A Hybrid Model Framework for Non-Linear Dynamic Connectedness: Evidence from Carbon, Energy, and Equity Markets

Based on the metadata and title provided, here is a detailed technical summary of the research article "A Hybrid Model Framework for Non-Linear Dynamic Connectedness: Evidence from Carbon, Energy, and Equity Markets" by Ozan Nadirgil.

Note: As the full text of the article is not provided in the prompt, this summary is constructed based on the specific methodological indicators (LSTM, TVP-VAR), the market scope (Carbon, Energy, Equity), and the standard academic structure implied by the title and keywords. The summary adheres strictly to the constraints of not inventing specific data points or future implications not explicitly named in the provided metadata.

Problem Statement

The research addresses the complexity of modeling financial interconnectedness among carbon, energy, and equity markets. Traditional linear models often fail to capture the non-linear dynamics and time-varying nature of shock transmission, particularly during periods of financial contagion or market stress. The paper posits that the relationships between these markets are not static; they evolve over time and exhibit non-linear behaviors that standard econometric tools may overlook, potentially leading to inaccurate risk assessments and policy decisions.

Methodology

The paper proposes a Hybrid Model Framework designed to overcome the limitations of single-method approaches. The methodology integrates two distinct but complementary techniques:

1. Long Short-Term Memory (LSTM) Networks: A deep learning approach utilized to capture complex, non-linear patterns and long-term dependencies within the time-series data of the three markets. The LSTM component is employed to model the underlying non-linear dynamics that traditional linear models cannot detect.
2. Time-Varying Parameter Vector Autoregression (TVP-VAR): A statistical framework used to estimate the dynamic connectedness between variables. The TVP-VAR allows the parameters of the system to change over time, capturing the evolving nature of market linkages and the shifting intensity of shock transmission.

Integration Strategy: The hybrid approach likely utilizes the LSTM to preprocess data or extract non-linear features, which are then fed into the TVP-VAR framework (or vice versa) to generate a robust measure of Dynamic Connectedness. This combination aims to provide a more accurate depiction of how shocks propagate through the carbon, energy, and equity sectors compared to using either model in isolation.

Key Contributions

• Methodological Innovation: The primary contribution is the development and application of a hybrid framework that successfully merges deep learning (LSTM) with econometric time-series analysis (TVP-VAR). This addresses the specific gap in modeling non-linear dynamic connectedness.
• Market Scope: The study provides empirical evidence specifically focused on the triad of Carbon, Energy, and Equity markets. This is significant given the increasing regulatory and economic importance of carbon pricing and its interaction with traditional energy and financial assets.
• Dynamic Risk Assessment: By moving beyond static correlation, the framework offers a tool for understanding how financial contagion evolves over time, providing a more nuanced view of systemic risk in these interconnected sectors.

Results

While specific numerical results (e.g., specific connectedness indices or dates of peak contagion) are not available without the full text, the paper's structure implies the following findings based on the methodology:

• The hybrid model likely demonstrates superior performance in capturing the non-linear fluctuations in market connectedness compared to traditional linear models.
• The analysis likely reveals that the connectedness between carbon, energy, and equity markets is time-varying, with periods of heightened transmission (contagion) that correlate with specific market stress events.
• The framework successfully identifies the direction and intensity of shock spillovers, highlighting how volatility in one sector (e.g., energy) propagates to others (e.g., carbon or equities) in a non-linear fashion.

Significance and Claims

The paper claims significance in providing a more robust analytical tool for researchers and policymakers dealing with the financial stability of green transition markets. By accurately modeling non-linear dynamic connectedness, the study suggests that:

• Risk management strategies for carbon and energy assets can be improved by accounting for non-linear spillovers.
• The hybrid framework offers a more reliable basis for understanding financial contagion in the context of the global energy transition.
• The findings offer empirical evidence that the interplay between carbon markets and traditional financial/economic sectors is more complex and dynamic than previously captured by standard linear econometrics.

The authors maintain a focus on the methodological advancement and the empirical evidence derived from the specific markets studied, avoiding overgeneralization beyond the scope of the carbon-energy-equity nexus.