of noise trader
  • smart guys can see the future pretty accurately
  • noise traders can't see the future, or don't want to see the future
  • In real world, there's a growing argument that the noise traders are large institutions. Partially because of the agency problem.

[edit] limits to arbitrage

• what is the limit to arbitrage?
  • EMH people: "if there's smart guys and dumb guys in the market, then the smart guys will take the money from the dumb guys and make prices efficient"
  • these guys are saying that there are limits to this
    • this is because you have to take risk in arbitrage
    • in the real world, you can lose money to the dumb guys
    • extreme example: royal dutch shell case. two securities that are practically the same security
    • another somewhat example: the closed end fund case - although it is harder to aribtrage this.

[edit] serial correlation

second most important part of the course

• EMH says there CAN'T BE SERIAL CORRELATION
• Behavorialists say you're wrong, there is predictability:
  • mean reversion in long run data
  • Data suggest mean reversion & value over growth represent the same thing
    • DEFINE value & growth: Value means low price to book net assets. Growth means high price to book. Stocks that have done horribly in the last 3 years are value stocks, opposite is true for growth
  • short run:
    • earnings momentum: stocks that have had good earnings news will continue to do well for the next 6 months. Reverse would also be true.
    • price momentum: stocks that have done poorly in the last 6 months will continue to do poorly for the next 6 months. Reverse would also be true.
  • at this point, you can't rule out if the serial correlation isn't still in the data
  • burton is interested in: will serial correlation also be true in foreign markets?
    • given that all these investors are different, are there common themes across all investors?

[edit] anomalies

burton considers this to be the least important part of the course