Scottish House Condition Survey: Key Findings 2011

4 Fuel Poverty

82. The term 'Fuel Poverty' refers to the situation where a household cannot afford to heat their home to an adequate level. The Scottish Government uses the following definition of fuel poverty as set out in the Scottish Fuel Poverty Statement (FPS)^[16] published in 2002:

"A household is in fuel poverty if it would be required to spend more than 10% of its income (including Housing Benefit or Income Support for Mortgage Interest) on all household fuel use."

The Scottish Government have pledged to ensure that by November 2016, so far as is reasonably practicable, people are not living in fuel poverty in Scotland.

83. Furthermore 'Extreme Fuel Poverty' is defined as a household having to spend more than 20% of its income on fuel.

84. Fuel poverty statistics are based on a model that calculates the cost of heating a dwelling according to a specified temperature regime and mid-year fuel prices. For a more detailed description of the definition of fuel poverty and analysis of previous years' SHCS fuel poverty statistics see the Fuel Poverty Statement, the SHCS 2002 Fuel Poverty in Scotland Report^[17] and the SHCS Fuel Poverty Report 2003/4^[18]. A technical note on the calculation of Fuel Poverty figures using SHCS data can be found on the SHCS website^[19].

85. An extensive review of the evidence on fuel poverty based on the SHCS data to 2010 is available from the Scottish Government website^[20]. In it we provide further analysis of the factors driving fuel poverty, a detailed profile of the fuel poor, and examine alternative definitions mooted by the Hills Review of Fuel Poverty^[21].

86. In each of the eight continuous survey years there were a number of cases where it was not possible to determine fuel poverty status. These have been apportioned on a pro-rata basis between the two categories (or three categories in the case of extreme fuel poverty) as was discussed in the 2004/5 SHCS Key Findings Report^[22]. 73 'missing' cases were reapportioned in 2011. Households with negative annual household income after council tax deduction are removed from the fuel poverty calculations but reapportioned in this way.

4.1 Trends and Drivers

87. Figure 9 and Table 23 show fuel poverty rates from 2002. The number of fuel poor households and the fuel poverty rate include those in extreme fuel poverty.

88. Latest figures for 2011 are shown with respect to two sets of fuel prices: mid-year July 2011, in line with previous annual statistics and for October 2011, showing the effects of large fuel price rises in the autumn of 2011. For more information see page 16.

89. From 2002 levels (293,000 or 13%) fuel poverty increased to a peak in 2009 (770,000 or 33%). Rates then decreased for 2 years between 2009 and July 2011, to a level of 582,000 households (25%). Following the sharp fuel price increases in autumn 2011, fuel poverty increased to 684,000 (29%) in October 2011.

90. Trends in the number and percentage of households in extreme fuel poverty follow a very similar pattern, with the latest estimate of 185,000 (8%) in October 2011.

Table 23 Fuel poverty and extreme fuel poverty 1996-2011 (000s and %)

Year	Fuel poverty		Extreme fuel poverty
	000's	%	000's	%
2002	293	13.4	71	3.2
2003/4	350	15.4	112	4.9
2004/5	419	18.2	119	5.2
2005/6	543	23.5	173	7.5
2007	586	25.3	172	7.4
2008	618	26.5	182	7.8
2009	770	32.7	243	10.3
2010	658	27.9	185	7.8
2011 (July Price)	582	24.6	153	6.4
2011 (Oct Price)	684	28.9	185	7.8

Source: SHCS 1996, 2002 and Continuous SHCS 2003/4 - 2011

Figure 9 Households in fuel poverty 1996-2011 (%)

Source: SHCS 1996, 2002 and Continuous SHCS

91. The 3 main factors driving changes fuel poverty rates are fuel prices, income and energy efficiency of the home^[23].

4.1.1 Fuel Prices

92. Table 24 shows the change in the 3 main fuel prices between May 1996 and July 2012 against changes in fuel poverty. This table includes an average fuel price figure weighted by the proportion of households using each of those fuels for their home heating as recorded by the SHCS.

93. Since 2003/4, fuel poverty rates have been estimated based on mid-year fuel prices. Due to a steep fuel price rise in autumn 2011 we have included additional estimates at this higher rate. Figure 10 shows how fuel prices moved on a quarterly basis since 2003.^[24]

Table 24 Retail Price Index fuel components, May 1996 to June 2012

Fuel price index numbers relative to the GDP deflator
Year	% Fuel poor	Gas	Electricity	Heating oils	Weighted average
2002	13.4	86.0	91.1	67.1	83.3
2004	15.4	88.4	91.9	79.4	85.9
2005	18.2	97.3	98.4	91.0	94.4
2006	23.5	126.5	117.7	115.1	121.1
2007	25.3	130.2	122.9	107.8	127.2
2008	26.5	142.9	134.5	187.5	145.3
2009	32.7	173.8	141.4	107.0	162.7
2010	27.9	158.9	137.0	137.1	153.9
July 2011	24.6	164.5	139.0	171.8	161.5
Oct 2011	28.9	191.0	152.1	168.8	183.5
2012		186.0	146.7	159.7	178.1

Source: Continuous SHCS; DECC RPI Fuel Components^[25]

Figure 10 RPI fuel price components from 2003 relative to the GDP deflator

Source: DECC RPI Fuel Components, Continuous SHCS

94. These figures show that, in general, between 2002 and 2009 fuel prices and fuel poverty rose in parallel. Between 2009 and 2010 fuel prices for all fuels except heating oils dipped, which is reflected in a fall in fuel poverty rates.

95. In July 2011, although the fuel price index swung back towards 2009 levels, there was no equivalent rise in fuel poverty. Until early autumn this was partly due to the disproportionate burden borne by users of other fuels, who saw a 34 point increase, compared to gas and electricity users (6 and 2 point increases respectively).

96. The autumn 2011 price rises were more widely distributed and include a 27 point increase in the cost of gas from July; a far more commonly used fuel. As a result, fuel poverty rates increased in October 2011.

4.1.2 Income and Energy Efficiency

97. Trends in fuel price and fuel poverty levels are shown alongside corresponding changes in income and energy efficiency in Figure 11^[26].

98. The SHCS samples drawn in 2010 and 2011 show an increase in average household income over this period, with the mean annual household income increasing from £23,000 to £24,500 and the median growing from £18,100 to £19,600^[27]. This increase is seen right across the income distribution. The estimated level of fuel poverty is primarily affected by changes in the bottom half of the distribution.

99. It is important to recognise some limitations to the income information collected in the SHCS. It covers net income from all sources for the highest income householder and his or her partner. It will therefore underestimate total income in households where more than 2 adults receive income from some source. This affects disproportionately the upper half of the distribution.^[28] The data collection methods and design of the survey are not geared to the production of the most robust estimates of household income change and because of incomplete income information some imputation of income components is usually undertaken^[29].

100. The examination of measures of central tendency shows that for the last decade the SHCS median has tracked well data from the Family Resources Survey (FRS), the official source for household income information, except for the period 2007 - 2009 where it showed slight reductions compared to the increase that would be expected in line with FRS trends.

Table 25 Change in mean annual income by income decile group, 2010-2011

Income Decile Group	Year		Percentage Change
	2010	2011
1	5,400	6,100	13%
2	9,300	9,900	6%
3	11,900	12,500	4%
4	14,400	15,000	4%
5	16,800	18,100	8%
6	20,000	21,600	8%
7	24,600	26,100	6%
8	30,900	32,100	4%
9	39,400	40,400	2%
10	60,900	63,000	4%
All	23,300	24,500	5%

Source: Continuous SHCS

101. As shown in Chapter 3, and alongside other fuel poverty drivers in Figure 11, we have seen continuing improvement in energy efficiency upgrades of the housing stock.

Figure 11 Fuel Poverty by energy efficiency and indexed fuel price and median income

Source: Continuous SHCS; DECC RPI Fuel Components

102. To understand the contribution of household income and the improvement of the energy efficiency of the housing stock we attempted to simulate the changes experienced between 2010 and 2011 separately for each set of factors and examined the resulting changes in the estimated level of fuel poverty. We used data from the 2010 and 2011 SHCS and, because each of those represents a fresh sample from the dwellings population of Scotland, our estimates will contain noise arising from sampling variation and should only be seen as an approximation for this particular combination of prices, energy efficiency and income levels. Annex A describes the methodology and results from this exercise, which are also illustrated in Figure 12.

103. There was a reduction in the estimated level of fuel poverty of 76,000 or 3.4% of households between 2010 and July 2011. The conclusion from our simulation is that:

• an increase in average household income contributed a 2.5 percentage points decrease in fuel poverty rates;
• improvements in the energy efficiency of housing contributed a 1.7 percentage points decrease in fuel poverty rates, and
• increases in fuel prices contributed a 0.8 percentage points increase in fuel poverty rates.

104. The corresponding analysis between 2010 and October 2011, when there was an increase in the estimated level of fuel poverty of 26,000 or 1% of households showed:

• income growth contributed a 2.2 percentage points decrease in fuel poverty rates;
• energy efficiency improvements contributed a 1.5 percentage points decrease in fuel poverty rates; and
• increases in fuel prices led to a 4.7 percentage points increase in fuel poverty rates.

105. Overall, our simulation found that for the period between 2010 and 2011 the growth in household income and the improvements in energy efficiency effectively contained the growth of fuel poverty against price increases up until the early autumn of 2011 when there was a step change in the price of fuel. Only about 80 per cent of the latest price increase could be offset by income growth and housing stock improvements, and the rate of fuel poverty went up. Around 60 percent of the offsetting effect was due to household income and around 40 percent can be attributed to changes in the housing stock, at least some of which will constitute energy efficiency upgrades.

Figure 12 Price, income and energy efficiency effects on fuel poverty rate

Source: SHCS 2010, SHCS 2011

4.2 Fuel Poverty, Dwelling and Household Attributes

106. The relative distribution of fuel poverty in 2011 remained very similar to previous years. The reduction in the prevalence of fuel poverty to July 2011 benefitted disproportionately lower income groups: those with weekly incomes below £300 and residents of social housing. These gains were reversed in the autumn and by October 2011 levels had returned close to those experienced in 2010. Groups who seem to have done consistently badly throughout 2011 include households living in higher council tax band properties (Bands F & G), those in dwellings rated 'poor' on the NHER scale and large families (although from a very low base). (Table 26 to Table 29)

107. Detached properties remain the most likely to suffer fuel poverty (157,000 or 33% by October prices). However, whereas in 2010 tenement flats and terraced houses had similar levels of fuel poverty (25%), the latter did not benefit from the reductions to July 2011 (tenements dropped 4 points) and both suffered from the subsequent rise to October. This put 30% of terraced dwelling occupants (141,000) into fuel poverty (Table 26 & Table 27).

108. Inter-war dwellings did not see a reduction in July 2011 but suffered a 5 point increase from 2010 to October 2011. This means that buildings from 1919-1944 are just as likely to be fuel poor as pre-1919 dwellings at 34% (153,000 pre-1919 and 106,000 from 1919-1944) (Table 26 & Table 27).

Figure 13 Fuel poverty rates by selected household and dwelling characteristics (%), 2010 and July and October fuel prices 2011.

Source: SHCS 2010-11

109. Pre-1919 dwellings are still more likely to be in extreme fuel poverty. In October 2011, almost twice as many were extreme fuel poor as for 1919-1944 constructions (52,000 or 12% compared with 27,000 or 9%).

110. This change, alongside a rise in fuel poverty in dwellings built between 1965 and 1982, means that the age dependency seen in previous year breaks down somewhat, with the exception of post-1982 dwelling which continue to perform well. This is likely a result of higher building standards in new buildings which require a minimum energy efficiency (Table 26,Table 27 and Figure 13).

111. Across-the-board price rises to October 2011 meant that changes in fuel poverty were relatively evenly spread. Oil users remain most likely to be fuel poor (50%) and gas-users least (26%) (Table 28).

112. As we might expect, dwellings with lower energy efficiency were more likely to be fuel poor. Households with a poor NHER rating increased in fuel poverty when there was a general downwards trend between 2010 and July 2011. This increase continued through October 2011, resulting in a 8 percentage point increase overall to 48,000 or 66% of households (Table 26, Table 27 and Figure 13).

Table 26 Fuel Poverty by dwelling characteristics (000s)

	July 2011 Prices			October 2011 Prices			Total	Sample size
	Not Fuel Poor	Fuel Poor	Extreme Fuel Poor	Not Fuel Poor	Fuel Poor	Extreme Fuel Poor
Type of Dwelling
Detached	350	157	67	340	166	73	506	779
Semi detached	376	117	22	350	143	29	493	687
Terraced	419	141	31	391	169	39	560	772
Tenement	407	106	18	386	127	26	512	530
Other flats	235	62	15	218	79	19	297	378
Age of Dwelling
Pre-1919	313	135	44	295	153	52	448	564
1919-1944	214	94	22	202	106	27	308	420
1945-1964	393	138	34	363	167	42	530	740
1965-1982	420	150	34	385	186	43	571	754
Post-1982	446	65	19	439	71	20	511	668
Extent of Central Heating
Full	1,714	538	138	1617	635	168	2,252	2,995
Partial	57	24	7	54	27	8	82	110
None	15	20	8	13	21	9	34	40
Primary Heating Fuel
Gas	1,461	397	89	1383	476	108	1,858	2,311
Oil	76	64	25	70	70	29	140	276
Electric	215	102	33	195	122	44	317	470
Other	33	19	6	36	16	4	52	88
NHER Band
Poor	28	45	22	25	48	27	73	134
Moderate	484	273	93	448	309	106	757	1,095
Good	1,274	263	39	1211	327	52	1,537	1,917
Urban/rural
Urban	1,530	446	107	1441	536	131	1,977	2,451
Rural	256	136	46	244	147	54	391	695
Mains Gas Grid
Not on gas grid	126	85	32	120	91	37	211	470
On gas grid	1,658	496	120	1562	592	148	2,154	2,673
All Scotland	1,786	582	153	1050	684	129	2,368	3,146

Source: SHCS 2011

Table 27 Fuel poverty rate by dwelling characteristics (Row %)

	2010	Jul-11	Oct-11	2011 Sample
Type of Dwelling
Detached	34	31	33	779
Semi detached	28	24	29	687
Terraced	25	25	30	772
Tenement	25	21	25	530
Other flats	28	21	27	378
Age of Dwelling
Pre-1919	35	30	34	564
1919-1944	29	31	34	420
1945-1964	31	26	31	740
1965-1982	28	26	33	754
Post-1982	17	13	14	668
Extent of Central Heating
Full	28	24	28	2,995
Partial	36	30	33	110
None	33	57	62	40
Primary Heating Fuel
Gas	25	21	26	2,311
Oil	45	45	50	276
Electric	37	32	38	470
Other	38	36	30	88
NHER Band
Poor	58	62	66	134
Moderate	40	36	41	1,095
Good	20	17	21	1,917
Urban/rural
Urban	26	23	27	2,451
Rural	38	35	38	695
Mains Gas Grid
Not on gas grid	45	40	43	470
On gas grid	26	23	27	2,673
All Scotland	28	26	29	3,146

Source: SHCS 2011

113. By tenure, fuel price rises in October 2011 had the smallest effect on Housing Association and private-rented households compared with 2010 levels. Both saw a three percentage point decrease in fuel poverty from 2010 to 26% for Housing Associations and 24% for private rented (Table 29). Local Authority households remain the most likely to be fuel poor at 33% followed by owner-occupied dwellings at 29%.

114. While most household types saw a fuel poverty reduction to June 2011, large families gained four percentage points on 2010 levels (7% to 11%), reaching 13% in October 2011 (21,000 households).

115. Pensioners households and single adults remain the most vulnerable to fuel poverty. 56% of single pensioners were fuel poor in October 2011, 40% of older smaller households and 39% of single adult households. Pensioner households accounted for 55% of the fuel poor in total (377,000) (Table 28 and Table 29).

116. After a 3 percentage point drop to July 2011, single adult fuel poverty rose again by 7 points to 39% in October 2011 (132,000 households).

117. As expected, there remains a clear correlation between income and fuel poverty. Fuel poverty reductions to July 2011 favoured the lower income groups, however those earning less than £100 per week are modelled to spend such a high proportion of their income on fuel that there was no significant effect. Those earning £100 to £200 and £200 to £300 per week saw a 2 and 6 percentage point reduction respectively, however the October 2011 fuel price increases hit them hardest (7 and 10 point increase), negating any advantage and leading to a net increase (Table 29).

Table 28 Fuel Poverty by household characteristics (000s)

	July 2011 Prices			October 2011 Prices			Total	Sample size
	Not Fuel Poor	Fuel Poor	Extreme Fuel Poor	Not Fuel Poor	Fuel Poor	Extreme Fuel Poor
Tenure
Owner-occupier	1,101	381	113	1050	433	129	1,483	2,022
LA/other public	263	89	17	236	115	23	352	461
HA/co-op	217	55	6	201	71	12	272	346
Private-rented	204	57	16	198	63	20	261	317
Private Sector	1,306	438	129	1247	497	150	1,744	2,339
Social Sector	481	144	23	437	187	35	624	807
Household Type
Single adult	232	109	31	209	132	37	341	429
Small adult	349	42	9	338	53	12	391	495
Single parent	86	26	4	79	33	4	112	151
Small family	319	22	4	315	26	6	341	440
Large family	143	17	6	139	21	9	160	217
Large adult	198	34	9	190	42	11	232	328
Older smaller	283	140	34	254	169	39	423	587
Single pensioner	177	192	55	161	208	67	369	499
Weekly Income Band
< £100 p.w.	1	51	38	-	52	42	52	65
£100 -199.99 p.w.	65	268	76	41	291	96	333	444
£200 -299.99 p.w.	333	159	28	285	207	30	492	652
£300 -399.99 p.w.	318	55	8	304	70	11	373	483
£400 -499.99 p.w.	253	21	1	246	28	4	274	366
£500 -699.99 p.w.	369	19	2	366	22	3	388	541
£700+	447	9	-	443	14	-	457	595
Council Tax Band
A	347	136	29	318	165	33	482	653
B	447	151	28	418	180	41	598	783
C	296	80	23	278	98	26	376	531
D	227	57	14	219	65	18	284	368
E	250	69	23	241	77	27	319	419
F	123	44	14	118	50	16	167	218
G	80	36	17	76	40	18	116	135
H	-	-	-	-	-	-	8	11
Unobtainable	-	-	-	-	-	-	19	28
All Scotland	1,786	582	153	1684	684	185	2,368	3,146

Source: SHCS 2011

118. While council tax band remains a poor indicator of fuel poverty, F- and G-band dwellings saw consistent increases in fuel poverty between 2010 and July 2011 while others improved. By October 2011 these groups had gained 7 percentage points in fuel poverty since 2010 (from 71,000 in 2010 to 90,000 in October 2011).

Table 29 Fuel poverty by households characteristics (Row %)

	2010	Jul-11	Oct-11	2011 Sample
Tenure
Owner-occupier	27	26	29	2,022
LA/other public	32	25	33	461
HA/co-op	29	20	26	346
Private-rented	27	22	24	317
Private Sector	27	25	28	2,339
Social Sector	30	23	30	807
Household Type
Single adult	35	32	39	429
Small adult	15	11	14	495
Single parent	31	23	29	151
Small family	6	6	8	440
Large family	7	11	13	217
Large adult	20	15	18	328
Older smaller	40	33	40	587
Single pensioner	55	52	56	499
Weekly Income Band
< £100 p.w.	98	98	100	65
£100 -199.99 p.w.	83	81	88	444
£200 -299.99 p.w.	38	32	42	652
£300 -399.99 p.w.	15	15	19	483
£400 -499.99 p.w.	7	8	10	366
£500 -699.99 p.w.	4	5	6	541
£700+	2	2	3	595
Council Tax Band
A	33	28	34	653
B	25	25	30	783
C	26	21	26	531
D	25	20	23	368
E	31	22	24	419
F	23	26	30	218
G	27	31	34	135
H	-	-	-	11
Unobtainable	-	-	-	28
All Scotland	28	25	29	3,146

Source: SHCS 2011