Houses, homes
and heating

While housing and the associated heating, water and sewerage infrastructure matter greatly for well-being, housing also accounts for a significant share of total energy use in the economy. Consequently, this chapter also considers the environmental footprint of housing, looking at the energy efficiency of the housing stock and heating systems in the EBRD regions. The chapter then concludes with a series of policy recommendations.

The analysis in this chapter draws on a range of different data sources, including the fourth round of the Life in Transition Survey (LiTS IV), which includes a special module on housing. That survey round, which was launched in 2022 and will conclude later in 2023, is being conducted in 37 economies, asking respondents in 1,000 randomly selected households per economy about a range of socio-economic outcomes (such as their employment and income), as well as their beliefs and attitudes (their views about the environment, for instance). For the first time, the survey also includes detailed questions on housing, with some of those questions being answered by the interviewer before the start of the interview (reporting on the condition of the building, for instance) and some being answered by respondents (providing information on the age of the building or the use of smart meters, for example).

Findings from the LiTS IV reveal that the quality of housing is closely linked to socio-economic outcomes. For instance, people living in higher-quality buildings tend to be healthier and less likely to experience mental distress than those living in buildings of inferior quality. Meanwhile, people who own their home are less likely to want to move. These associations hold even when taking into account people’s education, income and other characteristics.

The housing stock in the EBRD regions reflects the legacies of past policies. More than half of all people in those economies live in buildings constructed between the 1950s and 1980s – that is to say, before the transition from centrally planned to market economies. Large multi-apartment buildings are more common than in advanced economies. Indeed, in some EBRD economies, up to 40 per cent of all households live in prefabricated housing blocks. As a result of widespread mass privatisation in the early 1990s (whereby social housing tenants were able to buy their home from the state for a nominal fee), home ownership rates in most economies in the EBRD regions are much higher than elsewhere. Moreover, ownership rates in EBRD economies are not related to household income, with lower-income households just as likely to be homeowners as high-income households. (Elsewhere, those on lower incomes are more likely to rent.) On the other hand, there is now very little social housing in the EBRD regions, reflecting the state’s withdrawal from housing markets after the privatisation of the 1990s.

Housing inequality has started to increase, and vulnerabilities are emerging. Poorer households are more likely to (i) live in older buildings, (ii) live in housing that is in a worse condition, (iii) have more limited access to public transport and (iv) have less access to green space. Meanwhile, the percentage of the population who own their home outright has started to fall, and reliance on mortgages has been growing. At the same time, rents have increased as a share of income (and relative to average mortgage payments).

As far as the environmental footprint of housing is concerned, the residential sector accounts, on average, for 26 per cent of total emissions and 29 per cent of total energy use in the EBRD regions, compared with 22 per cent of total emissions and 26 per cent of total energy use in advanced European comparators. In some EBRD economies, emissions from the residential sector are the single largest contributor to total emissions, exceeding emissions from industry, transport or other services. In some cases, residential emissions remain high even as industry is becoming greener.

While total residential emissions per capita are lower in the EBRD regions than in advanced economies, the EBRD regions emit more per unit of residential energy use. Differences in countries’ fuel mix (particularly their reliance on coal) can explain around 40 per cent of all cross-country variation in residential emissions per capita. In particular, appliances (such as refrigerators and air conditioning units) are more emission-intensive in the EBRD regions than in advanced economies, reflecting differences in countries’ fuel mix. Differences in emission-intensity are smaller for heating. In addition to altering countries’ energy mix, findings from the latest round of the Life in Transition Survey suggest that there is also scope for significant emission reductions through improvements in insulation and metering of energy consumption (for instance, through energy-efficient upgrades to prefabricated housing blocks), even taking the building stock as given.

The link between (i) housing conditions and the characteristics of neighbourhoods, and (ii) economic outcomes and inequality has been well documented. Housing and the local environment affect access to healthcare, health (including mental health), educational attainment, employment and earnings in adulthood, as well as general well-being and intergenerational mobility.1 In the United States of America, housing instability (being behind on rent, moving multiple times or having been homeless in the past) is associated with adverse health outcomes for adults and children alike.2

More generally, housing segregation (the emergence of low-income neighbourhoods separate from higher-income areas) leads to a host of poor socio-economic outcomes and lower levels of well-being.3 Children growing up in poor-quality neighbourhoods perform less well in school and earn less as adults.4 Studies based on panel data which have tracked individuals over time in the United States have found that children who live in a crowded household at any time before the age of 19 are less likely to complete their secondary education and more likely to have lower educational attainment at the age of 25.5 Likewise, children living in poor-quality housing, in homes that have been in foreclosure, and in disadvantaged neighbourhoods tend to have lower nursery-readiness scores.6 In contrast, children who live in – or move to – better neighbourhoods and are thus exposed to better environments tend to see higher levels of educational attainment and earn more in adulthood.7

Chart 4.1 illustrates such correlations, drawing on the latest wave of the Life in Transition Survey. While it is possible that unobservable characteristics could be associated with both better outcomes and the quality of housing, people living in buildings that are in a better condition (that is to say, less in need of urgent repair) appear to be healthier, on average, than those living in buildings that are in a worse condition. They are also less likely to experience mental distress. Similar effects can be observed among people who own their homes. Furthermore, homeowners are also less likely to express a desire to move – whether within the country or abroad. These correlations are both economically and statistically significant, even when accounting for individual and household-level characteristics such as age, education, employment status and household income. For example, a person living in a better quality building is, on average, 30 per cent more likely to report being in good health. In contrast, if two people live in similar accommodation, but one earns twice as much as the other, that person is, on average, only around 3 per cent more likely to report being in good health.

The 1970s and 1980s, in turn, saw a paradigm shift in many economies with the commodification of housing, whereby the provision and allocation of housing were left to market forces. The public provision of housing (that is to say, social housing), which was common in many welfare states at the time, went through a process of commercialisation whereby housing provision was privatised, or responsibility for allocating it was passed on to non-profit organisations and housing associations, while some countries launched “right-to-buy” schemes enabling households to buy social housing rented from the state at a relatively low cost.18 As a result, the public sector retreated from the direct provision of housing in many countries (including in North America and the United Kingdom), focusing instead on indirect support for low-income households (through housing subsidies and allowances, for example). This policy shift resulted in a sharp rise in inequality, in terms of both the quality of housing and access to economic opportunities.19 Spatial segregation based on ethnicity, race, migrant backgrounds and other socio-economic differences increased.

In contrast, under central planning in central, eastern and south-eastern Europe and Central Asia, housing was seen as a political priority. Nikita Khrushchev, the Soviet leader in the 1950s and the early 1960s, famously promised: “To every family its own apartment”. The right to housing was often enshrined in constitutions,20 and considerable resources were devoted to large-scale housing projects – addressing growing housing needs in rapidly industrialising and urbanising economies and, in some cases, in response to the destruction of housing stock during the Second World War.21 In most economies, the state accounted for the bulk of the construction of housing and was the key provider of housing to the general population, typically in the form of multi-dwelling units.22 The Soviet Union constructed about 2 million dwellings annually over almost three decades, while the construction of new housing in Yugoslavia increased more than five-fold between 1955 and 1965 and remained at or above that level until the late 1980s.23

A key characteristic of housing policies under central planning was the extraordinary occupancy rights enjoyed by tenants: once they had occupied their unit, it was almost certainly theirs for life, and it could be passed on to successive generations of occupants as long as the successors were registered as living there before the previous occupants died or moved away. Housing was also universally affordable: in 1970, rent in the Soviet Union absorbed about 5 per cent of household income on average, with utilities accounting for another 5 per cent; in Yugoslavia, average spending on accommodation fell from about 5.4 per cent of total expenditure in 1969 to 3.4 per cent in 1980 (with tobacco and alcohol accounting for about 4.5 per cent of spending during that period).24

When the transition process began, housing lost its privileged status, leading to the state’s abrupt withdrawal from housing investment and the direct provision of housing services. Budget resources earmarked for housing were drastically reduced and construction was largely left to the private sector, with long-term effects on the supply of accommodation. Construction of new housing has fallen in all economies during the transition process, typically by half.25

On the demand side, the privatisation of social housing through right-to-buy policies dramatically transformed housing markets. Under those schemes, sitting tenants had the right to purchase their units from the local government or state enterprises, typically at a price that was substantially lower than the market value – in some cases, for a nominal fee covering the cost of administering the sale.26 Most of the housing involved was in multi-dwelling apartment buildings, with privatisation carried out on a unit-by-unit basis. The privatisation of housing proceeded rapidly. For instance, while 60 per cent of all units in Estonia were in state ownership before the transition process, by 1995 this share had fallen to just 10 per cent.27 As a result, over half of all people in major post-socialist cities live in system-built, high-density housing estates where housing was transferred to tenants through privatisation programmes – perhaps the most enduring legacy of socialist housing policies and something that continues to define those cities today.28

In principle, the large amounts of home equity that have been created by those mass privatisation programmes have given new owners the opportunity to purchase larger or better units using that equity and mortgages with low loan-to-value ratios. However, right-to-buy schemes have also been associated with large increases in inequality and a decline in security of tenure for some households. Many private renters and homeowners with mortgages are now less secure than they would have been as renters of a state unit with the standard lifetime “social” rental contract.29

In the southern and eastern Mediterranean (SEMED), the public sector played an important role in the provision of housing in the 1970s and 1980s, usually keeping housing affordable through subsidies. This was particularly true in countries that had adopted elements of the Soviet Union’s economic model, such as Egypt. Thousands of dwellings were built as part of major development plans. In the case of Egypt, the scarcity of agricultural land meant that new urban communities had to be built in the desert, resulting in the creation of 12 new towns. Mass housing policies resulted in large numbers of five-storey dwellings (typically without lifts). However, limited floor space often resulted in informal extensions being added to those buildings. The state’s subsequent withdrawal from the housing market further exacerbated the situation. Across the SEMED region, rapid population growth and urbanisation, combined with limited urban planning (with cities being built back to front – buildings first and services afterwards), led to severe housing shortages, high levels of informality and insecurity of tenure.30 In Egypt, the housing backlog is currently estimated at 3.5 million housing units, while Morocco is estimated to need an additional 600,000 units.31 Nearly 70 per cent of Cairo’s residents live in informal housing.32 The next section describes the current housing stock in the EBRD regions, drawing on the latest round of the Life in Transition Survey.

In many EBRD economies, construction rates have been low since the 1990s. Since 1995, the construction of dwellings has only accounted for 3 per cent of GDP in EBRD economies in the EU, 2 percentage points less than in advanced European economies, based on data from Eurostat. In both of those groups of economies, residential construction as a share of GDP increased temporarily in the boom years before the global financial crisis of 2008-09, fuelled by access to cheap credit, but is now back at levels comparable to those seen in the 1990s.33

This is despite demographic trends pointing to robust demand even in countries with shrinking populations (such as Bulgaria), as the number of individual households has continued to increase. Households have become smaller on average, as multi-generational households have fallen in relative terms, while single-person households have become more common. The increase in single-person households has actually been faster in emerging Europe than in advanced European economies, though that increase started at a lower level.

Reflecting the legacies of post-war housing construction, large multi-apartment buildings (defined as those with 10 dwellings or more) are more common in the EBRD regions than in Germany, being most prevalent in the Baltic states (see Chart 4.3). In contrast, small multi-apartment buildings (defined as those with less than 10 dwellings) are more common in the SEMED region and Türkiye.

Prefabricated housing blocks (assembled on-site using standard factory-made components) are common in many economies in the EBRD regions (see Box 4.1). Estimates from the latest round of the LiTS suggest that in some economies (such as Estonia, Latvia and Georgia) up to 40 per cent of households live in prefabricated housing blocks.

Many residential buildings in EBRD economies are in need of repair, partly reflecting their age. In the latest round of the LiTS, interviewers have been asked to rate the condition of the building in which the respondent lives on a four-point scale, with answers ranging from “very good” to “in urgent need of repair”. Interviewers’ responses suggest that in many economies in the Caucasus and the SEMED region, many dwellings are in need of repair (see Chart 4.4). This is especially true of small multi-apartment buildings. In general, older buildings are regarded as being in a worse condition than newer buildings (with the age of the building being estimated by the respondent as part of the interview).

Almost a third of all households living in multi-apartment buildings in the EBRD regions have no formal building management, with that percentage rising to more than 70 per cent in Albania and some economies in the SEMED region. In general, buildings with no central management or only informal management tend to be in a worse condition than buildings run by management companies or homeowner associations. For instance, at the level of the EBRD regions as a whole, 4.5 per cent of all multi-apartment buildings without formal management are regarded as being in urgent need of repair, compared with just 2.1 per cent of multi-apartment buildings with homeowner associations. Similarly, around half of all buildings with no formal management have seen no major refurbishments in the last decade, compared with around 32 per cent of buildings with homeowner associations (see Chart 4.5).

Today, advanced economies broadly follow one of two distinct approaches when it comes to the provision of housing: the unitary rental system (which is widespread in Germanic and Nordic countries) or the dualist rental system (which is common in Anglo Saxon countries).35 The unitary rental market model is characterised by higher numbers of renters and a larger non-profit (or otherwise heavily regulated) housing association sector, which complements the private housing market to ensure more equitable access to housing.36 Home ownership rates in Austria, Germany and Switzerland are around 40-50 per cent, and up to 20 per cent of households in these economies live in social housing (see Chart 4.6).

In contrast, the dualist rental system is characterised by a stronger role for free rental markets. It prioritises home ownership as a means of growing one’s assets and wealth. For instance, the home ownership rate in the United Kingdom is around 65 per cent. The private rental market is profit-making and typically lightly regulated, providing only limited security for tenants. It is based on the notion that competition among landlords can increase the overall quality of housing. Social housing providers are less common, do not form part of the competitive market and act primarily as a safety net for the poor.37

Home ownership rates in former centrally planned economies remain high by international standards as a result of mass privatisation, with over 90 per cent of households owning their homes in some economies in the Western Balkans (compared with just over half in Türkiye and some economies in the SEMED region; see Chart 4.6). These economies have some characteristics of the dualist housing system, with high home ownership rates, but limited social housing. (Even in Poland and Slovenia, where social housing is most common, it accounts for less than 10 per cent of total dwellings.) In these economies, housing security is mainly available through ownership. Rental markets do, however, play a role in facilitating residential mobility. Such mobility is especially important in the context of significant technological shifts, where jobs disappear in some regions (such as those dependent on coal) but are created in others (see Chapter 3 and Box 4.2).

Many economies in the EBRD regions combine high levels of home ownership with low take-up of mortgages, again reflecting widespread privatisation. Other economies, such as Greece (as well as Chile, Colombia, Italy and Mexico), owe that combination of high ownership and low mortgage debt to a long history of inheritance being used as an alternative way of building savings and gaining access to home ownership. In contrast, in countries such as Denmark, Germany, Sweden and Switzerland, most households owning a home have outstanding mortgage debt, while home ownership rates are relatively low.38

Outright home ownership has been falling in the EBRD regions, albeit from a high level, while use of mortgages is on the rise. The percentage of households owning their home outright declined by about 5.7 percentage points between 2010 and 2022, falling to 77.8 per cent, while the percentage of households with mortgages increased by about 4.2 percentage points, rising to 8.6 per cent39 (with a sharper increase in mortgage uptake being observed among richer households). In advanced European economies, home ownership and mortgage use are both in decline according to OECD data. The next section looks at how housing is related to socio-economic divides and access to economic opportunities.

In the EBRD regions, however, home ownership is largely unrelated to income – in stark contrast to Germany, where people in the top income decile are about 3.5 times as likely to own their home as those in the bottom decile (see Chart 4.7). (Similar patterns can be observed for other advanced economies, including those with higher home ownership rates (such as Italy), based on data obtained from earlier rounds of the LiTS and the OECD.) The lack of a relationship between home ownership and income levels in the EBRD regions reflects the high levels of home ownership in many economies on the back of right-to-buy schemes. Indeed, this pattern only holds for primary residences: ownership of other dwellings and land increases sharply with income. Take-up of mortgages also increases with household income (which is not surprising, as bank lending is conditional on households’ ability to repay loans).

While home ownership is largely unrelated to income in the EBRD regions, the link between the condition of housing and income is stronger than in Germany (see Chart 4.8), with poorer households more likely to live in dwellings that are in a worse condition. In urban areas in the EBRD regions, those in the top quintile of the income distribution are somewhat more likely to live in large multi-apartment buildings (and less likely to live in houses) than those in the bottom quintile of the income distribution. In Germany, by contrast, the urban poor are most likely to live in small multi-apartment buildings, while richer households are more likely to live in houses. While most rural households in the EBRD regions live in houses, regardless of their income, small multi-apartment buildings are more common in rural Germany, especially for poorer households.

Housing also affects access to economic opportunities. In general, access to public transport is more difficult in rural areas. However, while access to public transport in advanced European economies is largely unrelated to income, in most of the EBRD regions poorer households are also less likely to have access to public transport than richer households – something that holds in cities, towns, suburbs and rural areas alike (see Chart 4.9, which draws on Eurostat data for a subset of economies). This pattern is also reflected in households’ self-reported satisfaction with commuting time. Satisfaction with commuting time is significantly lower for the poorest 20 per cent in the EBRD regions – and this, again, holds across towns, suburbs and rural areas alike. In advanced Europe, meanwhile, differences between the poorest and richest quintiles are far smaller and, on average, only visible in rural areas – where the richest actually have more difficulty accessing public transport than the poorest.

It is also noticeable that urban areas in the EBRD regions (particularly in Central Asia and the SEMED region) lack access to green spaces relative to Germany. That lack of access to green spaces is more pronounced for lower-income households living in urban areas, who are more likely to live more than 30 minutes away from a public green space (see Chart 4.10). Eurostat data for a subset of economies in the EBRD regions point to similar patterns. Where the former centrally planned economies inherited generous public spaces in urban areas, those green spaces were often “lost in transition”, being used for new housing developments. Meanwhile, the large-scale green development projects that have been implemented in recent years have typically been concentrated in the largest and wealthiest cities.41

Monthly payments covering the cost of housing (mortgage or rent payments, plus the cost of utilities) account for a substantial share of household income, both in the EBRD regions and in advanced European economies (see Chart 4.11). In most EBRD economies, rent payments tend to be higher than mortgage payments as a share of income. In theory, renting and owning should be substitutes, but this is often not the case in practice – partly reflecting the segmentation of housing into renter-focused markets (typically consisting of smaller units closer to urban centres) and owner-focused markets (consisting of larger detached dwellings, often with higher maintenance costs).42 Furthermore, the gap between rent and mortgage payments has widened of late (see Chart 4.11), reaching a record high (though the time series is relatively short).43 Mortgage payments have fallen, both in the EBRD regions and in advanced European economies, reflecting the long-term decline in interest rates, while house prices have increased as mortgages have become more affordable. As house prices have risen, so has rent as a share of income, disadvantaging credit-constrained households, which are potentially unable to buy.

While the distributional implications of housing as an asset have received less attention, wealth inequality tends, in general, to be greater than income inequality, partly owing to the inherited nature of some wealth.48 On average, the bottom 40 per cent of households in OECD countries receive around 20 per cent of total disposable income, but account for only 3 per cent of net wealth. Wealth also tends to be much more concentrated at the top than income.49

However, greater home ownership is actually associated with lower wealth inequality, as housing, as an asset class, is more important to middle-class households than to people at the top of the income distribution. Consequently, inequality in net non-housing wealth (including business and financial wealth) tends to be higher than inequality in net housing wealth. Consistent with the high levels of home ownership in the EBRD regions and its more equal distribution relative to other economies, wealth inequality tends to be lower in post-communist economies for which data are available than in other economies with similar income inequality. For instance, in Hungary, the Slovak Republic and Slovenia, the bottom 40 per cent account for around 5-10 per cent of net wealth; in Austria, Denmark, Germany and the Netherlands, which have similar levels of income inequality, the equivalent figure is around 1 per cent or less. In other words, many homeowners in post-communist economies may be relatively asset-rich, but income-poor.50

Just as housing tends to be households’ largest asset, mortgages tend to be their largest liability. At the country level, mortgage debt makes up more than half of total household debt in almost all OECD countries. Among mortgage holders, mortgage debt represents more than 80 per cent of total debt. Young homeowners and homeowners at the bottom of the income distribution are the two groups where mortgages account for the largest share of household debt.51 The increased use of mortgages in the EBRD regions over time has widened the range of options available to buyers. At the same time, however, increased choice comes with greater vulnerability relating to the potential loss of income or increases in interest rates. Variable-rate mortgages, which expose borrowers to greater interest rate risk, are more prevalent in EBRD economies in the EU than in advanced European economies (being particularly common in the Baltic states, Greece and Poland). Mortgages denominated in foreign currency, which expose borrowers to foreign-exchange risk, are less common than they were a decade or two ago, as the associated risks have been laid bare by large currency movements (such as the appreciation of the Swiss franc – once a popular currency for loans owing to its low interest rates). Nonetheless, in some countries (Poland and Romania, for instance) foreign currency-denominated mortgages have continued to account for more than 10 per cent of total mortgages in 2023.52

Housing is also a source of pollution and greenhouse gas emissions. While energy use and the associated emissions in industry and transport tend to receive more attention, the environmental footprint of the residential sector is sometimes larger. This section puts emissions from housing into the broader context of overall emissions (including those produced by industry, transport and other services) and breaks residential emissions down into their various components. In particular, it distinguishes between (i) heating and hot water (including gas boilers) and (ii) domestic appliances (such as refrigerators or air conditioning units). While the former are often supplied through integrated systems that are part of the building stock (such as district heating), the latter are typically electrified individual units that are not built into housing structures and can be changed at the household level.

Over the last two decades, carbon emissions per capita in the EBRD regions have remained largely unchanged, broadly in line with the levels observed in other emerging markets. Panel regressions suggest that emissions in the EBRD regions are now comparable to those observed in other economies with similar levels of development, urbanisation and industrialisation. Emissions per capita in advanced European economies are higher on average, but they have declined rapidly since the mid 2000s. At the same time, the average for the EBRD regions conceals significant variation across individual economies. Emissions per capita are highest in Central Asia (particularly in Kazakhstan, Mongolia and Turkmenistan) and central Europe (particularly in the Czech Republic and Poland), where levels exceed those seen in advanced European economies. Total emissions per capita are lowest in the SEMED region, followed by the Western Balkans, and eastern Europe and the Caucasus.

Across the EBRD regions as a whole, the residential sector accounts for 26 per cent of total emissions and 29 per cent of total energy use (see Chart 4.13). However, there are several economies where the residential sector is the single largest contributor to total emissions, surpassing industry, transport and other services. This is the case, for instance, in Azerbaijan, Kosovo, the Kyrgyz Republic, Moldova and Serbia.

Differences in residential energy use explain only 22 per cent of total cross-country variation in residential emissions per capita (see Chart 4.15). For instance, residential emissions per capita in Kazakhstan are around 2.5 times the level seen in Estonia, despite Kazakhstan’s residential sector using only 4 per cent more energy per capita. Similarly, Bosnia and Herzegovina, Poland and Serbia emit about twice as much as Slovenia and the Slovak Republic while using similar amounts of energy.

While residential energy use per capita in the EBRD regions is less than half of the level seen in advanced European economies, the EBRD regions emit 43 per cent more carbon per unit of energy used (see Chart 4.16). As a result, total residential emissions per capita are only 3 per cent lower in the EBRD regions. The differential between advanced European economies and the EBRD regions in terms of the carbon efficiency of their energy supply is even larger for the industrial sector, with the EBRD regions emitting 57 per cent more carbon per unit of energy used.

In the EBRD regions, heating accounts for a smaller share of total residential energy use and emissions than in advanced European economies, partly reflecting the widespread use of district heating (see Box 4.3). However, it pollutes somewhat more per unit of energy used than in advanced European economies (see Chart 4.16). Appliances are even more emission-intensive in the EBRD regions than in advanced European economies, a fact that is largely explained by those economies’ reliance on fossil fuels (particularly coal) for the generation of electricity.

Demand for heating and cooling is largely determined by the climate of the relevant economy. In general, demand for heating tends to rise linearly when average monthly temperatures fall below 15°C.53 In Europe, a 1°C drop in the average winter temperature is associated with an increase of around 5 per cent in total consumption of natural gas, for instance.

Proper metering of heating and water, as well as cost-reflective pricing of gas and electricity, will strengthen incentives to use energy efficiently. Various characteristics of the building stock, such as its average age or the use of double or triple-glazed windows, will also affect energy consumption, while more spacious housing will consume more energy.

The following regression analysis looks at these various factors and their relative importance in explaining differences in the logarithm of residential emissions per capita. The contributions of various groups of factors are derived using a Shapley decomposition. The left-hand panel in Chart 4.17 focuses on key variables available for a larger sample of countries; the right-hand panel presents more detailed analysis based on a small sample of 20 economies,54 using information on building characteristics and metering derived from the Life in Transition Survey.

On average, differences in the prevalence of various fossil fuels in countries’ fuel mixes can explain around 25 per cent of total cross-country variation in heating-related emissions per capita, and over 40 per cent for emissions caused by the operation of domestic appliances.

Higher income per capita and larger dwellings per capita significantly increase demand for heating, accounting for close to 20 per cent of total variation in emissions (and, unsurprisingly, a much smaller share of variation in total emissions from appliances).

Colder and longer winters can explain some of the cross-country variation in heating-related emissions. While on average they only explain around 12 to 18 per cent, colder winters in Estonia, for instance, are estimated to result in 65 per cent more heating emissions per capita than colder winters in Croatia. As expected, the climate matters less for emissions from appliances: average temperatures explain only 2 to 6 per cent of total variation in the environmental footprint of appliances, even taking demand for cooling into account.

Incentives to use energy efficiently, captured here as the use of metering and fossil fuel energy subsidies, explain 5 per cent of total variation in heating-related emissions and 15 per cent of emissions from appliances. In a broad sample of countries, the doubling of fossil-fuel subsidies (as a percentage of GDP) is associated with a 40 per cent increase in heating-related emissions per capita. Smart meters provide additional incentives. Unlike traditional meters, which provide a running total of the amount of energy used, smart meters can record consumption at a high frequency, providing more information about energy use and automatically sending meter readings to the energy supplier, making it easier for residents (and energy suppliers) to monitor the cost of energy consumption in real time.

Older buildings are, on average, associated with a significant increase in emissions from heating (but not emissions from appliances). Building age and the percentage of buildings with double-glazed windows explain around a third of total variation in emissions from heating and a tenth of total variation in emissions from appliances. These variables are likely to reflect insulation that is not directly reported in the data (since newer buildings and those with double or triple-glazed windows are also likely to be better insulated). Around 10 to 40 per cent of total variation in emissions from heating and appliances remains unexplained.

Proper metering of water and heating, as well as cost-reflective pricing of gas and electricity, will strengthen incentives to use energy efficiently. This allows consumers to pay for the energy units that they actually use and can support more demand-driven provision of services.55 At present, 38 per cent of households across the EBRD regions report that they pay for heating on the basis of meter readings (see Chart 4.18). While there is some ambiguity in the way that respondents have interpreted the question about metered heating, given the variety of heating arrangements available (ranging from individual wood stoves to stand-alone gas boilers to district heating), the answers reflect substantial use of unmetered energy for heating. By way of comparison, 80 per cent of households across the EBRD regions (and 99 per cent in Germany) report having metered water. In addition to reflecting the metering of water, the figure for water meters also appears to be a good proxy for the prevalence of the metered supply of utilities in general and is therefore integrated into the regression analysis. Smart meters are not common in the EBRD regions: only 12 per cent of households have a smart meter for water, while 10 per cent have a smart meter for heating.

One of the key challenges when it comes to improving the energy efficiency of buildings is the availability of finance. The commercial banking sector is often unable to provide the necessary funding owing to its inability to assess projects and its focus on shorter-term lending. One example of a successful alternative financing vehicle is the Public Investment Development Agency (VIPA) in Lithuania, which was established in 2012. That agency oversees the Apartment Buildings Renovation Fund, a state-backed initiative targeting the renovation of older, energy-inefficient buildings. This initiative combines several different funding sources, including general government funding, government funding from earmarked emission levy revenues, EU structural funds and loans from multilateral development banks (including the EBRD). It works with Lithuania’s Housing Efficiency Agency (BETA) to provide low-interest loans and grants covering the cost of setting up energy-efficiency projects, for example. The use of blended funding comes with strict eligibility criteria: energy savings must total at least 40 per cent (indeed, they average 60 per cent), with an energy performance rating of C or above.

The concept of prefabricated housing originally dates back to pre-war France, where architects were focused on enabling large groups of people to live comfortably in the city while having easy access to green areas. It then came into its own in the aftermath of the Second World War, when the use of load-bearing panels made from pre-stressed concrete allowed large amounts of housing to be built all year round at lower cost (with savings of between 5 and 20 per cent) and at greater speed than with conventional techniques. Construction times dropped by 30 to 45 per cent, while labour requirements were reduced by 40 to 50 per cent. In contrast, bricklaying required skilled masons and could not easily be scaled up, while the use of cast-in-place concrete was limited by severe winters.59 Pre-casting also allowed for closer quality control.60

In the Soviet Union, standardised prefabricated housing accounted for 54 per cent of construction in 1980, up from just 1.3 per cent in 1959. In cities with a population of over a million, it accounted for more than 75 per cent of all construction by 1980.61

While some post-war housing estates have been demolished prematurely, others have been renovated and adapted to modern life. Their key shortcoming is poor thermal insulation. Retrofitting buildings can result in significant energy savings, create jobs and improve comfort. It is estimated to deliver €2 million in energy cost savings and create between 18 and 37 jobs for every €1 million invested.62

In the Baltic states, Bulgaria, the Czech Republic, Hungary, Poland, Romania and the Slovak Republic, the renovation of such housing estates has typically been heavily subsidised by the state (in some cases involving the use of EU funding). This has generally increased the value of such apartments, sometimes transforming the appearance of entire neighbourhoods through the use of coloured cladding panels.

In Hungary, for instance, about 788,000 prefabricated dwellings were built, predominantly in the 1960s and 1970s, and those dwellings still house about 18 per cent of the country’s population today. Indeed, there are some cities, such as Székesfehérvár, where they house almost the entire population.63 Almost a quarter of Hungary’s prefabricated buildings were renovated between 2000 and 2007, reducing energy use by between 8 and 50 per cent. That renovation work, some of which was paid for using state and municipal funding, included repainting, the replacement of doors and windows, and the upgrading of heating systems. In Szeged, where prefabricated housing blocks account for about 38 per cent of all dwellings, 75 per cent of all prefabricated dwellings were renovated between 2002 and 2009. Those renovations cost around €3,900 per apartment,64 with a third being covered by state funding and another third being paid for by the municipality, and achieved energy savings of 30 per cent for heating. Demand for renovated apartments has been high, resulting in a price premium of around 50 per cent for apartments located in refurbished buildings.65

Some cities have sought to turn their refurbished buildings into the next generation of smart homes. SmartEnCity, an EU-funded project focusing on the Estonian city of Tartu, has transformed 18 Khrushchev-era apartment blocks (a total of 664 apartments housing about 1,500 people) into self declared “smart blocks” (“smartkovki”). In addition to a thicker insulation layer, all buildings have been equipped with triple-glazed windows, a demand-based heat recovery ventilation system and roof-mounted solar panels. The hot water supply, which was previously generated by stand-alone boilers, has been integrated into district heating. The project also involved smart solutions in the areas of transport and street lighting through the creation of a city-level portal collecting city, building and apartment-level data, including information on energy use, lighting and vehicle traffic. Total energy demand has been reduced by 36 per cent overall (with a 54 per cent reduction for heating, corresponding to an average saving of €350 per year per tenant).66

Further east, in the Caucasus and Central Asia, such renovation work has been much less common, resulting in buildings becoming more dilapidated and eventually being demolished. Here, private renovation has been the norm, resulting in greater variation in the quality of buildings and the communal areas associated with them.

Those high mobility rates are, in part, a legacy of the transition away from central planning and the associated structural changes, such as the decline in the importance of industry (and one-company towns) and the rise of services. At the country level, mobility peaked in the 1980s and has declined sharply since 2000 (with a similar pattern being observed in Germany).

Older respondents and women are more likely to have moved, while education, current employment status and household income are not predictors of mobility. Those working in higher-skill occupations are somewhat more likely to have moved than people with medium or low-skill jobs, with medium-skilled respondents the least mobile on average (although differences across skill groups are not statistically significant). Mobility and preferences regarding mobility are closely linked: those who have always lived in their current location (or have lived there for longer) are, on average, significantly more satisfied with life and less likely to want to move. Similarly, homeowners are also less likely to express an intention to move in the future. These correlations hold when taking into account a range of individual and household-level characteristics, such as age, gender, education, employment status, household income, household size, urban or rural location, the condition of the building and country fixed effects. In contrast, those who have moved recently (within the last five years) are more likely to want to move again.

The rental market, while relatively modest in size in the EBRD regions, plays an important role in facilitating regional mobility. People who move are much more likely to rent (rather than own) accommodation in the years following the move, taking into account various individual and household-level characteristics (see Chart 4.2.2).

District heating systems are especially prevalent in emerging Europe, a legacy of central planning. In the average economy in the EBRD regions, 14 per cent of total residential energy use is accounted for by centralised heating, ranging from less than 1 per cent in Greece to 35 per cent in Latvia (see Chart 4.3.1). Some advanced European economies with cold winters – Scandinavian countries, Austria and Germany – have comparable levels of penetration for district heating. The five largest district heating systems in the world are found in Moscow, St Petersburg, Kyiv, Warsaw and Stockholm.68 Many smaller cities and towns rely on district heating systems connected to large industrial plants.

While district heating can be less emission-intensive than individual heating systems, this depends on the type of fuel used. On average, district heating systems in the EBRD regions emit around 70 per cent more CO₂ per unit of generated heat than those in advanced comparator economies. District heating is most emission-intensive in coal-dependent economies in the EBRD regions such as the Czech Republic, Kazakhstan, Poland and Serbia. In contrast, a number of economies in Scandinavia and the Baltic states have both high levels of district heating and low emissions, with more than 50 per cent of that heat being generated using renewables (see Chart 4.3.1).69

The largest comparative advantage of district heating lies in its network infrastructure, which enables the use of local renewable heat and electricity sources and large-scale heat pumps, combined with the recovery of excess or waste heat from industrial and urban sources.70 Combined heat and power plants generate electricity and use large amounts of waste heat as an input for the heating system. The Győrhő plant in Hungary, for example, achieves fuel efficiency of almost 82 per cent, converting about 43 per cent of fuel into electricity and around 38 per cent into heat.71 Coal-fired cogeneration plants produce around half of the emissions produced by conventional coal plants in terms of greenhouse gases per unit of energy generated, with similar efficiency gains being observed for other fuels.72 Cogeneration systems can also make use of industrial processes that generate large quantities of heat, such as steelmaking. For example, the Toplana Zenica project in Bosnia and Herzegovina – a joint venture between ArcelorMittal and local authorities in the city of Zenica (which involves funding from the EBRD) – uses waste gases from steelmaking to generate heat. Many initiatives at the frontier of waste heat technology are now seeking to recover heat from sources where heat recovery was previously considered uneconomical. For instance, the city of Odense in Denmark is using waste heat from a Facebook data centre, while the Hammarbyverket heat plant in Stockholm supplies 100 per cent of heating in the local area using heat from the treatment of wastewater.73 Stockholm is also currently trying to incorporate heat generated by data centres into its district heating network.74

However, district heating systems have clear limitations. Heat is lost in transmission, and district heating is uneconomical in many rural and mountainous regions. In many cases, the supply of heat to individual apartments cannot be adjusted, which results in excess supply (as indoor temperature preferences vary).75 In Hungary, for example, it was found that households in prefabricated housing blocks with district heating that could not be regulated at the level of individual apartments spent more of their income on energy and were more likely to accumulate energy-related debt than households with metered district heating or other heating systems. High levels of non-payment resulted in a vicious circle, with the supplier, in turn, having fewer resources and weaker incentives to modernise the district heating network.76 Consumers’ inability to change suppliers leads to a lack of customer focus, inefficiency and underinvestment.77 Resolving these issues is crucial in order to successfully leverage the existing cogeneration infrastructure in emerging Europe and Central Asia.

Warsaw’s district heating system, which was built during Poland’s post-war reconstruction, is the largest in the EU.78 Heating in Warsaw pollutes less than in Krakow, where coal-based heating of individual dwellings is still prevalent. At the same time, coal still powers around 70 per cent of Poland’s district heating systems, despite ongoing work to replace coal as a fuel source.79 At the same time, the installation of individual heating meters and apartment-level controls to regulate heat usage is under way.