Open research questions in Spatial and Panel Data Analysis

What the literature leaves open

• The paper identifies the challenge of distinguishing between the analysis model and the data generation types of spatial confounding. The authors note the challenge of determining when spatial confounding needs to be adjusted for in the analysis model. The paper highlights the challenge of addressing the flawed methodologies for alleviating spatial confounding.

  Re-Thinking Spatial Confounding in Spatial Linear Mixed Models · 2026 · DOI

• The current methodologies for alleviating spatial confounding are flawed. There is a need for a novel perspective on spatial confounding. The distinction between the analysis model and the data generation types of spatial confounding has not been clearly made.

  Re-Thinking Spatial Confounding in Spatial Linear Mixed Models · 2026 · DOI

• Further research is needed to validate the assumption of local unconfoundedness. The proposed method can be extended to other applications with spatially-dependent data. Sensitivity analysis can be conducted to gauge the strength of a missing local confounder that would alter the conclusions.

  A Spectral Confounder Adjustment for Spatial Regression with Multiple Exposures and Outcomes · 2026 · DOI

• The assumption of no unmeasured spatial confounding is often violated in observational studies. Existing methods do not account for multiscale confounding in multivariate spatial studies. There is a need for a statistical method to study the many-to-many relationship between social vulnerability and health outcomes.

  A Spectral Confounder Adjustment for Spatial Regression with Multiple Exposures and Outcomes · 2026 · DOI

• Identifying moments of all orders by induction and recovering the distribution via moment determinacy - Studying the full distribution of random coefficients -

  Identification and estimation in a time-varying endogenous random coefficient panel data model · 2026 · DOI

• The existing methods may not fully capture agents' optimization behavior. The correlation between random coefficients and regressors is not fully addressed by classical methods. There is a need for a new approach that allows for both a fixed effect and a time-varying random shock.

  Identification and estimation in a time-varying endogenous random coefficient panel data model · 2026 · DOI

• The method is limited to detecting at most one change point, - The selection of γ is data-driven and requires simulation results, - The analysis is based on a small sample of 5 countries

  Change Point Detection in Panel Linear Regression Models Based on Jump Information Criterion · 2026 · DOI

• The lack of a reliable method for change point detection in panel linear regression models. The need for a novel approach to reconstruct the traditional change point hypothesis testing problem into a parameter estimation problem. The requirement for a method that can efficiently solve the change point detection problem in panel data models.

  Change Point Detection in Panel Linear Regression Models Based on Jump Information Criterion · 2026 · DOI

• Spatial heterogeneity in human development. Limited understanding of regional disparities in education, health, and economic conditions. Need for spatially sensitive development policies.

  Haversine-Based Geographically Weighted Panel Regression of Human Development in Gorontalo (2016–2025) · 2026 · DOI

• The study suggests directions for future research, including the application of the integrated SAR-PCA approach to other spatial regression models.

  Spatial Regression Analysis using Queen Contiguity Weight Matrix and PCA Dimensionality Reduction · 2026 · DOI

• Conventional linear regression falls short in poverty analysis due to spatial interdependence and multicollinearity. Prior work overlooks spatial interdependence between neighboring regions.

  Spatial Regression Analysis using Queen Contiguity Weight Matrix and PCA Dimensionality Reduction · 2026 · DOI

• The difficulty in theoretically comparing the performance of the indicators with their associated tests. The mixed application orientations of the indicators with their associated tests. The need to design proper simulation experiments to assess the performance of the indicators.

  Comparison of Local Spatial Deviation Indicators with Their Associated Tests: Evidence from Simulations and Applied Cases · 2026 · DOI

• Existing techniques for trend modeling may not accurately capture spatial trends. There is a need for a methodology that can accurately model spatial distributions. The proposed methodology addresses this gap.

  Generalized ridge penalization for trend modeling and spatial prediction with generalized additive models · 2026 · DOI

• The study does not provide real-world data analysis, - The analysis is based on simulation evidence, - The results are limited to the proposed framework

  Estimation of Heterogeneous Panel Data Models With Mixed Sampling Frequencies · 2026 · DOI

• Examining whether the covariates are systematically correlated with the heterogeneities, - Investigating the properties of the CRE estimator

  Estimation of Heterogeneous Panel Data Models With Mixed Sampling Frequencies · 2026 · DOI

• The GCCM method is easily disturbed by false neighbors caused by spatial autocorrelation. Traditional causal inference methods have limitations in handling geographical data with spatial autocorrelation.

  Causal Inference for Spatial Cross-Sectional Data: A Geographical Convergent Cross Mapping Method Accounting for Spatial Autocorrelation · 2026 · DOI

• Although SA-GCCM performs well in reducing spatial auto- correlation interference, it has two main limitations. First, existing studies only consider bivariate causal inference. In complex Earth systems, multiple factors jointly influence a dependent variable. Future research could attempt to extend SA-GCCM to a multivariate causal inference framework using multivariate embedding. Furthermore, the core of SA-GCCM lies in calculating the spatial path roughness between two points and performing iterative screening. For large-scale, high-resolution raster data (such as global- scale remote sensing imagery), its computational cost grows exponentially. Future work should adopt spatial indexing or parallel computing to improve the computational efficiency of SA-GCCM for big spatial data.

  Causal Inference for Spatial Cross-Sectional Data: A Geographical Convergent Cross Mapping Method Accounting for Spatial Autocorrelation · 2026 · DOI

• The potential negative impacts of coal sector decline on population health are not well understood. There is a need for studies that examine the relation between coal production, working hours per miner, coal mining employment, and life expectancy in U.S. counties.

  A Not‐So‐Just Transition? Examining the Effects of Coal Sector Decline on Life Expectancy in U.S. Counties · 2026 · DOI

• Future research can build on the findings of this study. Further analysis can be conducted on the implications of population ageing for policymakers and practitioners.

  Assessment of population ageing using the statistical method of spatial autocorrelation: a case study of Nitra Region (Slovakia) · 2026 · DOI

• There is a gap in the analysis of population ageing at the local level. The study addresses the lack of research on population ageing in the Nitra Region.

  Assessment of population ageing using the statistical method of spatial autocorrelation: a case study of Nitra Region (Slovakia) · 2026 · DOI

• The study faces the challenge of estimating local Pareto exponents in a spatially varying manner. The exclusion of spatial heterogeneity in the analysis may lead to biased results. The study needs to account for other factors that may influence the spatial distribution of population.

  Spatial heterogeneity of population distribution in Poland: a geographically weighted regression approach to Zipf's law · 2026 · DOI

• Prior work has estimated unique Pareto exponents assuming spatial invariance, but this is unlikely to be the case. The exclusion of spatial heterogeneity in the analysis may lead to biased results.

  Spatial heterogeneity of population distribution in Poland: a geographically weighted regression approach to Zipf's law · 2026 · DOI

• The need for accurate treatment of regional site effects in partially non-ergodic Ground-Motion Models (GMMs), - The lack of comparison between Bayesian and frequentist mixed-effects models for period-dependent site amplification

  Bayesian vs. Frequentist Mixed-Effects for Regional Site Amplification · 2026 · DOI

• , spline or Gaussian-process priors on c1(T), c2(T), and the regional spreads—so information is shared across neighboring periods and Implementing uncertainty analogous cross-period structure in lmer is possible but cumbersome, and uncertainty propagation remains limited to asymptotic standard errors or bootstrap samples. These priors act as soft constraints: they do not force the answer, but they anchor estimates to physically plausible behavior and reduce variance where the data carry limited information—key advantages that lmer cannot provide natively. • Variance regularization: half-normal/half- Student-t priors on random-effect standard deviations to prevent over-dispersion of regional terms when data are sparse.

  Bayesian vs. Frequentist Mixed-Effects for Regional Site Amplification · 2026 · DOI

• The study suggests that future research should consider the complex relationships between economic growth and environmental quality. The study suggests that future research should examine the validity of the EKC hypothesis in different countries and contexts.

  ÇEVRESEL KUZNETS EĞRİSİ HİPOTEZİNİN GEÇERLİLİĞİ: TÜRKİYE ÖRNEĞİ · 2026 · DOI

Most-cited papers in Spatial and Panel Data Analysis

• Initial conditions and moment restrictions in dynamic panel data models · Journal of Econometrics · 1998 · 19,722 citations
• Testing for unit roots in heterogeneous panels · Journal of Econometrics · 2003 · 10,646 citations
• Local Indicators of Spatial Association—LISA · Geographical Analysis · 1995 · 10,389 citations
• Unit root tests in panel data: asymptotic and finite-sample properties · Journal of Econometrics · 2002 · 8,815 citations
• Consistent Covariance Matrix Estimation with Spatially Dependent Panel Data · The Review of Economics and Statistics · 1998 · 5,660 citations
• Testing slope homogeneity in large panels · Journal of Econometrics · 2007 · 5,335 citations
• The Analysis of Spatial Association by Use of Distance Statistics · Geographical Analysis · 1992 · 5,152 citations
• What To Do (and Not to Do) with Time-Series Cross-Section Data · American Political Science Review · 1995 · 4,799 citations
• Threshold effects in non-dynamic panels: Estimation, testing, and inference · Journal of Econometrics · 1999 · 4,415 citations
• Estimating long-run relationships from dynamic heterogeneous panels · Journal of Econometrics · 1995 · 3,850 citations

Most recent work

• Fine-Scale spatial simulation and interpretation of occupational population distributions using big data and machine learning · Habitat International · 2026
• Nickell bias in panel local projection: Financial crises are worse than you think · Journal of International Economics · 2026
• A Not‐So‐Just Transition? Examining the Effects of Coal Sector Decline on Life Expectancy in U.S. Counties · Rural Sociology · 2026
• Estimation and inference for unbalanced panel data models with interactive fixed effects · Journal of Econometrics · 2026
• Coarse‐to‐Fine Spatial Modeling: A Scalable, Machine‐Learning‐Compatible Framework · Geographical Analysis · 2026
• Sparse Warcasting · Scottish Journal of Political Economy · 2026
• Cointegration in Panel, Spatial and Spatio‐Temporal Models: Some Recent Advances and Applications · Oxford Bulletin of Economics and Statistics · 2026
• Causality in spatial economic analysis: with reference to the London Green Belt and house prices · Spatial Economic Analysis · 2026
• Time Invariant Variables in the Mundlak and Hausman–Taylor Panel Data Models · Oxford Bulletin of Economics and Statistics · 2026
• Adaptive group lasso penalized variable selection for high-dimensional varying-coefficient panel data models with fixed effects · Computational Statistics & Data Analysis · 2026

Related topics in Economics, Econometrics and Finance