Lecture 2:

Measurement levels:

• Nominal:  number only tells the category and there is no ranking (no logical order) there is also no difference between the values (Can be changed to binary and dummy variables)
• Ordinal: Ordered categories, but no equal distance
• Interval: information about differences between points on a scale, there are equal distances (also scale variable)
• Ratio: same as interval, but with an absolute zero e.g. weight

Not all calculations can be performed on all data

Descriptive/summary statistics:

• Step 1
• Quantitative description of main features of data
• Just a summary
• Before actual analysis
• Tells us: which are the players of the game, which is the nature of the variable (discrete or continuous) and do you see any problems or need to keep things in mind (like min, max and negative values)

Summary statistics:

1. Number of observations (choose minimum)
2. Measures of central tendency

• Mean: influenced by extreme observations (use for interval of ratio à depends on skewness)
• Median: middle point when values are ranked in order of magnitude. If the sample size is an even number, the median id average of the two middle observations. Relatively unaffected by extreme scores (use for ordinal always and for interval or ratio depends on skewness)
• Mode: Most frequent value, but there could be more than one mode (use for nominal variable)

3. Skewness says something about the shape of the distributions and the deviation from normal, the skew is 0. So it is compared to normal. If the value is positive than it is positively skewed, otherwise negatively. If the value is outside {-1,1} range than it is a substantially skewed distribution. Not skewed use mean otherwise use median
4. Kurtosis says something about the shape of the distribution and the deviation from normal. If it is normal than kurtosis is 3. Leptopkurtic: heavy tails (may also) pointy (>3) and platykurtic: light tail may also (flatter) (<3) in spss 0 is 3
5. Minimum and maximumà can define the range (measure of variability, defined as maximum value-minimum value and is affected by extreme scores). Also interquartile range (measure of variability q1 (lower quartile (25%) and q3 is upper quartile (75%)
6. Variance and standard deviationà square root of variance, how spread out the data are from the mean and heterogeneity of the sample (measures of variability)

Measure of association

1. Correlation (coefficient).

• Strength of relationship between 2 variable (can be positive or negative)
• [-1,1] magnitude says something about the strength of the relationship
• Coefficient of +1 or -1: two variables are perfectly positively (negatively) correlated. As one increases the other increases (decreases) by a proportionate amount
• Coefficient of 0: no linear relationship between two variables. As one variable changes the other stays the same
• Significance (significantly different from 0 or not)
• Word of warning: direction from causality (from A to B or the other way around) 3^rd variable problemà and unmeasured variable that influences the other two. So correlation is not always causality.

For graphs:

• Use an appropriate scaling
• Make sure you graph is self-explanator

Article:

What are we looking at:

• 433 cross border M&A announcements
• 58 EMMs
• 1991-2004
• Sample: Range of industries, mainly from latin America and Asia. Two firms from HUN and three from ZAF. 78.9% of the transactions initiated by Asian EMM’s
• Data: Thomsom SDC Platinum Database
• How to include the descriptive statistics,: First what you example then respectively mean, media, standard deviation and then Kurtosis and skewness.
• You can also include correlations
• Event study methodology. To see the impact of announcements on the value of the acquiring firm. Conflicting evidence (opportunities and  challenges. Diverse conditions and operation flexibility (+) Post acquisitions integration and liability of foreignness  (-). However Cross-border expansions of EMMs do not create value.
• Analysis of a cross-sectional sample of firms: To see what influences the market reaction. It has an effect on bidder value. Positive: target size, ownership structure of the target (private vs. public and structure of the bidder. Negative: high-tech nature of the bidder, pursuit of target in related industries
• Limitations: assumption: The market response is instantaneous, complete and unbiased. Regional concentration of the parent companies in Asia and Latin America à can we generalize  (external validity)? Focus on the bidder’s perspective à But what happens if we look at the combined value of the bidder and the target.