A Multiresolution Investigation of How Macro Uncertainties Influence Natural Gas Price Dynamics

The price dynamics within the energy complex appear independently state-dependent; for instance, it is well documented that the price of natural gas (crude) is especially susceptible to large swings in Winter (Summer), when demand is elevated. The influence flowing from crude price dynamics to the energy complex must therefore be considered a multiresolution problem. In this paper we use quantile-on-quantile (QOQ) regressions and wavelet coherence analysis to assess the response of natural gas price return volatility (HHNG) to two macro influences – economic polity uncertainty (EPU) and crude oil volatility (OVX). The QOQ regressions show that the response of natural gas volatility to uncertainty is strongly regime dependent. In tranquil states, when both volatility and uncertainty lie in lower or middle quantiles, shocks to EPU and OVX have little effect and may even slightly dampen volatility. In contrast, once markets enter turbulent regimes, even modest increases in uncertainty generate disproportionate and persistent volatility surges. Long-horizon dynamics show smoother but sustained effects, with high-quantile EPU and OVX jointly contributing to prolonged volatility elevation. Wavelet coherence analysis reinforces the QOQ regression findings by identifying periods of strong, in-phase co-movement between gas volatility and uncertainty indicators and by showing that leadership can shift: gas volatility often leads uncertainty at high frequencies, while EPU or OVX lead gas volatility at medium-term horizons. On the other hand, HHNG emerges as a regime-dependent process shaped jointly by domestic fundamentals, policy uncertainty, and oil-market sentiment. The evidence points to the importance of recognizing the nonlinear and state-contingent dynamics in the statistical evaluation of energy prices; consistent with our choice of methodology, a general and flexible form of state-contingent evaluation appears appropriate.