Lec 1 - research in finance and essentials of econometrics Flashcards
What is theoretical research?
A system of ideas trying to explain something, especially one based on general principles of the objects trying to be explained:
- On average
- In principle
- Probability
What are examples of theoretical research in the real world?
Relation between individual stock return and market return
CAPM
What is empirical research?
Empirical means verifiable by observations, experiences, and evidence (rather than theory)
What do you test in empirical analysis?
Predictions of a theory
- When findings are consistent with theory’s predictions, you have good level of confidence to believe the theory is right
- When findings are inconsistent with predicitions, a theory is rejected
What is a clinical study?
A study based on 1 case (you can have more than once company e.g a takeover has 2)
- Disney’s takeover of Marvel
- Marvel’s IPO
- Steve Job’s death in 2012
What are the benefits and drawbacks of clinical study?
Benefit:
- In-depth analysis is possible
Drawback:
Difficult to generalise to a broader group
What is a good example of a clinical study?
Lys and Vincent 1995
What is a large sample study?
Based on a large sample of similar cases
- Sample of IPO companies
- Sample of bidders in M&A transactions
Answers questions:
- What’s the average long-term performance of IPO firms?
- On average do bidders gain or lose?
What are the benefits and drawbacks of large sample study?
Benefit:
- Can make general conclusions
Drawback:
- Costly to conduct very detailed analyses of each case
What is a good example of a large sample study?
Gregory 1999 and Gao 2011
What is a Steve Jobs example of hypothesis testing?
The event would have negative impact on share price
Null: Apple’s (abnormal) share return on 5th Oct is non-negative
Alternative: Apple’s (abnormal) share return on 5th Oct is negative
How to test the Steve Jobs hypothesis?
Event study methodology
- Reject null in favour of alternative hypothesis, if Apple’s abnormal return is significantly negative upon the announcement
- Fail to reject the null hypothesis, if Apple’s abnormal return is insignificant or significantly positive upon the announcement
- One-tailed test
When do you perform a one-tailed test?
When the hypotheised value deviates from the null in only one direction
When do you perform a two-tailed test?
If the hypothesised value may deviate from the null in either direction
What are the characteristics of a two-tailed test?
- Stricter
- Fixed significance level - if a hypothesis is rejected in a two-tailed test, it must be rejected in a one-tailed test, ceteris parabus (this doesn’t work the other way)
When can a null hypothesis get rejected?
With a confidence level less than 100%
- Required level of confidence is 95% or 99% (99% being higher)
What is a random variable?
Numerical realisation of an underlying distribution
- Distribution that consists of the possible outcomes and their respective probabilities (stock return of the next day)
- Beta and abnormal returns are random variables
What is statistical significance?
Level of signifiance = 1 - level of confidence
E.g:
Statistical significance 1% = 99% level of confidence
Statistical signifiance 5% = 95% level of confidence
1% is a higher level of statistical significance
How do you test statistical significance?
Compare the test statistic and the critical value
What are the test statistics for a random variable?
A figure that summarises the information in the data on the random variable
- Reduces all information to a figure to draw inferences
- Is also a random variable
- Formula given by staticians
What is the critical value?
A specific value on the support of the distribution of the test statistic
What happens when a test statistic falls beyond a critical value?
- Reject the null hypothesis at the corresponding level of significance
Why does the critical value vary?
- Type of distributions random variable follows
- Type of test (one or two-tailed)
- Degrees of freedom of data
What is a type 1 error?
When you reject a null hypothesis that is true
What is a type 2 error?
When you fail to reject a null hypothesis that is false
What is the power of test?
Measures the power of test in picking up the hypothesised effect
If a test has high (low) power, it has a low (high) probability of making a type 2 error
How is the power of test measured?
Measured by 1 - Probability of making a type 2 error
