Bayesian inference for dynamic spatial quantile models with interactive effects
Tomohiro Ando, Jushan Bai, Kunpeng Li, Yong Song · 2025
Read Bayesian inference for dynamic spatial quantile models with interactive effects by Tomohiro Ando, Jushan Bai, Kunpeng Li, Yong Song (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2503-00772.txt.
Introduction
Tomohiro Ando and three co-authors build a Bayesian spatial model in twenty twenty-five. Large spatial panel data tracks many places over time. Quantile models describe more than the average response, including the tails. Their model is dynamic and spatial, with interactive effects capturing unobserved heterogeneity.
Cross-sectional dependence can be high-dimensional in such panels. Regression coefficients are allowed to differ across units. Estimation uses a new Bayesian Markov chain Monte Carlo algorithm. Novel pieces include quantile randomization and a Gibbs sampler for structural parameters. The tail behavior of an inverse Gaussian generator is stabilized as well. Consistency is proven as both panel dimensions grow. Simulations confirm the finite-sample behavior. Readers in econometrics get quantiles for maps that move.
Catalog record
| CREATOR | Tomohiro Ando, Jushan Bai, Kunpeng Li, Yong Song |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 62 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2503-00772
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Bayesian inference for dynamic spatial quantile models with interactive effects Thanks: The paper has benefited from constructive comments from participants at Australian spatial econometrics and statistics workshop 2023, held at Monash university. This research was supported by University of Melbourne Research Computing Services and the Petascale Campus Initiative. It is also supported by the Australian Research Council Discovery Grants DP230100959 and DP240101009. Tomohiro Ando, Jushan Bai, Kunpeng Li, Yong Song Abstract This paper proposes a dynamic spatial panel quantile model that accounts for unobserved heterogeneity. The Bayesian MCMC algorithm developed here introduces several novel features tailored to the challenges of high-dimensional dynamic spatial quantile models. Key innovations include quantile randomisation, which enables conditional conjugacy and facilitates efficient sampling, and a new Gibbs sampler specifically designed to handle structural spatial parameters. We also establish the Bayesian consistency of the proposed estimation method. Finally, we demonstrate the practical utility of the approach through an application to the quantile co-movement structure of the Australian gasoline market. Keywords : Dynamic panel, endogeneity, factor models, heterogenous spatial effects, high dimensional data. JEL classification : C31, C33, E44 1 Introduction Spatial panel data models explore spatial interactions across individuals or regions over time. These models have found wide-ranging applications in fields such as housing economics (e.g., Beenstock and Felsenstein (2015) ), marketing (e.g., Hunneman et al. (2022) ), and urban economics (e.g., Glaser et al. (2022) ). In this paper, we introduce a novel dynamic spatial panel quantile model with interactive effects. This model simultaneously accounts for spatial dependence, temporal dynamics, unobserved heterogeneity, and heterogeneous effects across quantiles—features that are often observed in economic and social data but rarely modeled together. There is a rich literature on linear panel models with interactive effects (e.g., Bai (2009) ; Bai and Li (2014) ; Ando and Bai (2017) ; Hallin and Liška (2007) ; Moon and Weidner (2015) ; Pesaran (2006) ; Stock and Watson (2002) ; Lu and Su (2016) ) and linear spatial panel models (e.g., Aquaro et al. (2021) ; Baltagi (2011) ; Bai and Li (2021) ; Kelejian and Prucha (2004) ; Lee (2004) ; Lin and Lee (2010) ; Li (2017) ; Lu (2017) ; Qu and Lee (2015) ; Reich et al. (2011) ; Shi and Lee (2017) ; Yu et al. (2008) ). However, research on quantile regression (see Koenker and Bassett (1978) ) in spatial and high-dimensional panel settings remains limited, especially for models that incorporate both interactive effects and dynamic structures. Recent work by Ando and Bai (2020) extended quantile methods to panel data with interactive effects, capturing quantile co-movements while allowing for heterogeneous slope coefficients. However, their model does not accommodate dynamic and spatial dependencies in quantiles. Spatial interactions and temporal dynamics are fundamental features of economics and the social sciences. Spatial dependence may arise from peer effects, spillover effects, or imitative behavior (see Manski (1993) ; Anselin (1988) , among others), while temporal dependence often reflects persistence in individual or group behavior over time. To capture these complex features, we propose a dynamic spatial quantile panel model with interactive effects. This framework facilitates quantile-specific analysis of both spatial and temporal dependence, accommodates heterogeneous slope coefficients, and incorporates latent factor structures. Due to the model’s high-dimensional parameter space—driven by the dynamic structure and large number of cross-sectional units—we develop a new Bayesian Markov Chain Monte Carlo (MCMC) estimation method tailored to this setting. Methodological challenges : Several methodological challenges arise in the estimation process. First, because the quantile function is defined implicitly in terms of its coefficients, a conventional Gibbs sampler is not directly applicable. To overcome this, we introduce randomised quantile dynamics that create conditional conjugacy, allowing for efficient sampling. Second, the spatial autoregressive parameter is structural in nature and does not permit a standard Gibbs step. Inspired by methods in structural vector autoregression, we construct a bimodal mixture distribution to incorporate this parameter within our MCMC framework. Finally, to alleviate the computational burden from high-dimensional matrix operations, we apply a breadth-first search algorithm from graph theory to block-diagonalise the spatial matrix, thereby significantly reducing computational complexity. These methodological innovations enable our MCMC algorithm to efficiently estimate large-scale dynamic spatial quantile panel models with latent factor structures. Theoretical Challenges: To establish the validity of our Bayesian approach, we prove Bayesian consistency—a theoretical result not previously achieved for panel data models with interactive effects. This requires overcoming several substantial hurdles: the dynamic structure of the model, high-dimensional incidental parameters from latent factors and their loadings, the nonsmooth nature of the quantile objective function, the nonlinearity introduced by the spatial term, and the rotational indeterminacy of the factor structure. Although recent work (e.g., Ando and Bai (2020) ; Chen et al. (2021) ) addresses some related issues, their results are not applicable to our setting. Consequently, we first derive new asymptotic properties within a frequentist framework (see Remark 7), which we then build upon to develop our Bayesian theory. Even with these foundations, our model demands additional theoretical innovation to rigorously establish Bayesian consistency. Our contributions are summarised as follows. First, we introduce a dynamic spatial panel quantile model with interactive effects, designed to accommodate key features such as spatial dependence and temporal dynamics. This specification enhances the model’s flexibility in capturing complex data patterns. Second, we develop a novel Bayesian estimation procedure tailored to high-dimensional parameter spaces. Third, we establish several asymptotic properties, including Bayesian consistency. Finally, we demonstrate the utility of our framework through an empirical analysis of the Australian gasoline market. The paper is organised as follows. Section 2 introduces a new dynamic spatial panel quantile model with interactive fixed effects. Section 3 outlines the proposed Bayesian MCMC estimation procedure. Section 4 establishes the Bayesian consistency of the method. In Section 5, we apply the proposed approach to the Australian gasoline market. Section 6 concludes the paper. To conserve space, all technical proofs are provided in the online supplementary material. The supplementary document also includes Monte Carlo simulation results, which demonstrate the effectiveness of the proposed estimation procedure. Notations: Let \|A\| = [\mathrm{tr}(A^{\prime}A)]^{1/2} be the Frobenius norm of matrix A , where “tr” denotes the trace of a square matrix, and let \|A\|_{2} be its spectrum norm (the largest singular value of A ). In addition, for any N\times N matrix \|A\|_{1} is defined as \|A\|_{1}=\max_{1\leq j\leq N} \sum_{i=1}^{N}|a_{ij}| where a_{ij} is the (i,j) -th element of A . Similarly, \|A\|_{\infty}=\max_{1\leq i\leq N}\sum_{j=1}^{N}|a_{ij}| . For sequences a_{n} and b_{n} , the notation a_{n}\lesssim b_{n} means a_{n}=O(b_{n}) , that is, there exists C>0 and for all n large enough, a_{n}\leq Cb_{n} . We write c_{n}=O_{p}(d_{n}) if c_{n}/d_{n} is stochastically bounded, and c_{n}=o_{p}(d_{n}) if c_{n}/d_{n} converges to zero in probability. 2 Dynamic spatial quantile models with interactive effects Suppose that, for the i -th unit (i=1,...,N) at time t (t=1,...,T) , its response y_{it} is observed together with a set of p explanatory variables \{x_{it,1},...,x_{it,p}\} . We consider the \tau -th quantile function of y_{it} by jointly modeling spatial effects, time effects and common shocks. To capture these effects simultaneously, we define the \tau -th quantile function of y_{it} as \displaystyle Q_{y_{it}}\Big(\tau|X_{t},F_{t,\tau},B_{\tau},\Lambda_{\tau},\mbox{\boldmath$\rho$}_{\tau},\mbox{\boldmath$\gamma$}_{\tau},\mbox{\boldmath$\delta$}_{\tau}\Big) \displaystyle\equiv\rho_{i,\tau}\sum_{i\neq j,j=1}^{N}w_{ij}Q_{y_{jt}}\Big(\tau|X_{t},F_{t,\tau},B_{\tau},\Lambda_{\tau},\mbox{\boldmath$\rho$}_{\tau},\mbox{\boldmath$\gamma$}_{\tau},\mbox{\boldmath$\delta$}_{\tau}\Big) \displaystyle\quad+\delta_{i,\tau}\sum_{i\neq j,j=1}^{N}w_{ij}Q_{y_{j,t-1}}\Big(\tau|X_{t-1},F_{t-1,\tau},B_{\tau},\Lambda_{\tau},\mbox{\boldmath$\rho$}_{\tau},\mbox{\boldmath$\gamma$}_{\tau},\mbox{\boldmath$\delta$}_{\tau}\Big) \displaystyle\quad+\gamma_{i,\tau}Q_{y_{i,t-1}}\Big(\tau|X_{t-1},F_{t-1,\tau},B_{\tau},\Lambda_{\tau},\mbox{\boldmath$\rho$}_{\tau},\mbox{\boldmath$\gamma$}_{\tau},\mbox{\boldmath$\delta$}_{\tau}\Big)+\mbox{\boldmath$x$}_{it}^{\prime}\mbox{\boldmath$b$}_{i,\tau}+\mbox{\boldmath$f$}_{t,\tau}^{\prime}\mbox{\boldmath$\lambda$}_{i,\tau} (1) for i=1,\dots,N and t=1,\dots,T . Here w_{ij} (i=1,2,\cdots,N;j=1,2,\cdots,N) are pre-specified