Recent debates on the cost-of-living crisis have rightly focused on differences in households’ exposure to rising prices. Existing work shows that low-income households spend a larger share of their budgets on necessities such as energy and food, and therefore tend to be more exposed to cost-of-living shocks (Pizer and Sexton 2019, Soldani et al. 2023, Menyhért 2022).

But exposure is only part of the distributional story. A cost-of-living shock combines a negative real-income shock with a shift in relative prices, so households must cope with lower purchasing power while also substituting away from goods whose prices have risen. The distributional consequences of a price shock thus depend on households’ capacity to adjust: some can shift consumption away from the more expensive good, draw on savings, or smooth the shock through additional labour income, while others have more limited opportunities to respond.

However, we know surprisingly little about those responses. Pinning them down requires household-level data on multiple adjustment margins, a large and plausibly exogenous price shock, and variation that exposes otherwise similar households to different price shocks.

A natural experiment from Finland’s energy crisis

In February 2022, Russia invaded Ukraine, triggering the European energy crisis. The crisis provides an unusually clean setting for overcoming these empirical obstacles (Ahlvik et al. 2026). First, the shock itself was large and unexpected. As Figure 1 shows, electricity prices increased by up to eightfold.

Figure 1 Prices of variable-price and 2-year fixed-price contracts for electricity in Finland

Second, rich administrative microdata from Finland allows us to directly observe household electricity use and to link it to earnings, benefits, and court-recorded payment defaults, making it possible to trace adjustment along several margins. Third, it was common for households to sign long-term contracts that fixed the electricity price for long periods. Predetermined differences in fixed-term contract expirations create quasi-experimental variation in price exposure among households facing the same broader economic environment. This setting allows us to compare otherwise similar households that differed only in whether they were exposed to the price shock.

The effects of the energy price shock differed sharply across income groups. Households adjusted along all four margins: electricity consumption, labour earnings, payment defaults, and other consumption and savings (Figure 2).

Figure 2 Household-level responses to expiration of electricity contracts

Three results stand out.

• Responses were heterogeneous. Figure 3 illustrates this heterogeneity by income groups. Higher-income households responded mainly by reducing electricity use, while lower-income households reduced it much less and were more likely to adjust through other margins instead. In particular, low-income households were more likely to increase labour earnings, accumulate payment defaults, and reduce other spending and savings. Among lower-income households, labour earnings per worker rose modestly, but there was no detectable increase in labour-force participation. This suggests that the earnings response came mainly from workers already in employment increasing their hours or effort, rather than from non-workers entering jobs.

Figure 3 Heterogeneous responses to the energy crisis, by income group

• Anticipation helped, but only partially. Those whose contracts expired later in the crisis began reducing electricity use before their contracts expired. On average, electricity use fell by 9%, and roughly one-quarter of the total reduction took place during the crisis, but in the months before contract expiration. This effect was larger for high-income households. Longer anticipation periods made adjustment easier, but did not fully eliminate the financial strain among more vulnerable households.
• Limited liquidity may explain the responses. Some middle-income households also had a rise in payment defaults, despite being less exposed than the lowest-income households. Especially those middle-income households that had a high debt-to-income ratio saw payment defaults rise, which is consistent with the idea of a liquidity-constrained middle class. Limited liquidity may partly explain the weaker response in electricity use among households at the bottom of the income distribution, as they may have had fewer opportunities to make energy-saving investments during the crisis than higher-income households.

What the results mean for relief policies

As governments responded to the energy crisis with packages combining lump-sum support and price-reducing measures, our results help clarify how such relief should be designed (e.g. Varga et al. 2022). The distributional and efficiency effects of subsidies, energy tax cuts, or targeted transfers depend on which households consume the subsidised goods and how they adjust to price changes. Classic public finance results imply that if adjustments to prices are similar across the income distribution, redistribution can be achieved through the tax-and-transfer system. If responses differ, commodity-specific interventions, such as targeted transfers and price caps, may be warranted. Our setting informs this distinction by estimating income-dependent adjustment responses.

Who bears the brunt of energy price shocks depends not only on who is most exposed, but also on who can adapt. Low-income households are doubly vulnerable, as they spend a larger share of their budgets on energy, but they are also less able to reduce electricity use. While labour income rose for those at work, we find no entry into the labour force among those whose income consists mainly of government benefits or pensions. This points to targeted relief rather than broad-based subsidies.

The results also suggest that compensating low-income households through subsidised prices need not come at a large efficiency cost. In our setting, these households reduced electricity use relatively little when prices rose, suggesting that the efficiency costs of targeted price support for low-income households may be smaller than is often feared.

References

Ahlvik, L, T Kaariaho, M Liski, and I Vehviläinen (2026), “Household-level responses to the European energy crisis”, American Economic Review: Insights, forthcoming.

Menyhért, B (2022), “The uneven effects of rising energy and consumer prices on poverty and social exclusion in the EU”, VoxEU.org, 13 December.

Pizer, W A, and S Sexton (2019), “The distributional impacts of energy taxes”, Review of Environmental Economics and Policy 13(1).

Soldani, E, O Causa, and N Luu (2023), “The cost-of-living squeeze: Distributional implications of rising inflation”, VoxEU.org, 24 April.

Varga, J, R Kasdorp, Å Johannesson Lindén, B Döhring, A Cima, and G Bethuyne (2022), “Targeted income support is the most social and climate-friendly measure for mitigating the impact of high energy prices”, VoxEU.org, 6 June.