ajafe@ajeb.edu.vn (+84) 28 3821 0238
Asian Journal of Applied Financial Econometrics Asian Journal of Applied Financial Econometrics
Đăng nhậpSign in Đăng kýSign up

Giỏ hàngCart

    MUACheckout
    Gửi bài viết Submit Article

    AIMS AND SCOPE

    The Asian Journal of Applied Financial Econometrics  (AJAFE) is an international, peer-reviewed journal dedicated to advancing the research frontier at the intersection of econometrics, finance, and economics. The journal publishes rigorous theoretical and empirical contributions that develop, evaluate, and apply econometric methods to address complex problems in modern financial and economic systems.

    The journal’s mission is to strengthen the link between econometric methodology and financial economics by promoting research that is both technically advanced and economically substantive. Submissions are expected to demonstrate clear contributions to empirical knowledge, methodological innovation, or both.

    The scope of the journal encompasses a wide range of topics, including financial markets, asset pricing, corporate finance, monetary policy, banking, investment and portfolio management, risk management, and behavioral finance. It also welcomes interdisciplinary contributions from areas such as industrial organization, labor economics, environmental and resource economics, and econophysics, particularly where they provide new insights into financial phenomena.

    Particular emphasis is given to research involving estimation, inference, prediction, and calibration in dynamic and high-dimensional settings. Relevant topics include, but are not limited to, volatility and dependence modeling, continuous-time and stochastic processes, dynamic conditional moments, extreme value methods, long memory, mixture models, endogenous sampling, and the econometrics of high-frequency and transaction-level data. Contributions addressing experimental and behavioral finance using advanced econometric techniques are also encouraged.

    By addressing the statistical and computational challenges posed by the evolution of global financial markets, particularly in Asia, which is one of the most dynamic financial regions globally, the Asian Journal of Applied Financial Econometrics aims to serve as a leading platform for innovative research that shapes both econometric practice and financial theory.

    SỐ MỚI NHẤT LATEST ISSUE

    Asian Journal of Applied Financial Econometrics Asian Journal of Applied Financial Econometrics

    Asian Journal of Applied Financial Econometrics
    Xem số mới nhất View current issue

    • Systemic shocks, sectoral vulnerability, and firm-level volatility: Bayesian evidence from Vietnam Systemic shocks, sectoral vulnerability, and firm-level volatility: Bayesian evidence from Vietnam

      Tran Hoang Truc Linh, Ahmad Shabir Faizi, Nguyen Van Diep

      DOI: 10.63065/ajafe.2026.1.003

      Tóm tắt Abstract | Bài viết Article | PDF | Mua ngay Buy now
      Tóm tắt Abstract

      Purpose – This study examines the impact of systemic shocks on firm-level stock return volatility and investigates whether sectoral heterogeneity persists during periods of heightened uncertainty. Focusing on the COVID-19 lockdown in Vietnam, the study evaluates the extent to which firm-specific characteristics remain informative under extreme market conditions.
      Design/methodology/ approach – Using daily stock return data from firms listed on the Ho Chi Minh Stock Exchange (HOSE) during the COVID-19 lockdown period, this study employs a Bayesian regression framework to examine firm-level volatility dynamics and sectoral heterogeneity.
      Findings – The results provide strong evidence of systemic dominance during the lockdown period. Lockdown measures significantly increase firm-level stock return volatility, while cross-sectional heterogeneity declines as firm-specific characteristics become less informative. However, firms in the communication and technology sector exhibit significantly stronger volatility responses than those in other industries.
      Originality/value – This study contributes to the crisis and volatility literature by proposing a dual-layer framework in which systemic shocks generate market-wide volatility synchronization while preserving sector-specific vulnerability. It also demonstrates the usefulness of Bayesian inference in evaluating volatility dynamics under extreme uncertainty, with implications for investors and regulators in emerging markets.

    • Ensemble learning for credit scoring: A comparative study of LightGBM and CatBoost Ensemble learning for credit scoring: A comparative study of LightGBM and CatBoost

      Do Thi Thanh Dieu, Ha Binh Minh, Phan Dinh Phung, Trinh Hoang Nam

      DOI: 10.63065/ajafe.2026.1.004

      Tóm tắt Abstract | Bài viết Article | PDF | Mua ngay Buy now
      Tóm tắt Abstract

      Purpose – This paper examines whether gradient-boosted tree ensembles can improve credit-scoring performance over a traditional logistic model while remaining usable within common banking governance practices.
      Design/methodology/approach – Using the UCI dataset (30,000 observations), we compare logistic regression, LightGBM, and CatBoost under consistent class weighting (w = 3.52). All models are tuned with Optuna (40 trials) using stratified 3-fold cross-validation. Evaluation includes ROC-AUC, PR-AUC, KS, and Brier metrics; DeLong’s test; SHAP interpretation; and calibration diagnostics via ECE and SSM regression. Gender fairness and PSI stability checks are also performed.
      Findings – Both ensembles outperform logistic regression (ROC-AUC ≈ 0.777 vs 0.710; KS ≈ 0.42 vs 0.36). DeLong’s test confirms no significant difference between CatBoost and LightGBM (p = 0.768), though both exceed the baseline (p < 0.001). SHAP identifies repayment status (PAY_0) and credit limit (LIMIT_BAL) as dominant drivers. Post-hoc isotonic calibration improves LightGBM’s ECE from 0.200 to 0.008. Finally, ensembles reduce fairness gaps and maintain high stability (PSI < 0.01).
      Originality/value – The paper offers a compact, reproducible evaluation workflow that integrates hyperparameter optimization, imbalance handling, statistical significance testing, explainability, fairness diagnostics, and stability monitoring for ensemble-based credit scoring.

    • Robustness of Bayesian coefficient alpha estimation under non-normal item distributions: Evidence from a Monte Carlo study using a normal posterior Robustness of Bayesian coefficient alpha estimation under non-normal item distributions: Evidence from a Monte Carlo study using a normal posterior

      Nguyen Ngoc Thach

      DOI: 10.63065/ajafe.2026.1.005

      Tóm tắt Abstract | Bài viết Article | PDF | Mua ngay Buy now
      Tóm tắt Abstract

      Purpose – This study evaluates whether a recently proposed Bayesian estimator of coefficient alpha based on a normal posterior distribution remains reliable when applied to non-normal item responses, a common feature of behavioral and social science data.
      Design/methodology/approach – A Monte Carlo simulation was conducted across 1,440 conditions defined by the number of items (5, 10, 15, and 20), correlation structure (two parallel and three congeneric models), sample size (50, 100, 150, 200, 250, and 300), item response categories (2, 3, 5, and 7), and three distribution types. Using non-informative priors and 2,000 posterior draws, the study examined three 95% Bayesian credible interval methods for coefficient alpha: percentile, normal-theory, and highest probability density intervals. Performance was assessed primarily through coverage probability, with interval width also considered.
      Findings – The results show that all three credible interval methods performed similarly and generally achieved acceptable coverage across most simulation conditions. However, performance deteriorated when data involved binary items, distribution type 2, and/or a parallel model with common loadings of .705, particularly when these conditions occurred jointly. The method was more dependable when items had at least three response categories and sufficient variability.
      Research limitations/implications: The findings are limited to the simulation settings examined and to the normal-posterior framework used to estimate coefficient alpha. The results suggest caution when applying this approach to binary, highly skewed, or range-restricted data, and indicate the need for future comparisons with alternative Bayesian and non-Bayesian estimators, as well as sampling-based approaches such as Gibbs sampling.
      Originality/value – This study extends recent Bayesian research on coefficient alpha by explicitly testing the robustness of a computationally convenient normal-posterior estimator under non-normal item distributions. It clarifies the empirical conditions under which the method performs well and identifies situations in which its use should be treated more cautiously.

    • How Granger’s time series methods can eliminate hypothesis testing & parameter-centric analyses How Granger’s time series methods can eliminate hypothesis testing & parameter-centric analyses

      William M. Briggs

      DOI: 10.63065/ajafe.2026.1.002

      Tóm tắt Abstract | Bài viết Article | PDF | Mua ngay Buy now
      Tóm tắt Abstract

      Purpose – Hypothesis testing and parameter-centric analyses ought to be replaced by the concept of relevance and usefulness. All models ought to be tested against reality, which almost never happens, and cannot when hypothesis testing or parameter statements are used.
      Design/methodology/approach – We review the state of evidence in time series models, specifically Granger causality and testing, and see why this was seen as an innovation. We show that the central idea of prediction applies to all models, not just time series.
      Findings – Granger so-called causality does not identify cause, nor does hypothesis testing or parameter statements. But the history of Granger’s idea show the shortcomings of classical analysis methods and point the way toward clear thinking on cause.
      Originality/value – If hypothesis testing and parameter-centric analysis could be abandoned in favor of predictive approaches, we could eliminate a major source of error and over-certainty.

    • Nonparametric econometrics with fractional and long-memory regressors Nonparametric econometrics with fractional and long-memory regressors

      Ciprian A. Tudor

      DOI: 10.63065/ajafe.2026.1.001

      Tóm tắt Abstract | Bài viết Article | PDF | Mua ngay Buy now
      Tóm tắt Abstract

      Purpose – This paper surveys advances in nonparametric regression models with long-memory regressors, particularly those driven by fractional Brownian motion, and examines how persistence affects econometric inference.
      Design/methodology/approach – A theoretical survey is conducted on kernel-based nonparametric estimators when regressors follow fractional stochastic processes. The analysis relies on asymptotic methods and incorporates tools such as local time techniques and Malliavin calculus to study estimator behavior under strong dependence.
      Findings – Long-range dependence substantially alters classical nonparametric results. The Hurst parameter plays a central role in determining consistency, convergence rates, and asymptotic distributions of estimators. Standard assumptions of independence or weak dependence are no longer applicable, and the dependence structure directly impacts bandwidth selection and limiting behavior. The analysis also highlights the importance of local time in understanding estimator properties under fractional dynamics.
      Originality – The paper provides a unified synthesis of recent theoretical developments at the intersection of nonparametric econometrics and fractional stochastic processes. It clarifies how long-memory features influence estimation theory and highlights key probabilistic tools for analyzing regression models with persistent regressors.

    KẾT NỐI VỚI CHÚNG TÔI QUA... CONNECT WITH US ON...

    CHỈ MỤC INDEXED IN

    gg scholar DOAJ — Directory of Open Access Journals EBSCO Discovery Service WorldCat IDEAS EconBiz Summon Scilit VJOL