Are Energy Bills Squeezing People’s Spending?

At the beginning of 2021, energy prices rose on the heels of a combination of supply and demand factors.¹ This structural increase was compounded by the Russian invasion of Ukraine on 24 February 2022, which resulted in an abrupt spike in energy prices on the global markets similar to that observed during the oil shocks of the 1970s.

Like other countries, Italy was hit by the increase in global energy prices, against a backdrop of this country’s heavy reliance on energy imports, which makes it especially vulnerable to energy price shocks. In fact, almost half of the electricity in 2021 and 60 per cent of all space heating in Italy were produced by burning natural gas, which, in the final months of 2021, was mostly imported (and mainly from Russia).² In 2022, gas and electricity prices on the Italian markets skyrocketed, reflecting uncertainties about the continuity and security of the supply stream due to the increasing Russian threats to Ukraine in late 2021 and the ensuing all-out war.³ The energy shocks have had both direct and indirect effects on inflation: in the fourth quarter of 2022, the contribution to headline inflation was in the order of 60 per cent, and the contribution to core inflation between 20 and 50 per cent (Neri et al., 2023).

The abrupt price increases struck households unevenly, with those that spend a greater share of their budget on items that were more heavily impacted by the surging prices or with a higher average propensity to consume suffering the most (Curci et al., 2022; Faiella and Lavecchia, 2022; Guan et al., 2023). By eating away at households’ purchasing power, the price hikes ultimately reduce disposable income, hampering households’ ability to repay their debts. Depending on the amount of strain placed upon them, families could become more financially vulnerable and risks to financial stability could materialize.

This paper aims to analyse to what extent energy price hikes affect the financial vulnerability of Italian households. To this end, understanding how household energy markets work in Italy and which data are available to use is key to laying the ground for the analysis.

Retail gas and electricity markets both have a two-tier structure. In each segment, the price of the energy component is determined according to specific rules, set separately for each market.⁴ In greater detail, on the regulated market (Mercato di Maggior Tutela), the price of the energy component is set by the Italian Regulatory Authority for Energy, Networks and the Environment (Autorità di Regolazione per Energia Reti e Ambiente - ARERA) and updated quarterly.⁵ On the free market (Mercato Libero), the energy price is determined by market rules. In this market, contracts can have different price-setting mechanisms and lengths.

Given this complex structure, determining the impact of the inflation shock on households is not straightforward as we lack granular information, including on contractual length and the clauses for the price of the energy component paid by each household. Therefore, we have to rely on alternative data sources to fill in the void and use economy-wide energy price inflation measures. We apply the price increase indistinctly to all households; the exercise, therefore, establishes by construction an upper bound for their financial vulnerability. The results can also be interpreted as stemming from a stress test.

Bearing this in mind, we first determine price elasticities of electricity and heating for Italian households; this is necessary to understand how households cope with rising energy prices and, consequently, with a possible reduction in disposable income. To do so, we apply an updated version of the methodology proposed by Faiella and Lavecchia (2021b) to estimate the energy demand and price elasticities for Italian households. We use a quasi-panel approach as in Deaton (1985) to estimate price elasticities for 32 different household groups or strata using the Italian Household Budget Survey (HBS).

Therefore, when an electricity or heating price shock hits the economy, all the households adjust their consumption bundle accordingly. The adjustment is, however, stratum-specific (although we apply a stochastic adjustment at the household level) and reflects households’ characteristics and consumption preferences on the item shocked, which prevail at each level.

To measure the effects of the energy price shocks on households’ financial vulnerability, we first combine the consumption adjustment calculated above with microdata from the Bank of Italy’s 2020 Survey on Household Income and Wealth (SHIW) using the 32 strata as merging keys. This new dataset allows us to calculate an adjusted measure of households’ disposable income which reflects changes in consumption at the group level, induced by the energy price shocks. We hold that household disposable income diminishes proportionately to the increase in energy expenditure, if any. Households’ ability to repay their debt is affected by the reduced disposable income.

To dig further into the issue of financial vulnerability of households in Italy, we exploit to their full extent the microsimulation models developed at the Bank of Italy (Michelangeli and Pietrunti, 2014; Attinà et al., 2020; Faiella et al., 2022). Within these models, macroeconomic variables evolve reflecting macroeconomic data and projections.

In particular, in addition to the evolution of household income and debt, we extend the existing models by adding consumption dynamics.⁶

We run the simulations to check for financial vulnerability over the period 2020-22 and introduce the energy-induced shocks (via the adjusted income) to take into account the 2021-22 surge in energy-related inflation.

We compare the evolution of households’ vulnerability both in the absence and in the presence of behavioural responses/consumption adjustments,⁷ to respond to energy inflation shocks. In addition to a baseline case where there are no energy-induced price shocks, we apply such shocks under three different price scenarios, which refer to different price index evolution. In particular, we apply the energy price increase observed between 2020 and 2022 to all of the households recorded in: a) the regulated market tariffs (Mercato di Maggior Tutela); b) the electricity and natural gas components of the NIC Index (Istat’s consumer price index for the entire national community); and c) the weighted average unit cost of energy from Eurostat.

The change in the final price of electricity (natural gas) between the average in 2020 and the average in 2022 was 172 (92) per cent according to the first source (regulated market tariffs), 142 (109) per cent according to the second source (NIC index), and 46 (47) per cent according to Eurostat. In absolute terms, this equates to an increase between 0.11 (Eurostat) and 0.34 (NIC) €/kWh for electricity and between 10.9 (Eurostat) and 23.4 (NIC) €/Gj for natural gas, with regulated market tariffs being in-between.

If households do not re-adjust their consumption choices after the energy price upsurge (i.e. assuming perfect inelastic demand or quantity invariance), energy expenditure rises proportionately. This leads to a proportionate decrease in disposable income. At the same time, household financial vulnerability has barely changed in 2021, reflecting the contained growth in energy expenses. In 2022, instead, debt at risk increases across the three scenarios considered (to between 8.15 and 8.62 per cent, from 7.9 in the baseline), with the stronger increase occurring when the energy price upsurge is measured by the NIC Index. The contained increase in 2021 reflects the fact that energy price variations are subdued in the years 2020-2021, while they are more marked in 2021-2022. While at first sight, such increases may appear to be low, they are not: in fact, they mark up between 3 and 11 per cent increase from the baseline in 2022. This non-monotonic pattern in vulnerability, however, does not correlate well with actual financial evidence, according to which the new non-performing loan rate has not been spiking in the wake of the energy price shock and, even more important, it has decreased between 2021 and 2022 (Bank of Italy, 2023). To properly assess the evolution of financial vulnerability, it is therefore crucial to take into account households’ reactions and consumption rebalancing. We, therefore, extend our setup to include energy price elasticities.

In this setting, facing price increases, households decide to re-adjust their consumption so that disposable income drops only to a limited extent. Hence, financial vulnerability is barely affected. Throughout all the scenarios under consideration, the share of vulnerable households and the debt at risk appear to be very similar to those prevailing in the baseline scenario. This confirms that vulnerable households adjust their consumption expenditure in such a way as to maintain the level of disposable income as stable as possible so that it is also available for debt repayment.⁸ Looking at the heterogeneity results, we find very minor differences across scenarios. We can thus conclude that households’ vulnerability continues to be driven by macroeconomic variables other than energy price variations which in contrast have a limited impact. On the other hand, the muted impact could come at the cost of consumption reduction, thus lowering household welfare.

This paper innovates the existing literature along three dimensions. To the best of our knowledge, it is the first to take on board the effects of energy price increases via the correction in disposable income obtained by considering consumption elasticities in the analyses of households’ financial vulnerability. To do so, we extend the existing microsimulation models of financial vulnerability (Michelangeli and Pietrunti, 2014; Attinà et al., 2020; Faiella et al., 2022) to account for consumption dynamics. In other exercises, aimed at evaluating the effects of the recent price upsurge on Italian households along with the multi-pronged measures implemented by the Italian government to combat it, elasticities are set to be equal to zero, and no financial vulnerability analysis is performed (Curci et al., 2022; UPB, 2022; Istat, 2022).

Secondly, the paper evaluates the effects of energy price increases on households’ financial vulnerability, intertwining it with the evolution of macroeconomic variables, including mortgages and consumer credit to households.

While the topic of the analysis of households’ financial vulnerability partly aligns this work with that by Faiella et al. (2022), our setting allows us to have a richer interaction between the variables and relax one of the hypotheses assumed therein.

Thirdly, the paper merges several datasets. Such data stem from an updated series for energy-related products (for the period 1997 – 2022) that makes it possible to calculate energy-demand elasticities based on observations that refer to both pre-pandemic and pandemic conditions; at the same time, by using the 2020 wave of the SHIW, we can include the households’ debt variables available in the Italian survey, which also reflect the impact of the COVID-19 pandemic.

Finally, we would like to point out some caveats on the results obtained. First and foremost, the exercise is based on the assumption of partial equilibrium and does not include spillover effects; second, the time horizon we focus on is short, though this appears to be less of a concern at the current stage since the recent bout of energy-price inflation seems to be already dissipating. At the same time, in light of the magnitude of the price shock, structural breaks in elasticity can not be ruled out.

The structure of the paper is as follows. Section 2 offers an overview of the existing literature on the topics. Section 3 describes our data and methodologies for simulating the energy shock; we also provide a discussion of estimated energy price elasticities. Section 4 shows the set of models and data used to assess how the energy shock affects households’ financial vulnerability. The results of the study are shown in Section 5, which also includes some heterogeneity analyses. Section 6 concludes the paper and sets the future research agenda.

This paper merges two different strands of literature.

The first is related to energy price elasticity; the second refers to microsimulation models. The latter strand can ideally be divided into three subset of papers: a) those dealing with methods to impute expenditures into an income data set; b) those connected to microsimulation models of households’ financial vulnerability; c) those, more recent, related to the uneven impact of rising energy prices on households’ standards of living.

Studies on the determinants of energy demand, especially electricity, are abundant both in temporal and spatial dimensions. From the first perspective, they date back several decades ago, kicking off with the seminal work of Houthakker (1951). Under the second viewpoint, there are, literally, hundreds of estimates of price elasticity, although most of them are based on US data. Espey and Espey (2004) report a meta-analysis of 36 papers, with more than 123 short-run and 96 long-run price elasticity estimates of residential electricity demand; in the short run, price elasticity ranges between -2.01 and -0.004 (mean: -0.35), while in the long run, elasticity ranges between -2.25 and -0.04 (mean: -0.85). Labandeira et al. (2017) carry out a meta-analysis for a dozen surveys on energy demand, again with mixed results, depending on the fuel considered and the time frame; as for electricity, elasticity is equal, on average, to -0.126 in the short run (-0.365 in the long run), while for natural gas it is equal to -0.180 in the short run (-0.684 in the long run). However, most of these studies are based, to different extents, on data from the United States, whose power sector is significantly different from Italy’s, along several dimensions, including in terms of market structure, prices and energy demand. In particular, the average American household consumes four times as much the amount of electricity as an average Italian household,⁹ with a significant contribution of electricity for space heating (one-third of all homes in the US appears to be fed by electricity for heating purposes, compared with an almost non-existing share in Italy).¹⁰

Estimates for energy demand price elasticity in the Italian case are just a few: Faiella (2011) finds that the effect of prices on the energy shares is negative for heating, while for electricity the effect is negative for the 1997-2004 period and positive for the 2005-2007 subsample. Bigerna (2012) observes that the price elasticity for electricity depends on the time of the day (due to the tariffs system in place up to 2016, which encourages off-peak use) and on the geographical zones, ranging between -0.03 and -0.10. Bardazzi and Pazienza (2019) observe that, concerning the age of the head of the households, electricity demand is hump-shaped, reaching a peak when the head of a household is 50 years old, while natural gas demand keeps increasing with age, as natural gas is the main component for heating in Italy, and the time spent at home increases. They also show that elasticities for electricity and natural gas (at the national level equal to -0.705 and -0.621, respectively) are higher in the Centre and the South. Finally, Favero and Grossi (2023) analyzes a sample of bills from Veneto, finding that the price elasticity for natural gas demand from residential customers is rather rigid (around - 0.5).

This paper also relates to the broad microsimulation literature. In this literature, the widespread methods to impute expenditures into an income data set are distinguished into two main categories: the so-called “explicit method”, based on the estimation of Engel curves, and the so-called “implicit method”, which uses statistical matching techniques. Decoster et al. (2020) cite the extensive literature on the topic and compares five different techniques.

In terms of financial vulnerability, our benchmark models are those developed by Attinà et al. (2020) and Michelangeli and Pietrunti (2014) at the Bank of Italy. Such models simulate the evolution of financial vulnerability of Italian households starting from SHIW data reconciled with macro-variables.

As for the impact of rising prices, including via carbon taxes, on households’ income and the uneven impact on their standards of living, there has been a blossoming of literature in recent years, also in relation to the recent inflation spike. We have no aim to cover it all and we will limit ourselves to papers more strictly related to the topic.

Faiella and Lavecchia (2021b) is, to the best of our knowledge, the only paper in Italy to put forward a methodology to estimate the demand and elasticity of energy-related items and to analyse the effects on consumption of a carbon tax. The authors combine the microdata of the Italian HBS with several external sources. They use a quasi-panel approach and estimate the demand elasticity via an autoregressive distributed lag (ARDL) model. They run several regressions applying many different estimation techniques and estimate the price elasticity for electricity and heating for 36 groups of Italian households in the short run (-0.44 and -0.54 on average for electricity and heating, respectively). In general, their results underscore that households’ price elasticity is greater in the long run and for transport fuels and electricity.

Faiella et al. (2022) apply the methodology developed in Faiella and Lavecchia (2021b) to determine the extent to which different levels of carbon taxes affect the financial vulnerability of households and firms. While taking into account the effects of energy price increases on the consumption bundle of households (and on the energy expenditure of firms), they run a static exercise and find that, within reasonable amounts of a carbon tax, the vulnerability of households and firms is only barely hit by energy-related shocks.

Curci et al. (2022) evaluate the sharp rise in prices registered in 2021-22 on the purchasing power of different types of Italian households. The authors maintain that household consumption (in terms of quantities) remains unchanged (i.e., they assume setting price elasticity equal to zero) in the face of the 2021-22 inflation bout. They use survey microdata on household expenditures, along with the tax and benefit microsimulation model of the Bank of Italy (BIMic), to gauge the impact of the inflationary shock on the distribution of households’ purchasing powers, while allowing government measures to kick in. They find that the government’s measures mitigated the distributional impact of the inflationary shock. Such measures curtailed inflation on average by slightly less than 2 percentage points, with a relatively more significant reduction for low-income households.

Similar results are obtained by UPB (2022). They use a microsimulation model fed by Istat’s HBS data, also taking into account fiscal and social contribution data. Like in Curci et al. (2022) price shocks - referring to the period between June 2021 and September 2022 - hit households’ consumption bundles but do not have an impact on quantities (i.e., price elasticity is again set equal to 0). According to the simulations, the actual increase in households’ expenses owing to inflation was contained thanks to the government measures. Such measures were also put in place to fight the regressive impact of the inflation shock; households belonging to the first income decile in terms of expenditure increase bear a much less burden than the average family. Also, it resulted in a smaller increase in energy poverty (OIPE, 2023).

Causa et al. (2022) provide a quantification of the impact of rising prices on households’ welfare over the past years for ten OECD countries. Drawing on micro-based HBS and CPI data, their paper tries to identify which households are more exposed and vulnerable to the recent rise in inflation, with a focus on energy and, to a lesser extent, food price inflation. By using it as a measure of purchasing power resulting from changes in consumer prices underlying inflation, the authors find that the decline in households’ purchasing power between August 2021 and August 2022 was driven by energy prices and is particularly relevant in Italy, among other countries. The plummeting of purchasing power has led to heterogeneity across countries and partly reflects differences in the rate of inflation, its diffusion across consumption items and the spending structure of the average household. Causa et al. (2022) also explore differences across household income groups and other relevant dimensions and find that the households most exposed to rising energy prices are low-income, senior and, in some countries, rural.

To assess the impact stemming from rising prices and interest rates on household debt affordability the Bank of England (2022) has produced a new cost of living adjusted debt-servicing ratio measure. The new measure adjusts income for taxes and an estimate of essential spending, which includes utility and council tax bills, housing maintenance, food and non-alcoholic beverages, motor fuels, vehicle maintenance, public transport and communication. However, consumption in the above-mentioned areas increases one-to-one with prices. The authors estimate that in 2022 the share of households with the high cost of living adjusted debt service ratios (DSRs) on either their mortgage or consumer credit remained significantly below pre-global financial crisis peaks and they were not projected to increase substantially in 2023. This reflects government support measures relieving some of the pressure on household finances, particularly from the rise in living costs, in the near term.

So far household vulnerability has been evaluated in the face of aggregate shocks to income (i.e. a recession), interest rates (i.e. contractionary monetary policy) or a carbon tax. In what follows we will bring together the different above-mentioned bodies of research to take into account the inflation impact (triggered by energy price shocks) on the households’ consumption bundle and the ensuing consumer expenditure adjustment. This, in turn, has a bearing on household income and hence on financial vulnerability. To carry out this exercise, we modify household disposable income in a heterogeneous way across households exploiting energy price elasticities. To this end, we analyse how energy prices impact Italian households’ consumption choices.

This section provides an overview of energy price data, the model used to estimate energy price elasticities and related results. It finally concludes with a discussion of the estimated values.

3.2. Estimation of energy price elasticities

To estimate the demand price elasticity, considering the absence of microdata on energy quantities, we adopt the approach developed by Faiella and Lavecchia (2021b). They utilized microdata on households’ energy expenditure from the Italian HBS) to construct a microsimulation model of households’ energy demand.¹⁸ Unfortunately, a centralized archive of electricity and natural gas consumptions from all Italian households and firms, known as the Sistema informativo integrato (SII) managed by Acquirente Unico, is currently unavailable. Additionally, although a survey on households’ energy consumption, including wood, pellet, and LNG, was conducted in 2021, the microdata are not yet accessible.

Faiella and Lavecchia (2021b) estimate the short- and long-run price elasticities of energy demand for electricity, heating and private transportation fuels for population subgroups of homogeneous households (strata).¹⁹ They use the HBS, integrated with other (aggregated) sources of information on energy prices, and calibrate the estimated quantities through official aggregated data.²⁰ As the HBS is not a panel, the authors use a quasi-panel approach (Deaton, 1985),²¹ and estimate the demand elasticity for each subgroup exploiting the change over time of energy prices and demand fitting the following Auto-Regressive Distributed Lag (ARDL) model (see Greene (2008) for further details):

(1) ${\displaystyle log{Q}_{s,t}^z={\lambda }_{s}log{Q}_{s,t-1}^z+{\beta }_{s}log{P}_t^z+{\gamma }_{s}log{E}_{s,t}+w+s+t+t^2+{ϵ}_{s,t}}$

where $Q_{s,t}^z$ is the average amount of fuel $z$, being electricity (in kWh) or natural gas (in $m^3$), consumed by each stratum $s$ in the month $t$; $P_t^z$ is the average price of fuel $z$ at time $t$; $E_{s,t}$ is the total expenditure of stratum $s$ at time $t$; $w$ and $s$ are seasonal dummies; $t$ and $t^2$ are time dummies; and ${\beta }_s$ is the coefficient of interest, the (short-run) price elasticity (see Table 2 for a description of the main variables used in the paper).

Table 2

Variable	Description	Source
	(a) Model to estimate price elasticities
$Q_{s,t}^z$	Electricity and heating quantity	HBS based on Faiella and Lavecchia (2021b)
$P_t^z$	Final retail price for electricity and natural gas	Eurostat, Istat and ARERA (own elaboration; see also Table 1)
$E_{s,t}$	Total expenditure of households	Own elaboration based on HBS
${\beta }_s$	Short run price elasticities for electricity and heating demand	Own elaboration based on HBS and (Faiella and Lavecchia, 2021b)
	(b) Model of households’ financial vulnerability
$y_{i,t}$	Income	National accounts and Bank of Italy macro model
$M_{q,i,t}$	Mortgage	Bank of Italy’s statistical data warehouse and macroeconomic model
$c_{i,t}$	Consumer credit	Bank of Italy’s Statistical data warehouse and macro model
$r_{q,i,t}$	Mortgage interest rate	10-year IRS (for fixed-rate)Euribor (for variable rate)

In our setting, we applied the following changes with respect to Faiella and Lavecchia (2021b): 1) we updated all data up to 2022; 2) we consider eight, instead of nine, groups of households (see Table 3 for a description of each stratum or subgroup); 3) we used Eurostat average unit prices to obtain an estimate of energy quantities.²²

Table 3

Stratum ID	Households’ type
x02	Single person under the age of 35 or aged 35-64
x03	Single person aged 65 and over
x05	Childless couple with contact person under the age of 35 years old or aged 35-64
x06	Childless couple with contact person aged 65 and over
x07	Couple with 1 child
x08	Couple with 2 or more children
x10	Single parent
x11	Other types

1. where x=1,2,3 or 4 are the quartiles of the total equivalised expenditure distribution.

Overall, information is available on 32 subgroups observed monthly between 1997 and 2022, for a total of almost 10,000 observations. We build on the least squares estimates of the short-run elasticity, available at the stratum level, which will be fed into a model of households’ financial vulnerability in Section 4.

3.2.1. Results on price elasticities

The estimates of the elasticities are reported, joined with their standard error, in Table 4 (see also Figures 1 and 2). On average, households’ price elasticity is smaller for electricity than for heating; for a 1 per cent increase in price, electricity demand decreases, on average, by 0.2 per cent (0.9 for heating). Moreover, the elasticity of electricity demand decreases as households become richer. This is not surprising, as electricity is a merit good, with a demand that is more rigid compared to space heating which, in turn, is easier to modify/adjust according to budget constraints.

Table 4

	Electricity		Heating
Strata	$\widehat{{\beta }_s}$	$\widehat{{\sigma }_{\beta }}$	$\widehat{{\gamma }_s}$	$\widehat{{\sigma }_{\gamma }}$
Single person under the age of 35 or aged 35-64, 1st quarter	-0.44	0.15	-1.04	0.18
Single person aged 65 and over, 1st quarter	-0.33	0.10	-0.76	0.13
Childless couple, < 35 years old or aged 35-64, 1st quarter	-0.67	0.16	-1.04	0.17
Childless couple, aged 65 and over, 1st quarter	-0.37	0.09	-0.90	0.13
Couple with 1 child, 1st quarter	-0.31	0.11	-0.69	0.12
Couple with 2 or more children, 1st quarter	-0.40	0.09	-0.70	0.11
Single parent, 1st quarter	-0.31	0.11	-0.98	0.12
Other types, 1st quarter	-0.23	0.13	-1.00	0.14
Single person under the age of 35 or aged 35-64, 2nd quarter	-0.24	0.12	-1.05	0.17
Single person aged 65 and over, 2nd quarter	-0.18	0.09	-0.87	0.13
Childless couple, < 35 years old or aged 35-64, 2nd quarter	-0.34	0.12	-0.84	0.15
Childless couple, aged 65 and over, 2nd quarter	-0.14	0.09	-0.83	0.14
Couple with 1 child, 2nd quarter	-0.16	0.09	-0.60	0.12
Couple with 2 or more children, 2nd quarter	-0.16	0.09	-0.60	0.11
Single parent, 2nd quarter	-0.13	0.11	-1.05	0.15
Other types, 2nd quarter	-0.29	0.16	-1.18	0.17
Single person under the age of 35 or aged 35-64, 3rd quarter	-0.16	0.11	-0.79	0.15
Single person aged 65 and over, 3rd quarter	-0.20	0.11	-0.91	0.12
Childless couple, < 35 years old or aged 35-64, 3rd quarter	-0.02	0.13	-1.03	0.14
Childless couple, aged 65 and over, 3rd quarter	-0.05	0.13	-0.98	0.15
Couple with 1 child, 3rd quarter	0.01	0.10	-0.79	0.17
Couple with 2 or more children, 3rd quarter	-0.04	0.09	-0.66	0.12
Single parent, 3rd quarter	-0.07	0.18	-0.95	0.16
Other types, 3rd quarter	0.05	0.20	-1.07	0.19
Single person under the age of 35 or aged 35-64, 4th quarter	-0.11	0.10	-0.77	0.11
Single person aged 65 and over, 4th quarter	-0.07	0.12	-0.92	0.14
Childless couple, < 35 years old or aged 35-64, 4th quarter	-0.07	0.14	-0.54	0.12
Childless couple, aged 65 and over, 4th quarter	-0.03	0.11	-1.01	0.15
Couple with 1 child, 4th quarter	-0.14	0.11	-0.62	0.13
Couple with 2 or more children, 4th quarter	-0.15	0.13	-0.84	0.16
Single parent, 4th quarter	-0.24	0.18	-1.20	0.16
Other types, 4th quarter	-0.11	0.21	-1.29	0.22

Figure 1

Download asset Open asset

Figure 2

Download asset Open asset

Higher-income households demonstrate greater heating elasticity as they usually have alternative heating systems (e.g., heating pumps) and/or better insulation.²³ As a result, they have the means to react more to an increase in heating prices. This includes switching fuels or reducing (already high) thermal comfort. Additionally, poor singles and single parents are very responsive to increases in heating prices.

3.2.2. Discussion on price elasticities

As highlighted in Section 2, there is significant heterogeneity across estimates of price elasticities, depending on the energy vector, type of clients (residential vs. industrial) or time frame (short vs. long run) at stake. Moreover, as already mentioned before, most of the existing literature is based on analyses referred to the US, which, for several reasons, are not very comparable to Italy - where, on the contrary, evidence is scant - from a point of view of retail energy markets. This implies that it is very difficult to gauge whether our elasticity estimates are “too high” or “too low”.

Peersman and Wauters (2023), working on Belgium data, find that the energy price elasticity is significantly higher for price increases rather than price decreases, while it diminishes heavily for greater price hikes like the ones recently experienced by Italian households.

Additionally, the energy price variations recorded in the period, and especially in 2020-22, are so large that they might have induced a structural break, i.e. households’ responses might have become more elastic. Unfortunately, the microdata for 2023 are not yet available but only aggregate demand which reports a reduction of households’ electricity and natural gas demand by, respectively, 3.7 and 13.5 per cent with respect to 2021.

To cope with the lack of information on energy prices at the household level, we used the available sources discussed in Section 3.1 and reported in Table 1. However, as previously pointed out, millions of households in Italy were provided electricity and natural gas at fixed prices when the price surge started (end of Q2-2021). While many of these contracts have expired over time and the supply of new fixed contracts almost disappeared over 2022, it is also true that for months, some households paid less for their electricity and natural gas in 2022 than in 2020 because of the government’s interventions.

All these assumptions do not come for free. Given the price variations and our estimates of (subgroups) price elasticities, we estimate that between 2021 and 2022 overall electricity demand would have fallen between 8 (using Eurostat prices) and 14.7 (NIC) percent (w.r. to an effective 3.7 percent decrease) while the estimated heating demand between 34.8 (Eurostat) and 53.2 (NIC) percent (w.r. to an actual cut of 13.5 percent).

While a fall in demand, following a price increase is prima facie evidence of a behavioural response (and, therefore, works assuming fixed demand should be reconsidered), the differences cited above in decreased demand are striking. Only when the fully-fledged official microdata are available this puzzle will possibly be solved. Nevertheless, as the goal of this paper is to run a stress test exercise, our estimates are very good for capturing the worst-case scenario.

In this section, we build the modelling structure of our exercise. A household is defined as financially vulnerable if loan instalments to income exceeds 30 per cent and its income is below the median of the population.

Comparing vulnerable households with the other indebted ones indicates that the former have slimmer net wealth, much thinner financial assets, much lower income in the face of substantially higher mortgage amount. This translates in higher annual instalment as well as debt-to-income ratio for vulnerable households (Table 5). In contrast, no significant differences emerge in term of either household size or median age of the main income earner. Consumption expenditure, which includes food consumption inside and outside the home, bills, all other consumption, and travel expenses, is lower for vulnerable households with respect to the other ones.

Financially vulnerable households are more likely to be late in their loan repayments by more than 90 days, which is the first stage of non-performing loans (Michelangeli and Rampazzi, 2016). This indicator is also highly correlated with the rate of new non-performing household loans based on Italy’s Central Credit Register data.^{24, 25} For these reasons, vulnerable households should be closely monitored to gain some insight into the threats potentially posed to the stability of the financial sector.

As for households, vulnerability from a financial stability standpoint does not necessarily mean default but refers to possible difficulties in meeting financial obligations when a negative shock occurs. Many vulnerable households are solvable if economic conditions do not change, but they could move from a state of vulnerability to one of default if their ability to repay their debts is hindered by a negative shock. We thus aim to identify indebted agents who could be problematic for their selves (for instance, they could lose their homes) as well as for the liquidity and solvability of financial intermediaries.

To analyse the problem at stake in a dynamic way, we need to make projections for several variables, namely mortgage and consumer credit instalments and debt, income and consumption, the latter variable being crucial for assessing the impact of a variation in energy prices.

This section presents the model results. We start by showing the baseline scenario, where there are no energy price shocks. We then take on board such shocks by applying the energy price changes. We repeat this exercise in two cases. In the first one, households do not adjust their consumption bundle to price increases and keep consuming the same quantities of energy goods despite the price shock, i.e. we assume no price elasticity; in the second case, households change their consumption habits to fight back against price hikes, i.e. we use previously estimated price elasticities per stratum. Given the lack of more granular information on prices, we assume that in each scenario all households face the same energy price increases, which is likely not the case in reality. This renders the results of the exercise an upper bound for households’ vulnerability since, in the spirit of a stress test, as we have seen above in Section 3, the majority of consumers have multi-year contracts.

5.1. Baseline

For the baseline scenario, the debt at risk decreases in 2021-22 and records a slight increase in 2023 (Figure 3). Such dynamics reflect the evolution of the macroeconomic drivers as described in Section 3.2: in 2021-22 the sustained recovery in nominal disposable income overcame the evolution of households’ liabilities, driving households’ vulnerability on a downward path.³¹ In 2023 the increased debt at risk would therefore be mainly attributable to the abrupt rise in loan rates following the change in monetary policy stance implemented to fight the inflation spike.

Figure 3

Download asset Open asset

5.2. Case 1 (no elasticity): households not re-adjusting their consumption after the shock

We consider a model where households’ income and debt evolve over time reflecting macroeconomic developments, but with households not readjusting their quantity of energy consumed following the yearly price increases. Therefore, households’ energy expenditure increases after the shock by an amount equal to the price change multiplied by the consumption level. Panel (a) in Figure 4 and Panels (A) and (B) in Table 6 show expenditure and income variations. The dots in Figure 4 represent the average ratio between the total household consumption expenses after and before the energy shock, $d_{i,t,{ϵ}_s}$, while the bars show the average disposable income adjusted to take into account the eventual re-composition of household expenses occurring as a consequence of the energy price shock. Averages are computed across different simulations.

Figure 4

Download asset Open asset

Table 6

	2020	2021	2022
	(1)	(2)	(3)
A. Expenses after /Expenses before
baseline		1.000	1.000
regulated market		1.016	1.049
NIC Index		1.011	1.055
Eurostat		1.003	1.024
B. Disposable income
baseline	26,882	27,884	29,652
regulated market	26,882	27,564	28,630
NIC Index	26,882	27,656	28,493
Eurostat	26,882	27,820	29,149
C. Share of vulnerable households
baseline	1.45	1.62	1.80
regulated market	1.45	1.54	1.82
NIC Index	1.45	1.55	1.85
Eurostat	1.45	1.53	1.74
D. Debt at risk
baseline	9.19	8.18	7.90
regulated market	9.19	8.25	8.62
NIC Index	9.19	8.18	8.75
Eurostat	9.19	8.07	8.15

1. Notes: Panel A shows households’ consumption average change after the energy price shock; Panel B reports disposable income in euros; Panels C and D include the share of vulnerable households and their debt at risk in percentage values.

In the baseline scenario, by construction, nothing changes. In 2021 and 2022 expenses are the same before and after the shock and the ratio $d_{i,t,{ϵ}_s}$ equals 1. In 2021, across the different scenarios that account for the energy price shocks, reductions in disposable income concerning baseline are contained (between 60 and 320 euros). This reflects the limited increase in energy prices and the consequent limited additional expenses to maintain constant the quantity consumed. On the opposite, the increase in energy expenses is rather pronounced in 2022, ranging from 2.4 to 5.5 per cent across scenarios. Households’ disposable income dropped markedly, up to 1,150 euros.

Panel (b) in Figure 4 and Panels (C) and (D) in Table 6 show the impact of energy price changes on households’ financial vulnerability in the baseline scenario and under the three alternative scenarios. The decrease in disposable income, by limiting the available financial resources that can be used to service the debt, hurts financial vulnerability. In 2021, in the three scenarios, financial vulnerability is about the same as the one recorded in the baseline scenarios, reflecting the contained growth in energy expenses. In 2022, instead, debt at risk increases with respect to 2021 across the three scenarios considered. The increase turns out to be stronger when the energy price upsurge is measured by the NIC Index - which recorded the highest increase - but, under any scenarios, we observe the same dynamics. This non-monotonic evolution of the debt at risk is inconsistent with the one for the new rate of non-performing loans, which continued to decrease in 2021 and 2022, highlighting that some important readjustment has taken place in actual data.

5.3. Case 2 (with price elasticity): households re-adjusting their consumption after the shock

We now consider a far more realistic, yet less studied, case where, in the face of an energy price shock, households re-adjust their yearly consumption depending on their price elasticity.

5.3.1. Aggregate results

As shown in Table 1, energy price variations are overall contained between 2020 and 2021 and are practically unchanged when they are computed using the Eurostat price measure. As a consequence, when the latter is employed, almost all households choose to accommodate the price raise with slightly higher expenses, bearing a negligible impact on disposable income (decreasing on average by less than 20 euros; Panel (a) of Figure 5 and Table 7). In the cases of the NIC Index or regulated market, larger price variations are associated with more marked expenses and a deeper, but still moderate, decline in average income (around 100 euros) with respect to the baseline.

Figure 5

Download asset Open asset

Table 7

	2020	2021	2022
	(1)	(2)	(3)
A. Expenses after /Expenses before
baseline		1.000	1.000
regulated market		1.005	1.019
NIC Index		1.004	1.020
Eurostat		1.001	1.008
B. Disposable income
baseline	26,882	27,884	29,652
regulated market	26,882	27,785	29,238
NIC Index	26,882	27,801	29,207
Eurostat	26,882	27,860	29,470
C. Share of vulnerable households
baseline	1.45	1.62	1.80
regulated market	1.45	1.54	1.71
NIC Index	1.45	1.54	1.72
Eurostat	1.45	1.59	1.72
D. Debt at risk
baseline	9.30	8.18	7.90
regulated market	9.30	8.10	8.02
NIC Index	9.30	8.10	8.06
Eurostat	9.30	8.06	7.91

1. Notes: Panel A shows households’ consumption average change after the energy price shock; Panel B reports disposable income in euros; Panels C and D include the share of vulnerable households and their debt at risk in percentage values.

Energy price variations are instead more pronounced over the period 2021-2022. Under any scenarios considered, households increase their average consumption expenditure in the range of 0.8-2.0 per cent.³² Notwithstanding differences in the extent of price hikes across scenarios, price spikes seem to have a linear impact on total expenditure. Hence, it appears that the price-triggered expenditure increase in the NIC and regulated market scenarios - which are more than twice as big as Eurostat’s (see Table 1) - drives a more pronounced consumption correction. The ensuing decrease in disposable income is in the range of 180 to 450 euros.

Household vulnerability does not go up following the rise in energy prices. Throughout any scenarios, the share of vulnerable households and the debt at risk appear to be about the same as that prevailing in the baseline (Panel (b) of Figure 5), suggesting that indebted households, in particular, adjust their consumption patterns in such a way as to keep a level of disposable income as unchanged as possible to be used also for debt repayments. Again, while the energy price increase does not lead to higher financial vulnerability, household welfare could have suffered a blow as a consequence of the cut in energy consumption.

The result on financial vulnerability measured by the debt at risk is very much in line with the evidence from the rate of new non-performing loan ratio based on supervisory reports (Figure 3). The debt at risk is confirmed to be an indicator with similar dynamics to the new non-performing loan ratio and, thus, it can be used to forecast the evolution of vulnerability.

5.3.2. Heterogeneity results

Tables 8 and 9show the results for the models by considering the heterogeneity across the three dimensions, namely geographical areas, number of household components and age classes. Against a backdrop of energy price shocks having a more marked impact in 2022, the higher the inflation shock, the more likely households to re-adjust their consumption habit. As a consequence of the consumption correction, disposable income is hit less. In general, taking on board consumption re-adjustment - which reflects the coming into play of the energy price elasticities - brings about limited effects in financial vulnerability. Concerning any of the heterogeneity dimensions considered, the energy price changes overall have a homogeneous impact.³³

Table 8

	Expenses after/Expenses before			Disposable income
	(1)	(2)	(3)	(4)	(5)	(6)
	Geographical area
	North	Center	South	North	Center	South
	2020
baseline				31,161	28,138	20,094
	2021
baseline	1.000	1.000	1.000	32,341	29,201	20,813
regulated market	1.003	1.004	1.006	32,239	29,107	20,717
NIC Index	1.004	1.004	1.005	32,255	29,123	20,734
Eurostat	1.001	1.001	1.001	32,316	29,178	20,790
	2022
baseline	1.000	1.000	1.000	34,414	31,063	22,097
regulated market	1.017	1.017	1.022	33,991	30,661	21,688
NIC Index	1.017	1.018	1.024	33,966	30,633	21,646
Eurostat	1.007	1.008	1.010	34,226	30,889	21,917
	Number of household components
	1-2	3	4+	1-2	3	4+
	2020
baseline				21,280	33,598	37,186
	2021
baseline	1.000	1.000	1.000	22,080	34,858	38,543
regulated market	1.005	1.005	1.005	22,000	34,738	38,412
NIC Index	1.004	1.004	1.004	22,016	34,754	38,429
Eurostat	1.001	1.001	1.001	22,062	34,828	38,510
	2022
baseline	1.000	1.000	1.000	23,488	37,066	40,969
regulated market	1.018	1.020	1.019	23,146	36,560	40,429
NIC Index	1.018	1.023	1.022	23,140	36,492	40,355
Eurostat	1.008	1.010	1.009	23,348	36,831	40,711
	Age classes
	15-39	40-65	66+	15-39	40-65	66+
	2020
baseline				27,611	30,218	21,803
	2021
baseline	1.000	1.000	1.000	27,970	31,569	22,959
regulated market	1.005	1.005	1.005	27,870	31,461	22,874
NIC Index	1.004	1.004	1.004	27,885	31,477	22,891
Eurostat	1.001	1.001	1.001	27,945	31,542	22,939
	2022
baseline	1.000	1.000	1.000	29,111	33,572	24,749
regulated market	1.018	1.018	1.019	28,703	33,120	24,382
NIC Index	1.020	1.020	1.020	28,666	33,068	24,378
Eurostat	1.008	1.008	1.008	28,925	33,364	24,599

1. Notes: Columns 1-3 show household consumption average change after the energy price shock. Columns 4-6 report disposable income in euros.

Table 9

	Share of vulnerable households			Debt at risk
	(1)	(2)	(3)	(4)	(5)	(6)
	Geographical area
	North	Center	South	North	Center	South
	2020
baseline	1.41	1.45	1.52	8.86	6.51	16.51
	2021
baseline	1.50	1.08	2.12	7.36	4.85	16.30
regulated market	1.30	1.11	2.14	7.36	4.41	16.43
NIC Index	1.30	1.11	2.14	7.36	4.31	16.42
Eurostat	1.43	1.08	2.12	7.34	4.36	16.32
	2022
baseline	1.50	1.38	2.49	6.52	5.25	16.03
regulated market	1.31	1.39	2.48	6.58	5.18	16.60
NIC Index	1.31	1.39	2.48	6.62	5.20	16.66
Eurostat	1.29	1.39	2.53	6.51	5.09	16.32
	Number of household components
	1-2	3	4+	1-2	3	4+
	2020
baseline	0.80	2.13	2.77	8.69	9.87	9.45
	2021
baseline	1.28	1.62	2.63	9.35	7.57	7.36
regulated market	1.11	1.64	2.73	8.99	7.61	7.48
NIC Index	1.11	1.64	2.73	8.99	7.61	7.48
Eurostat	1.21	1.63	2.67	9.01	7.58	7.38
	2022
baseline	1.40	2.16	2.68	9.68	7.37	6.47
regulated market	1.20	2.19	2.80	9.57	7.52	6.77
NIC Index	1.19	2.19	2.83	9.56	7.58	6.84
Eurostat	1.24	2.17	2.73	9.50	7.45	6.60
	Age classes
	15-39	40-65	66+	15-39	40-65	66+
	2020
baseline	2.66	1.95	0.26	12.36	7.70	13.38
	2021
baseline	2.17	2.29	0.56	10.81	7.34	6.36
regulated market	2.17	2.32	0.30	10.82	7.28	5.63
NIC Index	2.17	2.32	0.30	10.82	7.28	5.63
Eurostat	2.17	2.29	0.46	10.81	7.19	6.11
2022
baseline	2.20	2.56	0.73	11.27	7.05	5.62
regulated market	2.51	2.50	0.49	12.42	6.96	5.02
NIC Index	2.51	2.51	0.49	12.46	7.00	5.02
Eurostat	2.30	2.57	0.48	11.64	7.01	5.01

1. Notes: The share of vulnerable households and the debt at risk are in percentage values.

During 2021 retail prices of electricity and natural gas increased significantly owing to a mismatch of supply and demand; such dynamics were exacerbated by the Russian invasion of Ukraine in February 2022. The price surge could possibly jeopardize households’ welfare and financial soundness, especially in countries like Italy, where prices more than doubled in a short time.

Against this backdrop, we have developed a microsimulation model to evaluate the impact of the upsurge in energy prices on households’ financial vulnerability. This paper builds on previous works of Michelangeli and Pietrunti (2014); Attinà et al. (2020); Faiella and Lavecchia (2021b) and Faiella et al. (2022). For the period 2020-2022, we show that by turning a blind eye to consumers’ behavioural responses (i.e. assuming price inelasticity, a common assumption in the literature), households’ financial vulnerability may be significantly overestimated and its dynamics are not aligned with banks’ supervisory report data on the new rate of non-performing household loans. Conversely, by taking energy demand price elasticities into due account along with the evolution of the relevant macro variables, we find that the increase in the number of vulnerable households (and their associated indebtedness) is comparable to the rise recorded in a scenario where energy prices do not change and more in line to what has broadly occurred in reality. This is because households readjust their energy consumption levels (i.e., they trim their use of electricity and heating) to maintain a constant level of disposable income so that it can be used to service their debt.

Some caveats to the exercise results apply, though. First, and foremost, we performed a partial equilibrium analysis focused on energy markets only and the impact of energy price increases on household financial vulnerability; as a consequence, this exercise has to be interpreted as a stress test. Second, we are aware that there are differences in elasticity values in the presence of price rises or falls and that the 2021-22 episode marked an inflation spike not recorded in many years; at the same time, the nearness of the price surge to the current date prevents the availability of newer data; in addition, its short duration suggests that it would be safe to assume an unchanged energy price elasticity before and after the price spike. At the same time, in light of the magnitude of the price shock, structural breaks in elasticity can not be ruled out. Be as it may, in case of higher elasticity, we would converge towards a case with limited effect on households’ financial vulnerability, but heftier repercussions on welfare. Results would be still driven by the evolution of macro variables. Finally, we are focusing on energy expenditure, but other consumption items, such as food items, incurred large price increases. However, assessing a rebalancing in households’ consumption basket would require information on cross-price and income elasticities for any goods. This information is not available at the current stage and we leave it for future research.

We can then conclude that, if the energy price change did not lead to higher financial vulnerability, this is not a free lunch. Households forgo thermal comfort, potentially experiencing an increase in energy poverty (Faiella and Lavecchia, 2021a; OIPE, 2023), to keep within their budget constraints.

1. 1
   Modelling households’ financial vulnerability with consumer credit and mortgage renegotiations

   1. C.A. Attinà
   2. F Franceschi
   3. V Michelangeli

   (2020)

   International Journal of Microsimulation 13:67–91.

   https://doi.org/10.34196/ijm.00213

   • Google Scholar
2. 2
   Number 1 in Financial stability report. Bank of England

   1. Bank of England

   (2022)

   Number 1 in Financial stability report. Bank of England.

   • Google Scholar
3. 3
   Number 1 in Financial stability report

   1. Bank of Italy

   (2023)

   Number 1 in Financial stability report.

   • Google Scholar
4. 4
   Working papers, Universita degli Studi di Firenze

   1. R Bardazzi
   2. MG Pazienza

   (2019)

   Working papers, Universita degli Studi di Firenze.

   • Google Scholar
5. 5
   Electricity Demand Elasticity in Italy

   1. S Bigerna

   (2012)

   Atlantic Economic Journal 40:439–440.

   https://doi.org/10.1007/s11293-012-9331-5

   • Google Scholar
6. 6
   Cefin working papers no 92, UniMORE

   1. A Bonfatti
   2. E Giarda

   (2024)

   Cefin working papers no 92, UniMORE.

   • Google Scholar
7. 7
   Technical Report 1744, OECD

   1. O Causa
   2. E Soldani
   3. L Nhung
   4. C Soriolo

   (2022)

   Technical Report 1744, OECD.

   • Google Scholar
8. 8
   Working Paper 797, Banca d’Italia

   1. A Colabella
   2. E Guglielminetti
   3. C Rondinelli

   (2023)

   Working Paper 797, Banca d’Italia.

   • Google Scholar
9. 9
   Questioni di Economia e Finanza (Occasional Papers) 738, Banca d’Italia

   1. N Curci
   2. M Savegnago
   3. G Zevi
   4. R Zizza

   (2022)

   Questioni di Economia e Finanza (Occasional Papers) 738, Banca d’Italia.

   • Google Scholar
10. 10
    Panel data from time series of cross-sections

    1. A Deaton

    (1985)

    Journal of Econometrics 30:109–126.

    https://doi.org/10.1016/0304-4076(85)90134-4

    • Google Scholar
11. 11
    Comparative analysis of different techniques to impute expenditures into an income data set

    1. A Decoster
    2. J Loughrey
    3. C O’Donoghue
    4. D Verwerft

    (2020)

    International Journal of Microsimulation 13:70–94.

    https://doi.org/10.34196/IJM.00226

    • Google Scholar
12. 12
    Turning on the Lights: A Meta-Analysis of Residential Electricity Demand Elasticities

    1. JA Espey
    2. M Espey

    (2004)

    Journal of Agricultural and Applied Economics 36:1–17.

    https://doi.org/10.1017/S1074070800021866

    • Google Scholar
13. 13
    Working Paper 822, Banca d’Italia

    1. I Faiella

    (2011)

    Working Paper 822, Banca d’Italia.

    • Google Scholar
14. 14
    Energy poverty. How can you fight it, if you can’t measure it?

    1. I Faiella
    2. L Lavecchia

    (2021a)

    Energy and Buildings 233:110692.

    https://doi.org/10.1016/j.enbuild.2020.110692

    • Google Scholar
15. 15
    Questioni di Economia e Finanza (Occasional Papers) 614, Bank of Italy

    1. I Faiella
    2. L Lavecchia

    (2021)

    Questioni di Economia e Finanza (Occasional Papers) 614, Bank of Italy.

    • Google Scholar
16. 16
    Contenimento Dei Prezzi Dell’energia e Spesa Delle Famiglie

    1. I Faiella
    2. L Lavecchia

    (2022)

    Energia 1:36–39.

    • Google Scholar
17. 17
    A climate stress test on the financial vulnerability of Italian households and firms

    1. I Faiella
    2. L Lavecchia
    3. V Michelangeli
    4. A Mistretta

    (2022)

    Journal of Policy Modeling 44:396–417.

    https://doi.org/10.1016/j.jpolmod.2021.11.001

    • Google Scholar
18. 18
    Analysis of individual natural gas consumption and price elasticity: Evidence from billing data in Italy

    1. F Favero
    2. L Grossi

    (2023)

    Energy Economics 118:106484.

    https://doi.org/10.1016/j.eneco.2022.106484

    • Google Scholar
19. 19
    Econometric Analysis - Sixth Edition

    1. WH Greene

    (2008)

    Pearson Education.

    • Google Scholar
20. 20
    Burden of the global energy price crisis on households

    1. Y Guan
    2. J Yan
    3. Y Shan
    4. Y Zhou
    5. Y Hang
    6. R Li
    7. Y Liu
    8. B Liu
    9. Q Nie
    10. B Bruckner
    11. K Feng
    12. K Hubacek

    (2023)

    Nature Energy 8:304–316.

    https://doi.org/10.1038/s41560-023-01209-8

    • Google Scholar
21. 21
    Some Calculations on Electricity Consumption in Great Britain

    1. HS Houthakker

    (1951)

    Journal of the Royal Statistical Society. Series A (General) 114:359–371.

    https://doi.org/10.2307/2980781

    • Google Scholar
22. 22
    La Redistribuzione Del Reddito in Italia

    1. Istat

    (2022)

    Istat.

    • Google Scholar
23. 23
    A meta-analysis on the price elasticity of energy demand

    1. X Labandeira
    2. JM Labeaga
    3. X López-Otero

    (2017)

    Energy Policy 102:549–568.

    https://doi.org/10.1016/j.enpol.2017.01.002

    • Google Scholar
24. 24
    A Microsimulation Model to evaluate Italian Households’ Financial Vulnerability

    1. V Michelangeli
    2. M Pietrunti

    (2014)

    International Journal of Microsimulation 7:53–79.

    https://doi.org/10.34196/ijm.00107

    • Google Scholar
25. 25
    Questioni di Economia e Finanza (Occasional Papers) 369, Banca d’Italia

    1. V Michelangeli
    2. C Rampazzi

    (2016)

    Questioni di Economia e Finanza (Occasional Papers) 369, Banca d’Italia.

    • Google Scholar
26. 26
    Energyprice Shocks and Inflation in the Euro Area

    1. S Neri
    2. F Busetti
    3. C Conflitti
    4. F Corsello
    5. D Delle Monache
    6. A Tagliabracci

    (2023)

    mimeo.

    https://doi.org/10.2139/ssrn.4849009

    • Google Scholar
27. 27
    Annual report, Osservatorio italianosulla poverta energetica

    1. OIPE

    (2023)

    Annual report, Osservatorio italianosulla poverta energetica.

    • Google Scholar
28. 28
    SSRN, SSRN

    1. G Peersman
    2. J Wauters

    (2023)

    SSRN, SSRN.

    • Google Scholar
29. 29
    mimeo

    1. G Romanini
    2. E Tosti

    (2023)

    mimeo.

    • Google Scholar
30. 30
    Managing the liberalization of Italy’s retail electricity market: A policy proposal☆

    1. C Stagnaro
    2. C Amenta
    3. G Di Croce
    4. L Lavecchia

    (2020)

    Energy Policy 137:1–6.

    https://doi.org/10.1016/j.enpol.2019.111150

    • Google Scholar
31. 31
    Technical report, Ufficio parlamentare di bilancio

    1. UPB

    (2022)

    Technical report, Ufficio parlamentare di bilancio.

    • Google Scholar
32. 32
    Technical report, Ufficio parlamentare di bilancio

    1. UPB

    (2023)

    Technical report, Ufficio parlamentare di bilancio.

    • Google Scholar

Author details

1. Andrea Colabella

   Bank of Italy, DG for Economics, Statistics and Research - Financial Stability Directorate, Rome, Italy

2. Luciano Lavecchia

   Bank of Italy, Climate Change and Sustainability Hub, Rome, Italy

3. Valentina Michelangeli

   Bank of Italy, DG for Economics, Statistics and Research - Financial Stability Directorate, Rome, Italy

   For correspondence

   valentina.michelangeli@bancaditalia.it

4. Raffaella Pico

   Bank of Italy, DG for Economics, Statistics and Research - Financial Stability Directorate, Rome, Italy