Decoupling from gas

Victoria sits inside the National Electricity Market (NEM), a multi-region wholesale pool where generators bid every few minutes and the cheapest plant needed to meet demand sets the price (the “merit-order”).

Historically, brown coal supplied most Victorian energy, with gas often setting the marginal price. Several events reshaped this picture, including coal closures, Covid-19, Russia's invasion of Ukraine and the resulting global fuel crunch and domestic market interventions.

Our analysis sits on top of this institutional backdrop. We are not re-telling the standard merit-order story, but testing whether gas still anchors prices in this newer, high-renewables, interconnected system.

The dominant story in Australia’s electricity debate has been simple:

“If gas prices jump, electricity prices follow.”

Policy reports, think tanks and commentators all treat gas as the fuel that sets wholesale prices in the National Electricity Market (NEM). The Australia Institute’s 2025 briefing on gas and electricity prices puts it bluntly: “Gas and electricity prices are closely correlated because in the NEM the price of gas-powered electricity generation often sets the wholesale price of electricity.”

Earlier work from the same institute argued that linking east coast gas to export markets has “tripled Australian gas prices and doubled electricity prices”, cementing the idea that gas costs flow straight through to power bills.

Independent analysts make the same connection. In a widely cited piece on high power bills, IEEFA notes that gas prices have historically been a key driver of wholesale electricity price rises, with a very strong historical correlation between gas and electricity prices.

Academic work has reinforced this view. Nolan, Gilmore and Munro (2022), in a Griffith University working paper on the gas–electricity price relationship in the NEM, show that over 2012–2021 much of the mainland market behaved like a gas-indexed system: wholesale prices could be closely linked to Wallumbilla gas via an effective “grid heat rate”, with coal and hydro often shadow-pricing gas even when gas units were not literally setting the price.

The Climate Council’s explainer on why power prices are “sky high and rising” pushes the same story from another angle: “polluting and expensive fossil fuels like coal and gas are the main culprits” behind soaring bills.

Put together, think-tank commentary and academic research reinforce a clear picture: in the standard narrative, gas still sits at the centre of price formation, with renewables and weather playing supporting roles at the margins. The rest of this page tests that story directly using Victorian data from 2015–2025 — and shows how, in a high-renewables, interconnected system, that gas–power link has fundamentally changed.

This work builds on that standard gas–power story, but does three things differently:

• Victoria focus: It zooms in on Victorian wholesale prices rather than treating the mainland NEM as one homogeneous market.
• High-renewables period: It extends the data through 2025, capturing the post-2022 crisis years where renewables, storage and interconnectors play a much larger role.
• Time-series lens: It uses cointegration tests, structural break tests and VAR/FEVD analysis to look at regime shifts and variance shares, not just simple correlations or average “heat rates”.

Empirically, three findings drive the story:

• Pass-through collapses: rolling gas → power models show a clear positive relationship in 2015–19 that falls to around zero (and sometimes negative) in 2023–25.
• Renewables ≫ gas in variance shares: in the latest data, gas shocks explain only around 2% of two-year price variance, while renewables explain around 14–15%, and this result is robust to alternative VAR orderings.
• A decade-long shift: FEVD trends over time show renewables’ contribution to price variation rising and gas’s contribution shrinking, consistent with a structural move away from gas-anchored pricing.

Taken together, these results suggest that while gas historically anchored prices across the NEM, in Victoria's recent high-renewables regime wholesale prices have largely decoupled from international fossil fuel benchmarks. Geopolitical risk transmitted via global oil and gas prices has been replaced by meteorological and system risk — dunkelflaute, heatwaves and network conditions now sit at the centre of monthly price formation, with gas moving to the periphery.

This section pulls together the major structural shocks to Victoria's system – Hazelwood's closure, Covid-19, the Russia–Ukraine fuel shock and the June 2022 market intervention – as a single visual timeline before you look at prices and renewables in more detail.

Early in the decade, gas and electricity prices track one another closely. Around 2022–23, global fuel shocks and domestic interventions break that tight link: wholesale prices fall back faster than gas, consistent with a system that is less tightly anchored to fuel costs.

3.1 Data

We assembled monthly data for Victoria from early 2015 to late 2025:

• wholesale electricity prices,
• gas prices at the Wallumbilla gas hub,
• the share of generation from renewables,
• imports and exports between Victoria and other states,
• demand, temperature and key global fuel prices.

3.2 Methods

On top of this, we applied well-established econometric tools: unit root and cointegration tests, structural break tests, rolling models, and vector autoregressions (VARs) with forecast error variance decompositions (FEVDs), plus a standard spillover index to summarise cross-variable connectedness.

3.3 Breaks

These tests first ask whether any variable provides a stable long-run “anchor” for the others. In our data there is no cointegrating relationship tying electricity prices tightly to gas, renewables or demand; in plain terms, power prices do not reliably “track” any one of these over the long run. Around a quarter of overall volatility reflects shocks spilling across variables rather than each series moving on its own. That is one early sign that the old gas anchor has weakened.

You don't need the equations to follow the story. The important part is that we ask the same question in several independent ways: how important is gas today, compared with renewables, imports and demand?

When we line up the data with the model results, Victoria’s last decade looks like three broad phases. Across break tests, rolling regressions and variance decompositions, the same pattern emerges.

4.1 Gas-anchored prices (roughly 2015–2019)

In the early years of our sample, the old story holds. When gas prices move in our models, Victorian wholesale prices move with them. Over a two-year horizon, gas shocks explain roughly one in every eight movements in the wholesale price.

Renewables are growing, but they still explain only a modest share of price variation. Gas is central to price formation; renewables are supporting actors.

Phase 2: Crisis and transition (around 2020–2022)

Covid, global gas shocks and market interventions make prices volatile. In our models, this shows up as a clear rejection of a single, stable gas–price relationship over 2015–2025. Break tests on the rolling gas pass-through pick out shifts clustered around 2022–23 across different window lengths: shorter windows see more stepping, but the regime change is robust.

Phase 3: Weather and renewables in the driver's seat (2023 onwards)

By the most recent years, the picture has flipped. In our variance decompositions, gas shocks explain only around 2% of price variation, while renewables explain about 14–15% and imports a few percent more. The bulk – roughly four-fifths – is due to own-price shocks and other factors. In FEVD language, “other / own price” mostly captures shocks that hit electricity prices directly – policy shifts, bidding behaviour, outages or other market responses we haven’t modelled explicitly – plus the VAR’s mechanical own-shock term.

In practical terms, gas hasn’t disappeared, but it has moved from the centre of price formation to the periphery. In a high-renewables, interconnected system, prices behave less like a pure gas market and more like a weather-driven, renewable-rich system.

This is more than the familiar idea that “renewables push prices down” via the merit order. In our models, once we control for renewables, imports, demand and past prices, additional gas price movements explain only a small slice of remaining volatility – the gas–power link itself has weakened.

For a household or small business, the core question is simple:

“What am I actually exposed to now?”

Earlier in the decade, a significant share of the risk in your electricity bill really was about gas. If international gas prices spiked, there was a good chance your power bill would feel it.

Today, our analysis suggests that your bill is relatively less exposed to gas price shocks and more exposed to:

• weather and temperature, which drive demand,
• the availability of wind and solar, which shifts supply,
• and the health of transmission and interconnectors, which govern imports and congestion.

This doesn’t mean gas is irrelevant – during tight supply episodes gas-fired plant still often sets the marginal price – but on a month-by-month basis, weather, renewables and network conditions now explain much more of the ups and downs.

This shift has real consequences for hedging and market design. Hedging gas alone is no longer enough; weather, renewable generation and congestion risk now matter more for wholesale prices. Market participants increasingly need tools that manage scarcity and abundance risk – contracts linked to weather, flexible load, storage and interconnector availability, not just fuel.

For policymakers, a weaker gas anchor means traditional fuel-cost narratives will explain less and less of what happens to bills. Pricing, reliability and transition policy need to account for a system where renewables, storage and networks drive scarcity events. Market signals should reward flexibility – storage, demand response and transmission – as well as fuel.

The full working paper sets out the models and robustness checks in detail, including unit root tests, break tests and variance decompositions. This page is the accessible version of that story.

This analysis uses monthly data, so it smooths over five-minute and half-hour price spikes that can matter for some market participants. Our models also deliberately focus on a small set of key variables – prices, gas, renewables, imports, demand and weather – rather than the full richness of bidding strategies, contract positions or plant-level constraints.

The findings are specific to Victoria over 2015–2025 and rely on standard time-series assumptions. Different modelling choices, higher-frequency data or future structural changes in the market could shift the precise numbers, even if the broad story about weakening gas pass-through and stronger renewable influence remains.