PANEL DATA ANALYSIS
Introduction
In recent years, panel (or longitudinal) datasets have become more widely
available. This availability has, in turn, motivated the development of
appropriate econometric techniques for analyzing the datasets.
Panel data entails observing the same cross-section units (e.g., individuals,
households, firms, regions, countries, etc.) over time. This observations of
cross-section units (the cross-section units are also called the entities) over
time introduces a defined structure to the panel dataset, which is a source of
additional information that could potentially be positively exploited to study
a wider range of issues than would otherwise be possible if pure cross-
section or pure time series data are used. The purpose of panel data analysis
is to devise the best ways to extract the additional information that panel
datasets provide.
In light of the two-dimensional nature of a panel dataset, N is used to refer
to the number of cross-section units (e.g., individuals, households, firms,
regions, countries, etc.) comprising the panel dataset and T to and the
corresponding number of time periods.. Hence, a panel dataset with large N
and small T means that the cross-section dimension is larger than the time
dimension of the panel dataset, the so-called micro panel; and a panel
dataset with small N and large T means that the cross-section dimension is
smaller than the time dimension of the panel dataset, the so-called macro
panel.
1
, Generally, a panel model is given as:
𝑌𝑖𝑡 =𝛽0 + 𝛽1 𝑋𝑖𝑡 + 𝑢𝑖𝑡
Where i represents the cross section component, and t represents the time
series component.
Advantages of Panel Data
Gives a researcher a larger number of data points. This increases
degrees of freedom and reduces multicollinearity
Allows analysis of a number of important economic questions that
cannot be addressed using cross-section or time series data sets.
With panel data, we are able to identify and measure effects that are
simply not detectable in pure cross section and time series.
The use of panel data provides a means of resolve/reducing
econometric problems that arise in empirical analysis e.g
heteroscedasticity.
Usually gathered on micro-units e.g households, firms, country and
thus variables are more accurately measured at micro-level and biases
that arise from aggregation from firms and households are eliminated.
controls omitted variable bias, less multicollinearity, better analysis of
dynamics
Disadvantages of Panel Data
One has to deal with design and data problems. This includes
problems of coverage, incomplete account of population of interest.
Non-response due to lack of cooperation from the respondents.
2
Introduction
In recent years, panel (or longitudinal) datasets have become more widely
available. This availability has, in turn, motivated the development of
appropriate econometric techniques for analyzing the datasets.
Panel data entails observing the same cross-section units (e.g., individuals,
households, firms, regions, countries, etc.) over time. This observations of
cross-section units (the cross-section units are also called the entities) over
time introduces a defined structure to the panel dataset, which is a source of
additional information that could potentially be positively exploited to study
a wider range of issues than would otherwise be possible if pure cross-
section or pure time series data are used. The purpose of panel data analysis
is to devise the best ways to extract the additional information that panel
datasets provide.
In light of the two-dimensional nature of a panel dataset, N is used to refer
to the number of cross-section units (e.g., individuals, households, firms,
regions, countries, etc.) comprising the panel dataset and T to and the
corresponding number of time periods.. Hence, a panel dataset with large N
and small T means that the cross-section dimension is larger than the time
dimension of the panel dataset, the so-called micro panel; and a panel
dataset with small N and large T means that the cross-section dimension is
smaller than the time dimension of the panel dataset, the so-called macro
panel.
1
, Generally, a panel model is given as:
𝑌𝑖𝑡 =𝛽0 + 𝛽1 𝑋𝑖𝑡 + 𝑢𝑖𝑡
Where i represents the cross section component, and t represents the time
series component.
Advantages of Panel Data
Gives a researcher a larger number of data points. This increases
degrees of freedom and reduces multicollinearity
Allows analysis of a number of important economic questions that
cannot be addressed using cross-section or time series data sets.
With panel data, we are able to identify and measure effects that are
simply not detectable in pure cross section and time series.
The use of panel data provides a means of resolve/reducing
econometric problems that arise in empirical analysis e.g
heteroscedasticity.
Usually gathered on micro-units e.g households, firms, country and
thus variables are more accurately measured at micro-level and biases
that arise from aggregation from firms and households are eliminated.
controls omitted variable bias, less multicollinearity, better analysis of
dynamics
Disadvantages of Panel Data
One has to deal with design and data problems. This includes
problems of coverage, incomplete account of population of interest.
Non-response due to lack of cooperation from the respondents.
2
