Professor Josip Arnerić and Doctoral Student Mateusz Buczyński as Speakers at the QFRG and DSLab Seminar

We cordially invite you to attend the seminar titled „Recurrent neural architectures for nonlinear volatility modelling”. The presentation will explore the relative strengths, limitations, and potential synergies of two approaches to modelling nonlinear volatility: the Jordan neural network and the GARCHNet model. The lecture will be delivered by Professor Josip Arnerić (University of Zagreb) and doctoral student, Mateusz Buczyński.

The seminar, organised as part of the QFRG and DSLab seminar series, will take place on May 12, 2025, at 18:30 (please arrive/log in no later than 18:20).  

The event will be held in a hybrid format: in person in Room B002 at the Faculty of Economic Sciences and online.

Link to the meeting: https://uw-edu-pl.zoom.us/j/94024724682?pwd=4hzHO0N8bWKL1IMN7xIqwQvcfICw0c.1

(Joining a meeting remotely implies consent to recording. Please keep your camera and microphone turned off during the presentation and send the questions to the speaker in the chat.)

Below, we present the abstract of the presentation.

---

We bring together two advances in volatility modeling – the Jordan Neural Network (Arnerić et al.) and GARCHNet (Buczyński & Chlebus) – and critically examine their relative strengths, limitations and potential synergies.

We first introduce the JNN(1,1,1) approach, a parsimonious Jordan‐type recurrent network designed as a semi‐parametric analog to GARCH(1,1). Applied to daily CROBEX returns, the JNN outperforms the standard GARCH(1,1) in out‐of‐sample conditional variance forecasts, preserving interpretability of key parameters while capturing time‐varying nonlinearity.

Next, we present GARCHNet, which embeds an LSTM architecture within a classical GARCH framework to capture richer nonlinear dynamics in conditional variance. By combining maximum‐likelihood‐based GARCH estimation with an LSTM module, the model flexibly accommodates normal, t and skewed‐t innovations. An empirical study on WIG20, S&P 500 and FTSE 100 returns (2005–2021) demonstrates improved in‐sample fit and VaR performance, while suggesting further improvements for extending to alternative distributions and longer‐memory architectures.

A debate will follow to contrast these approaches and explore landscape for their integration.