Getting the measure of fuel poverty

Final Report of the Fuel Poverty Review Professor John Hills 15 March 2012

SLIDE 2

Approaching the Terms of Reference

Interim report

(1) Whether fuel poverty is a distinct problem (2) If so, how fuel poverty is best measured and does the current approach to measurement capture problems effectively?

Final report

Final conclusions on (1) and (2) above AND (3) Implications of measurement for the way we understand the effectiveness of the range of policy approaches to reducing it

SLIDE 3

The Review so far

Fuel poverty as a distinct problem

We found that fuel poverty is a distinct issue and of concern from at least three different perspectives. Fuel poverty is an additional problem for some low-income households.

Poverty

-Capital investments

out of reach for some

- Potential obstacle

to carbon mitigation policy delivery, especially where costs go on bills

Carbon / Energy

- Unequal ability to convert

cash to warmth - Pushed into poverty by high costs - Poor pay more

Health

-High rate of EWDs and

morbidity issues in general - Mental health and social well-being - Social isolation

SLIDE 4 THE REVIEW SO FAR

Warm Homes and Energy Conservation Act 2000

“A person is to be regarded as living “in fuel poverty” if he is a member of a household living on a lower income in a home which cannot be kept warm at reasonable cost.”

The Act establishes a duty to produce a strategy setting out how the following objective will be met:

“As far as reasonably practicable persons in England or Wales do not live in fuel poverty.”

This definition correctly focuses on the overlap of low income and unreasonable cost but this is

not

what we currently measure.

SLIDE 5 THE CURRENT OFFICIAL INDICATOR A fuel poor household is one that would need to spend more than 10% of its income on adequate warmth.

The indicator is rightly based on modelled needs. But it is fundamentally flawed because it misrepresents trends, includes some households that are not low income, does not show policy impacts very clearly and is sensitive to technical issues.

SLIDE 6 THE CURRENT OFFICIAL INDICATOR The current definition captures some relatively well-off households with high costs, especially as prices rise.

The answer is not simply to put an income threshold into the current official indicator.

SLIDE 7 THE CURRENT OFFICIAL INDICATOR We therefore recommend that the Government should change its approach to fuel poverty measurement away from the current ‘10 per cent’ ratio indicator.

SLIDE 8 LOW INCOME HIGH COSTS INDICATOR AND FUEL POVERTY GAP

Our alternative: Low Income and High Costs indicator and the fuel poverty gap

A fuel poor household is one that has both high modelled costs and low income.

The fuel poverty gap is the required reduction in modelled costs to take a household out of fuel poverty.

Together, the indicators show both the extent and depth of fuel poverty (rather than conflating them)

SLIDE 9 LOW INCOME HIGH COSTS INDICATOR AND FUEL POVERTY GAP

Setting the thresholds for low income and high costs

Each threshold is adjusted annually, making the indicator relative over time. It tracks the position of each household compared to current norms and creates a kind of moving target. The extent of fuel poverty depends on how costs change for all households not just the fuel poor.

This approach captures the risk that the poor could lose ground as housing improves in general.

SLIDE 10 LOW INCOME HIGH COSTS INDICATOR AND FUEL POVERTY GAP Our interim report was a consultative document.

There was broad

support

for: - our conclusions on the causes and impacts of fuel poverty - our view that the core problem is

correctly identified in WHECA

- the general approach represented by LIHC indicator and the fuel

poverty gap

- how to set the income threshold under any LIHC approach There were strong

concerns

about: - the way we measured required costs, especially how we adjusted

costs for household size

- the level of the costs threshold - the relative nature of the costs threshold In preparing our final report we have considered these criticisms very carefully.

SLIDE 11 LOW INCOME HIGH COSTS INDICATOR AND FUEL POVERTY GAP

Adjusting modelled costs for household size and composition

We want to adjust modelled costs so that we can compare all households in England against the threshold. BUT the way we adjusted for household size in our interim report was wrong.

We also received a carefully-developed suggestion to measure costs as unit costs (£/m 2 ) as a way of side-stepping the need to adjust for household size.

On balance we prefer the equivalisation approach and have derived specific new factors to improve the calculation. One reason for preferring our equivalisation approach is that it more closely correlates with SAP. Another is that it is fairer to small households.

Equivalisation Unit Costs

SLIDE 12 LOW INCOME HIGH COSTS INDICATOR

Concerns Threshold is too low

This level is driven by the high levels of energy inefficiency in the housing stock

Threshold is too high

It is very difficult to ensure that zero low-income households have higher-than-typical costs

Response

We consider a range of different ways of setting the threshold, including absolute and relative approaches.

Although after careful reflection we retain our initial approach, the analysis is set out for those who would prefer an alternative.

Target setting

The key indicator should be the scale of the fuel poverty gap. If this is reduced to a low level then no household can be left very far above the threshold. This is preferable to using a fixed standard that is easier to beat, but becomes out of date.

Before After

SLIDE 13 LOW INCOME HIGH COSTS INDICATOR

Fuel poverty under twin indicators, 1996-2009

Under the LIHC indicator, the number of fuel poor households has remained broadly stable over this period.

The fuel poverty gap increased by three quarters between 2003 and 2009.

SLIDE 14 LOW INCOME HIGH COSTS INDICATOR AND FUEL POVERTY GAP The Government should adopt a new indicator of the extent of fuel poverty under which households are considered fuel poor if: - they have required fuel costs that are above the median level; and - were they to spend that amount they would be left with a residual income below the official poverty line.

The Government should count the number of individuals in this position as well as the number of households they live in.

The Government should adopt a new indicator of the depth of fuel poverty as represented by the average and aggregate fuel poverty gap, defined as the amounts by which the assessed energy needs of fuel poor households exceed the threshold for reasonable costs.

SLIDE 15

Applications: Finding the fuel poor

Characteristics

The LIHC indicator lends itself to understanding the distribution of fuel poor households by household and dwelling characteristics.

The example here shows EPC rating. It can be seen that the majority of fuel poor households occupy EFG homes – and account for 90% of the fuel poverty gap.

SLIDE 16 FINDING THE FUEL POOR

Proxies

Understanding household characteristics is only part of the picture.

In practical terms, proxies are needed to identify specific households for assistance. Proxies will hit some of the right people and some of the wrong ones.

The chart shows the hit rate of means-tested benefits as an eligibility criterion. 62% of LIHC households are on such benefits representing 62% of the fuel poverty gap.

SLIDE 17 FINDING THE FUEL POOR

Using the fuel poverty gap

The fuel poverty gap can provide a bridge between targeting and the measurement of fuel poverty.

Importantly, the fuel poverty gap also helps identify those who are deepest in fuel poverty who are priorities for assistance.

SLIDE 18

Applications: Understanding policy

≈ Warm Front

Principles

The impact of a given policy on fuel poverty will depend on three factors: 1. The type of policy (i.e. whether it addresses energy efficiency, income or prices) 2. Who pays for the policy (i.e. customers or taxpayers) 3. Who benefits (i.e. fuel poor households or all households)

≈ CERT ≈ Winter Fuel Payments

SLIDE 19 UNDERSTANDING THE IMPACT OF POLICY

Existing climate and energy package

Current policy spans the three key drivers of fuel poverty. The picture is changing between 2009 and 2016 as shown.

2009 2016

(2009 prices) Taxpayer-funded Consumer-funded

SLIDE 20 UNDERSTANDING THE IMPACT OF POLICY

Winners and losers

The net effect of a policy and group of policies on fuel poverty will depend on precisely who benefits and who pays. There is also a distributional impact.

This kind of trade-off is a live issue with ECO (see Figure) which is currently expected to have a regressive impact. To remove this, one would have to spend more than half (rather than one quarter) of ECO on Affordable Warmth.

SLIDE 21

Applications: Projecting fuel poverty

LIHC indicator and fuel poverty gap

Our projections show an increase under both indicators by 2016 – but one that is lower than it would be in the absence of policies.

On our central projection, the fuel poverty gap is more than 50% higher in 2016 than in 2009 and nearly three times what it was in 2003.

This is 10% lower (but only 10% lower) than it would be in the absence of policies.

SLIDE 22

Applications: Projecting fuel poverty

LIHC indicator and fuel poverty gap

Our projections show an increase under both indicators by 2016 – but one that is lower than it would be in the absence of policies.

On our central projection, the fuel poverty gap is more than 50% higher in 2016 than in 2009 and nearly three times what it was in 2003.

This is 10% lower (but only 10% lower) than it would be in the absence of policies.

SLIDE 23 PROJECTIONS OF FUEL POVERTY

The impact of individual policies

The projections provide details on the impacts of individual policies.

Some (shown in red) move in the wrong way from a fuel poverty perspective either in terms of the fuel poverty gap of the headcount measure or both. Others (shown in green) move in the right way.

The overall effect is in the right direction but the scale is disappointing.

Note: vertical axes have been adjusted.

SLIDE 24 PROJECTIONS OF FUEL POVERTY

Comparing indicators

Compared to the official indicator, the LIHC indicator is much more stable in the number of households affected and less unduly sensitive to fuel price assumptions.

The fuel poverty gap (as shown previously) is sensitive to price changes, like the current indicator. This seems appropriate. The main impact of sharp price rises is to deepen fuel poverty rather than make the core problem much more widespread.

SLIDE 25

Applications: Making further progress

Policy archetypes

- Bill rebate* - Narrowly-targeted energy efficiency policy* - Broadly-targeted energy efficiency policy* - Increase in means-tested benefits - Increase in Winter Fuel Payment

* For these policies we have modelled both Exchequer- and supplier-funded variants

Modelling We spend £500 million on each intervention in 2016.

These are stylised scenarios with standardised inputs to allow comparison of effects against key indicators.

SLIDE 26 MAKING FURTHER PROGRESS

Results - eligibility by household status

Narrowly targeted supplier-funded energy efficiency archetype

This is analogous in some ways to ECO affordable warmth.

Broadly targeted supplier-funded energy efficiency archetype

This is analogous in some ways to ECO carbon.

Energy bill rebate Winter Fuel Payments

SLIDE 27 MAKING FURTHER PROGRESS

Results – fuel poverty gap impacts

Archetype Supplier-funded, narrowly targeted energy efficiency Exchequer-funded, narrowly targeted energy efficiency Exchequer-funded, broadly targeted energy efficiency Supplier-funded broadly targeted energy efficiency Exchequer-funded rebate policy Short term change in fuel poverty gap (£ million) -50 -70 -20 +20 -70 Supplier-funded rebate policy -40 Increase in Means-Tested Benefits Increase in Winter Fuel Payment -3 <-1 Life-time change in fuel poverty gap (£ million) -2,930 -2,630 -680 -390 -70 -40 -3 <-1

The results show that narrowly targeted energy efficiency policies have the biggest life-time effect.

In the short term, rebates can make a difference. But these only have one off value.

SLIDE 28 MAKING FURTHER PROGRESS

Results – carbon emissions and cost benefit analysis

Archetype Supplier-funded, narrowly targeted energy efficiency Exchequer-funded, narrowly targeted energy efficiency Exchequer-funded, broadly targeted energy efficiency Supplier-funded broadly targeted energy efficiency Exchequer-funded rebate policy Supplier-funded rebate policy Increase in Means-Tested Benefits Increase in Winter Fuel Payment Total change in carbon emissions (MtCO 2 ) -4.92

-3.40

-3.76

-6.76

+0.58

+0.35

<+0.01

+0.58

Non equity weighted NPV (£ million) 590 310 360 990 50 100 <10 60 Equity weighted NPV (£ million) 1,900 1,730 860 1,360 600 490 550 420

The results show that energy efficiency policies have the best impact on carbon emissions. Bill support leads to increased emissions.

The highest net returns come from narrowly targeted energy efficiency policies when equity-weighted. Even without such weighting, all energy efficiency policies provide substantial net benefits to society.

SLIDE 29 MAKING FURTHER PROGRESS

Results of analysis of interventions

Archetype Supplier-funded, narrowly targeted energy efficiency Exchequer-funded, narrowly targeted energy efficiency Exchequer-funded, broadly targeted energy efficiency Supplier-funded broadly targeted energy efficiency Exchequer-funded rebate policy Proportion of recipients that are LIHC (%) 55 Short term change in fuel poverty gap (£ million) -50 55 18 13 28 -70 -20 +20 -70 Life-time change in fuel poverty gap (£ million) -2,930 -2,630 -680 -390 -70 Supplier-funded rebate policy 28 -40 -40 Increase in Means-Tested Benefits Increase in Winter Fuel Payment 28 10 -3 <-1 -3 <-1 Total change in carbon emissions (MtCO 2 ) -4.92

-3.40

-3.76

-6.76

+0.58

+0.35

<+0.01

+0.58

Non equity weighted NPV (£ million) 590 310 360 990 50 100 <10 60 Equity weighted NPV (£ million) 1,900 1,730 860 1,360 600 490 550 420

SLIDE 30

Conclusion

Fulfilling our ToR

1. Is fuel poverty distinct from general poverty? Yes – it is a serious problem and appears set to rise 2. What does this imply for measurement?

3. How can measurement help effective policy making?

The current indicator is

flawed. It would be better to focus directly on the overlap of having both low income and high costs and to separate the measurement of extent and depth.

The LIHC indicator provides a framework for

analysis. It flags priorities for action, opens up tools for targeting and highlights risks and trade-offs.

We also make a number of technical recommendations.

SLIDE 31 CONCLUSIONS AND RECOMMENDATION

RECOMMENDATION

The Government – not just DECC but also other Departments – should set out a renewed and ambitious strategy for tackling fuel poverty, reflecting the challenges we lay out in the review’s reports and the framework we have set out for understanding them.