The remaining potential for energy savings in UK households

Progress on improving energy efficiency of UK homes has stalled in recent years and the question arises how much more potential for further energy savings exist across the housing stock. Whilst there are some high-level estimates of the potential for buildings energy efficiency in the UK, a more granular assessment is needed to understand exactly where this potential lies and what form it takes. Our analysis fills this gap. It is based on the best available evidence on the remaining potential for energy efficiency improvements within UK residential buildings. Using UK government criteria for investment appraisal, we demonstrate that there is a significant resource of untapped energy-saving opportunities in UK homes. Specifically, our estimates suggest that: one quarter of the energy currently used in UK households could be cost effectively saved by 2035; and this could increase to one half if allowance is made for falling technology costs and the wider benefits of energy efficiency improvements. However, these estimates are sensitive to the assumptions made about capital,


Introduction
Improved energy efficiency has played a pivotal role in reducing the UK's energy use and carbon emissions. On a temperature corrected basis, total UK household energy use decreased Evaluations of these policies have shown them to be highly cost-effective, both in terms of the cost savings to participating households and in terms of broader societal welfare (Lees 2006(Lees , 2008Rosenow and Galvin 2013). This experience supports the argument that market forces alone cannot deliver all cost-effective investment in residential buildings, owing to multiple and overlapping market failures. Instead, policy intervention can be used to improve economic efficiency.
The UK Climate Change Act (2008) establishes a long-term target of reducing greenhouse gas (GHG) emissions to 80 percent below 1990 levels by 2050. To meet this target, the government has established five-yearly carbon budgets which restrict the amount of greenhouse gases UK can legally emit in each five-year period. In October 2017, the government published its Clean Growth Strategy (BEIS, 2017a) that set its approach to meeting the 4 th (2023)(2024)(2025)(2026)(2027) and 5 th (2028)(2029)(2030)(2031)(2032) carbon budgets and to ensuring that the UK maximizes the economic and environmental benefits of green technology. However, the full set of policies to deliver on those carbon budgets has yet to be developed (CCC, 2018). Since residential buildings account for approximately one quarter of UK carbon emissions, the remaining potential for improving the energy efficiency of those buildings is of considerable importance. Accurately assessing and effectively unlocking this potential will be critical to ensuring the UK follows the most costeffective path to meeting its targets. Underestimating and neglecting this potential could cost the UK billions in unnecessarily high energy bills and higher cost energy supply.
In the past, multiple studies have indicated a large potential for improved energy efficiency in UK residential buildings. And this remains the case, despite the substantial improvements over the last 15 years. For example, in its impact assessment for the 5 th Carbon Budget, the UK government concluded that residential buildings had the highest technical potential for carbon abatement across the economy, offering 32% of the total (DECC 2016a). Much of this is due to the potential for improved energy efficiency. The government's analysis is corroborated by other studies, including the Infrastructure Transitions Research Consortium (ITRC 2016) whose evidence fed into the National Needs Assessment. 1 However, a more granular assessment is needed to understand exactly where this potential lies and what form it takes. Our analysis fills this gap. It is based on the best available evidence on the remaining potential for energy efficiency improvements within UK residential buildings.
Using UK government criteria for investment appraisal, we demonstrate that there is a significant resource of untapped energy-saving opportunities in UK homes. Specifically, our estimates suggest that: one quarter of the energy currently used in UK households could be cost effectively saved by 2035; and this could increase to one half if allowance is made for falling technology costs and the wider benefits of energy efficiency improvements. However, these estimates are sensitive to the assumptions made about capital, energy and carbon costs, and capturing this potential will require both significant policy change and large-scale investment.
The paper is structured as follows. The following section presents our methodology for estimating the remaining technical and economic potential for energy efficiency measures in UK households, while Section 3 summarises the data sources employed and the assumptions made. Section 4 presents the results of the modelling exercise as three scenarios with varying levels of ambition, and indicates the potential development of energy and cost savings over time. Section 5 discusses some of the uncertainties in the analysis, as well as the type of measures required to unlock the identified potential. Section 6 concludes with some policy recommendations.
1 The National Needs Assessment brought together a coalition including industry, investors, environmental, legal and professional bodies, and politicians and opinion formers to deliver a 35 year view of the changing demands on infrastructure services.

Methodology
We estimate the lifetime energy savings associated with different levels of deployment of energy efficiency technologies in UK households over the period to 2035, together with the net present value of those energy savings. In valuing these energy savings, we consider both the avoided energy costs and the wider benefits of reduced energy consumption, such as reduced greenhouse gas emissions and improved air quality.
Our estimates for the technical and economic potential of different types of energy efficient technology are based upon analyses by the UK Committee on Climate Change (CCC), while our estimates of the costs and benefits associated with those technologies are informed by policy appraisal guidance (called the Green Book) provided by the UK Treasury and UK Department of Business, Energy and Industrial Strategy (BEIS, 2015). In particular, our calculation of the costs and benefits of energy savings employ a spreadsheet toolkit provided by the Interdepartmental Analysis Group (IAG, 2015). Hence, our analysis does not rely on new modelling tools, but instead explores the implications of the assumptions and tools employed by UK government departments and UK government agencies.
We model four groups of technologies: heating efficiency; appliance efficiency; heat networks; and individual low-carbon heating systems. Heating efficiency encompasses building fabric, heating controls and efficient boilers. Efficient appliances include fridges and freezers, washing machines, dryers, ovens, televisions and lighting. 2 Heat networks include schemes in high density areas that are powered from a number of sources (e.g. gas-fired combined heat and power, industrial waste heat, energy from waste). Finally, low carbon individual heating systems include air-source and ground-source heat pumps (with and without heat storage) and biomass boilers. Heat networks and low carbon heating systems are included in the analysis, since they can significantly reduce final energy demand as well as carbon emissions 3 . For example, heat networks may be driven by gas-fired combined heat and power (CHP) schemes that are more efficient than separate provision of electricity and heat. Similarly, they may be driven by waste heat from power plants and industrial facilities, or by latent heat from sewage or rivers.
All scenarios are modelled relative to a baseline scenario. The baseline shows total household energy demand increasing to 2035, owing to the assumption of no energy efficiency improvements in the existing stock (27 million dwellings), together with the construction of five million new dwellings over that period. The baseline scenarios published by the UK government and CCC include the impact of replacing existing boilers with more efficient condensing boilers when the former are retired. Here, these energy savings have been removed from the baseline and shown separately within our constructed scenarios. The latter also include the energy savings from energy efficiency measures, together with those from deploying heat pumps and heat networks in new dwellings -relative to the use of gas heating in the baseline scenario.
We develop three scenarios representing progressively more ambitious levels of technology deployment. These are termed the limited ambition, cost-effective and technically possible scenarios respectively. Each scenario assumes a different level of technology deployment in existing buildings, but the energy performance of new buildings is assumed to be the same. replaced before the end of its natural life. On the whole in this scenario, estimated energy cost savings (based on BEIS' central scenario for energy prices), discounted over the lifetime of the measures, exceed the associated capital costs. We use a discount rate of 3.5% in line with official guidance (BEIS, 2015). In that year, the Committee on Climate Change (2016b) estimated that 85% of the carbon savings in buildings from the government's projections were 'at risk' (i.e. may not materialise) owing to low take-up of measures, less than full implementation of policies or poor enforcement of standards.
2. Cost-effective: This includes all energy efficiency measures deployable to 2035 that are estimated to be cost-effective according to criteria used by the UK government to appraise public policies and projects (DECC/BEIS 2015). The level and rate of deployment is based closely on the CCC's central scenario for meeting the 5 th Carbon Budget, which means technologies are deployed at a rate informed by their historical deployment and follow a traditional S-curve trajectory for market diffusion. The cost-effective scenario uses a discount rate of 3.5% and accounts for energy cost savings in the same way as the 'limited ambition' scenario, but also places a monetary value on improvements in comfort and air quality, as well as on reductions in greenhouse gas emissions. Measures are cost effective when the discounted sum of these private and social benefits exceeds the associated capital and other costs. This level of deployment is broadly consistent with an overall approach taken across all sectors of the economy that meets the 5th Carbon Budget at least cost. For illustration, we also estimate the value of some additional benefits from energy efficiency measures, such as the improvements in health that result from warmer and drier homes, but we do not include these in our assessments of cost effectiveness. In order to decarbonise the housing stock by 2050 a large share of this technical potential will need to be delivered.

Data and assumptions
This section summarises the sources and assumptions used for: a) the potential energy savings associated with energy efficiency improvements; c) the capital cost of those improvements; and d) the social benefits of those improvements, expressed in monetary terms.

Energy savings from energy efficiency improvements
Estimates of potential energy savings are based upon estimates of the remaining potential for installing different types of energy efficiency measure within UK residential buildings. Our primary source for estimates of the technical and economic potential for improved energy efficiency is the 'Fifth Carbon Budget Dataset' provided by the CCC (2016a). This dataset combines information from a variety of sources. For its 'central scenario' (which meets the 4 th and 5 th carbon budgets), the CCC estimate the annual deployment of different technologies in existing and newly built dwellings between now and 2035, together with the associated energy and emission savings. We use this data to estimate the impact of alternative scenarios for the scale and mix of technology deployment over time.
We supplement the CCC dataset with additional sources, such as those providing estimates of the remaining technical potential for efficiency improvements, together with government policy impact assessments. Table 1 summarises our assumptions for the remaining number of each type of measure, and the fuel and electricity savings associated with each. The fuel and electricity savings are based principally on a study by Element Energy (2013) and assume that all technical potential is fulfilled and are thus incremental for every measure across the existing housing stock. The energy savings associated with district heating are scaled from the CCC dataset to match the potential for heat supplied through heat networks, in this way (perhaps crudely) that the same proportionate mix of district heat sources applies at larger scale (including for waste heat as a source). * TWh for heat networks as number of measures not appropriate as a metric

Costs and lifetimes of energy efficiency improvements
Tables 2-4 summarise our assumptions for the capital cost and lifetime of each measure. The assumptions for capital costs are derived from a variety of sources (ACE 2016;BEIS 2016b;CCC 2015CCC , 2016b. We assume that the capital cost of solid wall and floor insulation fall at the rates set out in Guertler (2014) -leading to a 30% reduction by 2030. Assumptions about the lifetime of different technologies are taken from Ofgem (2015) and are consistent with those used in the most recent supplier obligation. We assume, as households would replace them anyway, that energy-efficient electrical appliances and replacement condensing gas boilers involve no additional capital cost.
For heat networks, we rely upon a study for the CCC by Foster et al. (2015) which provides estimates of the levelised cost of delivered heat (£/MWh). From this, we isolate the capital costs of the central plant and network and express the sum of these in levelised form (Table 3).
For heat pumps and biomass boilers, we adopt capital cost estimates from the CCC (2015; 2016b) and assume that these fall by 20% by 2030 (Table 4) -which is broadly in line with the reductions assumed by DECC (2016d, 2016e) in their 'mass market' scenario.
In addition, Table 5 presents our assumptions for the take-up of behavioural measures to save energy in both the cost-effective and limited ambition scenarios. While public campaigns to increase their prevalence and persistence are likely to carry a cost, this has not been estimated here.   Sources: Ofgem (2015); CCC (2015; 2016b) For illustrative purposes, we also estimate some wider benefits that do not form part of formal policy appraisals, although they are sometimes included in annexes to policy impact assessments. These benefits include the improved health of homeowners and electricity system benefits such as avoided network investment. We also consider some potential macroeconomic benefits, including the contribution of the capital investment to economic output and employment. However, we do not use these wider benefits in our assessment of costeffectiveness owing to the uncertainties associated with each. . For simplicity, we ignore broader market impacts (e.g. lower energy prices), since these can only be captured with a macroeconomic model.
Since investments in improved energy efficiency reduce the variable cost of energy services such as heating and lighting, they encourage increased consumption of those services. This is normally referred to as the direct rebound effect (Chitnis and Sorrell 2015;Sorrell et al. 2009).
We assume a direct rebound effect of 15 percent for fabric improvements and heating system upgrades, in line with the estimates used by the UK government in impact assessments (DECC 2014b). This implies that only 85 percent of the potential energy savings from these improvements are achieved in practice.
We discount the estimated monetary benefits at a rate of 3.5% for the first 30 years, followed by 3% for the next 45 years. This represents a social discount rate that is relevant for policy appraisal and hence is lower than the implicit discount rates applied by consumers for energy efficiency investments.

Valuing reduced greenhouse gas emissions
Our estimates of the greenhouse gas reductions associated with energy savings are based upon standard emission factors (100-year GWP) for the relevant energy commodities, together with assumptions about future changes in the electricity generation mix. These reductions are then assigned a monetary value. In accordance with UK government practice, the carbon emissions associated with electricity consumption are valued differently from those associated with fuel consumption, since the former are covered by the EU emissions trading system (EU ETS).
Traded sector emissions are initially assigned a low value (£4/tCO2e in 2016) to reflect the low price of EU ETS allowances, while non-traded emissions are assigned a much higher value (£64/tCO2e in 2016). Following UK government practice, these values are assumed to converge by 2030 (at £77/tCO2e) to reflect the growth of a global carbon market (BEIS 2015) -although in practice, EU-ETS allowance prices appear unlikely to rise to that level.

Valuing improved air quality
Energy efficiency measures in residential buildings should improve local and regional air quality. This is due to fewer nitrogen oxides (NOx) and particulate matter (PM) being emitted from both fossil-fuel electricity generation and from the direct combustion of fuels in buildings.
The IAG spreadsheet uses standard emission factors to estimate the size of these reductions over time, allowing for changes in the electricity generation mix. The avoided damage associated with these emissions is then valued at £1052/tonne for NOx and £30,225/tonne for PM (DECC, 2016d).

Valuing improved thermal comfort
Improved comfort is an important benefit of and motivator for undertaking fabric and heating improvements. Where buildings are under-heated, energy efficiency improvements can allow occupants to increase indoor temperatures without using additional energy. In addition, draught-proofing can make buildings feel more comfortable even if indoor temperatures remain unchanged. As simple approximation for the value of these benefits is to use the market value of the energy savings that homeowners are willing to forego for improved comfort. These foregone energy savings represent the direct rebound effect mentioned above. We therefore use the retail price of energy to value improved comfort, rather than the long run variable cost of energy. There are potential overlaps between the value of thermal comfort and the value of health benefits associated with higher internal temperatures (below), but at present we treat these separately.

Valuing health benefits
Energy efficiency measures that lead to higher indoor temperatures may also deliver important health benefits such as lower rates of respiratory illness and fewer winter deaths. The UK Government's Impact Assessment of the last phase of the Energy Company Obligation includes an estimate of the monetary value of health impacts, but this is excluded from their net present value calculations owing to the potential overlap with comfort benefits. However, the Government is refining its approach to quantifying these benefits, with the aim of including them in future assessments.
For the purpose of this paper, we estimate health benefits by taking the midpoint between: a) a lower set of estimates derived from the impact assessment of the 'ECO Help to Heat' scheme (BEIS 2016b); and b) a higher set of estimates derived from the 2013 Fuel Poverty Strategy for England (DECC 2013b). The resulting figures are applied to cavity, loft and solid wall insulation, and boiler upgrades only on the basis of the average health benefits per kWh saved.

Valuing benefits to the electricity system
Reductions in electricity consumption can deliver benefits to the electricity system including avoided line losses, avoided expenditure on both generation capacity and electricity networks and reducing the need for some categories of reserve. However, these benefits are uncertain and they can be very context specific. In the absence of reliable estimates for the UK, we employ per kWh estimates (Table 5) from a US case study (Lazar and Colburn 2013) and multiply these by our estimated electricity savings to obtain an estimate of the total benefit to the electricity system. investments on the costs associated with each of the different components of electric system (generation, transmission and distribution), as well as its combined or total impact on the German electric system. Most of the benefits identified in the study were associated with the reduced need for investment in renewable energy production in a more efficient power system that is subject to long-term targets for a high share of renewable energy generation.
Many of those benefits are, of course, associated with reduced peak load as well as a reduction in overall electricity consumption. In addition, more efficient and smarter appliances are able to provide demand response services and could provide further benefits if part of a demand response programme. In the absence of data and the need to make a lot of assumptions we have not quantified those effects in this study and future research should investigate more precisely how improved energy efficiency of the building stock would contribute to cost reductions in this area.

Valuing macroeconomic impacts
Investments in energy efficiency can have complex impacts on economic activity and employment. For example, there will be increased activity by manufacturers of energy-efficient technologies (e.g. heat pumps), as well as by suppliers of those firms. The former is a direct impact of the energy efficiency investment, while the latter is an indirect impact. There may also be induced impacts as a result of increased expenditure by both the additional employed workers (direct and indirect) and by consumers whose real income has increased as a result of the energy efficiency improvements (Blyth et al. 2014;Rosenow et al. 2014). This increased spending may increase sales and hence economic activity in many sectors in the economy.
However, these gains may be partly offset by displaced activity in other sectors-for example, in the energy supply sector as a result of reduced energy demand. Similarly, if the investment is funded through a levy on domestic energy bills, the resulting reduction in household expenditure could have negative economic effects.
For simplicity, we confine attention to the impact of energy efficiency investments on overall economic activity, and estimate this in a relatively crude way. We use gross value added (GVA) as a metric for describing the contribution to gross domestic product (GDP, as measured by the production method, rather than income or expenditure methods 5 ) resulting from energy efficiency investments. GVA is defined by the ONS (2016) as the total value of output of goods and services produced less the intermediate consumption (goods and services used up in the production process in order to produce the output). It is the main component of GDP as measured by the production method.
To estimate GVA, we take values from ONS (2015) of the ratio of GVA to turnover in the relevant industry sectors, averaged over three-years (2010 to 2012). This gives a ratio 39.8% for the efficiency activities and 46.3% for the heat activities. In order to calculate a figure for GVA we use the installed costs of the different technologies, indicated earlier in the paper, as a proxy for turnover.

Results
In this section, we summarise the results of modelling the potential for energy efficiency improvements within UK residential buildings in the period to 2035. First, we show the level of deployment of each technology within each of our scenarios and the associated energy savings. Second, we model a plausible pathway over time for delivering all cost-effective energy savings in the period to 2035. Third, we calculate the costs and benefits of this scenario, including both 'narrow' and 'wide' benefits. However, only one third of the technical potential for solid wall insulation is achieved (2.4 million dwellings out of a total of 7.6 million with solid walls), and only 12% of the potential for heat pumps (2.7 million dwellings).

Figure 1: Proportion of the technical potential in the existing housing stock that is achieved by 2035 in the 'limited ambition' and 'cost-effective' scenarios
Note: Diagram indicates the relevant number of measures, as defined in Table 1 4.2 Energy savings in existing building stock Using UK government criteria for investment appraisal and taking into account broader social benefits, we estimate that half of this investment would be cost-effective in the period to 2035 (the 'cost-effective' scenario), and would lead to a 25% reduction in energy consumption. At Wet appliances -39.2m Cold appliances -37.1m Efficient lighting (lamps) -321.6m Heat pumps -23.0m Heat networks (2050 heat supplied) -40TWh Heating controls and upgrades -12.4m Boiler upgrades -11.7m Other fabric measures -39.7m Double glazing -17.9m Floor insulation -19.5m Solid wall insulation -7.6m Loft insulation -7.1m Cavity wall insulation -5.2m Limited ambition Cost-effective Remaining current technical potential current energy prices (Ofgem 2017), this equates to energy cost savings of approximately £270 per household per year -split between a 36% saving on heating fuels and a 15% electricity saving. In contrast, if cost effectiveness is evaluated solely on the basis of energy savings to households (the 'limited ambition' scenario), a 12% reduction in energy consumption could be achieved. Hence, taking a broader view of the costs and benefits of energy efficiency investments leads to a doubling of the estimated cost effective potential -with the majority of this additional potential deriving from the value placed on carbon emissions (see below).
Notably, most low carbon heating investments only appear cost effective over this time period when these broader benefits are taken into account.
The measures offering the greatest technical potential for energy savings over this period are low carbon heating systems, building fabric improvements and upgrades to condensing boilers ( Figure 3). These contribute reductions in energy consumption of 28%, 15% and 7% respectively compared to the 2015 baseline. But while boiler upgrades are fully exploited in all scenarios by 2035, the limited ambition scenario includes only 4% of the potential from low carbon heating and 13% of the potential from building fabric improvements -since both of these are relatively expensive measures. Including the wider benefits of energy savings (the cost-effective scenario), increases these figures to 25% and 40% respectively.

Energy savings in the future housing stock
The energy savings of the 'cost-effective' scenario can be achieved in a more or less linear fashion over time, although supply chain constraints will mean that initially the roll-out rate   From a policy perspective, it is also important to understand the costs and benefits of this scenario. These are explored in the next section.

Costs and benefits over time
The estimated costs and benefits associated with the cost-effective scenario are shown in green in Figure 4. Taken together, we estimate the net present value (discounted benefits minus discounted costs) of this scenario to be approximately £7.5 billion. Energy cost savings account for 54% of the benefits in this scenario, with the broader, social benefits accounting for the remainder. Hence, the value placed upon those external benefits can have a significant influence on assessed cost-effectiveness. Greenhouse gas savings account for 37% of the total benefits, while air quality and comfort improvements together account for less than 10%.
Hence, the NPV will be sensitive to the assumptions made for the variable cost of energy and the value of GHG emissions, together with the capital cost of the relevant investments. The 'wider' benefits of this scenario are shown in purple. These are less certain and are not necessarily additional because of uncertainties around alternative uses of the funds and the associated GDP effects. Their net value depends on the composition of other investments in the economy. These benefits are dominated by the (partially assessed) impact of the energy efficiency investments on GDP, which account for 81% of the total.

Discussion
Current UK policy is geared towards delivering the 'limited ambition' scenario by 2035, but much of this delivery is uncertain and at risk. In terms of the UK's carbon emissions targets, achieving the 'cost-effective' scenario is consistent with meeting the 5 th Carbon Budget as well For 2050, going further and capturing most of the technical potential will also be necessary.
We anticipate that much of this will be more cost-effective than reducing emissions in electricity supply, industry and transport. Also, a combination of rising energy prices and technological improvements will push the envelope of what is cost-effective in the future.
However, it is clear that the scale of the task is immense. Millions of properties will need to receive measures and this requires developing a supply chain that can deliver at scale and to high quality. For example, the rate of installation of heat pumps will need to increase from around 20,000 per year at present to over one million per year in the 2030s (CCC, 2018). It is also important to treat energy efficiency opportunities in a similar manner to low carbon energy supply. The simplistic assumption that deployment can be left to 'the market' because householders benefit from the energy savings is incorrect. Energy efficiency delivers significant social benefits ('positive externalities') that are not captured by individual households. Hence, even in a well-functioning market, investment in energy efficiency will be less than socially optimal unless these externalities can be reflected through public policy.
Figure 2 also shows that the technical potential is significantly larger than the current cost-effective potential. But experience with mass market, low carbon technologies (most obviously solar photovoltaics) shows that increased deployment leads to falling costs and improving performance which makes such technologies increasingly cost-effective. Again, these 'learning economies' have wider social benefits that are not captured by individual investors. Policy support for deployment can therefore be justified and will bring the cost-effective potential closer to the technical potential. In short, the main rationales for policy support for low carbon energy supply are equally relevant to energy efficiency, but are given insufficient weight in current UK Government policy.
In its 2017 Clean Growth Strategy (BEIS, 2017a), the UK government set out a range of policies, proposals and intentions to reduce UK carbon emissions and to build a greener economy. But many of these proposals are lacking in either detail or ambition, and at the time of writing the 'firm' policies that have been announced are insufficient to deliver on the fourth and fifth carbon budgets (CCC, 2018). For households, the Strategy set out an aspiration to upgrade as many homes as possible to Energy Performance Certificate (EPC) Grade C by 2035, provided this is "practical, cost-effective and affordable" (BEIS, 2017). With 19 million homes (70%) currently below this standard, achieving this aspiration would go a long way towards delivering the energy savings identified above, together with the associated private and social benefits. But much hinges upon how the 'cost effective and affordable' caveat is interpretedand in particular, whether the broader social benefits of energy efficiency investments are recognized. For example, the government is currently consulting on energy efficiency standards for the private rental sector (BEIS, 2017b), but the proposals include a £2500 cap on the costs incurred by landlords. This would greatly restrict the level of investment and hence the amount of energy savings achieved, as well as failing to remove many tenants from fuel poverty (ACE, 2017). If similar restrictions are placed upon other policy initiatives, the UK will run the risk of either missing its 4 th and 5 th Carbon Budgets or relying upon traded or banked carbon credits to comply -which in turn would make it much harder to deliver deeper emission reductions in the future. This paper has deliberately not attempted to develop policies for delivering on the potential we identified. An interesting and important questions is how energy efficiency at scale can and should be funded, how much will be public and how much private investment, and whether public investments should be funded through energy bills (for example and Energy Efficiency Obligation) or general taxation (for example a loan or grant programme). A recent working paper by the UK Energy Research Centre looks at some of those questions in more detail (Barrett et al. 2018).

Conclusions and policy recommendations
It is clear that there remains large untapped potential for energy efficiency improvements within UK housing. Using the best available evidence, our analysis demonstrates that: • one quarter of the energy currently used in UK housing could be cost-effectively saved by 2035; • allowing for falling equipment costs and including the wider, social benefits of energy efficiency improvements, it should be possible to cost-effectively reduce energy demand in UK homes further. Technically, energy demand of UK homes could be halved. With innovation in technology and delivery, appropriately supported by public policy, it is likely that more ambitious reductions can be achieved.
The evidence shows that the benefits of improved energy efficiency in UK homes are considerable and justify significant investment from both the public and private sectors. In addition to energy savings, upgrading homes delivers a wide range of persistent benefits to the economy and society, such as improved health, better comfort, increased productivity, more skilled employment and reduced investment in electricity networks -all of which are hallmarks of a modern, low carbon infrastructure. These in turn can contribute to broader policy objectives, such as relieving pressure on the NHS, supporting households struggling to make ends meet, delivering more cohesive communities and reducing fuel poverty. These benefits reduce the cost of the transition to a low carbon economy. More so than carbon abatement in other sectors, the benefits from investing in homes accrue directly to people everywhere in the UK.
To leverage the necessary private investment for these benefits to be captured, there needs to be significant policy change and public investment. The ambitions set out in the Clean Growth Strategy need to be backed up with strong policies in a range of areas, including: an effective energy efficiency programme for the 'able to pay' sector; more ambitious standards for private rented buildings; regulations that ensure new buildings meet high standards of energy efficiency and are suitable for low carbon heating; and greater support for the deployment of heat pumps, particularly in dwellings that do not have access to natural gas. To achieve the necessary cross-departmental focus, UKERC (2017) has proposed an Energy Efficiency White Paper that, amongst other things, would review the relevant cost benefit, tax and accounting frameworks, establish a dedicated heat regulator, set out the contribution of energy efficiency to achieving climate targets and propose a policy mix that will deliver this deployment. The proposed National Infrastructure Assessment plan for 2018 also provides an opportunity to ensure that energy efficiency and low carbon heat are central to future plans for UK infrastructure.
The best available evidence shows that upgrading the UK's existing buildings can provide substantial energy savings while delivering large benefits to society. It should therefore be a