# Panel and Grouped Data Course

This workbook is based upon the content of the RATS e-course on Panel and Grouped Data, offered in spring 2012. While it covers the basic techniques of panel data econometrics, the emphasis is on the "time-series" aspects, with Dynamic Panels (Chapter 7), Unit Root Tests (Chapter 9), Cointegration (Chapter 10) and VAR's (Chapter 13). It also includes several examples of the use of Gibbs sampling for panel data including linear and non-linear random effects in Chapter 11 and random coefficients models in Chapter 12 and applied to VAR's in Chapter (13).

This makes use of a number of important features for dealing with panel data added with RATS versions 8 and 8.1. The improvements to the core instructions PREGRESS, PANEL and PFORM provide greater flexibility for dealing with both panel and general grouped data.

### Workbook Contents

(195 pages, 18 examples)

#### 1 Introduction

1.1 Heterogeneous vs Homogeneous parameters
1.2 Panel vs Grouped Data
1.3 Balanced vs Unbalanced Panels
1.4 PANEL date scheme

#### 2 Preparing Data

2.1 Combining Multiple Time Series
2.2 Changing the Blocking
2.3 Mixed Time-Varying and Time-Invariant Data

#### 3 Computational Tools for Panel Data

3.1 SET, SSTATS and related instructions
3.2 PANEL
3.3 SWEEP
3.4 PSTATS
3.5 BOOT and Bootstrapping
3.6 LWINDOW=PANEL and Clustered Standard Errors

#### 4 Fixed Effects

4.1 A More Realistic Example
4.2 Balanced vs Unbalanced Samples
4.3 Implementing with RATS
4.4 Testing Fixed Effects
4.5 METHOD=BETWEEN
Example 4.1 Fixed Effects

#### 5 Random Effects

5.1 The Random Effects Estimator
5.2 Estimating the Component Variances
5.3 Using RATS
5.4 Hausman Test
5.5 Direct Calculation of Component Variances
5.6 Random Effects Transformations
Example 5.1 Random Effects

#### 6 Two-Way Effects

6.1 Balanced vs Unbalanced
6.2 Using RATS
6.3 Hausman Tests
6.4 PANEL instruction
Example 6.1 PREGRESS with Two-Way Effects

#### 7 Dynamic Panels

7.1 The Bias in Fixed Effects Estimators
7.2 The Examples
7.3 First Difference Instrumental Variables Estimators
7.4 Expanded Instrument Sets
7.5 Bias Correction
Example 7.1 Dynamic Panel-Instrumental Variables Estimation
Example 7.2 Dynamic Panel-Biased Correction Estimation

#### 8 Non-Linear Models

8.1 Non-linear Least Squares
8.2 Probit and Logit Models
8.2.1 Conditional (Fixed Effects) Logit
8.2.2 Random Effects Probit
Example 8.1 Non-linear Least Squares with Fixed Effects
Example 8.2 Logit and Probit with Individual Effects

#### 9 Unit Root Testing

9.1 The Example
9.2 Levin-Lin-Chu test
9.3 Harris-Tzavalis Test
9.4 Im-Pesaran-Shin Test
9.5 Breitung Test
Example 9.1 Panel Unit Root Tests

#### 10 Cointegration and Error Correction Models

10.1 Pedroni Tests
10.2 Estimation with Heterogeneous Cointegrating Vectors
10.2.1 Fully-Modified Least Squares
10.2.2 Dynamic OLS
10.3 Estimation with Homogeneous Cointegrating Vectors
10.3.1 Mark and Sul DOLS
10.3.2 Pesaran-Shin-Smith Pooled Mean Group
Example 10.1 Panel Cointegration with Heterogeneous Cointegrating Vectors
Example 10.2 Panel Cointegration with Homogeneous Cointegrating Vector
Example 10.3 Panel Error Correction Model

#### 11 Simulation and Bootstrap Methods

11.1 Linear Random Effects
11.2 Random Effects Probit
11.3 Bootstrapping
Example 11.1 Random Effects by Gibbs Sampling
Example 11.2 Random Effects Probit by Gibbs Sampling
Example 11.3 Bootstrapping a Unit Root Test

#### 12 Mean Group and Related Estimators

12.1 Mean Group Estimator
12.2 Swamy Random Coefficients Models
12.3 MCMC Estimation for Random Coefficients
Example 12.1 Mean Group and Random Coefficients Models

#### 13 Vector Autoregressions

13.1 Instrumental Variables Estimators
13.2 Shrinkage Estimators, Univariate Autoregressions
13.3 Shrinkage Estimators, VAR's
13.4 Causality Tests
Example 13.1 Panel VAR with instrumental variables
Example 13.2 Univariate Autoregression: Shrinkage Estimator
Example 13.3 Vector Autoregression: Shrinkage Estimator
Example 13.4 Causality Test

#### A Random Effects Component Estimators

A.1 Individual Effects Only
A.1.1 Wallace-Hussain (OLS-based)
A.1.2 Amemiya or Wansbeek-Kapteyn (Fixed Effects-based)
A.1.3 Swamy-Arora (Fixed and Between)
A.2 Two-Way Effects
A.2.1 Wallace-Hussain
A.2.2 Amemiya (Fixed Effects-based)
A.2.3 Swamy-Arora (Fixed and Between)

#### C Probability Distributions

C.1 Chi-Squared Distribution
C.2 Univariate Normal
C.3 Multivariate Normal
C.4 Gamma Distribution
C.5 Inverse Gamma Distribution
C.6 Wishart Distribution