A silver lining to the European energy crisis: Energy efficiency, productivity, and potential output

European natural gas prices surged in the summer of 2022 to more than ten times their early 2021 levels, and by 2025 remained about twice their historical average. How has the shock affected the economy’s productive capacity, and what does it imply for longer term potential output?

A large literature studying the macroeconomic effects sprung up immediately after the initial energy price shock in 2022, with a consensus now emerging that output held up somewhat better than some initially feared, as Russian gas was replaced with alternative supplies, gas-dependent production processes partially switched to other inputs, and broader substitution along the production chain took place (e.g. Bachmann et al. 2024, Di Bella et al. 2024, Lan et al. 2022).

But most of the existing studies focus on shorter-term dynamics and take aggregate productivity as given. In a new paper (Lan et al. 2026), we take a step back and explicitly focus on how the surge in energy prices has impacted productivity in the euro area and hence potential output. If high energy prices spur innovation in energy efficiency, the costs of the shock could be smaller than feared. This idea draws on the theory of directed technical change (Acemoglu 2002), which predicts that changes in relative prices steer innovation toward scarce inputs. As energy becomes more expensive, firms have an incentive to innovate in energy-saving technologies. For a given level of R&D, however, this reallocation comes at the cost of lower innovation in capital labour-augmenting technologies.

Energy efficiency increased significantly following the sharp increase in energy prices

Natural gas consumption fell durably following the sharp rise in energy prices, as shown in Figure 1, which compares monthly EU gas use before and after the price shock.

Figure 1 Natural gas consumption in the EU (million cubic metres)

Sources: Eurostat and authors’ calculations.

In principle, lower natural gas consumption after the price increase may reflect several factors. In addition to energy efficiency gains, these include behavioural changes, reduced output – via reduced capacity utilisation of capital and labour – and fuel switching, such as substituting oil for natural gas.

As a first step to isolating efficiency effects, Figure 2a shows that output per unit of gas rose sharply as gas prices spiked and then remained at this higher level even as prices declined. For context, a proxy for labour productivity is also plotted, which remained flat or declined. This aligns with the directed technical change model discussed below: as gas prices rose, firms shifted resources toward improving energy productivity at the expense of improvements in labour or capital productivity.

Because a drop in gas use per unit of output could partly reflect fuel switching rather than true efficiency improvements, Figure 2b shows that aggregate energy use per unit of output fell in 2022 and 2023 (latest data), dropping below both the pre-pandemic trend and the pandemic average. The varying magnitudes of the declines across energy sources indicate some fuel switching – especially away from gas given its price spike – but overall energy productivity did increase.

Survey evidence reinforces the role of firm decision-making in driving these gains. For example, a 2024 European Investment Bank (EIB) survey reported that 50% of firms were investing in energy efficiency, with such investments accounting for 12% of total investment – the highest share since the EIB began tracking it: “Europe’s high energy costs have driven efficiency investments” (EIB 2025: 1).

Figure 2 Energy intensity in the EU

The impact on potential output and the silver lining

To assess the macroeconomic impact of higher energy prices, we use a model of endogenous technical change in which firms allocate R&D between capital labour technologies and energy-related technologies. We study the impact of the energy price shock of the past years, with real fossil fuel prices peaking in 2022 at nearly 400% above pre-pandemic levels and returning to their pre-pandemic average by 2026/27 (Figure 3).

We estimate that euro area potential output declined by 0.8% by 2026 relative to a no-shock counterfactual while energy efficiency is around 3% higher. The potential growth effect is strongest early on, when the price shock is largest, and fades as prices normalise and firms reallocate resources to improve energy efficiency. By 2026, the growth effect has dissipated.

Figure 3 Real euro area fossil fuel energy prices (index, average 2016/19=100)

Sources: IMF World Economic Outlook and authors’ calculations. Note: Nominal energy prices are deflated by the euro area GDP deflator. Prices based on July 2025 IMF World Economic Outlook (WEO).

To understand these results, it helps to begin with a counterfactual without directed technical change. In that case, the energy price shock simply raises the cost of a key production input. Because energy is difficult to substitute (it has a low elasticity of substitution with the capital/labour bundle), output drops sharply. With directed technical change, however, higher energy prices induce firms to shift R&D from capital labour-augmenting technologies toward improving energy efficiency. In the short to medium term, these gains cannot fully offset the higher energy price or the transitional cost of redirecting innovation. Thus, while directed technical change mitigates the shock relative to a scenario without it, potential growth and output still fall compared to a no-shock baseline.

How large was the cushioning effect of increased energy efficiency? To gauge the magnitude, we drastically reduce the responsiveness of energy efficiency to price shocks in the model. Had energy efficiency reacted by 80% less, for example, then (as shown in Figure 4) the euro area’s potential output loss from the energy price shock would have been around two-thirds larger (i.e. there would have been a drop of around 1.3% in potential output by 2026). ¹

Figure 4 Impact of change in energy prices over 2022-27 on potential output in the euro area (percent)

Source: Authors’ calculations.

Italy and Germany were more impacted than Spain and France

These effects are heterogenous across large euro area economies. The drop in potential output by 2027 is estimated to be highest in Italy at around 1.2%, followed by Germany at 0.9%, and then Spain and France at 0.6% and 0.4%, respectively, due to a different energy mix (and hence price shock) and different estimated elasticities of substitution and investment efficiencies (Figure 5). ²

Figure 5 Impact of change in energy prices over 2022-27 on potential output in individual euro area countries (deviation from no shock scenario in 2027)

Sources: Authors’ calculations.

Sharp energy price increases are significantly more costly than gradual price increase

We also compare the impact of the observed price shock to the impact of a counterfactual shock that would have increased fossil fuel prices more smoothly over several years to yield the same energy price endpoint by 2027. We find that the impact on potential output would have been more than four times smaller with the gradual price increase. Large and abrupt price changes, coupled with the transitional costs of reallocating investment from labour/capital saving to energy saving, entail larger deviations from the balanced growth path. This suggests that, over and above alleviating a trend rise in energy prices, there is an output gain to be reaped from containing their lower-frequency fluctuations – particularly from avoiding disruptive spikes that distort the allocation of firms’ R&D expenditures.

Policy conclusions

The European energy crisis illustrates a general principle: shocks that change relative prices can steer innovation and reshape growth paths. Our analysis suggests that this response offered a modest silver lining to the crisis: durable energy efficiency gains cushioned the 2022 shock and provided some protection against possible future fossil fuel price shocks.

At the same time, we find that abrupt price spikes are particularly harmful. But rather than policies that subsidise retail energy prices persistently, market reforms that lower wholesale price levels and volatility – such as further integration of the European energy system – should be prioritised (Kammer 2025, D’Arcangelo et al. forthcoming). Another way to put this is to say that energy price volatility should be lowered by removing economic distortions (e.g. the artificial fragmentation of European energy markets along national lines) and not by introducing new economic distortions (e.g. price subsidies).

The productivity trade-off at the heart of our model also calls for a focus on labour/capital productivity in Europe, since the diversion of R&D resources from labour/capital-augmenting technology accentuates Europe’s longstanding productivity challenges. Ambitious policy action is needed to boost Europe’s productivity. This includes national structural reforms (Budina et al. 2025), deepening the EU single market (Arnold et al. 2025) and joint provision of key public goods at the EU level (Busse et al. 2025). Additional resources for R&D, which would loosen the trade-off between labour/capital and energy innovation and reduce the spending gap vis-à-vis the US, will also be critical.

Source: cepr.org