Study Name

“Stock Market Returns: A Temperature Anomaly”

Authors:

Melanie Cao and Jason West

Field*:

External Factors

Capsule Review:

Purports to show that stock market returns are related to

temperature. They find that low temperature, and high sunshine are associated with high returns and that high temperature are associated with low returns. They show this is true for eight international markets including the US, and Europe during the period from 1962 to 1989.  They attribute their finding to the higher aggression in investors that lower temperatures elicit.

For the United States, where the effect is the strongest, the average return differences between the low temperature days and the high temperature days is 0.0013  per day, or 13  pennies on a hundred dollar stock. The chances of a rise during the low temperature days is 67 percent versus a 60 percent chance of a rise on the high-temperature days. Considering that there are some 37 years of daily data considered 9,375 separate observations these differences strike us as very large from both a practical and statistical point of view.

We have two quibbles with their finding, however.  First, no out of sample tests are made. What s more, the authors implicitly considered many different hypotheses since there is no good reason that if they had found that high temperatures are associated with high returns, that they couldn’t equally have formulated an explanation for this phenomenon based no mood. By that low the authors have apparently tapped into the recently listless months of July through October in United States stock markets.  

SpecDuo Ranking (Scale of 0-10 with 10 high)

Practical significance

10

Point proved?

10

Tested in real world?

1

Possibility of regime change noted?

1 (since the authors could just as well have concluded

hypothesized that high temperatures are associated with high returns due to the higher sunshine quota).

Rater:

Victor Niederhoffer

Study Name

Abnormal Returns to a Fundamental Analysis Strategy

Authors:

Jeffery Abarbanell and Brian Bushee

Field*:

Fundamentals

Capsule Review:

Using traditional rules of fundamental analysis, such as changes in

inventories and accounts receivable, gross margins, selling expenses, tax

rates, and so on, the authors developed a collection of signals with which to

trade a portfolio.  The average annual abnormal return was 13.2 percent, a

significant portion of which was generated around subsequent earnings

announcements.  These abnormal returns, however, largely tapered off after one year from the date the signals were generated.

SpecDuo Ranking (Scale of 0-10 with 10 high)

Practical significance

6

Point proved?

5

Tested in real world?

1

Possibility of regime change noted?

1

Rater:

Adam Robinson

Study Name

Long-Run Stock Returns Following Bond Ratings Changes

Authors

Illia Dichev and Joseph Piotroski

Field*

Fundamentals

Capsule Review

The authors studied virtually all the Moody’s bond rating changes

between 1970 and 1997.  They discovered that stocks of companies whose bonds had been upgraded outperformed those of companies whose bonds were downgraded. The outperformance, which averaged 12 percent, lasted for up to a year following the announcement of the upgrade or downgrade, but there was little evidence of a difference after that point.

SpecDuo Ranking (Scale of 0-10 with 10 high)

Practical significance

10

Point proved?

10

Tested in real world?

1

Possibility of regime change noted?

6

Rater

Adam Robinson

Study Name

Shock Effects on Stocks, Bonds and Exchange Rates

Authors

Ray C Fair

Field*

External Factors

Capsule Review

The approach here is to look at 221 large changes in 5 minutes prices of more than 0.75 for 6 contracts from march 1982 to march 2000. The authors divided the events into monetary, price (CPI), real ( GDP, NAPM), wage, fiscal, and company earnings, trade gaps, internal conflicts, central bank intervention, specific Japanese, and specific German. Next propositions to explain the co-movements in prices are constructed ad hoc. The main effect noted is that bond and stock prices are related. These relations are tested for each of 3 ad hocisms such as a monetary event decreases bonds and this has a first order effect decreasing stocks because of future discount rates higher and a second order effect because expected earnings are slower . The study only looks at 1,/30000 of the 5 minute interval , say 64 of 2 million intervals and then examines the 15 retrospective co-movement relations between each possible pair.

SpecDuo Ranking (Scale of 0-10 with 10 high)

Practical significance

1

Point proved?

1

Tested in real world?

1

Possibility of regime change noted?

0. Since the effect is completely opposite after march 2000 without note, multiple comparisons since the statistical effects are at the 10% level and an implicit thousands of hypotheses are considered.

Rater

Victor Niederhoffer

Study Name

Consumer Confidence Announcements

Authors

David Gulley and Jahangir Sultan, Applied Financial Economics 1998.

Field*

External Factors

Capsule Review

The authors examine monthly conference board figures from 1980-1993 and conclude there through regression analysis that there is no predictive impact on stocks but a small statistical impact on British Pound and Deutsch Mark and Swiss Franc, with negative changes in the Dollar for exchange having a larger impact.

SpecDuo Ranking (Scale of 0-10 with 10 high)

Practical significance

1

Point proved?

1

Tested in real world?

1

Possibility of regime change noted?

1. If  looking at 100 different independent variable dependent variable combinations, you'll find some market that shows a week statistical but useless practical effect

Rater

Victor Niederhoffer