Tài liệu về hồi quy với dữ liệu bảng

Nội dung

• • Definition and motivation for panel data analysis.

  • Advantages of panel data over cross-sectional and time series data.

  • Examples of panel data in economics, finance, and social sciences.

  • Understanding the dimensions of panel data: Cross-sectional units and time periods.

  • Notation and terminology for panel data models.

  • Sources of panel data: Surveys, longitudinal studies, and administrative records.

  • Types of panel data: Balanced vs. unbalanced, short panels vs. long panels.

  • Overview of fixed effects, random effects, and mixed models.

Sách

• Baltagi, B. H. (2008). Econometric analysis of panel data (Vol. 4, pp. 135-145). Chichester: Wiley. (Chapter 1)

• Hsiao, C. (2022). Analysis of panel data (No. 64). Cambridge university press.

•  Greene, W. (2012) Econometric Analysis. 7th Edition (Chapters 11)

• Cameron, A. C. (2007, October). Panel data methods for microeconometrics using Stata. In West Coast Stata Users’ Group Meetings (Vol. 13).

Bài báo

• Kropko, J., & Kubinec, R. (2020). Interpretation and identification of within-unit and cross-sectional variation in panel data models. PloS one, 15(4), e0231349.

• Stimson James, A. (1985). Regression in space and time: A statistical essay. American Journal of Political Science, 29(4), 914-947.

• Zhu, L. (2012). Panel data analysis in public administration: Substantive and statistical considerations. Journal of Public Administration Research and Theory, 23(2), 395-428.

Nội dung

Section 1: Introduction to Linear Models for Panel Data

• • Advantages of panel data over cross-sectional and time-series data.

  • Applications of linear models in analyzing micro and macro panel data.

  • Types of linear panel data models: Pooled OLS, fixed effects, random effects, and mixed models.

Section 2: Pooled Ordinary Least Squares (OLS) for Panel Data

• Structure of Pooled OLS Model

  • General form of the pooled OLS model

  • Assumptions of pooled OLS: Homogeneity across cross-sectional units and over time.

• Estimation of Pooled OLS

  • Estimating pooled OLS and interpreting the coefficients.

  • Limitations of pooled OLS in the presence of unobserved heterogeneity.

• Example of Pooled OLS in Practice

  • Application of pooled OLS in wage determination across workers.

  • Empirical estimation using real panel data.

Section 3: The Fixed Effects (FE) Model

• Introduction to Fixed Effects

  • Motivation for fixed effects: Controlling for unobserved individual-specific characteristics.

  • General form of the fixed effects model

• Fixed Effects Estimation (Within Estimator)

  • The within transformation: Removing individual-specific effects via de-meaning.

  • Estimation of fixed effects: Least Squares Dummy Variable (LSDV) approach vs. within estimator.

• Interpretation of Fixed Effects Coefficients

  • Interpreting the coefficients: How individual-specific effects are captured in fixed effects.

• Advantages and Limitations of Fixed Effects Models

  • Controlling for time-invariant unobserved heterogeneity.

  • Limitations: Inability to estimate time-invariant variables.

Section 4: The Random Effects (RE) Model

• Introduction to Random Effects

  • The concept of random effects: Assuming individual effects are uncorrelated with regressors.

  • General form of the random effects model

• Estimation of Random Effects Model

  • Estimating random effects via Generalized Least Squares (GLS).

  • Interpretation of coefficients in random effects models.

• Advantages and Limitations of Random Effects Models

  • Advantages: Estimating both time-variant and time-invariant variables.

  • Limitations: Potential bias if individual effects are correlated with regressors.

• Example of Random Effects Model in Practice

  • Application: Estimating the determinants of firm productivity using random effects.

  • Empirical example with interpretation of results.

    1. 1. • Choosing Between Fixed Effects and Random Effects Models: The Hausman Test for Model Selection

• • Introduction to the Hausman test: Fixed effects vs. random effects.

  • Null and alternative hypotheses: Consistency of random effects estimator.

  • Interpretation of Hausman test results and model choice.

• Hausman Test in Practice

  • Application: Choosing between fixed and random effects in panel data on economic growth.

  • Practical steps for implementing the Hausman test in statistical software.

Section 6: Two-Way Fixed Effects Model

• Incorporating Time Fixed Effects

  • General structure of two-way fixed effects model.

  • Estimating two-way fixed effects: Controlling for both individual and time effects.

• Interpretation of Two-Way Fixed Effects Coefficients

  • Capturing time-varying factors common to all cross-sectional units 

• Example of Two-Way Fixed Effects in Practice

  • Application: Estimating the impact of a national policy on employment using two-way fixed effects.

  • Empirical estimation with interpretation of results.

Section 7: Panel Data Models with Time-Invariant Variables

• Dealing with Time-Invariant Variables in Fixed Effects Models

  • The problem of estimating time-invariant variables in fixed effects models.

  • Solutions for estimating time-invariant variables using random effects or hybrid models.

• Example of Time-Invariant Variables in Practice

  • Application: Estimating the effect of geography on trade flows using random effects.

Section 8: Testing Assumptions and Diagnostics for Linear Panel Data Models

• Testing for Serial Correlation in Panel Data

  • Methods for detecting serial correlation: Wooldridge test.

  • Impact of serial correlation on panel data models and corrections.

• Testing for Heteroskedasticity

  • Tests for heteroskedasticity in panel data: Breusch-Pagan test.

  • Adjusting for heteroskedasticity using robust standard errors.

• Testing for Cross-Sectional Dependence

  • Pesaran’s CD test for cross-sectional dependence.

  • Consequences of cross-sectional dependence and correction methods.

• Example of Diagnostics in Practice

  • Empirical example: Running diagnostic tests on a panel data model of firm investment.

Section 9: Extensions of Linear Panel Data Models

• 9.1 Dynamic Panel Data Models

  • Introduction to dynamic panel data models: Incorporating lagged dependent variables.

  • Estimation techniques: Arellano-Bond GMM.

  • Applications of dynamic models: Economic growth, investment, consumption.

• Panel Data with Endogeneity: Instrumental Variables

  • Dealing with endogeneity in linear panel data models.

  • Instrumental variables (IV) estimation in panel data.

• Nonlinear Extensions of Linear Panel Data Models

  • Introduction to nonlinear panel data models: Logit, Probit, Tobit.

Section 10: Applications of Linear Panel Data Models in Practice

• Case Study 1: Estimating Wage Dynamics Using Panel Data

  • Applying fixed and random effects models to study individual wage changes over time.

• Case Study 2: Analyzing Firm Performance with Panel Data

  • Estimating the impact of R&D spending on firm productivity using random effects.

• Case Study 3: Evaluating Policy Impact with Two-Way Fixed Effects

  • Example: Estimating the effect of a healthcare policy on population health using two-way fixed effects.

Section 11: Software Implementation of Linear Panel Data Models

• Implementing Fixed and Random Effects Models in Stata

  • Step-by-step guide to using Stata for fixed and random effects estimation.

• Implementing Linear Panel Data Models in R

  • Overview of R packages (plm, lme4) for linear panel data models.

• Implementing Linear Panel Data Models in EViews and MATLAB

  • Code and procedures for estimating linear models in EViews and MATLAB

Section 12: Common Challenges and Best Practices in Linear Panel Data Analysis

• Addressing Common Pitfalls in Panel Data Models

  • Issues with multicollinearity, heteroskedasticity, and serial correlation.

  • Handling missing data and unbalanced panels.

• Best Practices for Reporting and Interpreting Panel Data Results

  • Guidelines for interpreting coefficients, diagnostics, and robustness checks.

• Example of Best Practices in Reporting Results

  • Real-world example: Presenting panel data results in research papers.

Section 13: Conclusion and Further Reading

• Summary of Key Concepts in Linear Panel Data Models

  • Recap of fixed effects, random effects, pooled OLS, and two-way fixed effects models.

• Suggested Further Reading

  • Foundational papers and textbooks for deeper exploration of linear panel data models.

Sách

• Hsiao, C. (2022). Analysis of panel data (No. 64). Cambridge university press. (Chapter 3 )

• Wooldridge, J. M. (2010). Econometric analysis of cross section and panel data. MIT press. (Chapter 6 + 10)

• Baltagi, B. H., & Baltagi, B. H. (2008). Econometric analysis of panel data (Vol. 4, pp. 135-145). Chichester: Wiley. (Chapter 2+4)

• Greene, W. (2012) Econometric Analysis. 7th Edition (Chapters 11)

• Pesaran, M. H. (2015). Time series and panel data econometrics. Oxford University Press. (Chapter 26)

• Cameron, A. C., & Trivedi, P. K. (2010). Microeconometrics using stata (Vol. 2). College Station, TX: Stata press. (Chapter 8)

• Arellano, M. (2003). Panel data econometrics. OUP Oxford. (Chapter 2) 

Other ways of modeling heterogeneity in panel data models

• • Pesaran, M. H. (2015). Time series and panel data econometrics. Oxford University Press. (Chapter 28 + 29)

Bài báo

• Almond, D., Chay, K. Y., & Lee, D. S. (2005). The costs of low birth weight. The Quarterly Journal of Economics, 120(3), 1031-1083.

• Beck, N. (2001). Time-series–cross-section data: What have we learned in the past few years?. Annual review of political science, 4(1), 271-293.

• Bell, A., & Jones, K. (2015). Explaining fixed effects: Random effects modeling of time-series cross-sectional and panel data. Political Science Research and Methods, 3(1), 133-153.

• Bleakley, H. (2010). Malaria eradication in the Americas: A retrospective analysis of childhood exposure. American Economic Journal: Applied Economics, 2(2), 1-45.

• Clark, T. S., & Linzer, D. A. (2015). Should I use fixed or random effects?. Political science research and methods, 3(2), 399-408.

• Dieleman, J. L., & Templin, T. (2014). Random-effects, fixed-effects and the within-between specification for clustered data in observational health studies: a simulation study. PloS one, 9(10), e110257.

• Finkel, S. E. (2008). Linear panel analysis. Handbook of longitudinal research: Design, measurement, and analysis, 475-504.

• Jordan, S., & Philips, A. Q. (2023). Improving the interpretation of random effects regression results. Political Studies Review, 21(1), 210-220.

• Kittel, B., & Winner, H. (2005). How reliable is pooled analysis in political economy? The globalization‐welfare state nexus revisited. European Journal of Political Research, 44(2), 269-293.

• Mummolo, J., & Peterson, E. (2018). Improving the interpretation of fixed effects regression results. Political Science Research and Methods, 6(4), 829-835.

• Mundlak, Y. (1978). On the pooling of time series and cross section data. Econometrica: journal of the Econometric Society, 69-85.

• Plümper, T., & Troeger, V. E. (2019). Not so harmless after all: The fixed-effects model. Political Analysis, 27(1), 21-45.

Cross-sectional effect heterogeneity in static models

• Beck, N., & Katz, J. N. (2007). Random coefficient models for time-series—cross-section data: Monte Carlo experiments. Political Analysis, 15(2), 182-195.

Hypothesis testing

• Ma, J., & Vijverberg, W. P. (2010). Five Diagnostic Tests for Unobserved Cluster Effects. Communications in Statistics-Simulation and Computation, 39(6), 1212-1227.

Other ways of modeling heterogeneity in panel data models

•  Bai, J. (2009). Panel data models with interactive fixed effects. Econometrica, 77(4), 1229-1279.

• Coakley, J., Fuertes, A. M., & Smith, R. (2006). Unobserved heterogeneity in panel time series models. Computational Statistics & Data Analysis, 50(9), 2361-2380.

• Guvenen, F. (2009). An empirical investigation of labor income processes. Review of Economic dynamics, 12(1), 58-79.

• Pesaran, M. H. (2006). Estimation and inference in large heterogeneous panels with a multifactor error structure. Econometrica, 74(4), 967-1012.

• Plümper, T., & Troeger, V. E. (2007). Efficient estimation of time-invariant and rarely changing variables in finite sample panel analyses with unit fixed effects. Political analysis, 15(2), 124-139.

• Plümper, T., & Troeger, V. E. (2011). Fixed-effects vector decomposition: properties, reliability, and instruments. Political Analysis, 19(2), 147-164.

• Schunck, R. (2013). Within and between estimates in random-effects models: Advantages and drawbacks of correlated random effects and hybrid models. The Stata Journal, 13(1), 65-76.

Nội dung

Section 1: Introduction to Model Assumptions in Linear Panel Data Models

• Importance of Testing Assumptions in Panel Data Models

  • The role of assumptions in ensuring consistent and unbiased estimators.

  • Consequences of violating assumptions in linear panel data models.

• Overview of Key Assumptions in Panel Data Models

  • Homoskedasticity.

  • No serial correlation.

  • Cross-sectional independence.

  • Exogeneity of regressors.

Section 2: Testing for Homoskedasticity in Panel Data Models

• Assumption of Homoskedasticity

  • Definition and importance of homoskedasticity: Constant variance of errors across individuals and time.

  • Consequences of heteroskedasticity: Inefficient estimates and biased standard errors.

• Breusch-Pagan Test for Heteroskedasticity

  • Overview of the Breusch-Pagan test for detecting heteroskedasticity.

  • Assumptions and testing procedure for the Breusch-Pagan test.

• Likelihood Ratio Test for Heteroskedasticity

  • Using the likelihood ratio test to detect heteroskedasticity.

• Adjusting for Heteroskedasticity

  • Robust standard errors: Huber-White sandwich estimator.

  • Feasible Generalized Least Squares (FGLS) to correct for heteroskedasticity.

• Example: Testing for Heteroskedasticity in Wage Data

  • Practical example of detecting and correcting for heteroskedasticity in panel data.

Section 3: Testing for Serial Correlation (Autocorrelation) in Panel Data Models

• Assumption of No Serial Correlation

  • Definition and importance of serial correlation in panel data.

  • Consequences of serial correlation: Biased standard errors and incorrect inference.

• Wooldridge Test for Serial Correlation

  • Overview of the Wooldridge test for serial correlation in panel data.

  • Assumptions and testing procedure.

• Durbin-Watson Test for Serial Correlation

  • Applying the Durbin-Watson test in fixed effects models.

• Correcting for Serial Correlation

  • Use of robust standard errors (clustered standard errors) to correct for serial correlation.

  • Feasible Generalized Least Squares (FGLS) for correcting autocorrelation.

• Example: Detecting and Correcting Serial Correlation in Firm-Level Productivity Data

  • Empirical application: Correcting for autocorrelation in firm-level panel data.

Section 4: Testing for Cross-Sectional Dependence in Panel Data Models

• Assumption of Cross-Sectional Independence

  • Importance of cross-sectional independence in panel data models.

  • Consequences of cross-sectional dependence: Biased parameter estimates and inefficient estimators.

• Pesaran's Cross-Sectional Dependence (CD) Test

  • Overview of Pesaran’s CD test for detecting cross-sectional dependence.

  • Assumptions and interpretation of Pesaran's CD test results.

• Breusch-Pagan LM Test for Cross-Sectional Dependence

  • Applying the Breusch-Pagan LM test to detect cross-sectional dependence.

• Correcting for Cross-Sectional Dependence

  • Robust standard errors clustered at the cross-sectional level.

  • Estimating models with Common Correlated Effects (CCE) to address cross-sectional dependence.

• Example: Detecting Cross-Sectional Dependence in Trade Data

  • Empirical example of detecting and addressing cross-sectional dependence in international trade panel data.

Section 6: Testing for Endogeneity in Panel Data Models

• Assumption of Exogeneity

  • Definition of exogeneity: No correlation between regressors and the error term.

  • Consequences of endogeneity: Biased and inconsistent parameter estimates.

• Instrumental Variables (IV) and Generalized Method of Moments (GMM)

  • Using IV or GMM to correct for endogeneity in linear panel data models.

  • Selecting valid instruments in panel data.

• Example: Detecting and Correcting for Endogeneity in a Panel Data Model of Investment

  • Empirical application: Detecting endogeneity in a model of firm investment and correcting using IV or GMM.

Section 7: Testing for Model Misspecification in Panel Data

• Ramsey RESET Test

  • Using the Ramsey RESET test for functional form misspecification in panel data.

  • Assumptions and testing procedure for the RESET test.

• Lagrange Multiplier Test for Random Effects

  • Testing whether random effects are necessary using the Lagrange Multiplier test.

• Testing for Omitted Variables

  • Detecting omitted variable bias using the Hausman specification test.

• Example: Testing for Model Misspecification in Wage Determination Models

  • Practical example of detecting omitted variables and functional form issues.

Section 8: Model Diagnostics for Linear Panel Data Models

• Checking for Outliers and Leverage Points

  • Identifying influential observations in panel data: Cook's distance, DFBETAS.

  • Dealing with outliers: Robust regression techniques.

• Example: Performing Residual Diagnostics for Firm-Level Data

  • Step-by-step example of running residual diagnostics for a panel data model of firm productivity.

Section 9: Practical Application of Diagnostics in Software

• Implementing Diagnostic Tests in Stata

  • Commands for testing heteroskedasticity, serial correlation, cross-sectional dependence, and endogeneity.

  • Example: Implementing diagnostics in a panel data model of firm investment.

• Implementing Diagnostic Tests in R

  • Packages and functions in R for running diagnostic tests (plm, lmtest, sandwich).

  • Example: Using R to test for model misspecification and heteroskedasticity.

• Implementing Diagnostic Tests in EViews and MATLAB

  • Practical guide to running diagnostics in EViews and MATLAB for panel data models.

Section 10: Best Practices for Diagnostics and Reporting Results

• Interpreting Diagnostic Test Results

  • How to interpret diagnostic test outcomes and their implications for model choice.

• Adjusting Models Based on Diagnostics

  • Choosing the appropriate corrections based on test results (e.g., robust standard errors, clustered standard errors, FGLS).

• Best Practices for Reporting Diagnostics in Research

  • Guidelines for reporting test results in papers and reports.

• Example: Reporting Diagnostics for a Study on Environmental Policy and Economic Growth

  • Example of a well-presented diagnostics section in an academic paper.

Section 11: Conclusion and Further Reading

• Summary of Key Diagnostics in Linear Panel Data Models

  • Recap of key assumptions and diagnostic tests.

  • Importance of proper diagnostics for valid econometric analysis.

• Suggested Further Reading

  • Foundational papers and resources for deepening understanding of diagnostics in panel data.

Sách

• Baltagi, B. H., & Baltagi, B. H. (2008). Econometric analysis of panel data (Vol. 4, pp. 135-145). Chichester: Wiley. (Chapter 5)

• Hsiao, C. (2022). Analysis of panel data (No. 64). Cambridge university press. (Chapter 3 ) 

• Greene, W. (2012) Econometric Analysis. 7th Edition (Chapters 11)

Bài báo

• Beck, N., & Katz, J. N. (1995). What to do (and not to do) with time-series cross-section data. American political science review, 89(3), 634-647.

• King, G., & Roberts, M. E. (2015). How robust standard errors expose methodological problems they do not fix, and what to do about it. Political Analysis, 23(2), 159-179.

Nội dung

Section 1: Introduction to Cross-Sectional Dependence in Panel Data

• • Explanation of cross-sectional dependence: Correlation across cross-sectional units in panel data.

  • Causes of cross-sectional dependence: Common shocks, spatial spillovers, global trends.

  • Consequences of ignoring cross-sectional dependence: Biased and inconsistent estimators, inefficiency in hypothesis testing.

  • Key examples of cross-sectional dependence in economic, financial, and environmental data.

  • Overview of different methods for detecting and correcting cross-sectional dependence.

Section 2: Causes of Cross-Sectional Dependence

• Common Shocks and Global Factors

  • Explanation of common shocks: Economic, financial, or policy events affecting all cross-sectional units.

  • Example: Global financial crisis and its impact on countries.

• Spatial Interdependence

  • Spatial spillovers and their contribution to cross-sectional dependence.

  • Example: Contagion effects in regional or neighboring economies.

• Unobserved Common Factors

  • Impact of unobserved common factors driving cross-sectional dependence.

  • Example: Unobserved technological progress affecting all firms in an industry.

Section 3: Testing for Cross-Sectional Dependence

• Overview of Cross-Sectional Dependence Tests

  • Key tests for detecting cross-sectional dependence: Pesaran’s CD test, Breusch-Pagan LM test, and other methods.

• Pesaran’s Cross-Sectional Dependence (CD) Test

  • Definition and mathematical formulation of the CD test.

  • Null hypothesis and alternative hypothesis of the CD test.

  • Step-by-step guide to calculating and interpreting Pesaran’s CD test results.

• Breusch-Pagan LM Test

  • Introduction to the Breusch-Pagan LM test for cross-sectional dependence.

  • Applicability to large and small panels.

  • Testing procedure and limitations of the Breusch-Pagan LM test.

• Pesaran Scaled LM Test

  • Overview of the scaled LM test for large panel datasets.

  • Applications and advantages in panels with a large cross-sectional dimension.

• Bias-Corrected Scaled LM Test

  • Correction for potential bias in the scaled LM test.

  • Example: Applying the bias-corrected scaled LM test in practice.

• Example of Cross-Sectional Dependence Testing in Practice

  • Application: Detecting cross-sectional dependence in global stock market data.

Section 4: Consequences of Cross-Sectional Dependence

• Impact on Estimation

  • Consequences for pooled OLS, fixed effects, and random effects models.

  • Bias and inefficiency in parameter estimates due to cross-sectional dependence.

• Impact on Hypothesis Testing

  • Incorrect inference from standard tests (e.g., t-tests, F-tests) when cross-sectional dependence is present.

  • Example: How ignoring cross-sectional dependence affects policy analysis.

Section 5: Addressing Cross-Sectional Dependence in Panel Data Models

• Robust Standard Errors

  • Introduction to heteroskedasticity and autocorrelation consistent (HAC) standard errors.

  • Using clustered standard errors to account for cross-sectional dependence.

  • Example: Implementing robust standard errors in a panel data model of regional GDP growth.

• Feasible Generalized Least Squares (FGLS)

  • Overview of FGLS as a method to correct for cross-sectional dependence.

  • Estimation procedure for FGLS: Addressing both heteroskedasticity and cross-sectional dependence.

  • Limitations and potential bias in small samples.

  • Example: Applying FGLS to estimate inflation persistence across countries.

• Driscoll-Kraay Standard Errors

  • Explanation of Driscoll-Kraay standard errors for correcting cross-sectional dependence.

  • Advantages of Driscoll-Kraay in panels with both cross-sectional dependence and serial correlation.

  • Example: Using Driscoll-Kraay standard errors to analyze the effect of education on earnings.

Section 6: Common Correlated Effects (CCE) Estimator

• Introduction to Common Correlated Effects (CCE)

  • Explanation of the CCE approach for dealing with cross-sectional dependence.

  • Incorporating common factors to account for cross-sectional dependence.

• Estimating the CCE Model

  • CCE pooled estimator vs. CCE mean group (CCEMG) estimator.

  • Applications in panels with large cross-sectional dimensions.

• Example: Applying the CCE Estimator in Global Macroeconomic Data

  • Example: Using the CCE estimator to model the impact of global oil prices on economic growth.

Section 7: Factor Models for Cross-Sectional Dependence

• Introduction to Factor Models

  • Explanation of factor models: Modeling cross-sectional dependence with unobserved common factors.

• Estimating Factor Models

  • Estimation techniques: Principal components analysis (PCA) for extracting common factors.

  • Examples of factor models in practice: Macro-financial linkages, global trade flows.

• Applications of Factor-Augmented Panel Data Models

  • Factor-augmented VAR (FAVAR) models in macroeconomic analysis.

  • Example: Estimating a factor model for cross-country interest rate differentials.

Section 8: Dealing with Spatial Dependence in Panel Data

• Introduction to Spatial Panel Data Models

  • Explanation of spatial dependence: Spatial autocorrelation and its role in cross-sectional dependence.

• Spatial Lag Models (SLM) and Spatial Error Models (SEM)

  • Structure of spatial lag and spatial error models.

  • Estimating spatial panel data models: Maximum Likelihood (ML) and Generalized Method of Moments (GMM).

• Example: Applying Spatial Panel Data Models in Regional Economics

  • Application of spatial models to regional income and growth dynamics.

Section 9: Dynamic Panel Models with Cross-Sectional Dependence

• Introduction to Dynamic Panel Models

  • Overview of dynamic panel models: Incorporating lagged dependent variables.

  • The challenge of cross-sectional dependence in dynamic panels.

• System GMM with Cross-Sectional Dependence

  • Using System GMM for dynamic panel models with cross-sectional dependence.

  • Testing for cross-sectional dependence in dynamic panels.

• Example: Estimating Dynamic Panel Models for Investment and Growth

  • Application of dynamic models with cross-sectional dependence in corporate finance.

Section 10: Practical Implementation of Cross-Sectional Dependence Tests and Corrections

• Implementing Cross-Sectional Dependence Tests in Stata

  • Commands for Pesaran’s CD test, Breusch-Pagan LM test, and robust standard errors.

  • Example: Running cross-sectional dependence diagnostics in a panel data model of firm performance.

• Implementing Cross-Sectional Dependence Tests in R

  • Using R packages (plm, lmtest) for cross-sectional dependence testing.

  • Example: Applying factor models and CCE estimators in R for macroeconomic analysis.

• Implementing Cross-Sectional Dependence Tests in EViews and MATLAB

  • Practical guide to running diagnostics and corrections for cross-sectional dependence in EViews and MATLAB.

Section 11: Best Practices for Addressing Cross-Sectional Dependence in Research

• Interpretation of Test Results

  • How to interpret test results and decide on corrective actions.

  • Choosing between different correction methods: Robust standard errors, FGLS, CCE, and factor models.

• Reporting Cross-Sectional Dependence in Research Papers

  • Best practices for presenting cross-sectional dependence diagnostics and corrections.

  • Example of reporting cross-sectional dependence tests in empirical research.

Section 12: Conclusion and Further Reading

• Summary of Key Methods for Detecting and Correcting Cross-Sectional Dependence

  • Recap of testing methods: Pesaran’s CD test, Breusch-Pagan LM test, robust standard errors.

  • Overview of estimation methods: FGLS, Driscoll-Kraay, CCE, and factor models.

• Suggested Further Reading

  • Foundational papers and textbooks for deeper exploration of cross-sectional dependence in panel data.

Sách

• Baltagi, B. H., & Baltagi, B. H. (2008). Econometric analysis of panel data (Vol. 4, pp. 135-145). Chichester: Wiley.

Tài liệu/Bài giảng tham khảo

• General Diagnostic Tests for Cross Section Dependence in Panels (link)

Bài báo

• De Hoyos, R. E., & Sarafidis, V. (2006). Testing for cross-sectional dependence in panel-data models. The stata journal, 6(4), 482-496.

• Pesaran, M. H. (2004). General diagnostic tests for cross section dependence in panels. Cambridge Working Papers. Economics, 1240(1), 1.

• Parks, R. W. (1967). Efficient estimation of a system of regression equations when disturbances are both serially and contemporaneously correlated. Journal of the american statistical association, 62(318), 500-509.

• Phillips, P. C., & Sul, D. (2003). Dynamic panel estimation and homogeneity testing under cross section dependence. The econometrics journal, 6(1), 217-259.

• Baltagi, B. H., & Pinnoi, N. (1995). Public capital stock and state productivity growth: Further evidence from an error components model. Empirical Economics, 20, 351-359.

• Frees, E. W. (1995). Assessing cross-sectional correlation in panel data. Journal of econometrics, 69(2), 393-414.

Nội dung

Section 1: Introduction to Two-Stage Least Squares (2SLS) in Panel Data

• Review of 2SLS Estimation

  • Definition of Two-Stage Least Squares (2SLS): Purpose and applications.

  • The problem of endogeneity in econometric models: Omitted variable bias, measurement error, and reverse causality.

  • Endogeneity and the challenges of dealing with endogeneity in panel data.

  • Role of panel data structure in improving the precision of instrumental variable estimation.

Section 2: 2SLS in Panel Data

• Endogeneity in Panel Data Models

  • Causes of endogeneity in panel data: Unobserved heterogeneity, simultaneity, and measurement error.

  • Example: Endogeneity in wage determination models due to unobserved ability

• Linear Panel Data Models and Endogeneity

  • Fixed effects (FE) and random effects (RE) models with endogenous regressors.

  • Example: Simultaneous determination of labor supply and wages in panel data.

• Fixed Effects 2SLS (FE-2SLS)

  • Incorporating 2SLS into fixed effects models: Addressing individual-specific unobserved heterogeneity.

  • First-stage regression for fixed effects: Accounting for within-individual variation.

• Random Effects 2SLS (RE-2SLS)

  • 2SLS in random effects models: Accounting for between- and within-variation.

  • Adjusting for correlation between individual effects and regressors in random effects models.

    1. 1. 1. • The Two-Stage Least Squares (2SLS) Estimation Procedure

• First Stage of 2SLS: Instrumenting Endogenous Regressors

  • Selecting appropriate instruments: Criteria for instrument relevance and exogeneity.

  • Estimating the first-stage regression for the endogenous variable(s).

  • Interpretation of the first-stage regression results: Predicted values of endogenous variables.

• Second Stage of 2SLS: Estimation of the Structural Equation

  • Using the predicted values of endogenous variables in the main regression.

  • Estimating the coefficients of the structural model using predicted values.

  • Interpretation of second-stage results.

• Example of 2SLS Estimation in Panel Data

  • Step-by-step example: Using 2SLS to estimate the impact of education on wages with endogenous education variable.

Section 3: Testing Assumptions and Diagnostics for 2SLS

• Weak Instruments

  • Definition and consequences of weak instruments: Bias in 2SLS estimates.

  • Detection of weak instruments: Stock-Yogo test for weak instruments.

  • Example: Assessing instrument strength in a panel data model of trade and economic growth.

• Overidentifying Restrictions

  • Testing the validity of instruments with more instruments than endogenous variables.

  • Sargan-Hansen test for overidentifying restrictions.

• Hausman Test for Endogeneity

  • Testing for the presence of endogeneity in panel data: Comparison between OLS and 2SLS estimates.

  • Interpretation of Hausman test results and model choice.

Section 4: Extensions of 2SLS in Panel Data Models

• Generalized Method of Moments (GMM)

  • Overview of GMM as an extension of 2SLS for panel data.

  • Differences between 2SLS and GMM: Efficiency gains from using additional moment conditions.

  • Example: Applying GMM to estimate dynamic panel models with endogenous regressors.

• Dynamic 2SLS in Panel Data Models

  • Estimation of dynamic panel models using 2SLS: Dealing with lagged dependent variables.

  • Example: Dynamic wage models with lagged wage as an endogenous regressor.

Section 5: Practical Applications of 2SLS in Panel Data

• 2SLS in Microeconomic Panel Data

  • Application of 2SLS to individual or firm-level panel data.

  • Example: Estimating labor supply models with endogenous hours of work.

• 2SLS in Macroeconomic Panel Data

  • Application of 2SLS to cross-country macroeconomic panel data.

  • Example: Estimating the effect of trade openness on GDP growth with endogenous trade variable.

• 2SLS in Policy Evaluation

  • Application of 2SLS for evaluating the impact of policy interventions using panel data.

  • Example: Estimating the effect of government subsidies on firm-level productivity using panel data and 2SLS.

Section 6: Implementation of 2SLS in Statistical Software

• 2SLS Estimation in Stata

  • Step-by-step guide to implementing 2SLS in Stata: xtivreg command.

  • Practical example: Estimating a panel data model with endogenous education variable in Stata.

• 2SLS Estimation in R

  • Using R packages (plm, AER) for 2SLS estimation in panel data.

  • Example: Applying 2SLS to panel data on healthcare spending and outcomes in R.

• 2SLS Estimation in EViews and MATLAB

  • Implementation of 2SLS estimation in EViews and MATLAB.

  • Example: Estimating a panel data model of environmental regulation and firm productivity in EViews.

Section 7: Common Challenges and Best Practices in 2SLS with Panel Data

• Dealing with Weak Instruments in 2SLS

  • Strategies for addressing weak instruments: Stronger external instruments, using more instruments.

  • Example: Correcting for weak instruments in a panel data model of financial development and growth.

• Reporting 2SLS Results

  • Best practices for reporting 2SLS estimates: First-stage results, tests for instrument strength, and overidentification tests.

  • Example: Structuring a well-reported empirical study using 2SLS in panel data.

Section 8: Conclusion and Further Reading

• Summary of Key Concepts in 2SLS with Panel Data

  • Recap of 2SLS, endogeneity, instrumental variables, and extensions like GMM.

• Suggested Further Reading

  • Foundational papers and textbooks for further study of 2SLS and IV methods in panel data econometrics.

Sách

• Baltagi, B. H., & Baltagi, B. H. (2008). Econometric analysis of panel data (Vol. 4, pp. 135-145). Chichester: Wiley. (Chapter 7)

• Greene, W. (2012) Econometric Analysis. 7th Edition (Chapters 13)

Bài báo

• Ahn, S. C., & Schmidt, P. (1999). Modified generalized instrumental variables estimation of panel data models with strictly exogenous instrumental variables. Analysis of Panels and Limited Dependent Variable Models, 171-198.

• Amemiya, T., & MaCurdy, T. E. (1986). Instrumental-variable estimation of an error-components model. Econometrica: Journal of the Econometric Society, 869-880.

• Baltagi, B. H., & Khanti‐Akom, S. (1990). On efficient estimation with panel data: An empirical comparison of instrumental variables estimators. Journal of Applied econometrics, 5(4), 401-406.

• Breusch, T. S., Mizon, G. E., & Schmidt, P. (1989). Efficient estimation using panel data. Econometrica: Journal of the Econometric Society, 695-700.

• Cornwell, C., & Rupert, P. (1988). Efficient estimation with panel data: An empirical comparison of instrumental variables estimators. Journal of Applied Econometrics, 3(2), 149-155.

• Hausman, J. A., & Taylor, W. E. (1981). Panel data and unobservable individual effects. Econometrica: Journal of the Econometric society, 1377-1398.

• Im, K. S., Ahn, S. C., Schmidt, P., & Wooldridge, J. M. (1999). Efficient estimation of panel data models with strictly exogenous explanatory variables. Journal of Econometrics, 93(1), 177-201.

• Wilson, S. E., & Butler, D. M. (2007). A lot more to do: The sensitivity of time-series cross-section analyses to simple alternative specifications. Political analysis, 15(2), 101-123.

Nội dung

Section 1: Introduction to Dynamic Panel Data Models

• What Are Dynamic Panel Data Models?

  • Definition and characteristics of dynamic panel data models.

  • Key features of small T (short time periods) and large N (large cross-sectional units) settings.

  • Importance of lagged dependent variables: Capturing dynamic behavior over time.

  • Overview of real-world applications: Modeling firm profitability, economic convergence, and labor market dynamics.

Section 2: Challenges in Estimating Dynamic Panel Data Models (Small T, Large N)

• Nickell Bias in Dynamic Panel Models

  • Introduction to Nickell bias: Bias in fixed effects models when T is small.

  • Causes and consequences of Nickell bias.

  • Examples of bias in short panel settings: Labor market dynamics and firm-level investment.

• Endogeneity in Dynamic Panel Data

  • Problem of endogenous regressors in dynamic models: Lagged dependent variables and endogenous covariates.

  • Why OLS and within estimators are inconsistent in dynamic models.

Section 3: Generalized Method of Moments (GMM) for Dynamic Panel Data Models

• Introduction to GMM Estimation

  • Overview of the Generalized Method of Moments (GMM) framework.

  • Why GMM is suitable for small T, large N settings: Handling endogeneity and Nickell bias.

  • Moments and instruments in dynamic panel data models.

Section 4: Arellano-Bond Estimator (Difference GMM)

• The Arellano-Bond (1991) Estimation Method

  • Step-by-step explanation of the Arellano-Bond difference GMM estimator.

  • Transforming the model into first differences to eliminate individual effects.

  • Using lagged levels of the endogenous variables as instruments.

• Assumptions of the Arellano-Bond Estimator

  • Key assumptions: Absence of second-order serial correlation and exogeneity of instruments.

  • Role of predetermined and endogenous variables.

• Testing for Serial Correlation in Difference GMM

  • AR(1) and AR(2) tests for serial correlation in residuals.

  • Importance of passing the AR(2) test in Arellano-Bond estimation.

• Instrument Validity in Arellano-Bond Estimation

  • Sargan and Hansen tests for overidentifying restrictions.

  • Detecting weak instruments and instrument proliferation.

• Example: Estimating Investment and Growth with Arellano-Bond Estimator

  • Application: Step-by-step estimation of dynamic investment models using difference GMM.

Section 5: Blundell-Bond Estimator (System GMM)

• Introduction to Blundell-Bond (1998) Estimation

  • Motivation for system GMM: Addressing weak instruments in difference GMM.

  • Combining levels and differences in a system of equations to improve efficiency.

  • Using lagged differences as instruments for levels and vice versa.

• Assumptions of the Blundell-Bond Estimator

  • Key assumptions: Stationarity of the variables and validity of additional moment conditions.

  • Benefits of system GMM over difference GMM in finite samples.

• Example: Estimating Employment Dynamics Using System GMM

  • Step-by-step guide to estimating a dynamic panel model of employment with system GMM.

Section 6: Instrument Selection and Weak Instruments in GMM

• Challenges of Instrument Proliferation in GMM

  • Definition and impact of instrument proliferation: Overfitting of the model.

  • Trade-offs between the number of instruments and model efficiency.

• Instrument Selection Strategies

  • Collapsing instruments to avoid over-identification.

  • Using deeper lags as instruments to reduce the number of instruments.

  • Example: Instrument selection in a dynamic panel model of firm profitability.

• Dealing with Weak Instruments

  • Detection of weak instruments: Stock-Yogo weak instrument test.

  • Consequences of weak instruments for GMM estimates.

Section 7: Testing and Diagnostics in Dynamic Panel Data Models

• AR(1) and AR(2) Tests for Serial Correlation

  • Understanding the role of serial correlation tests in GMM models.

  • Correct interpretation of AR(1) and AR(2) results.

• Sargan and Hansen Tests for Overidentifying Restrictions

  • Using the Sargan test to detect overidentification in the model.

  • Hansen test: A robust alternative to Sargan for overidentifying restrictions.

  • Example: Interpretation of Sargan and Hansen test results in a model of economic growth.

Section 8: Alternative Estimation Methods for Dynamic Panel Data Models

• Anderson-Hsiao (AH) Estimator

  • Introduction to the Anderson-Hsiao estimator: A simpler approach to dynamic panel estimation.

  • When to use the Anderson-Hsiao method: Conditions and limitations.

• Maximum Likelihood Estimation (MLE) for Dynamic Panel Models

  • Explanation of the MLE approach for dynamic panel models.

  • Comparing MLE with GMM in terms of efficiency and bias.

• Empirical Example: Using Anderson-Hsiao and MLE for Dynamic Firm-Level Data

  • Example: Estimating a dynamic model of firm growth using alternative estimators.

Section 9: Practical Applications of Dynamic Panel Data Models

• Dynamic Investment Models

  • Application of dynamic models to firm investment decisions.

  • Example: Estimating dynamic investment models with lagged investment as an endogenous variable.

• Dynamic Models in Economic Growth

  • Use of dynamic panel models to estimate long-run and short-run growth relationships.

  • Example: Modeling economic convergence using GMM in cross-country growth data.

• Labor Market Dynamics

  • Application of dynamic panel models to labor demand and wage determination.

  • Example: Estimating wage dynamics using a system GMM approach.

Section 10: Software Implementation of Dynamic Panel Data Models

• Implementing Dynamic Panel Data Models in Stata

  • Step-by-step guide to using Stata commands (xtabond, xtabond2) for dynamic panel data models.

  • Example: Estimating a dynamic model of investment with Arellano-Bond and Blundell-Bond estimators in Stata.

• Implementing Dynamic Panel Data Models in R

  • Using the plm and pgmm packages in R for dynamic panel data analysis.

  • Example: Estimating a dynamic wage model using GMM in R.

• Implementing Dynamic Panel Data Models in EViews and MATLAB

  • Practical guide to running dynamic panel models in EViews and MATLAB.

  • Example: Using system GMM to estimate a dynamic model of firm productivity in MATLAB.

Section 11: Best Practices and Common Challenges in Dynamic Panel Data Models

• Addressing Instrument Proliferation in GMM

  • Guidelines for choosing the right number of instruments: Collapsing instruments, using deeper lags.

  • Best practices for reporting GMM results: AR(1), AR(2), and Hansen tests.

• Interpretation of Coefficients in Dynamic Panel Models

  • Interpreting short-run and long-run coefficients in dynamic models.

  • Practical tips for presenting results in research papers.

Section 12: Conclusion and Further Reading

• Summary of Key Concepts in Dynamic Panel Data Models

  • Recap of the Arellano-Bond and Blundell-Bond estimators.

  • Importance of addressing endogeneity and Nickell bias in small T, large N panels.

• Suggested Further Reading

  • Foundational papers and textbooks for deeper exploration of dynamic panel data econometrics.

Sách

• Hsiao, C. (2022). Analysis of panel data (No. 64). Cambridge university press. (Chapter 4 )

  • Baltagi, B. H., & Baltagi, B. H. (2008). Econometric analysis of panel data (Vol. 4, pp. 135-145). Chichester: Wiley. (Chapter 8)

  • Pesaran, M. H. (2015). Time series and panel data econometrics. Oxford University Press. (Chapter 27 + 28 + 29)

  • Arellano, M. (2003). Panel data econometrics. OUP Oxford. (Chapter 7 + 8) 

  • Bun, M. J., Sarafidis, V., & Baltagi, B. H. (2013). Oxford Handbook of Panel Data. (Chapter 3)

Bài báo

• Ahn, S. C., & Schmidt, P. (1999). Estimation of linear panel data models using GMM. Generalized methods of moments estimation”, Cambridge.

• Ahn, S. C., & Schmidt, P. (1995). Efficient estimation of models for dynamic panel data. Journal of econometrics, 68(1), 5-27.

• Ahn, S. C., & Schmidt, P. (1999). Modified generalized instrumental variables estimation of panel data models with strictly exogenous instrumental variables. Analysis of Panels and Limited Dependent Variable Models, 171-198.

• Arellano, M., & Bond, S. (1991). Some tests of specification for panel data: Monte Carlo evidence and an application to employment equations. The review of economic studies, 58(2), 277-297.

• Arellano, M., & Honoré, B. (2001). Panel data models: some recent developments. In Handbook of econometrics (Vol. 5, pp. 3229-3296). Elsevier.

• Arellano, M., & Bover, O. (1995). Another look at the instrumental variable estimation of error-components models. Journal of econometrics, 68(1), 29-51.

• Beck, N., & Katz, J. N. (2011). Modeling dynamics in time-series–cross-section political economy data. Annual review of political science, 14, 331-352.

• Balestra, P., & Nerlove, M. (1966). Pooling cross section and time series data in the estimation of a dynamic model: The demand for natural gas. Econometrica: Journal of the econometric society, 585-612.

• Blundell, R., & Bond, S. (1998). Initial conditions and moment restrictions in dynamic panel data models. Journal of econometrics, 87(1), 115-143.

• Bond, S. R. (2002). Dynamic panel data models: a guide to micro data methods and practice. Portuguese economic journal, 1, 141-162.

• Hahn, J. (1999). How informative is the initial condition in the dynamic panel model with fixed effects?. Journal of econometrics, 93(2), 309-326.

• Hauk, W. R., & Wacziarg, R. (2009). A Monte Carlo study of growth regressions. Journal of economic growth, 14, 103-147.

• Kiviet, J. F., Pleus, M., & Poldermans, R. W. (2017). Accuracy and efficiency of various GMM inference techniques in dynamic micro panel data models. Econometrics, 5(1), 14. 

• Williams, L. K., & Whitten, G. D. (2011). Dynamic simulations of autoregressive relationships. The Stata Journal, 11(4), 577-588.

Nickell bias; inconsistency; instrumental variable approaches

• Anderson, T. W., & Hsiao, C. (1981). Estimation of dynamic models with error components. Journal of the American statistical Association, 76(375), 598-606.

• Nickell, S. (1981). Biases in dynamic models with fixed effects. Econometrica: Journal of the econometric society, 1417-1426.

• Roodman, D. (2009). A note on the theme of too many instruments. Oxford Bulletin of Economics and statistics, 71(1), 135-158.

• Wawro, G. (2002). Estimating dynamic panel data models in political science. Political Analysis, 10(1), 25-48.

• 
  

Transformed-likelihood, quasi- and full-maximum likelihood estimators

• Pickup, M., & Hopkins, V. (2022). Transformed-likelihood estimators for dynamic panel models with a very small T. Political Science Research and Methods, 10(2), 333-352.

 Some guidance on Model/Estimator Choices for small T, large N

• Campos, J., Ericsson, N. R., & Hendry, D. F. (2005). General-to-specific modeling: an overview and selected bibliography. FRB International Finance Discussion Paper, (838).

Nội dung

Section 1: Introduction to Panel Unit Root Tests

• • Definition of unit roots and their importance in econometric analysis.

  • Advantages of panel unit root tests over individual time series tests.

  • Handling non-stationarity in macroeconomic and financial panel data.

  • Applications in cross-country, regional, and firm-level panel data.

• Overview of Panel Unit Root Test Methods

  • First-generation vs. second-generation tests.

  • Accounting for cross-sectional independence vs. dependence.

Section 2: Concepts and Assumptions of Panel Unit Root Tests

• Individual Time Series Unit Root Tests

  • Review of key time series unit root tests: Augmented Dickey-Fuller (ADF), Phillips-Perron (PP), and KPSS.

  • Limitations of individual unit root tests: Low power in small samples.

• Assumptions Underlying Panel Unit Root Tests

  • Cross-sectional independence vs. dependence.

  • Homogeneity vs. heterogeneity across cross-sectional units.

  • Balanced vs. unbalanced panels: Implications for unit root tests.

Section 3: First-Generation Panel Unit Root Tests (Cross-Sectional Independence)

• Levin, Lin, and Chu (LLC) Test

  • Overview of the LLC test: Testing for common unit roots across all panel units.

  • Assumptions: Homogeneity in the autoregressive parameters.

  • Null and alternative hypotheses.

  • Testing procedure and limitations.

  • Example: Applying the LLC test to GDP growth data across countries.

• Im, Pesaran, and Shin (IPS) Test

  • Overview of the IPS test: Allowing for heterogeneous autoregressive parameters.

  • Null and alternative hypotheses: Individual unit roots vs. some stationary processes.

  • Testing procedure: Average of individual ADF tests.

  • Example: Using the IPS test for unemployment rate data across states.

• Breitung Test

  • Introduction to the Breitung test: Transformation-based unit root test for panel data.

  • Assumptions and testing procedure.

  • Example: Applying the Breitung test to inflation rate data across countries.

• Fisher-Type Tests: Maddala and Wu, Choi

  • Overview of Fisher-type tests: Combining p-values from individual unit root tests.

  • Maddala and Wu, and Choi Z-statistic methods.

  • Testing procedure and interpretation.

  • Example: Using Fisher-type tests to assess stock market index data across countries.

• Hadri Lagrange Multiplier (LM) Test

  • Overview of the Hadri test: Testing for stationarity (null hypothesis of stationarity).

  • Assumptions and testing procedure.

  • Example: Using the Hadri LM test to analyze firm-level productivity data.

Section 4: Second-Generation Panel Unit Root Tests (Cross-Sectional Dependence)

• Cross-Sectional Dependence in Panel Data

  • The problem of cross-sectional dependence in panel unit root tests.

  • Common shocks, global trends, and spillovers leading to cross-sectional dependence.

• Pesaran’s CIPS (Cross-sectional Im, Pesaran, and Shin) Test

  • Overview of the CIPS test: Accounting for cross-sectional dependence.

  • Testing for unit roots in the presence of cross-sectional dependence.

  • Null and alternative hypotheses.

  • Example: Applying the CIPS test to global stock market data.

• Bai and Ng Unit Root Test

  • Introduction to the Bai and Ng test for panel unit roots with unobserved common factors.

  • Estimating common factors and adjusting for cross-sectional dependence.

  • Testing procedure and interpretation.

  • Example: Applying the Bai and Ng test to panel data on international trade flows.

• Cross-sectional Augmented Dickey-Fuller (CADF) Test

  • Overview of the CADF test: Augmented Dickey-Fuller test with cross-sectional augmentation.

  • Adjusting for cross-sectional dependence using cross-sectional averages.

  • Example: Using the CADF test to examine industrial output data across regions.

Section 5: Unit Root Tests with Structural Breaks in Panel Data

• The Need for Structural Breaks in Panel Unit Root Tests

  • Why structural breaks matter: Economic crises, policy changes, and technological innovations.

  • Consequences of ignoring structural breaks in unit root testing.

• Carrion-i-Silvestre Unit Root Test with Multiple Structural Breaks

  • Overview of unit root tests with structural breaks in panel data.

  • Procedure for identifying structural breaks in the series.

  • Example: Testing for unit roots in exchange rates with structural breaks.

• Other Tests with Structural Breaks

  • Perron test with structural breaks.

  • Testing for structural breaks in macroeconomic and financial panel data.

  • Example: Using structural break tests in financial markets during crisis periods.

Section 6: Interpretation and Implications of Panel Unit Root Test Results

• Interpreting Panel Unit Root Test Results

  • Understanding null and alternative hypotheses in various tests.

  • Practical implications of finding unit roots or rejecting unit roots.

  • Implications for further econometric modeling: Cointegration and error correction.

• Choosing the Appropriate Unit Root Test

  • Guidelines for selecting the right test: Cross-sectional independence, sample size, heterogeneity.

  • Strengths and limitations of different panel unit root tests.

  • Example: Selection of unit root tests for cross-country income convergence analysis.

Section 7: Practical Applications of Panel Unit Root Tests

• Macro Panel Data Applications

  • Example: Testing for unit roots in panel data on GDP, inflation, and trade across countries.

  • Example: Analyzing financial time series (e.g., stock prices, interest rates) using panel unit root tests.

• Micro Panel Data Applications

  • Example: Using panel unit root tests to assess firm-level productivity and profitability.

  • Application to individual panel data: Household income, consumption, and employment.

Section 8: Software Implementation of Panel Unit Root Tests

• Implementing Panel Unit Root Tests in Stata

  • Commands for Levin-Lin-Chu, IPS, and Fisher-type tests in Stata.

  • Example: Step-by-step guide to running panel unit root tests in Stata using macroeconomic data.

• Implementing Panel Unit Root Tests in R

  • Using R packages (plm, punitroot) for panel unit root testing.

  • Example: Applying panel unit root tests in R for international trade data.

• Implementing Panel Unit Root Tests in EViews and MATLAB

  • Step-by-step implementation of unit root tests in EViews and MATLAB.

  • Example: Running the Pesaran CIPS test in MATLAB for financial data.

Section 9: Best Practices and Common Pitfalls in Panel Unit Root Testing

• Dealing with Cross-Sectional Dependence

  • When to account for cross-sectional dependence in panel unit root tests.

  • Best practices for addressing common shocks or spillovers in panel data.

• Reporting Unit Root Test Results

  • Guidelines for reporting test results: First and second-generation tests, structural breaks.

  • Example of reporting panel unit root tests in academic research.

• Common Pitfalls in Panel Unit Root Testing

  • Misinterpretation of results.

  • Ignoring structural breaks or cross-sectional dependence.

  • Selecting inappropriate tests for panel data characteristics.

Section 10: Conclusion and Further Reading

• Summary of Key Concepts in Panel Unit Root Tests

  • Recap of the importance of panel unit root tests for handling non-stationarity in panel data.

  • Overview of first- and second-generation tests and their applications.

• Suggested Further Reading

  • Foundational papers and textbooks for further exploration of panel unit root tests.

Sách

• Hsiao, C. (2022). Analysis of panel data (No. 64). Cambridge university press. (Chapter 10)

• Baltagi, B. H., & Baltagi, B. H. (2008). Econometric analysis of panel data. (Chapter 12)

• Pesaran, M. H. (2015). Time series and panel data econometrics. Oxford University Press. (Chapter 31)

Bài báo

Spurious regression with panel data

• Baltagi, B. H., Kao, C., & Liu, L. (2008). Asymptotic properties of estimators for the linear panel regression model with random individual effects and serially correlated errors: the case of stationary and non‐stationary regressors and residuals. The Econometrics Journal, 11(3), 554-572.

• Baltagi, B. H., Kao, C., & Na, S. (2011). Test of hypotheses in panel data models when the regressor and disturbances are possibly non-stationary. AStA Advances in Statistical Analysis, 95, 329-350.

• Hsiao, C. (2014). Panel Macroeconometric Modeling☆ This paper is dedicated to PCB Phillips for his creative and lasting contributions to econometrics. In Essays in Honor of Peter CB Phillips (pp. 205-239). Emerald Group Publishing Limited.

• Kao, C. (1999). Spurious regression and residual-based tests for cointegration in panel data. Journal of econometrics, 90(1), 1-44.

• Phillips, P. C., & Moon, H. R. (1999). Linear regression limit theory for nonstationary panel data. Econometrica, 67(5), 1057-1111. 

Testing for stationarity in panel data

• Bai, J., & Ng, S. (2004). A PANIC attack on unit roots and cointegration. Econometrica, 72(4), 1127-1177.

• Bai, J., & Ng, S. (2010). Panel unit root tests with cross-section dependence: a further investigation. Econometric Theory, 26(4), 1088-1114.

• Peter, P. (2004). Asymptotic and Finite Sample Properties of Pooled Time Series Tests with an Application to the PPP Hypothesis.

• Hlouskova, J., & Wagner, M. (2006). The performance of panel unit root and stationarity tests: results from a large scale simulation study. Econometric Reviews, 25(1), 85-116.

• Maddala, G. S., & Wu, S. (1999). A comparative study of unit root tests with panel data and a new simple test. Oxford Bulletin of Economics and statistics, 61(S1), 631-652.

Nội dung

Section 1: Introduction to Panel Vector Autoregression (PVAR)

• What is Panel Vector Autoregression (PVAR)?

  • Definition of Vector Autoregression (VAR) models and their extension to panel data.

  • Key features of PVAR: Modeling dynamic interactions between multiple variables across units 

  • Why use PVAR in panel data: Joint modeling of cross-sectional and time-series dependencies.

• Differences Between Standard VAR and PVAR

  • Comparison of standard time series VAR and PVAR models.

  • Advantages of using PVAR in economic and financial panel data analysis.

• Applications of PVAR in Empirical Research

  • Overview of key applications: Macroeconomic policy analysis, financial shocks, international trade, firm-level dynamics.

  • Examples: Cross-country analysis of economic growth and inflation dynamics, firm-level investment and productivity interactions.

Section 2: Theoretical Foundations of PVAR

• Dynamic Interactions in Panel Data

  • The concept of dynamic interactions between multiple variables in a system.

  • How PVAR models capture feedback loops and endogenous relationships in panel data.

• Assumptions in PVAR Models

  • Stationarity and stability conditions for PVAR models.

  • Time series and cross-sectional dimensions in PVAR: Balanced and unbalanced panels.

  • Assumptions about error terms: Cross-sectional dependence, heteroskedasticity, and autocorrelation.

• Identifying the Order of the PVAR Model

  • Selection of lag length in PVAR models: Information criteria (AIC, BIC, HQIC).

  • Testing for the appropriate lag order in PVAR: The impact of lag structure on dynamic interactions.

Section 3: Estimation Methods for PVAR

• GMM Estimation for PVAR

  • Generalized Method of Moments (GMM) for estimating PVAR models.

  • Why GMM is widely used in PVAR: Handling endogeneity in dynamic systems.

  • Steps to implement GMM in PVAR: Moment conditions, instrumental variables, and assumptions.

• Fixed Effects and Random Effects in PVAR

  • Fixed effects in PVAR: Controlling for unobserved individual heterogeneity.

  • Random effects and their applicability in PVAR models.

  • Comparison of fixed vs. random effects in dynamic systems.

• System GMM Estimation

  • Using system GMM to estimate PVAR models with lagged dependent variables.

  • Advantages of system GMM over difference GMM in PVAR settings.

  • Practical considerations for applying system GMM: Instrument proliferation and weak instruments.

• Bayesian Estimation for PVAR

  • Introduction to Bayesian methods for PVAR estimation.

  • Bayesian priors and posteriors in the context of PVAR.

  • Example: Using Bayesian methods to estimate a PVAR model for macroeconomic variables.

Section 4: Impulse Response Functions (IRFs) and Forecast Error Variance Decomposition (FEVD)

• Impulse Response Functions (IRFs) in PVAR

  • Definition and interpretation of impulse response functions.

  • Estimating IRFs in PVAR models: Tracking the dynamic effects of shocks on all variables in the system.

  • Examples.

• Forecast Error Variance Decomposition (FEVD) in PVAR

  • Decomposing forecast error variance to assess the contribution of each variable in explaining future variation.

  • Examples

• Confidence Intervals for IRFs and FEVD

  • Bootstrapping and Monte Carlo methods to obtain confidence intervals for IRFs and FEVD in PVAR models.

  • Practical issues in interpreting IRF and FEVD results.

Section 5: Diagnostic Testing in PVAR Models

• Testing for Stationarity in PVAR

  • Unit root tests in the context of panel data.

  • Stationarity testing across cross-sectional units.

• Cross-Sectional Dependence and Serial Correlation

  • Testing for cross-sectional dependence in PVAR models.

  • Wooldridge test for serial correlation in panel data.

• Model Stability Testing

  • Checking for stability conditions in PVAR models.

  • Stability diagnostics: Eigenvalues and their role in ensuring valid dynamic systems.

• Testing for Granger Causality in PVAR

  • Granger causality tests in the PVAR framework.

  • Practical examples of using Granger causality in economic and financial applications.

Section 6: Handling Endogeneity and Heterogeneity in PVAR

• Endogeneity in PVAR

  • Addressing endogenous relationships between variables in a dynamic system.

  • Instrumental variable techniques for addressing endogeneity in PVAR models.

• Heterogeneity Across Cross-Sectional Units

  • Allowing for heterogeneity in short-run dynamics and long-run relationships in PVAR models.

  • Dealing with heterogeneous slope coefficients across units: Mean group and pooled mean group approaches.

Section 7: Structural PVAR Models

• Structural Identification in PVAR

  • Structural PVAR (SVAR) models: Imposing economic theory-based restrictions to identify shocks.

  • Short-run and long-run identification schemes in structural PVAR models.

• Applications of Structural PVAR

  • Example: Identifying demand and supply shocks in a structural PVAR model of international trade.

  • Example: Using SVAR to identify monetary policy shocks in a panel of countries.

Section 8: Practical Applications of PVAR

• Macroeconomic Applications

  • Example: Estimating the dynamic relationship between inflation, GDP growth, and interest rates using PVAR.

• Financial Applications

  • Example: Modeling the interaction between stock prices, exchange rates, and interest rates using PVAR.

• Firm-Level Applications

  • Example: Analyzing the dynamic effects of R&D investment on firm productivity using PVAR.

• Policy Evaluation with PVAR

  • Example: Evaluating the effects of fiscal and monetary policy on economic growth and employment using PVAR.

Section 9: Software Implementation of PVAR

• PVAR Estimation in Stata

  • Step-by-step guide to using Stata's xtvar command for PVAR estimation.

  • Example: Estimating a PVAR model of firm profitability and investment with Stata.

• PVAR Estimation in R

  • Using the panelvar package in R for PVAR estimation.

  • Example: Implementing PVAR for cross-country macroeconomic data in R.

• PVAR Estimation in MATLAB and EViews

  • Practical guide to implementing PVAR in MATLAB and EViews.

  • Example: Estimating PVAR for financial data using MATLAB.

Section 10: Best Practices and Common Pitfalls in PVAR Analysis

• Best Practices in PVAR Modeling

  • Guidelines for selecting the appropriate lag structure and ensuring model stability.

  • Handling endogeneity and weak instruments in PVAR.

• Common Pitfalls in PVAR Analysis

  • Misinterpretation of impulse response functions and forecast error variance decompositions.

  • Avoiding overfitting and instrument proliferation in GMM estimation.

Section 11: Conclusion and Further Reading

• Summary of Key Concepts in PVAR

  • Recap of key elements in PVAR modeling, estimation, and interpretation.

• Suggested Further Reading

  • Foundational papers and textbooks for further study of PVAR models and applications.

Sách

• Hamilton, J. D. (2020). Time series analysis. Princeton university press.

• Lütkepohl, H. (2005). New introduction to multiple time series analysis. NY: Springer.

Bài báo

• Carpenter, S., & Demiralp, S. (2012). Money, reserves, and the transmission of monetary policy: Does the money multiplier exist?. Journal of macroeconomics, 34(1), 59-75.

• Head, A., Lloyd-Ellis, H., & Sun, H. (2014). Search, liquidity, and the dynamics of house prices and construction. American Economic Review, 104(4), 1172-1210.

• Love, I., & Zicchino, L. (2006). Financial development and dynamic investment behavior: Evidence from panel VAR. The Quarterly Review of Economics and Finance, 46(2), 190-210.

• Mora, N., & Logan, A. (2012). Shocks to bank capital: evidence from UK banks at home and away. Applied Economics, 44(9), 1103-1119.

• Neumann, T. C., Fishback, P. V., & Kantor, S. (2010). The dynamics of relief spending and the private urban labor market during the New Deal. The Journal of Economic History, 70(1), 195-220.

• Canova, F., & Ciccarelli, M. (2013). Panel Vector Autoregressive Models: A Survey. In VAR models in macroeconomics–new developments and applications: Essays in honor of Christopher A. Sims (pp. 205-246). Emerald Group Publishing Limited. (link)

Nội dung

Section 1: Introduction to Granger Causality in Panel Data

• What is Granger Causality?

  • Definition of Granger causality: Concept of predictive causality based on temporal precedence.

• Importance of Testing for Granger Causality in Panel Data

  • Advantages of panel data for Granger causality: Combining cross-sectional and time series dimensions.

  • Applications in economics, finance, and social sciences.

  • Examples: Testing the relationship between GDP and energy consumption, or monetary policy and inflation.

• Challenges in Granger Causality Testing for Panel Data

  • Heterogeneity across cross-sectional units.

  • Dynamic interactions and feedback effects.

  • Dealing with non-stationarity, endogeneity, and cross-sectional dependence.

Section 2: Theoretical Foundations of Granger Causality

• Granger Causality in Time Series Models

  • Definition of Granger causality in time series: Using lagged values of one variable to predict another.

  • Stationarity, lag selection, and model specification in time series Granger causality.

  • Example: Granger causality in a simple VAR model between GDP and investment.

• Extending Granger Causality to Panel Data

  • Definition of Granger causality in a panel context: Testing for causality in cross-sectional units over time.

  • Assumptions for panel Granger causality: Homogeneous vs. heterogeneous causal relationships across units.

  • Fixed effects vs. random effects in Granger causality testing.

Section 3: Panel Data Models for Granger Causality Testing

• VAR Models in Panel Data

  • Introduction to panel VAR (PVAR) models for Granger causality testing.

  • Dynamic interactions and the role of lags in panel data.

  • Example: Using panel VAR to test for Granger causality between inflation and interest rates across countries.

• Error Correction Models (ECM) in Panel Data

  • Cointegration and error correction models (ECM) for Granger causality testing in panel data.

  • Short-run and long-run causality in cointegrated panel data models.

  • Example: Using ECM to test for short-run and long-run causality between exports and GDP in a panel of countries.

• Dynamic Panel Models for Granger Causality

  • Dynamic panel data models: Estimating dynamic interactions with lagged dependent variables.

  • Using GMM (Arellano-Bond) to estimate dynamic panel models for Granger causality.

  • Example: Testing for Granger causality between investment and productivity in a dynamic panel model for firms.

Section 4: Granger Causality Tests in Panel Data

• Dumitrescu-Hurlin Test for Panel Granger Causality

  • Overview of the Dumitrescu-Hurlin (2012) test for panel Granger causality.

  • Assumptions: Heterogeneous causal relationships across cross-sectional units.

  • Null hypothesis and alternative hypothesis: No Granger causality vs. presence of Granger causality for at least one cross-sectional unit.

  • Testing procedure and interpretation of results.

  • Example: Applying the Dumitrescu-Hurlin test to test causality between renewable energy consumption and economic growth across regions.

• Holtz-Eakin, Newey, and Rosen (HENR) Test

  • Overview of the HENR (1988) approach to testing Granger causality in panel data.

  • Using GMM to estimate dynamic panel models for Granger causality testing.

  • Example: Applying the HENR test for Granger causality between firm R&D expenditure and profitability.

• Hurlin and Venet Test for Homogeneous Causality

  • Introduction to Hurlin and Venet’s (2001) test for homogeneous causality in panel data.

  • Null hypothesis: Homogeneous Granger causality across all cross-sectional units.

  • Testing procedure and interpretation.

  • Example: Using the Hurlin and Venet test to check for homogeneity in Granger causality between education expenditure and economic growth across countries.

Section 5: Steps for Implementing Granger Causality Tests in Panel Data

• Pre-Testing for Unit Roots and Cointegration

  • Importance of checking for unit roots in panel data before testing for Granger causality.

  • Unit root tests: Levin-Lin-Chu (LLC), Im-Pesaran-Shin (IPS), Fisher-type tests.

  • Testing for cointegration in panel data: Pedroni, Westerlund, Kao tests.

• Lag Selection in Panel Granger Causality Tests

  • Choosing the appropriate lag length for Granger causality testing.

  • Information criteria (AIC, BIC) for selecting optimal lag lengths.

  • Example: Selecting lag lengths for testing Granger causality between unemployment and inflation in a panel of countries.

• Testing for Granger Causality in Non-Stationary Data

  • Addressing non-stationarity and cointegration in Granger causality tests.

  • Error correction mechanisms (ECM) for non-stationary panel data.

  • Example: Testing for short-run and long-run causality between financial development and income inequality in a cointegrated panel.

• Interpreting Granger Causality Test Results

  • How to interpret the rejection or acceptance of the null hypothesis.

  • Practical interpretation of Granger causality in economic relationships.

  • Example: Interpreting the results of a Granger causality test between public spending and economic growth.

Section 6: Extensions of Granger Causality in Panel Data

• Causality in Heterogeneous Panels

  • Allowing for heterogeneity across units in panel Granger causality testing.

  • Mean group and pooled mean group estimators for heterogeneous causal relationships.

  • Example: Testing for heterogeneous causality between trade and GDP in developing and developed countries.

• Testing for Granger Causality in the Presence of Cross-Sectional Dependence

  • Cross-sectional dependence in panel data: Accounting for common shocks or unobserved factors.

  • Using the Pesaran CD test to check for cross-sectional dependence.

  • Handling cross-sectional dependence in panel Granger causality testing.

  • Example: Testing for Granger causality in financial contagion across stock markets with cross-sectional dependence.

• Panel Granger Causality in Nonlinear Models

  • Extending Granger causality to nonlinear panel data models.

  • Nonlinear Granger causality testing using threshold models.

  • Example: Testing for nonlinear Granger causality between energy consumption and economic growth in a panel of countries.

Section 7: Practical Applications of Granger Causality in Panel Data

• Macroeconomic Applications

• Financial Applications

• Firm-Level Applications

Section 8: Software Implementation of Granger Causality Tests in Panel Data

• Implementing Panel Granger Causality Tests in Stata

  • Step-by-step guide to using xtgcause for Dumitrescu-Hurlin Granger causality testing in Stata.

  • Example: Testing Granger causality between inflation and unemployment with Stata.

• Implementing Panel Granger Causality Tests in R

  • Using the plm and pgrangertest packages for panel Granger causality testing in R.

  • Example: Implementing the Dumitrescu-Hurlin test in R for a panel of firm-level data.

• Implementing Panel Granger Causality Tests in EViews

  • Practical guide to using EViews for panel Granger causality testing.

  • Example: Estimating Granger causality in EViews for macroeconomic panel data.

Section 9: Best Practices and Common Pitfalls in Granger Causality Testing

• Best Practices for Granger Causality Testing

  • Guidelines for ensuring robustness in Granger causality tests: Pre-testing, lag selection, and addressing non-stationarity.

  • Handling potential biases and limitations in Granger causality interpretation.

• Common Pitfalls in Panel Granger Causality Tests

  • Misinterpreting Granger causality: Temporal precedence does not imply causality.

  • Ignoring cross-sectional dependence, non-stationarity, or cointegration in Granger causality testing.

Section 10: Conclusion and Further Reading

• Summary of Key Concepts in Granger Causality Testing

  • Recap of Granger causality testing methods in panel data and their importance in empirical research.

• Suggested Further Reading

  • Foundational papers and textbooks for deeper exploration of Granger causality testing in panel data.

Sách

•  Arellano, M. (2003). Panel data econometrics (Vol. 231). Oxford University Press.

Bài báo

• Dumitrescu, E. I., & Hurlin, C. (2012). Testing for Granger non-causality in heterogeneous panels. Economic modelling, 29(4), 1450-1460.

• Paramati, S. R., Apergis, N., & Ummalla, M. (2017). Financing clean energy projects through domestic and foreign capital: The role of political cooperation among the EU, the G20 and OECD countries. Energy economics, 61, 62-71.

• Paramati, S. R., Ummalla, M., & Apergis, N. (2016). The effect of foreign direct investment and stock market growth on clean energy use across a panel of emerging market economies. Energy economics, 56, 29-41.

• Salahuddin, M., Alam, K., & Ozturk, I. (2016). The effects of Internet usage and economic growth on CO2 emissions in OECD countries: A panel investigation. Renewable and Sustainable Energy Reviews, 62, 1226-1235.

Nội dung

Section 1: Introduction to Cointegrating Panel Data Models

• What is Cointegration in Panel Data?

  • Definition of cointegration: Long-term relationships between non-stationary variables.

  • Differences between single-equation and panel cointegration.

  • Importance of modeling cointegrating relationships in panel data.

  • Examples 

• Applications of Cointegrating Panel Data Models

  • Real-world applications: Long-run relationships in economic growth, trade, energy consumption, and policy evaluation.

Section 2: Basic Concepts of Cointegrating Panel Data Models

• • Concepts of stationarity and unit roots.

  • The need for cointegration models when dealing with non-stationary data.

  • Stationarity and Unit Roots in Panel Data

  • Overview of testing for Cointegration in Panel Data

Section 3: Estimation Techniques for Cointegrating Panel Data Models

• Dynamic Ordinary Least Squares (DOLS)

  • Definition and role of DOLS in cointegrating panel data.

  • Dealing with endogeneity and serial correlation.

  • Testing procedure and interpretation.

  • Example

• Fully Modified OLS (FMOLS)

  • Overview of FMOLS: Addressing serial correlation and endogeneity.

  • Estimation procedure and interpretation of coefficients.

  • Example

• Pooled Mean Group (PMG) Estimator

  • Introduction to the PMG estimator: Allowing for heterogeneous short-run dynamics with homogeneous long-run relationships.

  • Application of PMG in macroeconomic models.

  • Example

Section 4: Error Correction Models (ECM) for Cointegrating Panel Data

• Error Correction Representation

  • Introduction to the Error Correction Model (ECM) in the context of cointegration.

  • Estimating short-run dynamics and long-run equilibrium adjustments.

• Panel ECM

  • Application of ECM to panel data: Capturing speed of adjustment across panel units.

  • Example: Estimating the adjustment speed in the financial development and economic growth relationship across Southeast Asia.

Section 5: Models with Cross-Sectional Dependence in Cointegrating Panels

• Cross-Sectional Dependence in Cointegrating Relationships

  • Importance of accounting for cross-sectional dependence in cointegrating models.

  • Examples

• Common Correlated Effects (CCE) Estimator

  • Overview of the CCE estimator: Accounting for unobserved common factors.

  • Estimation procedure and application to cointegrating panel data.

  • Examples

Section 6: Heterogeneity in Cointegrating Panel Data Models

• Heterogeneous Panel Data Models

  • The importance of allowing for heterogeneity in short-run dynamics across cross-sectional units.

  • Heterogeneous dynamics in regional growth models and climate change impacts.

• Mean Group (MG) and Pooled Mean Group (PMG) Estimators

  • Differences between MG and PMG in terms of handling heterogeneity.

  • When to use MG and PMG estimators in panel data.

  • Example: Analyzing heterogeneous growth patterns across Vietnam’s provinces using MG and PMG.

Section 7: Structural Breaks in Cointegrating Panel Data Models

• Structural Breaks in Long-Run Relationships

  • The role of structural breaks in modeling cointegrating relationships.

  • Structural breaks due to economic policy changes, crises, or technological shifts.

• Modeling Structural Breaks in Cointegrating Panels

  • Tests for detecting structural breaks in panel data.

  • Estimating models with structural breaks in long-run equilibrium.

  • Example: Modeling the effect of the 2008 global financial crisis on long-term trade relationships between Vietnam and its trading partners.

Section 8: Nonlinear Cointegrating Panel Data Models

• Introduction to Nonlinear Cointegration

  • Why nonlinearities may arise in long-run relationships.

  • Nonlinear dynamics in macroeconomic and financial relationships.

• Threshold Cointegration Models

  • Overview of threshold cointegration models: Allowing for regime shifts.

  • Applications of threshold models in policy analysis.

  • Example: Threshold cointegration between interest rates and inflation across ASEAN countries.

Section 9: Testing and Diagnostics for Cointegrating Panel Data Models

• Diagnostic Tests for Cointegrating Panel Models

  • Residual diagnostics: Serial correlation, heteroskedasticity, and cross-sectional dependence.

  • Testing for stability of long-run relationships.

• Overidentification and Instrument Validity

  • Checking the validity of instruments in dynamic cointegrating models.

  • Example: Assessing instrument validity in fiscal policy models across Southeast Asia.

Section 10: Practical Applications of Cointegrating Panel Data Models

• Long-Term Relationships in Macroeconomic Data

  • Examples

• Financial Applications

  • Examples

• Environmental and Energy Applications

  • Examples

Section 11: Software Implementation of Cointegrating Panel Data Models

• Implementing Cointegrating Panel Models in Stata

  • Step-by-step guide to FMOLS, DOLS, and ECM estimation in Stata.

• Implementing Cointegrating Panel Models in R

  • Using R packages (plm, urca) for cointegration analysis in panel data.

• Implementing Cointegrating Panel Models in EViews and MATLAB

  • Practical guide to running dynamic models for cointegrating panels in EViews and MATLAB.

Section 12: Best Practices and Common Pitfalls in Cointegrating Panel Data Models

• Dealing with Cross-Sectional Dependence and Structural Breaks

  • How to handle common shocks and changes in long-term relationships.

• Reporting Results and Interpretation

  • Guidelines for reporting cointegrating panel data models in research papers.

  • Common misinterpretations and how to avoid them.

Section 13: Conclusion and Further Reading

• Summary of Key Concepts

  • Recap of models for cointegrating panel data and their importance in empirical research.

• Suggested Further Reading

  • Foundational papers and textbooks for further study of cointegration in panel data models.

Sách

•  Arellano, M. (2003). Panel data econometrics (Vol. 231). Oxford University Press.

• 
  

Bài báo

• Blackburne III, E. F., & Frank, M. W. (2007). Estimation of nonstationary heterogeneous panels. The Stata Journal, 7(2), 197-208.

• Pesaran, M. H., Shin, Y., & Smith, R. P. (1999). Pooled mean group estimation of dynamic heterogeneous panels. Journal of the American statistical Association, 94(446), 621-634.

• Pesaran, M. H., & Smith, R. (1995). Estimating long-run relationships from dynamic heterogeneous panels. Journal of econometrics, 68(1), 79-113.

• Balke, N. S., & Wohar, M. E. (2006). What drives stock prices? Identifying the determinants of stock price movements. Southern Economic Journal, 73(1), 55-78.

• Coers, R., & Sanders, M. (2013). The energy–GDP nexus; addressing an old question with new methods. Energy Economics, 36, 708-715.

• Costantini, V., & Martini, C. (2010). The causality between energy consumption and economic growth: A multi-sectoral analysis using non-stationary cointegrated panel data. Energy Economics, 32(3), 591-603.

• Mark, N. C., & Sul, D. (2003). Cointegration vector estimation by panel DOLS and long‐run money demand. Oxford Bulletin of Economics and statistics, 65(5), 655-680.

Nội dung

Section 1: Introduction to Unbalanced Panel Data

• • Difference between balanced and unbalanced panel data.

  • Causes of unbalanced panel data: Missing observations, entry and exit of units, gaps in data collection.

  • Examples of unbalanced panels in economics, finance, and social science research.

• Why Study Unbalanced Panel Data?

  • Advantages of working with unbalanced panels: Leveraging incomplete data.

  • Challenges in estimating models with unbalanced panels: Biases, inefficiencies, and potential data loss.

  • Practical examples: Firm-level data with missing periods, country-level panel data with entry and exit of countries.

Section 2: Types of Unbalanced Panels

• Randomly Unbalanced Panels

  • Definition: Missing observations due to random factors.

  • Implications for estimation and interpretation.

• Systematically Unbalanced Panels

  • Definition: Systematic patterns in missing data (e.g., attrition, firm exit, policy changes).

  • Implications for model estimation and possible biases.

• Panel Attrition

  • Causes of attrition in panel data: Firm bankruptcy, country withdrawal from a study.

  • Handling attrition through estimation techniques and robustness checks.

Section 3:Models for Unbalanced Panel Data

• Key Assumptions in Panel Data Models

  • Assumptions for balanced vs. unbalanced panels.

  • Consequences of unbalancedness for model consistency and efficiency.

• Unobserved Heterogeneity in Unbalanced Panels

  • Importance of accounting for unobserved factors in unbalanced data.

  • Fixed effects vs. random effects in the context of unbalanced data.

Section 4: Estimation Techniques for Unbalanced Panel Data Models

• Fixed Effects (FE) Model for Unbalanced Panels

  • Fixed effects estimation with unbalanced data.

  • Dealing with missing data and incomplete panels.

  • Example: Estimating firm profitability using a fixed effects model with unbalanced panel data.

• Random Effects (RE) Model for Unbalanced Panels

  • Random effects estimation when panel data is unbalanced.

  • Conditions under which random effects remain valid in unbalanced settings.

  • Example: Random effects model for cross-country panel data with missing observations on GDP.

• Mixed Effects Models

  • Introduction to mixed effects models combining fixed and random effects.

  • Advantages for unbalanced panels: Flexibility in handling random missing data.

  • Example: Using mixed effects models to estimate student test scores in education panel data with gaps.

• Maximum Likelihood Estimation (MLE) for Unbalanced Panels

  • Application of MLE in unbalanced panels: Handling missing data efficiently.

  • Example: Estimating healthcare expenditure models using MLE on an unbalanced panel of OECD countries.

• Generalized Method of Moments (GMM) for Unbalanced Panels

  • Dynamic panel data models with GMM: Arellano-Bond estimator for unbalanced panels.

  • Addressing endogeneity and missing data through GMM.

  • Example: Estimating investment dynamics using GMM in firm-level unbalanced panel data.

Section 5: Dealing with Missing Data in Unbalanced Panels

• Missing Data Mechanisms

  • Missing Completely at Random (MCAR), Missing at Random (MAR), and Missing Not at Random (MNAR).

  • How different missing data mechanisms affect model estimation.

• Methods to Handle Missing Data

  • Listwise Deletion: When to use listwise deletion in unbalanced panels.

  • Imputation Methods: Simple imputation, multiple imputation, and advanced techniques.

  • Expectation-Maximization (EM) Algorithm: Handling missing values with EM in unbalanced panels.

  • Examples

• Selection Models for Unbalanced Panel Data

  • Selection models for handling non-random missing data (e.g., attrition).

  • Heckman selection models for unbalanced panel data.

  • Examples.

Section 6: Heterogeneity and Dynamic Models for Unbalanced Panels

• Heterogeneous Panels with Missing Data

  • Modeling individual heterogeneity in the presence of unbalancedness.

  • Allowing for heterogeneity in short-term and long-term dynamics in unbalanced panels.

  • Example: Estimating heterogeneous price elasticity across industries with unbalanced firm-level data.

• Dynamic Models for Unbalanced Panels

  • Dynamic models with lagged dependent variables in unbalanced panels.

  • GMM methods for dynamic unbalanced panel data: Handling missing lags and leads.

  • Example: Modeling wage dynamics with dynamic GMM for unbalanced panel data.

Section 7: Structural Breaks and Non-Stationarity in Unbalanced Panel Data

• Structural Breaks in Unbalanced Panels

  • Detecting and accounting for structural breaks in unbalanced panel data.

  • Example: Structural breaks in energy consumption models across countries with missing data.

• Unit Root Tests and Cointegration in Unbalanced Panels

  • Panel unit root tests for unbalanced data: Adjusting for missing observations.

  • Cointegration models in unbalanced panels: Estimating long-run relationships with missing data.

  • Example: Cointegration analysis in trade and GDP for ASEAN countries with incomplete time series.

Section 8: Cross-Sectional Dependence in Unbalanced Panel Data Models

• Addressing Cross-Sectional Dependence in Unbalanced Panels

  • Common correlated effects (CCE) estimator for unbalanced panel data.

  • Handling unobserved common factors in the presence of missing data.

  • Example: Estimating cross-sectional dependence in financial markets with unbalanced firm data.

Section 9: Practical Applications of Unbalanced Panel Data Models

• Macroeconomic Applications

  • Examples

• Firm-Level Applications

  • Examples

• 9.3 Health and Education Applications

  • Examples

Section 10: Software Implementation of Unbalanced Panel Data Models

• 10.1 Stata Commands for Unbalanced Panel Data

  • Overview of Stata commands for handling unbalanced panels

  • Example: Step-by-step guide to implementing fixed effects and random effects models in Stata for unbalanced panels.

• Packages for Unbalanced Panels

  • Using plm, lme4, and nlme packages in R for unbalanced panel data models.

  • Example: Implementing random effects and dynamic models in R for incomplete firm-level data.

• EViews and MATLAB for Unbalanced Panel Data

  • Guide to modeling unbalanced panels in EViews and MATLAB.

  • Example: Using EViews for cointegration models with unbalanced panel data.

Section 11: Best Practices and Common Pitfalls in Unbalanced Panel Data Models

• Best Practices for Dealing with Missing Data

  • Guidelines for deciding when to drop or impute missing observations.

  • Choosing the right model for unbalanced panels based on data structure.

• Reporting Results and Addressing Limitations

  • Best practices for reporting results in research papers.

  • Handling potential biases and limitations when working with unbalanced panels.

Section 12: Conclusion and Further Reading

• Summary of Key Concepts in Unbalanced Panel Data

  • Recap of estimation techniques and models for unbalanced panels.

• Suggested Further Reading

  • Foundational papers and textbooks for further study on unbalanced panel data models.

Sách

• Baltagi, B. H., & Baltagi, B. H. (2008). Econometric analysis of panel data (Vol. 4, pp. 135-145). Chichester: Wiley. (Chapter 9)

• Pesaran, M. H. (2015). Time series and panel data econometrics. Oxford University Press. (Chapter 26)

• Jushan Bai, Yuan Liao, and Jisheng Yang (2015). The Oxford handbook of panel data. Oxford University Press. (Chapter 5)