Panel Data Econometrics

Question 1

What types of data are there?

Answer 1

Time series

Cross-sectional

Pooled

Panel

Question 2

What is time series data?

Answer 2

Time series is a set of observations on the values that a variable takes at different times (daily, weekly, monthly, quarterly, annually, etc.).

Question 3

What is cross-sectional data?

Answer 3

Cross-section are data on one or more variables collected at the same point in time.

Question 4

What is pooled data?

Answer 4

•Pooled data combine both time series and cross-sectional data. (Note: the cross sectional units differ i.e. in Year 1 we view firms A, B, C. in Year 2 we observe firms D, E, F)

Question 5

What is panel data?

Answer 5

Panel or Longitudinal data are a special type of pooled data in which the same cross-sectional unit (a family or a firm) is surveyed over time.

Question 6

What are Micro panels?

Answer 6

• Data collected for a large number of individuals (in 100,000’s)
• Collected over a short period of time
• Can vary from a minimum of 2 years to a maximum rarely exceeding 10-20 years

Question 7

What are Macro panels?

Answer 7

• Usually include a number of countries over time
• This may have a moderate number of countries (i.e. 20 OECD countries). Doesn’t normally exceed 100 countries
• Observed annually over 20-60 years
• Very high frequency data (i.e. stock indices varying over time)

Question 8

What are the differences between macro and micro panels? (2)

Answer 8

Macro panels have to deal with such issues as non-stationarity, unit roots, structural breaks, cointegration (micro panels don’t have these issues)
Macro panels also may include cross-country dependence, which micro panels often don’t have to deal with due to firms/households being randomly sampled

Question 9

What are the benefits of using panel data? (11)

Answer 9

• Controlling for individual heterogeneity
• Give more informative data, more variability, less collinearity between the variables, more degrees of freedom and more efficiency
• Dynamics of adjustment (change, duration of economic stance, speed of adjustment to economic policy changes, intertemporal relations)
• Identify and measure effects that are not detectible in pure-time series or pure-cross sectional data (i.e. does union membership increase wages? in panel data we can hold all other factors than job equal to see how much the wage changes)
• Construct and test more complicated behavioural models than pure-time series or pure-cross sectional data
• Biases resulting from aggregation over firms or individuals may be eliminated (micro panel data)
• Give more informative data;
• More variability (variation between states and variation within states);
• Less collinearity among the variable;
• More degrees of freedom;
• More efficiency.

Question 10

What are the limitations of using panel data? (5)

Answer 10

-Design and data collection problem (problems of coverage, nonresponse, recall, etc.)

• Distortions in measurement error (i.e. incorrect recording of data, participants lying about habits etc.)
• Selectivity problem (Self-selectivity, nonresponse, attrition)

–Self-Selectivity can be censored (one piece of info is missing i.e. wages), or truncated (multiple pieces of info are missing)

• Short time-series dimension for micro panel
• Cross-section dependence for macro panel

Question 11

What are the sources of endogeneity? (3)

Answer 11