Scholarly article on topic 'Innovative financing models for low carbon transitions: Exploring the case for revolving funds for domestic energy efficiency programmes'

Innovative financing models for low carbon transitions: Exploring the case for revolving funds for domestic energy efficiency programmes Academic research paper on "Economics and business"

CC BY
0
0
Share paper
Academic journal
Energy Policy
OECD Field of science
Keywords
{Finance / Investment / Carbon / Energy / Retrofit / "Revolving fund"}

Abstract of research paper on Economics and business, author of scientific article — Andy Gouldson, Niall Kerr, Joel Millward-Hopkins, Mark C. Freeman, Corrado Topi, et al.

Abstract The IEA has estimated that over the next four decades US$31 trillion will be required to promote energy efficiency in buildings. However, the opportunities to make such investments are often constrained, particularly in contexts of austerity. We consider the potential of revolving funds as an innovative financing mechanism that could reduce investment requirements and enhance investment impacts by recovering and reinvesting some of the savings generated by early investments. Such funds have been created in various contexts, but there has never been a formal academic evaluation of their potential to contribute to low carbon transitions. To address this, we propose a generic revolving fund model and apply it using data on the costs and benefits of domestic sector retrofit in the UK. We find that a revolving fund could reduce the costs of domestic sector retrofit in the UK by 26%, or £9 billion, whilst also making such a scheme cost-neutral, albeit with significant up-front investments that would only pay for themselves over an extended period of time. We conclude that revolving funds could enable countries with limited resources to invest more heavily and more effectively in low carbon development, even in contexts of austerity.

Academic research paper on topic "Innovative financing models for low carbon transitions: Exploring the case for revolving funds for domestic energy efficiency programmes"

Contents lists available at ScienceDirect

Energy Policy

journal homepage: www.elsevier.com/locate/enpol

Innovative financing models for low carbon transitions: Exploring the case for revolving funds for domestic energy efficiency programmes

Andy Gouldsona'*, Niall Kerra, Joel Millward-Hopkins a, Mark C. Freeman b, Corrado Topic, Rory Sullivan a

a School of Earth and Environment, University of Leeds, Leeds LS2 9JT, UK b School of Business and Economics, Loughborough University, UK c Environment Department, University of York, UK

HIGHLIGHTS

• Examines the need for substantially higher levels of low carbon investment.

• Explores the need for innovative financing mechanisms such as revolving funds.

• Shows that revolving a fund could reduce the cost of UK retrofit by £9 billion or 26%.

• Also shows that a revolving fund could make retrofit cost-neutral in the long term.

• Concludes that revolving funds could dramatically increase low carbon investment.

ARTICLE INFO ABSTRACT

The IEA has estimated that over the next four decades US$31 trillion will be required to promote energy efficiency in buildings. However, the opportunities to make such investments are often constrained, particularly in contexts of austerity. We consider the potential of revolving funds as an innovative financing mechanism that could reduce investment requirements and enhance investment impacts by recovering and reinvesting some of the savings generated by early investments. Such funds have been created in various contexts, but there has never been a formal academic evaluation of their potential to contribute to low carbon transitions. To address this, we propose a generic revolving fund model and apply it using data on the costs and benefits of domestic sector retrofit in the UK. We find that a revolving fund could reduce the costs of domestic sector retrofit in the UK by 26%, or £9 billion, whilst also making such a scheme cost-neutral, albeit with significant up-front investments that would only pay for themselves over an extended period of time. We conclude that revolving funds could enable countries with limited resources to invest more heavily and more effectively in low carbon development, even in contexts of austerity.

© 2015 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license

(http://creativecommons.org/licenses/by/4.0/).

Article history: Received 26 March 2015 Received in revised form 6 August 2015 Accepted 10 August 2015

Keywords:

Finance

Investment

Carbon

Energy

Retrofit

Revolving fund

1. Introduction

1.1. The importance of the climate finance gap

Tackling climate change undoubtedly represents an enormous challenge, but at the global scale the economic case for tackling it is compelling. Stern famously estimated that the costs of avoiding

n Corresponding author.

E-mail addresses: a.gouldson@leeds.ac.uk (A. Gouldson), n.kerr@leeds.ac.uk (N. Kerr), jmhopkins@gmail.com (J. Millward-Hopkins), m.c.freeman@lboro.ac.uk (M.C. Freeman), corrado.topi@york.ac.uk (C. Topi), rory@rorysullivan.org (R. Sullivan).

climate change could be between 1% and 2% of global GDP, but the costs of suffering climate change could amount to between 5% and 20% of GDP per year (Stern, 2007). Even with such a compelling global case for action, however, it is clear that an effective response still requires enormous levels of investment. It is also clear that the general, long term, social case for action on climate change does not always translate into a specific, short term, private case for investment, and that the availability of public funds is frequently constrained in contexts of austerity. Particularly in some settings, these factors have led to levels of financing for low carbon development that are much lower than many estimates of what is necessary. The IPCC (2014) estimated that global investment in climate mitigation and adaptation was in the range of USD

http://dx.doi.org/10.1016/j.enpol.2015.08.012

0301-4215/© 2015 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).

343-385 billion per year in the period between 2009 and 2012, and Buchner et al. (2013) suggested that global climate finance flows have plateaued at USD 359 billion. Both of these estimates equate to roughly 0.5% of global GDP; approximately one third of the upper end of the investment needs as estimated by McKinsey (2010), GEA (2012), WEF (2013), McCollum et al. (2013) and IEA (2013a) and one quarter of the upper end of the investment needs as set out in the Stern Review (Stern, 2007).

The need for an effective response to under-investment in climate mitigation is pressing. As the years pass, decisions are made that will lock the world in to high carbon development paths for years to come, whilst at the same time long lived emissions continue to accumulate in the atmosphere and the opportunity to make investments that will help to avoid dangerous climate change diminishes. Indeed, the IEA (2013a, p3) reported that 'the goal of limiting warming to 2 °C is becoming more difficult and more costly with each year that passes'. In assessing the scope to avoid dangerous levels of climate change by limiting atmospheric emissions to no more than 450 ppm, a level that is associated with a good chance of avoiding dangerous climate change (IPCC, 2014), the IEA (2013a, p3) finds that 'almost four-fifths of the CO2 emissions allowable by 2035 are already locked-in by existing power plants, factories, buildings, etc. If action to reduce CO2 emissions is not taken before 2017, all the allowable CO2 emissions would be locked-in by energy infrastructure existing at that time'.

But the conditions for investment in low carbon development have hardly been ideal in the last few years. The failure to reach a global agreement on climate change in Copenhagen in 2009 coincided with the financial crisis and the start - in many countries -of a period of instability, uncertainty, recession and austerity. In many settings, for the past few years at least, more emphasis has been placed on these economic and financial issues than on tackling climate change. Indeed, as market instability and policy uncertainty limit private investment, and budget deficits and austerity limit public investment, it seems appropriate to explore some innovative ways of substantially increasing investment in low carbon development.

1.2. The potential role of revolving funds

With this in mind, this paper explores the case for the creation of an innovative financing mechanism - the revolving fund -where the savings from investments in energy efficiency and other forms of low carbon development are captured and reinvested to either reduce the need for new finance or to increase the impact of what finance there is. Such funds have been discussed before (EC, 2011; Forum for the Future, 2011; DECC, 2012a; IEA, 2013b) and have been adopted in different contexts to fulfil a range of objectives including energy efficiency upgrades, promotion of renew-ables, the provision of clean water and the clean up of contaminated land. Examples of such revolving funds include:

• The US Clean Water State Revolving Fund that was established in the 1990s and has provided over 33,000 loans with a total value of over $100 billion (USEPA, 2015).

• The Thai Energy Efficiency Revolving Fund that since its creation in 2003 has invested c$470m in 294 energy efficiency projects, mostly in factories (Gruning et al., 2012).

• The UK SALIX revolving fund that since its establishment in 2004 has invested £339 million in over 12,000 energy efficiency and renewable energy projects in the public sector with estimated fuel cost savings of £1.2 billion (SALIX, 2015).

• The US Sustainable Endowments Initiative that was set up in 2005 and has since helped to create 79 revolving funds that invested over $100 million of investment in energy efficiency and renewable energy projects in higher education institutes (SEI, 2015).

Various other revolving funds have also been created for urban regeneration, infrastructure provision and economic development. However, although there were evaluations of revolving funds for water and infrastructure provision in the 1990s (see Holcombe, 1992; O'Toole, 1996), as far as we are aware there has never been a formal academic evaluation of the contribution that such funds can make either to reducing the cost of achieving particular carbon reduction targets or to enhancing the impacts of scarce low carbon investment funds. This lack of academic analysis on the potential of revolving funds to help mitigate climate change is not unusual -indeed the IPCC (2014) notes that the scientific literature on investment and finance to address climate change is still very limited and that knowledge gaps are substantial.

1.3. The need for investment in energy efficiency in buildings

These factors are particularly significant for the buildings sector. Globally, over one-third of all final energy and half of electricity are consumed in buildings that are therefore responsible for approximately one-third of global carbon emissions (IEA, 2013b). Energy use in buildings is therefore of critical importance, and many reports highlight the presence of cost-effective opportunities to improve their energy efficiency (IPCC, 2014). However, the IPCC (2014) noted that many potentially attractive energy efficiency investments do not meet the short-term financial return criteria of businesses, investors, and individuals. As a result, the IEA (2013b) predicted that without a concerted push from policy, two-thirds of the economically viable potential to improve energy efficiency in buildings will remain unexploited by 2035.

The reasons for this inertia relate to the presence of strong barriers to change. The IPCC (2014) cited imperfect information, split incentives, lack of awareness, transaction costs, inadequate access to finance, industry fragmentation, the need for new delivery mechanisms and the absence of pipelines of bankable energy efficiency projects as significant barriers. Focusing specifically on the financial barriers, the IEA (2013a) highlighted the importance of up-front costs, levels of risk, issues with interest and discount rates and the inadequacy of traditional financing mechanisms for energy-efficient projects. New forms of policy support, new institutional arrangements, new forms of finance, and new business models are therefore required if the energy efficiency opportunities in buildings are to be exploited (DECC, 2012a; GEA, 2012; IEA, 2013a; IPCC, 2014).

The scale of the challenge is formidable - the IEA (2013b) estimated that over the next four decades USD 31 trillion will be required to promote energy efficiency in buildings at a rate that gives the world a good chance of limiting the temperature increases associated with climate change to 2 °C. Whilst the IEA (2013a) suggests that 'it is widely recognised that mobilising huge investment into energy efficiency is essential' it also argues that 'offering advantageous financing mechanisms is likely to require public funds and these may be harder to justify with tighter public budgets' and that as a result mobilising private as well as public sector financing will be essential. In 2008, the IEA argued that one way of doing this might be to establish revolving funds for building refurbishment and retrofit (IEA, 2008).

1.4. The European context

These issues are particularly relevant in Europe. The European Commission has set a target of reducing energy consumption by 20% by 2020, with performance assessed relative to business as usual projections that include assessments of background trends in energy use and energy efficiency (EU, 2012). It has also recognised that €100 billion a year will be needed to reach this target, and it has set aside €27 billion to support the transition to a

low carbon economy through the European Structural and Innovation Funds and €265 million for a European Energy Efficiency Fund. It has also recently established the Energy Efficiency Financial Institutions Group to find new ways of overcoming barriers and scaling up investment in support the search for new ways of supporting energy efficiency investments, particularly in buildings renovation (EEFIG, 2015). In each of these areas, there is a pressing need not only to make more funds available for energy efficiency and low carbon transitions, but also to significantly enhance understanding of the ways in which those funds could be most effectively and efficiently deployed.

The European buildings sector is a central part of the wider drive to improve energy efficiency. Energy use in buildings accounted for 34% of total final energy demand in Europe in 2007, with the residential sector accounting for 23% and the commercial sector 11% (GEA, 2012). The European Commission (2011, p8) stated that 'In Europe, the built environment provides low-cost and short-term opportunities to reduce emissions, first and foremost through improvement of the energy performance of buildings ... emissions in this area could be reduced by around 90% by 2050'. It also stated that the buildings sector provides the second largest untapped and cost-effective potential for energy savings after the energy sector itself (EC, 2011). Like many others, the EC recognised the importance of finance and investment if the transition to more energy efficient buildings is to be made. There is recognition that there needs to be a marked improvement in financial incentive structures and that 'Innovative programmes will be needed to eliminate information barriers, reduce transaction costs and mobilise investment capital' and that smart financing schemes are needed that can leverage private sector investments (ECF, 2010 p16). Indeed, the European Union has stated that 'Public finance through innovative financing instruments, such as revolving funds, preferential interest rates, guarantee schemes, risk-sharing facilities and blending mechanisms can mobilise and steer the required private finance' (EC, 2013, p11).

1.5. The focus and structure of this paper

Within the broader context of the issues and knowledge gaps relating to climate finance in general and revolving funds in particular that have been introduced above, this paper explores the case for the creation of revolving funds that could be used to increase levels and enhance the performance of investments in energy efficient and low carbon buildings. Based on the development of a model designed to explore and illuminate the workings of a revolving fund, the paper considers the impacts that such a fund could have on the financing of a large-scale energy efficiency programme for the domestic sector in the UK. Data are drawn from various formal assessments of the costs, performance and scope for deployment of different energy efficiency and low carbon measures that could be adopted across the UK housing stock. In order to test the sensitivity of the model to changes in real world conditions, provision is made to adjust variables including household repayment levels, energy prices, administrative costs, incentives for participation, performance gaps and rebound effects. Results are presented for a number of different scenarios, each with a slightly different design. The findings of the analysis for each specific case are presented and key issues in the governance of revolving funds are discussed before the wider implications for policy and practise are discussed.

2. Methods-creating a basic revolving fund model

Revolving funds could come in many different forms, with different structures, scales, business models and governance

Fig. 1. The basic structure of the revolving fund.

arrangements. In this paper, we explore the case for a large-scale revolving fund that could be adopted - by actors in the public, private or civic sectors on either a for-profit or not-for-profit basis-to fund energy efficiency or low carbon investments in domestic buildings.

2.1. Model design

The basic structure of the revolving fund assessed in this paper is set out in Fig. 1. This figure shows the flows of finance into, around, and out of the revolving fund, and creates the basis for our analysis of different variables within the fund. At the heart of the revolving fund as conceived here is a new entity (here called a Special Purpose Vehicle or SPV) that is established to receive new investment funds from different sources and to invest these in energy efficiency and low carbon measures in households.

Our baseline model calculates the total investment needs associated with the widespread adoption of a range of energy efficiency and low carbon measures in a particular country, region, or city. For simplicity, we assume that the up-take of such measures is entirely financed through the revolving fund and that impacts are additional to wider background trends. To enable us to adjust the investment needs to reflect the available investment resources, the model sets a maximum proportion of this total that can be financed through the fund each year. Funds - that can be provided by the government, banks, institutional investors or local communities - are borrowed by the SPV at a real (i.e. after inflation) interest rate. The model then assumes that a proportion of the total capital loaned to the SPV will be repaid to the original funders each year. The SPV, after making these repayments and covering its administrative costs, then invests any remaining funds in energy efficiency and low carbon measures.

The SPV then makes money available to fund a wide range of

energy efficiency or low carbon measures at the household level. It is assumed that the SPV offers a menu of measures that households could opt to apply, but that the SPV ultimately decides which measures it will invest in. In taking this decision, the SPV has some limited scope for optimising its investments, and the model allows for measures to be prioritised according to their Net Present Value (NPV), pay-back period, or carbon savings per pound invested. To reflect constraints that may arise due to a limited availability of local installers, for example, the maximum annual deployment potentials of any measure can also be set. Once the maximum available potential of the top priority measure is exploited in any one year, any available funds remaining are diverted to the next highest priority. Allowance is also made for the costs and the profits of the installers.

Households are encouraged to participate by being offered a share of the savings generated by measures whilst investment costs are being recovered, and all of the savings thereafter for the remainder of the life of the measure. Other inducements to encourage participation, such as cash-back vouchers and government subsidies are also allowed for by the model. We do of course recognise that financial incentives alone may not stimulate significant levels of participation and that any scheme will need to be designed and run in ways that raise awareness, minimise risk and disruption and build trust and confidence in the scheme (Dowson et al., 2012; Rosenow and Eyre, 2013). These non-financial aspects are critically important and will determine levels of take-up by households and thus the scale of the investments that can be made and the carbon savings that can be stimulated by the revolving fund.

Where participation can be promoted, households make regular periodic repayments to the SPV that allow it to recoup its financial investment. Repayments are based on a set percentage of the expected cost savings that arise from implementing the carbon efficiency measures after taking into account the impacts of performance gaps and direct rebound effects. Performance gaps reflect the difference between the technical savings that could be generated by a measure and the real savings that are likely to occur in practice with imperfect installation or operation. Direct rebound effects reflect the fact that householders may consume some of the efficiency benefits through increased comfort levels so that the measures achieve lower savings than predicted.

After taking these factors into account, an agreed proportion of the savings realised from each of the different measures funded is recovered and fed back to the SPV by the companies that supply energy to the building. The energy companies are paid a small fee for acting in this way. The period over which the payments are made is then determined in a way that ensures that the Internal Rate of Return (IRR) to the SPV from the original investment reaches a pre-determined threshold. This will ensure the financial resilience of the SPV. A consequence of this funding mechanism is that, the more effective the measure, the shorter the time period over which repayments are made (see Table 1). In some cases, the repayment period for a measure is longer than its effective lifetime. In this case, the household makes repayments until the measure is no longer effective, with the SPV then absorbing any financial shortfall. Although the SPV does not achieve the threshold IRR on these economically unviable projects, these measures are still funded because of the carbon savings that they generate.

As some householders may renege upon their financial obligations, the total flow of savings fed back to the SPV also has to be reduced to take defaults into account. After adjusting to account for these defaults, any recovered funds are used to repay the SPVs original funders/investors with interest, and after this any remaining funds are recycled and reinvested in further energy efficiency measures in households. Once all available potential for the

Table 1

List of energy efficiency and low carbon measures assessed.

Measure Lifetime Payback period for mea-

(years) sures installed in 2014 (years)

Loft insulation from 0 to 270 mm 40 2.4

Loft insulation from 25 to 270 mm 40 4.2

Loft insulation from 50 to 270 mm 40 6.4

Loft insulation from 75 to 270 mm 40 8.0

Loft insulation from 100 to 270 mm 40 22.3

Cavity wall insulation for homes 40 3.6

built before 1976

Cavity wall insulation for homes 40 7.4

built between 1976 and 1983

Cavity wall insulation for homes 40 15.8

built after 1983

DIY floor insulation for suspended 40 4.1

timber floors

Solid wall insulation 40 33.6

Paper type solid wall insulation 40 176.1

Best practise standard windows 25 28.7

Uninsulated hot water cylinder to 20 4.7

high performance cylinder

Modestly insulated hot water cy- 20 51.0

linder to high performance

cylinder

Primary pipework insulation 15 4.1

Improve airtightness 25 6.3

Thermostatic radiator valves 15 15.1

Room thermostats 15 6.1

Hot water cylinder thermostat 15 176.1

Efficient lighting 10 3.2

Photovoltaic generation with FIT 20 20.4

Micro wind turbines (1 kW) with 10 16.9

Mini wind turbines (5 kw) with FIT 15 9.1

deployment of energy efficiency and low carbon measures has been exploited, and all debts to original funders have been repaid, the SPV can generate a revenue stream for its owners (who could be central or local government, local communities or private entities). This revenue stream could be taken as income or reinvested in other low carbon initiatives.

2.2. Application and data sources

To apply the model of the revolving fund described above, and to make use of available datasets, we focus on the scope for energy efficiency and low carbon measures to be installed in the domestic sector in the UK. The UK has 26 million residential buildings, most of which were constructed before 1980 and have relatively low levels of energy efficiency (Sweatman and Managan, 2010; Webber et al., 2015). Domestic energy use and carbon emissions have fallen in recent years as a result of a range of market and non-market factors, but domestic energy use still accounts for around 25% of UK carbon emissions (DECC, 2013b). The retrofitting of much of the existing housing stock is still required therefore, but current rates of retrofit are unlikely to meet current targets and a substantial increase in the rate of their adoption is required (CCC, 2014).

To assess the contribution that a revolving fund could make to the financing of an extensive retrofit programme in the UK, we draw input data on the costs, performance and scope for deployment of a range of these measures from a model that was developed by the UK Building Research Establishment for the UK Committee on Climate Change (BRE, 2008). This is the most comprehensive publicly available data set that is currently available; in the analysis below some key aspects of it (particularly energy prices, carbon intensities of electricity and levels of subsidy and feed-in-tariffs) are updated to reflect current conditions and

the most recent forecasts. We also test the sensitivity of the findings to key assumptions within the data, including performance gaps and rebound effects.

We use input data on 23 energy efficiency and low carbon measures including different levels of loft, cavity, solid wall and floor insulation, improved lighting and appliances and microgeneration from solar photo-voltaics and small wind turbines. A full list of the measures - including their estimated lifespans and payback periods - is included in Table 1. Input data on the costs of the different measures takes into account the purchase, installation, running and maintenance costs and the different lifespans of each measure.

To assess the performance of each measure, the input data considers their impact in a standard or 'average' UK house that has already been upgraded to a good standard of energy efficiency. As many UK houses are not yet at this level of efficiency, and as energy prices could be higher than predicted, each measure could have a greater impact than is incorporated in our calibrations. Savings are then calculated using up-to-date energy prices for multiple scenarios integrating low, central and high projections of future energy prices (DECC, 2012b). We assume that both technology prices and current levels of feed-in-tariff stay constant throughout the analysis. Carbon savings are calculated using the most recent assessments and forecasts of the carbon intensities for electricity generation (DECC, 2012b). The scope for deploying each of the measures in suitable houses across the entire UK housing stock is also assessed in the model, based on data from the UK Committee on Climate Change. This data takes into account the proportion of homes that may be hard to reach, it adjusts for background trends in the take-up of different measures and it has been up-dated to 2012 levels. By considering the total scope for deployment of each measure across the UK, assessments of each individual measure can be scaled up to consider aggregated costs and benefits if all measures are installed in every suitable property in the UK.

2.3. Default settings and variations

To run the model using the input data described above, we use a series of default settings for each of the key variables. These are presented in Table 2. These settings are based on a series of assumptions that have been informed by, for example, the UK Green Deal impact assessment (DECC, 2012c) and the adoption of domestic energy efficiency schemes in cities such as the Birmingham Energy Savers scheme (BES, undated) and the Leeds City Region Domestic Energy Efficiency Programme (LCR, undated). We therefore believe that these settings realistically reflect the

Table 2

Default settings.

Variable Default setting

SPV maximum borrowing level from external sources (as % of 5%

total investment need)

SPV borrowing rate (real) 3%

SPV annual repayment of loan (as % of total loan) 5%

SPV running costs (as % of overall investment) 3%

SPV required internal rate of return (real) 4%

Measures considered All

Maximum exploitation of any measure in any year 20%

Installer costs 10%

Household repayment to SPV as proportion of total savings 75%

during repayment period

Performance gaps 41.6%

Direct rebound effects 15%

Household default levels 5%

Cost recovery administration cost 0.1%

conditions under which a large-scale revolving fund might operate in the UK, whether at the national or the city scales.

The default settings assume that the SPV at the heart of the revolving fund can, in any one year, borrow a maximum of 5% of the total investment required to exploit the full potential of all measures. This is at a real interest rate of 3%. We also assume that the SPV would repay 5% of its existing total borrowing in any year, until the point where all investments had been made when any surplus funds generated through on-going savings would be used to clear any remaining debts. We also assume that the SPVs running costs would equate to 3% of its annual level of investment and that its threshold IRR is 4% real. We assume that the SPV may, over time, invest the full amount required for the full deployment of all 23 measures included in the assessment. Because of limited availability of installers and limited levels of take-up, we assume that in any year a maximum of 20% of the potential take-up of any one measure can be exploited. We also assume that installers charge 10% of the capital cost of any measure. It is assumed that, as an incentive to participate, householders retain 25% of the savings arising from any measures installed during the repayment period, with the remaining 75% diverted to the SPV. Thereafter, households retain all of the savings. Reflecting the base settings in the model that generated the input data, it is assumed that performance gaps are 41.6% and that rebound effects are 15% - in other words they assume that 56.6% of the technical energy savings potential of any measure is lost. As more recent assessments have found that performance gaps and rebound effects are much lower than these levels (see Webber et al., 2015), this means that the model may generate conservative predictions both of the financial viability of the revolving fund and the carbon savings that it might generate. The default settings also assume that 5% of all households default on their financial obligations, but that the SPV is still responsible for repaying any finance provided to these households. Finally, it is assumed that energy companies are paid 0.1% of the savings to pay for the administrative costs associated with collecting repayments and feeding them back to the SPV.

While these values are adopted in the baseline run of the model, we also examine some of the sensitivities in the analysis to help us to understand the most significant factors shaping these outcomes. Specifically, we examine the effects of altering 5 key variables both individually, whilst holding all other factors constant, and in combination;

1. Increasing (and reducing) energy prices from the central to the high and low energy price forecasts provided by DECC (2012b);

2. Increasing (and reducing) the difference or spread between the SPV's borrowing and threshold IRR to 2% (and 0%);

3. Increasing (and reducing) performance gaps from 41.6% to 50% (and 30%);

4. Increasing (and reducing) the level of savings retained by householders whilst loans to the SPV are being repaid from 25% to 30% (and 20%) of estimated savings; and

5. Increasing (and reducing) the level of household defaults from 5% to 7% (and 3%).

Finally, to explore the potential for the fund to run in different ways, using the default settings we examine the impacts of three key variants in the design and operation of the fund on investment needs, financial viability and on the carbon savings it could achieve.

The first variant considers the implications of only funding measures with a particular payback period. Rather than funding all of the 23 measures mentioned above, we limit the measures that can be funded to those with 15, 10 and 5 year payback periods. Whilst a fund seeking maximum carbon savings might fund all measures, a fund that seeks to minimise investment needs or

maximise payback may focus only on measures with short payback periods.

The second variant considers the impacts of changing the af-fordability of different inducements to householders. As participation in domestic retrofit schemes may be restricted by short-termism and risk aversity, we consider the implications of a voucher scheme that gives householders 10% of the total costs of the measures installed as 'cash-back' when they agree to participate. This is in addition to the 25% of the value of the energy savings that they are assumed to retain during the repayment period and the 100% of savings that they retain thereafter. To find ways of paying for such inducements, we also consider the implications of asking householders to pay back 10% of the energy savings estimated after the repayment period but for a maximum of ten years from the date of the initial investment.

The third variant considers the implications of assuming that households only pay back the capital that they have received, with no additional financing component. In this case, the government provides a subsidy to cover the financing costs of the SPV. In each period, the government thus pays interest on the outstanding capital component of each measure to the SPV at the threshold IRR rate. The case for such a subsidy comes from the public interest benefits that such investments would realise - for example in helping to reduce fuel poverty, improve public heath and reduce carbon emissions.

3. Results

3.1. Main findings

Under the default settings set out above, the results show that the total funding required to exploit the full realistic potential of all 23 measures shown in Table 1 across the UK would be £33.7 billion. Obviously this is a very substantial level of investment, but critically the results suggest that while £24.8 billion of this total would need to come from new capital, £8.9 billion could come from recycled investment based on savings that were recovered and reinvested. The results show that recycled investment could therefore make up 26.4% of the total investment needs over the lifetime of the fund.

The capital flows associated with this calibration of the model are presented in Fig. 2. This shows that new investment of £1.5 billion of new capital would be required every year for the first 12 years, but that the savings that these investments generate - that for a period of nearly 30 years from year 5 would reach £3 billion per year-would allow the proportion of the total savings captured and reinvested by the SPV to grow to a peak of nearly £1.2 billion in year 5. Critically, all available opportunities for investment in the 23 energy efficiency and low carbon measures in the domestic sector would be fully exploited within 18 years. Annual savings increase after 15 years once expensive measures such as solid wall insulation have been fully exploited. After this point, investment is

10000 9000 8000 j? 7000

a 6000

I 5000 3 4000 J 3000

a 2000 1000

I Annual ^^

400,000

350,000

300,000 £

250,000 & c

200,000 3

150,000 J

100,000 3 50,000

Fig. 2. Profile of investments and savings with default settings.

Fig. 3. Profile of annual and accumulated carbon savings with default settings.

diverted to more cost-effective options with more rapid economic paybacks. Once these investment opportunities had been fully exploited, the SPV would not need either any new capital or to retain and reinvest any of the on-going savings of its earlier investments. This would then allow it to significantly accelerate its repayments to its original investors and to become debt free after 38 years.

The levels of carbon savings arising from these investments are depicted in Fig. 3. The graph shows that these investments would reduce domestic carbon emissions in the UK by up to 9.3 megatonnes (MT) per year, which equates to c6.7% of their 2012 level. In 2012, the average level of emissions per home was 5 t (DECC, 2013a), and so the peak annual savings generated from these investments would be equivalent to the 2012 level of emissions from 1.86 million homes. This level of reduction would be approached after 12 years when most investments have been made - however it would decline over time as the carbon intensity of electricity falls and as some of the measures reach the end of their operating life; a feature that is particularly evident after 40 years. The results also show that over the lifetime of the investments, total carbon savings of 353 MT would be generated.

3.2. Sensitivity analysis

The results of the sensitivity analysis described above are set out in Table 3. As can be seen, increasing energy prices, widening the spread between borrowing and the SPV threshold IRR, and lowering performance gaps all result in shorter payback times from the SPV to the investors. These are now 32, 34 and 30 years respectively. However, increasing the threshold IRR may have a significant effect on householder participation. There may therefore be a case for public support for the government to make financial support that enhances the viability of the revolving fund without reducing levels of participation. We consider this further below. Reducing performance gaps reduces new lending requirements by £3 bn, and increases recycled investment levels by the same amount. In contrast, lowering energy prices, decreasing the spread between borrowing rate and the threshold IRR, or increasing performance gaps all have the effect of leaving the SPV unable to fully repay investors.

Each of the other variations has a smaller impact on all outputs, but varying them all together towards a best business case scenario reduces new lending requirements by £6.6 billion, increases recycled investment by the same level, increases peak carbon savings by 2.8 MT per year and reduces the time taken for the SPV to repay all loans to investors by 15 years. Comparative outputs for the worst-case scenario are an increase of £4.6 billion in new capital requirements, a decrease in recycled investment of the same level, a decrease in peak carbon savings of 2 MT per year and a failure of the SPV to fully repay its debts to investors. This highlights the need for the adoption of risk mitigation measures, perhaps in the form of government underwriting or the introduction of supporting policies.

Table 3

Sensitivities and impacts of changing key variables.

Total investment New lending Recycled investment Peak carbon sav- Lifetime carbon Time for SPV to become

(£bn) (£bn) (£bn) ings (MT) savings (MT) debt free (yrs)

Default settings 33.7 24.8 8.9 9.3 353 38

High (low) energy prices 33.7 22.5 (27) 11.2 (6.7) 9.4 (9.2) 354 (353) 32 (-)

Increasing (reducing) spread to 2% 33.7 24.5 (25.1) 9.2 (8.6) 9.3 (9.3) 353 (353) 34 (-)

Reducing (increasing) performance 33.7 21.8 (27.2) 11.9 (6.5) 11.9 (7.4) 448 (284) 30 (-)

gaps to 30% (50%)

Reducing (increasing) incentives to 33.7 24.2 (25.7) 9.5 (8) 9.3 (9.3) 354 (353) 35 (42)

20% (30%)

Reducing (increasing) defaults to 3% 33.7 24.4 (25.5) 9.3 (8.2) 9.3 (9.3) 353 (353) 36 (41)

Overall best (worst) case scenario 33.7 18.2 (29.4) 15.5 (4.3) 12.1 (7.3) 449 (283) 23 (-)

3.3. Wider options

After reverting to the default settings, we also examine the impacts of the other changes outlined above. The impacts of these changes are presented in Table 4. Total investment costs could be reduced if the range of measures that could be funded was restricted to those with paybacks of 15,10 or 5 years. If the fund only invested in measures with a 5 year payback period or less, total investment costs could be cut from £33.7 bn to £8.8 bn, but the lifetime carbon savings generated would be more than halved from 354 to 175 MT. If a voucher scheme was adopted to encourage participation, total investment costs would increase by £3.4 billion, and new lending needs would increase by £3.8 bn, however this increase could be partly offset by asking householders to pay back 10% of the energy savings realised after the repayment period but for a maximum of ten years from the date of the initial loan as this would save £2 billion in new investment needs.

If the government were to fully subsidise the financing requirements of the SPV, this would significantly decrease new lending requirements (by £4.3 bn) whilst increasing recycled investment by an equal amount. The time for the householders to repay the SPV would decrease to 24 years, at a cost to the government of £6.3 bn, as can be seen in Fig. 4. However, this would have unintended consequences for the long-term viability of the fund. This is because the subsidies from government to the SPV would be conditional on the repayments from householders to the SPV not including a financing component. This would shorten payback times and reduce income to the SPV, and this would alter the cash flow of the SPV to such an extent that it could not cover its debts. At its lowest level after 24 years, the SPV's outstanding debt would be £7 bn. If this debt was also to be paid by the government, then the total cost of the subsidy from government would approach £14 bn. This is obviously a very considerable amount, both in aggregate and as a proportion of the £33.7 bn total investment needed to fund the retrofit scheme in its entirety.

Fig. 4. Profile of investments and savings with government paying interest.

4. Discussion

The results presented above have important implications that relate to the specific case of revolving funds, to wider discussions about climate change finance and investment, and to much broader debates on the changing roles and limited capacities of the state, particularly in an age of austerity.

Specifically, the analysis has highlighted the need, or alternatively the opportunity, for very significant levels of investment to be undertaken that improve the energy efficiency and reduce the carbon emissions of the UK housing stock. The results show that the total level of funding needed to deploy all 23 measures included in the analysis across the entire UK housing stock is £33.7 billion. Obviously this is a vast amount of money, but it is worth pointing out that it represents 1.4% of the UK's 2012 level of GDP and that the investment could be spread over many years.

Perhaps more importantly, the analysis has shown that with a revolving fund such an extensive domestic sector retrofit scheme could be made essentially cost-neutral, albeit with significant up-front investments that would only pay for themselves over an extended period of time as energy savings come through. It has also shown that the up-front investment costs of such a scheme could be significantly reduced (in this case by £8.9 billion) through

Table 4

Variations in programme design.

Total investment (£bn) New lending (£bn) Recycled investment (£bn) Peak carbon savings (MT) Lifetime carbon savings (MT) Time for SPV to become debt free (yrs)

Default settings 33.7 24.9 8.8 9.2 354 38

Payback <15yrs 32.7 22.9 9.8 9.2 345 40

Payback <10yrs 15 7.9 7.1 5.7 186 33

Payback < 5yrs 8.8 3.9 4.9 5.2 175 20

Voucher scheme 37.1 28.7 8.4 9 352 45

Voucher scheme plus 10yr 10% 37.1 26.7 10.4 9.1 353 41

payback

Government subsidised interest 33.7 20.6 13.1 9.4 354 24

(cost to Govt. is £6.2bn)

the creation of a revolving fund. The analysis has also revealed the extent of the reduction in carbon emissions that could be generated through such investments. Although in this case a 6.7% drop in domestic emissions is a relatively small reduction, the carbon savings are equivalent to the emissions of 1.86 million houses. Previous research has found that the impacts of implementing the key energy efficiency and small scale renewable measures included in this study have been higher than forecast (see Webber et al., 2015).

It is also important to point out that the results are based on figures that assume the financing and deployment of a sub-set of the wider range of energy efficiency and low carbon measures available for the domestic sector and that a large proportion of the savings that are technically achievable from these measures are lost through significant performance gaps and rebound effects. Indeed, we see elsewhere in the analysis that addressing performance gaps can have a very significant effect on the financial viability of the scheme by reducing new investment needs and replacing them with recycled investments. If initiatives to reduce performance gaps were combined with, for example, subsidised lending from the SPV, then the results suggest that further substantial reductions in new investment costs could be achieved but at a very substantial cost to the government.

The results have also highlighted the ways in which the design of a revolving fund could be varied to make it more financially viable and less risky, and to further incentivise participation in such a scheme. We have shown that focusing a fund so that it only invests in measures with short payback periods might make the scheme much more financially attractive, but that this would be achieved at the expense of significant cuts in the carbon savings that can be realised. More positively, the results show that there could be innovative ways of incentivising householder participation that again could be cost-neutral over time. Other forms of policy support for a revolving fund could also be envisaged; governments could enable the emergence and effective operation of revolving funds by underwriting their borrowing to reduce default risk and creating stable policy climates to further reduce levels of uncertainty. They could also establish supporting incentive structures and provide guarantees, assurance and supporting regulations that encourage participation and ensure that public interest objectives are met.

All of these findings are also relevant to broader policy debates on climate finance and investment in low carbon development. The results suggest that the investment needs and the associated investment gaps discussed at the beginning of this paper might be significantly reduced if revolving funds were widely deployed. However, given the lack of research on climate finance in general and revolving funds in particular, more work needs to be done to consider the contexts in which and the extent to which revolving funds can make a useful contribution. An important issue relates to whether revolving funds have the ability to fund 'deep retrofit' initiatives that deliver much more significant cuts in energy use and carbon emissions than those demonstrated here. There are also questions about whether, how and to what extent revolving funds could be applied to new buildings as well as the retrofit of existing buildings, to commercial and industrial as well as domestic buildings and to other sectors such as energy or transport. And more broadly there are important issues about the extent to which revolving funds could be adopted, the ways in which they should be designed and governed and the extent to which they can perform in different social, political and economic conditions.

More generally, the broad finding that revolving funds could enable major public interest programmes to be undertaken in ways that both significantly reduce the need for new investment and that even render substantial programmes cost-neutral over time could be of great importance. This finding suggests that

innovative financing arrangements such as revolving funds could enable states with limited capacities and resources to act in contexts and on issues where action might otherwise be impossible. Pragmatically, therefore, it seems that revolving funds could have massive potential, particularly in an era of austerity.

The model presented here allows us to start to consider the potential contribution of different sources and types of finance. For example, if the revolving fund is financed by the private sector, this is likely to require the interest rate to be higher than if the capital was provided by the public sector and it may also require the capital to be repaid over a shorter period (Sullivan et al., 2013). In relation to this latter point, we note that there is significant institutional investor interest in long-term (20 + years) investment structures and opportunities (see, for example, OECD, 2013), and so there may be good alignment between these investors' timeframes and the investment and repayment periods of such a fund. Similarly, when we look at the role of government funding, the model allows us to explore how different funding interventions may be deployed. For example, government funding could be used to provide some level of insurance against defaults, to reduce the cost of capital, to reduce the level of capital required (through debt or equity financing), to provide direct incentives to households, or to cover some of the setup and transaction costs for the SPV or the installers. In each case, the effect of these on the revolving fund's cash flows, revenues, profits and repayments can be explicitly modelled and assessed. It also enables the financial impacts on the other actors involved, which may include central and local government, capital providers, energy providers and service companies, businesses and households, to be assessed. Finally, the model enables the effect of transaction costs such as SPV fees and set-up costs on investment returns and capital deployment to be evaluated.

Returning to our earlier comment about the financing of energy efficiency in an era of austerity, the analysis presented here suggests that central government can help stimulate these investments in a whole host of ways and at much lower cost than central government looking to actually fund such investments directly. It suggests that governments can encourage and provide practical support to local authorities, community groups or other bodies looking to invest in energy efficiency within their areas, and that government can help address some of the key investment risks associated with these investments (see Sullivan, 2011, Sullivan et al., 2013) and thereby encourage the private sector to invest in energy efficiency. Within this, it is important to recognise that many investment risks relate to, or are informed by, investors' views on the dependability of public policy measures such as carbon taxes and their views on governments' long-term commitment to action on climate change and energy efficiency. The fact that central and/or local government provides some level of financial support for the revolving fund is, in itself, likely to be seen as mitigating some of the financial risks and, in turn, may see private investors being happy to receive lower rates of return on their investments.

While the broad case for using revolving funds to support the financing of energy efficiency is clear, they also raise profound questions around governance, accountability and legitimacy. Some of these issues are critically important in public policy debates: measures to mitigate financial risks may be seen as effectively being a form of subsidy to the private sector; a limited focus on economically attractive options may mean that governments do not achieve their wider climate change commitments; the involvement of central government or of the private sector may limit the autonomy of local government or community groups to develop funds that deliver on wider local priorities and concerns.

More generally, there is also what many would see as a danger that revolving funds - and the greater involvement of non-state

actors in providing the capital needed for the delivery of public goods benefits such as reducing greenhouse gas emissions or improving energy efficiency and strengthening energy security and resilience-could be used by some to argue for the further curtailment of the powers of the state. This is particularly relevant in contexts where, in a climate of austerity or neoliberalism, many governments are moving away from being the provider of public goods and instead acting as facilitators or enablers for other private or civic actors to deliver these public goods (Gouldson and Sullivan, 2014). Whether or not this is a politically desirable outcome is to some extent open to debate, but it is important to note that some important variables considered in the analysis above are likely to vary depending on how a scheme is both funded and governed. A privately funded, profit-driven scheme that is poorly supported or loosely regulated is likely to have a narrower focus and to generate lower levels of saving than a publicly or civically funded, not-for-profit scheme that is more actively supported or more closely regulated to ensure that it generates wider social and environmental benefits. One of the contributions of this article -and a focus of our on-going research-is that the revolving funds model presented enables the financial costs and benefits and the public goods outcomes (in terms of energy savings or greenhouse gas emission reductions) to be explicitly identified and assessed, thereby enabling a more informed debate on the costs, benefits and implications of this type of innovative financing mechanism.

5. Conclusions and policy implications

Although there is a strong economic case for climate action at the global scale, in many settings the ability of both public and private actors to invest in low carbon development has been curtailed in recent years. Globally, levels of investment fall a long way short of those required if dangerous climate change is to be avoided. There is therefore a pressing need to explore the potential of innovative financing arrangements that might stimulate flows of capital into, and ideally also to reduce the costs and increase the effectiveness of, climate action.

This paper has explored the contribution that revolving funds might make in this respect, with a particular focus on their ability to finance improvements in the energy performance of the buildings sector. To explore the practical contribution that revolving funds could make to solving this problem, the paper proposed a generic model of a revolving fund and applied it based on the specific case of the domestic sector in the UK. The finding that a revolving fund could reduce the cost of an extensive retrofit programme by 26%, or £9 billion, is significant in itself as it increases the chance that governments with scarce resources will intervene to promote retrofit. Obviously we should be very cautious when extrapolating from one case, but if a similar scale of reduction could be made to the US$31 trillion that the IEA forecasts will be required to promote energy efficiency in buildings over the next 40 years, then the savings would equate to US$8 trillion. This figure gives a broad indication of the scale of the contribution that revolving funds could make if they were widely adopted to finance the retrofit of the buildings sector.

But the finding that the entire retrofit programme could be make cost-neutral over an extended period perhaps has more profound policy implications as it suggests that retrofit could be enabled by government but financed and delivered by private or civic actors. This finding is very much in line with broader debates on governance that argue that in many settings we are witnessing a move away from the position where the state is relied upon to provide public services and toward a position where states facilitate or enable private or civic actors to provide those services. This implies a radical shift in the role of government and in the nature

of the relationship between the public, private and civic sectors.

The enabling role of government could be pivotal if revolving funds and other innovative financing mechanisms are to fulfil their apparent potential. The experience to date with revolving funds is limited, and there is therefore a need to strengthen the evidence base on how they function, on their costs and benefits, and on the financial risk management and mitigation measures that may be adopted to improve their performance and on how these affect the financial and energy saving outcomes that are seen. There is also a need to look much more explicitly at the governance of these funds and the associated delivery vehicles, at both the macro level (i.e. do they accelerate the withdrawal of the state from certain types of public good provision) and at the local level (i.e. how do they affect local needs and interests, how do they affect local governance). These factors deserve urgent attention given the apparent potential of revolving funds to close the climate finance gap - and government support for pilot and demonstration schemes is probably required given the reluctance of private and civic actors to invest in novel and perhaps more risky schemes. If they can be properly proven, then the analysis presented in this paper suggests that revolving funds clearly have the potential to enable intensified and accelerated action on climate change, even in an era of austerity.

Acknowledgements

The authors gratefully acknowledge the financial support for the research that led to this paper that was provided by the EU (Grant No. 266800) through the FESSUD project and by the EPSRC and ESRC under the IBuild Programme (Grant No. EP/K012398/1). Support was also provided by the ESRC through the Centre for Climate Change Economics and Policy (Grant No. ES/K006576/1).

References

BES, undated. Birmingham Energy Savers, Birmingham City Council and Carillion Energy Services. Available at: <http://www.energysaverspartnerships.co.uk/ birmingham).

BRE, 2008. MAC Curves for the Domestic and Non-Domestic Building Sectors: Technical Documentation, Buildings Research Establishment, Watford.

Buchner, B., et al. 2013. The Global Landscape of Climate Finance 2013, Climate Policy Initiative.

CCC, 2014. Meeting Carbon Budgets-Progress Report to Parliament 2014. UK Committee on Climate Change.

DECC, 2012a. Low Carbon Communities Challenge Evaluation Report, UK Department for Energy and Climate Change.

DECC, 2012b. Updated Energy and Emissions Projections, UK Department for Energy and Climate Change.

DECC, 2012c. Final Stage Impact Assessment for the Green Deal and Energy Company Obligation, UK Department for Energy and Climate Change.

DECC, 2013a. United Kingdom Housing Energy Fact File, UK Department for Energy and Climate Change.

DECC, 2013b. Energy Consumption in the UK, UK Department for Energy and Climate Change.

Dowson, M., Poole, A., Harrison, D., Gideon, S., 2012. Domestic UK retrofit challenge: barriers, incentives and current performance leading into the Green Deal. Energy Policy 50, 294-305.

EC, 2011. A Roadmap for Moving to a Competitive Low Carbon Economy in 2050, C0M/2011/0112 final, European Commission.

EC, 2013. Financial Support for Energy Efficiency in Buildings, Report from the Commission to the European Parliament and the Council, C0M(2013)225 final, European Commission.

ECF, 2010. Roadmap 2050: A practical guide to a prosperous, low carbon Europe, European Climate Foundation.

EEFIG, 2015. Energy Efficiency - The First Fuel for the EU Economy: How to Drive new Finance for Energy Efficiency Investments, Energy Efficiency Financial Institutions Group.

EU, 2012. Directive 2012/27/EU of the European Parliament and Council of 25 October 2012 on energy efficiency, European Union.

Forum for the Future, 2011. Funding Revolution: A guide to Establishing and Running Low Carbon Community Revolving Funds, Forum for the Future.

GEA, 2012. Global Energy Assessment-Toward a Sustainable Future, Cambridge

University Press, Cambridge, UK and New York, NY, USA and the International Institute for Applied Systems Analysis, Laxenburg, Austria.

Gouldson, A., Sullivan, R., 2014. Understanding the governance of corporations: an examination of the factors shaping UK supermarket strategies on climate change. Environ. Plan. A 46, 2972-2990.

Gruning, C., Menzel, C., Panofen, T., Shuford, L., 2012. Case Study: The Thai Energy Efficiency Revolving Fund, UNEP/Frankfurt School.

Holcombe, R., 1992. Revolving fund finance: the case of wastewater treatment. Public Budg. Finance 12 (3), 50-65.

IEA, 2008. Promoting Energy Efficiency Investments: Case studies in the Residential Sector, International Energy Agency, Paris.

IEA, 2013a. World Energy Outlook 2012 - Executive Summary, International Energy Agency, Paris.

IEA, 2013b. Transition to Sustainable Buildings: Strategies and Opportunities to 2050, International Energy Agency, Paris.

IPCC, 2014. 5th Assessment Report-Climate Change 2014: Mitigation of Climate Change, Intergovernmental Panel on Climate Change.

LCR, undated. Housing, Regeneration and Planning, Leeds City Region, Available at: <http://www.leedscityregion.gov.uk/our-work/housing/>.

McCollum, D., Nagai, Y., Riahi, K., Marangoni, G., Calvin, K., Pietzcker, R., van Vliet, J., van der Zwaan, B., 2013. Energy investments under climate policy: a comparison of global models. Clim. Change Econ. 4 (4), 3-32.

McKinsey, 2010. The Impact of the Financial Crisis on Carbon Economics-GHG Cost Curve v2.1., McKinsey and Company.

OECD, 2013. G20/OECD High-Level Principles of Long-Term Investment Financing by Institutional Investors, Organisation for Economic Cooperation and Development.

O'Toole, L., 1996. Hollowing the infrastructure: revolving loan programs and network dynamics in the American States. J. Public Adm. Res. Theory 6 (2), 225-242.

Rosenow, J., Eyre, N., 2013. The Green Deal and the energy company obligation. Proc. ICE - Energy 166 (3), 127-136.

SALIX, 2015. Solving Energy Efficiency Finance in the Public Sector, Salix Finance. Available at: <http://salixfinance.co.uk/about-us/successes>.

SEI, 2015. Sustainable Endowments Institute. Available at: <http://www.en dowmentinstitute.org>.

Stern, N., 2007. The Economics of Climate Change. Cambridge University Press, Cambridge.

Sullivan, R., 2011. Investment-Grade Climate Change Policy: Financing the Transition to the Low-Carbon Economy. Institutional Investors Group on Climate Change, London, UK.

Sullivan, R., Gouldson, A., Webber, P., 2013. Funding low carbon cities: local perspectives on opportunities and risks. Clim. Policy 13 (4), 514-529.

Sweatman, P., Managan, K., 2010. Financing Energy Efficiency Building Retrofits: International Policy and Business Model Review and Regulatory Alternatives for Spain, Climate Strategy and Partners. Available at: <http://www.climatestrategy. es/index.php?id = 19>.

USEPA, 2015. Clean Water State Revolving Fund: 25 years of Investment in our Nation's Water Infrastructure, United States Environmental Protection Agency.

Webber, P., Gouldson, A., Kerr, N., 2015. The impacts of household retrofit and domestic energy efficiency schemes: a large scale, ex-post evaluation. Energy Policy 84, 35-43.

WEF, 2013. The Green Investment Report: The ways and means to unlock private finance for green growth. World Economic Forum, Geneva.