Lecture 6 Flashcards
Overview of selected financial performance measures:
Sales ie through household scanner data: sales is a measurement of demand. Data coming from retail upc scanners or from company itself
Purchase intnerion: market research companies track the perception of firms and brands
Does not measure actual behaviour, indication for how strong a brand in the hearts and minds of consumers
Stock price data: refers to the current price that a share of stock is trading for on the market
A company’s stock price reflects investors perception of its ability to earn and grow its profits in the future
Typical events used in event studies
New product information, alliance formation, channel restructuring, outsourcing, conversion of non voting shares into voting shares, mergers and acquisitions
Efficiënt market hypothesis
efficiënt market hypothesis is mostly used as an underlying assumption of the event study methodology
The effect of the event is incorporated instantaneously used into stock prices. Thus stock prices reflect all available info
Whether markets are efficient has been extensively researched and remains controversial. Market may be considered as weak, Semi strong, or strong (ie stock price reflects all public and private info)
Three main premises of efficient markets
A large number of competiting profit maximising participant analyse and value securities, each independently of the others
New info regarding securities comes to the markets in a random fashion
Profit maximising investors adjust security prices rapidly to reflect the effect of new info
Design of event studies
1, event definiton
- Treatment of confounding effects
- Selection of an appropriate model
- Test for significance and their power
- Moderating analysis
- Event definition and sampling
Is the event unambiguously defined and visible to investours
Determine and define the type of the event
Determine the selection criteria eg only consider one firm, selection of competitors of a focal firm (all firms in the same industry)
- Treatment of confounding events
How are confounding observations being handled, should they be excluded from analysis?
One common concern in event studies is how to handle cases when multiple announcements by the same entity occur in close proximity
Confounding events are events that may overlap with the effect of the focal event
Eg a firm may announce the intro of a new product a day af he r they announced that they will not meet their earnings estimates
- Selection of an appropriate model: model estimation
Calculate the mean of the returns of firm over the estimation period
Most popular model
For calculating expected returns
- Tests for significance and their power
Are the (cumulative) abnormal returns at or around the time of the event statistically significant?
Null hypothesis: h0 = the event has no impact on firm value and observed returns are different from zero only by chance
Alternative hypothesis: h1 doesn’t equal 0 (non-directional hypothesis) h1:car>0 (directional hypothesis) h1:car<0 (directional hypothesis)
The event has an impact of firm value
Tests for significance:
T tests / standardised residual tests / corrado rank test / generalised sign test / skewness adjusted test