DOI: 10.17233/sosyoekonomi.2026.03.06 ISSN: 1305-5577

Innovations in Monte Carlo-Based Markov Switching Tests: An Application to Türkiye’s GDP Data

Anıl Güzel
Show PDF Cite
This study explores the application of Monte Carlo-based methodologies, specifically the “Local Monte Carlo Likelihood Ratio Test (LMC-LRT)” and the “Maximised Monte Carlo Likelihood Ratio Test (MMC-LRT)”, within the context of Markov switching models to analyse Türkiye’s quarterly GDP data from 1998Q1 to 2024Q2. These tests, developed by Rodríguez-Rondon and Dufour (2024), address challenges associated with non-standard asymptotic distributions and nuisance parameters, offering robust solutions for identifying regime shifts. The empirical analysis demonstrates that the “Markov Switching Autoregressive (MSAR)” model significantly outperforms the baseline Autoregressive (AR) model, capturing two distinct economic regimes characterised by differing means and variances. The Monte Carlo-based tests strongly reject the null hypothesis of no regime-switching (p < 0.05), validating the presence of structural breaks in Türkiye’s GDP dynamics. The findings emphasise the utility of advanced methodologies for analysing economic time series, providing policymakers with valuable insights and reinforcing the theoretical contributions of Monte Carlo-based approaches. This study highlights the practical and theoretical significance of regime-switching models in understanding economic dynamics and structural changes over time.

More from our Archive