Detecting Cointegrating Relations In Non Stationary Matrix Valued Time

Happy to say that my first Linkedin post is to announce that our new paper, “Detecting cointegrating relations in non-stationary matrix-valued time series,” co-authored with Alain Hecq and Ines Wilms, is now published in... In this work, we propose a Matrix Error Correction Model (MECM) to uncover long-run relationships across multiple dimensions of a time series—such as macroeconomic indicators and countries—while keeping the model parsimonious. Using information criteria for rank selection, we show how the MECM provides reliable estimations of cointegrating relations via Monte Carlo simulations and a real-world macroeconomic application. If you’re interested in cointegration or matrix-valued time series, we hope you’ll find our approach valuable. The paper is available open access here: https://lnkd.in/e2aTC5KZ Special thanks to everyone who offered insights and support throughout this project! Congratulations on the publication.

This research adds valuable depth to complex data analysis. 📊 #DataScience 📘 New Video Tutorial: Bivariate Morlet Wavelet Analysis in R – Understanding Coherency, Lead–Lag, and Phase Relationships In this comprehensive tutorial, Peterson Owusu Junior, PhD explore how to estimate and interpret Bivariate Morlet Wavelet... This method allows researchers to detect how two time series move together over time and frequency — helping uncover lead–lag dynamics, co-movement strength, and phase relationships that traditional correlation analysis often misses. 💡 What You’ll Learn: ✅ The theory behind wavelet coherency and why it matters in time-frequency analysis. ✅ Step-by-step guide on how to use analyze.coherency() for bivariate time-series analysis.

✅ How to interpret the coherency plot, including power zones, arrows, and the cone of influence. ✅ Understanding in-phase vs out-of-phase movements and identifying which variable leads or lags the other. ✅ Application of these insights in financial, macroeconomic, and climate research. This tutorial bridges theory and application, helping data scientists, economists, and quantitative researchers move beyond static correlations to uncover dynamic interdependencies within complex systems. 🎥 Watch now: https://lnkd.in/eq-5AXTp Follow https://lnkd.in/ebi6tRb7 for more videos on Wavelet Analysis, Connectedness Models, QVAR, TVP-VAR, and Advanced R-based Econometrics. #RStats #Econometrics #WaveletAnalysis #FinancialEconometrics #CoherencyAnalysis #DataScience #TimeSeries #QuantitativeResearch #RProgramming #FinancialModeling #AcademicResearch

Thrilled to share our new publication in Statistical Methods & Applications! 📄 Quasi-maximum likelihood estimation for non-stationary stochastic volatility models: diffuse Kalman filtering approach. In this paper, we developed a quasi-maximum likelihood estimation (QMLE) procedure using the diffuse Kalman information filter to estimate parameters of non-stationary stochastic volatility (SV) models. The key contributions include: -Capturing high volatility persistence as a random walk; -Accounting for leverage effects and heavy-tailed return distributions; -Demonstrating robust and efficient estimation through Monte Carlo simulations and empirical validation on NASDAQ... This work provides a reliable and computationally efficient tool for modeling complex financial time series with heavy tails and non-stationary volatility. Congratulations to my co-authors Abdeljalil Settar and Sara Chegdal for this fruitful collaboration!

🎉 #Statistics #TimeSeries #StochasticVolatility #KalmanFilter #QMLE #FinancialEconometrics #Research #Publication Take a look at our article! https://lnkd.in/dMm-nYxG In this recently published paper - "Generalization of the Feynman-Kac formula for Markov processes", published in Physical Review E (https://lnkd.in/eXy6krqy), we demonstrate how the forward and backward Feynman-Kac equations can be extended to any... To illustrate the power of the extended Feynman-Kac formalism, we study the model with mean reversion and uncorrelated pre- and post-jump states. This model is antipodal to both continuous diffusion and random walks. It is most appropriate when reality involves strong exogenous shocks with exponential relaxation to the mean value between these shocks, leading to a significant departure from conventional modeling assumptions.

Also, we demonstrate how naturally the generalized Feynman-Kac equations can be specialized to treat resetting (as a particular case of the uncorrelated states model) in stochastic hybrid systems. Importantly, the pricing process within the extended Feynman-Kac formalism developed in this paper has the martingale property, which opens new ways to account for deviation from the conventional diffusion model in valuations of derivative...

People Also Search

• Detecting cointegrating relations in non-stationary matrix-valued time ...
• Reproduction material for Detecting Cointegrating Relations in Non ...
• Ines Wilms - Publications - Google Sites
• Ivan Ricardo - Maastricht University
• Research - Ivan Ricardo
• Alain Hecq, Ivan Ricardo arXiv:2411.05601v2 [econ.EM] 24 Jan 2025