Stats NZ has a new website.

For new releases go to

www.stats.govt.nz

As we transition to our new site, you'll still find some Stats NZ information here on this archive site.

  • Share this page to Facebook
  • Share this page to Twitter
  • Share this page to Google+
Measuring energy hardship / fuel poverty

While there is no international consensus around measurement of energy hardship, some indicators have been developed and are in use in Europe, the United Kingdom and Australia. Fuel poverty or energy hardship can be measured using subjective or objective indicators, or through a composite measure where these indicators are combined.

Surveys designed specifically to measure fuel poverty ask a number of questions designed to directly measure difficulties in affording heating as well as the comfort of the internal temperature in the home.

Range of indicators used to estimate energy hardship

This section shows a range of different indicators used to estimate the numbers of households experiencing fuel poverty or energy hardship. We have attempted to be as comprehensive and accurate as possible when compiling this information, but there may be other indicators in use that we have not included here. The use of ** indicates an official measure, while the other indicators are used by researchers.

Objective measures: Income/expenditure

Indicator: More than 10 percent of household income required to heat home to an adequate level
Rationale: Regarded as excessive expenditure on heating, based on Boardman (1991). 10 percent was twice the median at that time.
Where indicator used: United Kingdom** and New Zealand. Indicator modelled by researchers in New Zealand (Lloyd, 2006; Howden-Chapman et al, 2012)

Indicator: Households with required fuel costs above the median level; if they were to spend that amount, they would be left with a residual income below the official poverty line (AHC and adjusted by household size and composition) – number of households, number of people
Rationale: Measures thermal efficiency of dwelling as well as cost burden of heating a home adequately. Attempts to measure the overlap between low income and fuel poverty, as well as the depth of fuel poverty.
Where indicator used: United Kingdom** and New Zealand. Indicator modelled by researchers in New Zealand (Lloyd, 2006; Howden-Chapman et al, 2012)

Indicator: More than 10 percent of AHC household income required to heat a home to an adequate level
Rationale: AHC expenditure is considered to more truly reflect the distribution of fuel poverty, as some households may have high fuel expenditure but low housing costs, and are therefore not fuel poor.
Where indicator used: United Kingdom

Indicator: Twice the median proportion of income spent on domestic energy
Rationale: Attempts to measure the cost burden of excessive energy expenditure required to heat a home
Where indicator used: World Health Organization

Objective measures: Other

Indicator: Electricity disconnection due to non-payment of bills (extreme energy hardship)
Rationale: Disconnection results in the inability to heat dwelling at all, and could also result in other issues, such as lack of hot water and food spoilage, if household disconnected for more than a few hours.
Where indicator used: New Zealand

Indicator: Households using prepayment metering
Rationale: Indicative of households struggling to pay power bills
Where indicator used: New Zealand

Indicator: Indoor room temperature (living room)
Rationale: Direct indicator of adequacy of heating
Where indicator used: Europe, New Zealand

Indicator: Excess winter mortality
Rationale: Indicator of indoor cold strain
Where indicator used: Europe, Japan, New Zealand

Consensual measures / self-reported information

Indicator: No heating used in dwelling
Rationale: Due to New Zealand’s temperate climate and the thermal inefficiency of much of the building stock, most New Zealand housing requires heating. Low indoor temperatures in winter occur even in the Far North (O’Sullivan et al, 2016).
Where indicator used: New Zealand Census of Population and Dwellings

Indicator: Interruptions in gas, oil, or electricity delivery because of non-payment. Days without heating or cooling because of non-payment.
Rationale: Indicative of energy uncertainty. Households that have been unable to pay utility bills on time are more likely to find it difficult heating a home adequately. Disconnection results in the inability to heat the dwelling at all, and could also result in other issues, such as lack of hot water and food spoilage, if household disconnected for more than a few hours.
Where indicator used: Children’s HealthWatch (United States)

Indicator: Use of unsafe substitute heating methods
Rationale: These can have health consequences – for example, the use of portable, unflued gas heaters can result in dampness and the release of gases into a dwelling.
Where indicator used: Children’s HealthWatch (United States)

Indicator: Could not pay electricity, gas, rates, or water bills on time ‘not at all’, ‘once’, ‘more than once’
Rationale: Indicative of energy uncertainty. Households that have been unable to pay utility bills on time are more likely to find it difficult heating a home adequately and may be at risk of disconnection (Healy, 2004).
Where indicator used: Australia, part of composite indicator in Europe, New Zealand (part of ELSI)

Indicator: Dwelling too cold or hard to heat
Rationale: Indicates energy hardship. Dwelling is unlikely to be energy efficient and may be inadequately heated or continuously unheated.
Where indicator used: New Zealand (used in some form in HES and GSS)

Indicator: How long respondent shivers from the cold when indoors on a cold winter night, or how many days/nights did respondent shiver last winter.
Rationale: Cold strain – useful physiological measure that indicates that the body is not adequately protected from a cold environment and is experiencing some level of thermos-regulation (Healy, 2004; O’Sullivan, 2016)
Where indicator used: European Household Panel Survey

Indicator: Households unable to heat home adequately
Rationale: Direct indicator of fuel poverty
Indicator: Dwelling damp, dwelling has rotten window frames, dwelling has leaky roof
Rationale: These indicate that a dwelling is unlikely to be energy efficient, and may be inadequately heated or continuously unheated.
Indicator: Dwelling without central or electric storage heating, household declaring a lack of adequate heating facilities
Rationale: Inadequate heating – household unlikely to be able to heat dwelling sufficiently
Indicator: Housing satisfaction
Rationale: Indicative of housing issues
Where indicators used: Composite indicator in European Household Panel Survey. Dampness and mould asked in Stats NZ’s GSS and HES.

Different indicators may identify different at-risk groups

Research has found that the different indicators may identify different groups of households experiencing energy hardship, although all were experiencing some kind of deprivation. For example, Australian researchers found very little overlap in the measurement of fuel poverty depending on which measure was used. In particular, there was a difference between income-expenditure and self-reported measures (based on households’ self-reported ability to heat their homes and to pay bills on time) (Azpitarte, Johnson and Sullivan (2015).

They also found that the measure used made a difference to both the prevalence of fuel poverty and the characteristics of households. A recent study by Lawson et al in New Zealand (2015), found that a household's spending on fuel was only weakly related to self-reported fuel deprivation. Again, differences emerged among household types. Their study found that people aged 65 and over commonly spent more than 10 percent of their household income on domestic energy but did not go without adequate energy. In contrast, families who spent less than 10 percent did not heat their homes adequately.

Jonathan Healy (2004), who studied the incidence of fuel poverty in Europe, also found this disparity, and recommended the use of composite indicators to identify households experiencing fuel poverty. However energy hardship is manifested, there are underlying structural factors that indicate deprivation.

The new United Kingdom measure is dependent on having information about the energy efficiency of a dwelling, which is obtained through the English Housing Survey.

Income-expenditure measures have the advantage of being easy to calculate, but researchers have argued that these are arbitrary measures and cannot fully capture the proportion of the population experiencing energy hardship (Healey, 2004).

Self-reported or ‘consensual’ indicators, such as the ability to heat the house adequately and an inability to pay utility bills on time, are more direct indicators of fuel poverty. Healy notes that households that have been unable to pay utility bills on time are more likely to find it difficult to heat a home adequately. In more extreme cases, people may be subject to electricity disconnections. In New Zealand, households in this position may be encouraged to get prepayment meters, which will automatically disconnect when credit runs out (O’Sullivan et al, 2015, 2016).

The advantages of consensual measures is that they more directly address the issue, but they may also be subject to recall error (Healy, 2004). In Information on energy hardship from the Household Economic Survey we discuss which of these measures are used in this paper.

The role of domestic energy prices in energy hardship

Energy hardship / fuel poverty occurs due to a relationship between housing quality, energy prices, and household incomes. The following section looks at the change in energy prices over time, the percentage of income spent on domestic energy, and changing energy use.

Domestic energy prices, particularly gas and electricity prices, have risen at a greater rate than average weekly earnings since 1989. Household firewood, however, has risen at approximately the same rate as income.

Figure 3
Graph, Household energy price indexes and change in average weekly earnings index, 1989–2016.

Household Economic Survey: Year ended 30 June 2007 showed that 94.0 percent of households reported expenditure on electricity compared with 45.3 percent using gas and 18.7 percent using solid fuel. Research has shown that households have been consuming less electricity, with figures from the Ministry of Business, Innovation and Employment showing that residential electricity consumption has fallen slightly between 2006 and 2012 (cited in Allan and Kerr, 2016).

Over time, however, the percentage of household expenditure used for domestic fuel and power has not changed very much. In 1998, households spent, on average, $24 a week on fuel (the equivalent of $34 in 2013). Around 3.5 percent of households’ net weekly spending was on domestic energy. By 2013, the percentage of household expenditure used for fuel had increased slightly to 4.2 percent ($46). By 2015/16 average weekly expenditure on domestic energy was $47, but it had fallen slightly as a proportion of total expenditure (to 3.6 percent). Spending on household energy is slightly above the levels in Australia, where household energy has remained fairly constant at around 2.7 percent of household income (Burke and Ralston, 2013).

However, domestic energy expenditure has increased significantly for households in the lowest equivalised income quintile (see figure 4). For households in the lowest disposable income quintile, domestic energy as a percentage of expenditure had increased from 4.3 to 6.0 percent between 1988 and 2015/6, compared with 2.0 to 2.5 percent for the highest quintile.

Figure 4
Graph, Percent of total expenditure on domestic energy by equivalised disposable household income quintiles, HES 1988–2015/16.

Information on energy hardship from the Household Economic Survey

In this paper we have included as many of the measures as it is possible to obtain data for. We have concentrated on the Household Economic Survey as this survey collects all household expenditure and also collects some information about subjective housing quality. We look at the numbers and proportions of households experiencing these indicators and compare the characteristics of each population. We have also included a measure that looks at percentage of expenditure used for domestic energy in order to see how it compares with percentage of income spent on domestic energy. A higher proportion of expenditure on fuel can also mean that a household may have to curtail expenditure in other areas.

Information from the Household Economic Survey shows that, depending on which measure is used, between 1 in 4 and 1 in 22 New Zealand households experienced an energy hardship indicator in 2015/16. When we look at domestic energy as a proportion of total expenditure, 424,000 households (25.1 percent) spent 6.3 percent or more. For 354,000 households (21.0 percent) the proportion of ‘after-housing costs’ (AHC) household income spent on domestic energy was twice the median or more. Having a major problem with dampness and mould was the least common problem, with 72,000 households (4.3 percent) having this issue.

However, experience of an energy hardship indicator was much higher for low-income households. We found that low-income households paid a higher proportion of their household income on domestic energy. They were approximately twice as likely to experience difficulty paying a utility bill on time and to experience cold and/or damp housing conditions in both 2012/13 and 2015/16.

We have defined low-income households as households in the lowest equivalised income quintile (when household income was adjusted by the number and ages of people in the household).

Table 2 shows the numbers and proportions of households experiencing an energy hardship indicator. There was no significant change in the proportion of households whose energy costs were high in relation to household income between 2012/13 and 2015/16. However, while there was a small but significant decrease in the proportion of all households experiencing a subjective indicator of energy hardship, there was no significant change in the proportion of low-income households experiencing an energy hardship indicator.

Table 1 explains some of the indicators and terms used in table 2.

Table 1

 Glossary of terms used for energy hardship measures
 Term  Definition
 Domestic energy  Energy used in the home for cooking, heating, running appliances, and heating water. Includes expenditure on electricity, wood, gas, and other domestic fuels.
 AHC  Residual household income left after housing costs. Housing costs include rent, mortgage, rates, and dwelling insurance (not contents insurance).
 Domestic energy costs greater than 10 percent of household income  Households that paid 10 percent or more of their gross income on domestic energy.
 Domestic energy costs greater than 10 percent of AHC household income  Households that paid 10 percent or more of their gross residual income after housing costs on domestic energy.
 Twice the median proportion of household income spent on energy  Households that spent twice the median proportion of gross income on domestic energy. In 2015/16 this was 5.6 percent or more. In 2012/13 this was 6.4 percent or more.
 Twice the median proportion of AHC household income spent on energy  Households that spent twice the median proportion of gross residual income after housing costs on domestic energy. In 2015/16 this was 6.6 percent or more. In 2012/13 this was 7.2 percent or more.
 Domestic energy expenditure as a percentage of total expenditure in highest quartile  Households where domestic energy expenditure was in the highest quartile as a percentage of all expenditure. In 2015/16 this was 6.3 percent or more. In 2012/13 this was 6.7 percent.

Table 2

 Measures of energy hardship
HES 2012/13 and 2015/16
 Energy hardship measure

 Low-income households (1)

 All households

 2012/13

 2015/16

 2012/13

 2015/16

 Number
[95% CI]

 Percent
[95% CI]

 Number
[95% CI]

 Percent
[95% CI]

 Number
[95% CI]

 Percent
[95% CI]

 Number
[95% CI]

 Percent
[95% CI]

 Domestic energy costs greater than 10 percent of household income

 94,000
[75,000–113,000]

 28.9
[24.4–33.4]

 86,000
[70,000–101,000]

 25.5
[21.7–29.4]

 117,000
[96,000–138,000]

 7.2
[5.9–8.5]

 103,000
[86,000–120,000]

 6.1
[5.1–7.1]

 Domestic energy costs greater than 10 percent of AHC household income

 143,000
[118,000–169,000]

 44.2
[38.6–49.8]

 121,000
[104,000–136,000]

 35.7
[31.5–39.9]

 199,000
[170,000–221,000]

 12.0
[10.5–13.6]

 175,000
[154,000–197,000]

 10.4
[9.1–11.7]

 Proportion of household income spent on domestic energy costs twice the median or more

199,000
[169,000–229,000] 

 61.3
[55.5–67.0]

 203,000
[180,000–226,000]

 60.5
[55.6–65.5]

 309,000
[278,000–340,000]

19.0
[17.1–20.9] 

 333,000
[305,000–360,000]

 19.7
[18.1–21.3]

 Proportion of AHC household income spent on domestic energy costs twice the median or more

215,000
[183,000–246,000] 

 66.1
[60.5–71.6]

 198,000
[179,000–217,000]

 58.9
[54.9–62.9]

347,000
[315,000–379,000] 

 21.4
[19.4–23.4]

 354,000
[326,000–382,000]

 21.0
[19.3–22.6]

 Domestic energy expenditure as % total expenditure in highest quartile

 156,000
[134,000–179,000]

 48.2
[43.1–53.2]

 175,000
[153,000–197,000]

 52.0
[46.8–57.2]

 409,000
[379,000–438,000]

 25.2
[23.4–27.0]

 424,000
[390,000–457,000]

 25.1
[23.1–27.1]

 Total number of households(2)

325,000

 ..

 336,000

 ..

 1,624,000

 ..

1,686,000 

 ..

 Could not pay electricity, gas, rates, or water bills on time:

 Not at all

 242,000
[212,000–271,000]

 60.7
[55.5–65.9]

227,000
[202,000–252,000] 

69.1
[64.7–73.5] 

1,270,000
[1,238,000–1,302,000] 

78.3
[76.3–80.3] 

 1,379,000
[1,352,000–1,406,000]

 82.9
[81.3–84.6]

 Once

 27,000
[15,000–38,000]

8.3
[4.8–11.7] 

19,000
[12,000–26,000] 

5.6
[3.5–7.8] 

 81,000
[64,000–98,000]

5.0
[3.9–6.0] 

57,000
[44,000–69,000] 

 3.4
[2.7–4.1]

 More than once

 54,000
[38,000–71,000]

 16.8
[12.2–21.4]

 35,000
[26,000–45,000]

 10.8
[8.2–13.4]

 119,000
[95,000–144,000]

 7.4
[5.9–8.9]

 77,000
[64,000–90,300]

 4.6
[3.9–5.4]

 Put up with feeling cold:
 Not at all

 196,800
[170,000–224,000]

 60.7
[55.5–65.9]

 227,000
[202,000–252,000]

 69.1
[64.7–73.5]

 1,270,000
[1,238,000–1,302,000]

 78.3
[76.3–80.3]

 1,379,000
[1,352,000–1,406,000]

 82.9
[81.3–84.6]

 A little

 62,000
[47,000–77,000]

 19.1
[14.5–23.7]

 59,000
[46,000–72,000]

 18.0
[14.5–21.5]

 215,000
[182,000–248,000]

 13.3
[11.2–15.3]

 196,000
[172,000–221,000]

 11.8
[10.3–13.3]

 A lot

 65,000
[47,000–84,000]

 20.2
[15.3–25.1]

 43,000
[33,000–52,500]

 12.9
[10.0–15.9]

 137,000
[111,000–163,000]

 8.5
[6.8–10.1]

 88,000
[75,000–101,000]

 5.3
[4.5–6.0]

 Heating accommodation and/or keeping it warm in winter:
 No problem

 189,000
[163,000–214,000]

 59.2
[52.9–65.4]

 222,000
[197,000–247,000]

 68.1
[63.2–73.0]

 1,176,000
[1,141,000–1,212,000]

 72.9
[70.7–75.2]

 1,277,000
[1,244,000–1,310,000]

 77.4
[75.4–79.3]

 Minor problem

 77,000
[58,000–96,000]

 24.1
[19.1–29.2]

 62,000
[49,000–75,700]

 19.1
[15.3–22.9]

 289,000
[256,400–321,600]

 17.9
[15.9–19.9]

 264,000
[237,000–291,000]

 16.0
[14.3–17.6]

 Major problem

 53,300
[39,000–68,000]

 16.7
[12.5–21.0]

 42,000
[30,000–53,000]

 12.8
[9.5–16.1]

 147,000
[126,000–169,000]

 9.1
[7.8–10.5]

 110,000
[92,000–129,000]

 6.7
[5.5–7.8]

 Dampness and/or mould:
 No problem

 197,000
[174,000–221,000]

 60.8
[55.6–66.0]

 239,000
[213,000–265,000]

 72.9
[68.4–77.2]

 1,166,000
[1,133,000–1,199,000]

 71.9
[69.8–73.9]

 1,317,000
[1,280,000–1,354,000]

 79.4
[77.2–81.5]

 Minor problem

 84,000
[64,000–104,000]

 25.8
[20.9–30.6]

 60,000
[47,000–73,000]

 18.3
[14.8–21.9]

 337,000
[306,000–368,000]

 20.8
[18.9–22.7]

 270,000
[240,000–300,000]

 16.3
[14.5–18.1]

 Major problem

 43,000
[30,000–57,000]

 13.4
[9.4–17.4]

 29,000
[20,000–38,000]

 8.8
[6.1–11.5]

 120,000
[97,000–143,000]

 7.4
[6.0–8.8]

 72,000
[58,000–86,000]

 4.3
[3.5–5.2]

 1. Equivalised household income in lowest quintile
2. Total number of households where economic standard of living index questions were asked.

In the year ended June 2015/16, over 370,000 households had a problem heating their dwelling or keeping it warm in winter. For 110,000 households this was a major problem. Of households where heating was a major problem, over half (57 percent) had equivalised household incomes in the lowest two income quintiles. Households in the lowest equivalised income quintile were significantly more likely to report that heating was a major problem, compared with all households.

When we look at fuel-to-income measures, around one-fifth of households paid twice the median on domestic energy (5.6%) as a proportion of household income. Fewer households paid 10 percent or more of their income on domestic energy. In 2015/6, around 1 in 17 households paid 10 percent or more of their before-tax income on domestic energy but this rose to around 1 in 10 when we looked at the percentage of AHC residual income.

In figure 5 we have used the more extreme end of each indicator, such as when dampness or mould are considered a major problem.

Figure 5
Graph, Households with selected energy hardship indicators by equivalised household income quintile, HES 2015/16.

While we have largely used quintiles because of the small sample size in HES, we have included one example here with equivalised income deciles to illustrate the considerable difference in energy hardship measures by decile. We can see that households in the lowest income deciles were much more likely to spend 10 percent or more of their income on electricity.

Figure 6
Graph, Expenditure on energy as a proportion of household income, by JEAH income decile, HES 2015/16.

Characteristics of households

This section explores some of the characteristics of fuel-poor households. We look at how they are different from each other and from all households.

Table 3

 Life satisfaction, income adequacy, and income quintiles
HES 2015/16
 Energy hardship measure

 Percent of households where income not enough
[95% CI]

 Percent of households dissatisfied or very dissatisfied with life now
[95% CI]

 Percent of households in lowest equivalised income quintile
[95% CI]

 Percent of households in lowest income quintile (not equivalised)
[95% CI]

 Domestic energy costs greater than 10 percent of household income 23.7
[17.1–30.3] 
 14.5
[8.5–20.5]
83.3
[76.1–90.6] 
 82.6
[75.1–90]
 Domestic energy costs greater than 10 percent of AHC household income 26.9
[21.8–32.1]
 14.0
[9.9–18.1]
 68.4
[62.7–74.1]
 63.4
[57.0–69.9]
 Proportion of household income spent on domestic energy costs twice the median or more  22.3
[18.5–26.2]
10.9
[8.4–13.3] 
61.2
[57.0–65.3] 
 59.5
[54.9–64.1]
 Proportion of AHC household income spent on domestic energy costs twice the median or more 23.0
[19.2–26.9] 
11.9
[8.9–14.8] 
 56.0
[51.8–60.1]
 52.3
[47.7–56.8]
 Domestic energy expenditure as percentage of total expenditure in highest quartile 18.9
[15.5–22.2] 
 10.0
[7.8–12.2]
 41.2
[37.6–44.8]
40.5
[36.3–44.7] 
 Could not pay electricity, gas, rates, or water bills on time more than once (ELSI) 45.3
[35.0–55.5] 
21.5
[14.0–29.1] 
 45.8
[36.6–55.1]
30.9
[23.1–38.7] 
 Put up with feeling cold a lot (ELSI) 34.9
[27.4–42.4] 
 25.0
[18.2–31.8]
 48.4
[39.1–57.8]
42.0
[33.9–50.2] 
 Heating and/or keeping dwelling warm in winter a major problem (ELSI) 31.8
[25.2–38.3] 
24.1
[18.0–30.1] 
37.8
[29.6–46.1] 
 31.2
[25.0–37.4]
 Dwelling damp or mouldy a major problem (ELSI) 32.9
[22.8–43.1] 
 22.2
[13.8–30.7]
 40.1
[29.7–50.4]
 26.9
[19.0–34.8]
 All households 10.2
[9.0–11.4] 
 6.1
[5.2–7.1]
 19.9
[18.2–21.7]
19.9
[18.4–21.3] 
 Note: For around 22,000 households there is no ELSI information.

Table 3 shows that fuel-poor households were more likely to be in the lowest income quintiles and the lowest equivalised income quintile. These households were more likely to report that their income was not enough or that they were dissatisfied or very dissatisfied with life now. These were significant differences. Life dissatisfaction was highest in households where people experienced problems with heating, dampness and/or mould, and feeling cold. Households with an energy hardship indicator were also three or four times more likely than all households to say their income was not enough.

There were some variations between indicators, however. Households with high energy costs in relation to income were much more concentrated in the lowest income quintile. These differences appear more marked when household income is equivalised (adjusted for the number and ages of people in the household). However, this was not a significant difference. While problems with dampness and/or mould and heating the home were not just confined to the lowest income quintiles, households in the lowest income quintile were still around five times more likely to report that heating their home adequately was a major problem.

Table 4

 Socio-economic characteristics of selected fuel-poor households in New Zealand
HES 2015/16
 Characteristics  Energy costs twice median
[95% CI]
Energy costs 10 percent or more of AHC income
[95% CI] 
Energy costs twice median AHC income
[95% CI] 
 Could not pay bills on time more than once
[95% CI]
 Put up with feeling cold a lot
[95% CI]
 Heating a major problem
[95% CI]
Damp and/or mould a major problem
[95% CI] 
All household
[95% CI] 

 Percentage of people in households

 Sex
 Male  47.7
[45.5–49.8]
48.2
[45.0–51.3] 
48.1
[45.6–50.6] 
 50.6
[46.3–54.9]
 47.7
[43.8–51.6]
49.8
[46.3–53.4] 
54.1
[50.5–57.8] 
49.6
[49.0–50.1] 
 Female  52.3
[50.2–54.5]
 51.8
[48.7–55.0]
 51.9
[49.4–54.4]
49.4
[45.1–53.7] 
 52.3
[48.4–56.2]
 50.2
[46.6–53.7]
45.9
[42.2–49.5] 
 50.4
[49.9–51.0]
 Total  100.0  100.0  100.0  100.0 100.0  100.0%  100.0%   100.0%
 Age
 Under 15 21.4
[19.1–23.6] 
 25.1
[21.4–28.8]
 22.9
[20.4–25.5]
35.9
[30.7–41.2] 
 24.2
[18.7–29.7]
 26.5
[22.4–30.7]
 23.3
[17.7–29.0]
 20.1
[20.0–20.1]
 15–24  13.3
[11.0–15.6]
 15.8
[12.0–19.7]
 14.2
[11.2–17.3]
16.6
[11.6–21.7] 
 21.2
[14.0–28.3]
 16.0
[11.1–20.9]
 18.4
[12.5–24.2]
 14.1
[14.0–14.1]
 25–44  19.3
[16.7–22.0]
21.9
[18.3–25.5] 
 21.4
[19.1–23.7]
 27.8
[23.8–31.8]
 25.6
[20.3–30.8]
29.4
[24.6–34.2] 
 31.3
[25.1–37.6]
 26.0
[26.0–26.1]
 45–64  20.3
[17.5–23.1]
21.6
[17.5–25.7] 
 20.4
[17.7–23.0]
 16.0
[11.4–20.6]
 20.1
[14.6–25.7]
21.0
[16.3–25.8] 
 21.1
[15.5–26.8]
25.6
[25.5–25.6] 
 65 and over  25.7
[22.3–29.1]
 15.6
[11.8–19.3]
 21.1
[18.3–23.8]
 3.7
[1.7–5.6]
 8.9
[5.3–12.5]
 7.0
[4.4–9.6]
 5.9
[3.2–8.5]
 14.2
[14.2–14.3]
 Total people  100.0 100.0   100.0  100.0 100.0  100.0  100.0   100.0
 Percentage of households
 Household composition
 Couple only  25.4
[22.5–28.3]
 18.5
[13.6–23.3]
 22.7
[19.5–26.0]
 10.8
[5.0–16.7]
 17.1
[9.6–24.5]
 15.1
[9.6–20.6]
 14.4
[5.8–23.1]
 27.3
[25.9–28.6]
 Couple with children  20.0 [16.5–23.6]  21.2
[16.3–26.1]
 22.3
[18.7–25.8]
 36.4
[26.2–46.6]
18.6
[11.5–25.7] 
27.7
[19.9–35.5] 
 32.2
[21.8–42.6]
 34.1
[32.4–35.8]
 One parent with children 12.6
[9.6–15.7] 
16.6
[11.9–21.3] 
 14.1
[10.9–17.2]
 22.7
[13.9–31.5]
22.2
[14.6–29.9] 
 19.8
[13.6–25.9]
 24.3
[15.5–33.1]
7.4
[6.2–8.5] 
 Other one family  3.4
[1.5–5.3]
4.3
[1.1–7.5] 
 3.4
[1.5–5.4]
8.2
[2.1–14.3] 
3.0
[0.2–5.8] 
 6.8
[2.3–11.3]
 5.4
[1.0–9.8]
 4.7
[3.6–5.8]
 One person household  34.7
[30.9–38.6]
 32.5
[26.8–38.1]
 32.6
[28.5–36.7]
14.2
[8.3–20.2] 
 29.8
[22.5–37.0]
 22.9
[17.1–28.7]
15.2
[9.2–21.1] 
21.8
[20.8–22.7] 
 Other  3.8
[1.9–5.6]
 7.0
[3.3–10.8]
 4.9
[2.0–7.7]
7.7
[0.5–14.8] 
 9.3
[3.1–15.5]
7.7
[2.9–12.5] 
8.6
[2.1–15.1] 
4.
[3.8–5.9] 
 Total households  100.0  100.0  100.0 100.0   100.0 100.0  100.0   100.0
 Dwelling owned or not
 Owner-occupied dwelling  60.5
[56.0–64.9]
48.3
[41.9–54.7] 
 56.7
[52.5–60.9]
 32.4
[24.8–40]
 30.7
[22.4–39.0]
 28.8
[21.3–36.3]
 22.8
[13.9–31.6]
67.3
[65.5–69.1] 
 Dwelling not owned  39.5
[35.1–44]
51.7
[45.3–58.1] 
43.3
[39.1–47.5] 
 67.6
[60.0–75.2]
69.3
[61.0–77.6] 
71.2
[63.7–78.7] 
 77.2
[68.4–86.1]
 32.7
[30.9–34.5]
 Other 100.0  100.0   100.0  100.0  100.0  100.0  100.0  100.0
 Who dwelling is rented from
 Renting from private person, business, or trust 72.9
[66–79.9] 
80.5
[73.3–87.7] 
79.0
[73.1–85.0] 
 81.2
[72.2–90.2]
 64.1
[52.5–75.8]
72.5
[63.1–81.9] 
80.6
[70.4–90.9] 
 87.2
[84.2–90.2]
 Renting from local government or Housing New Zealand Corporation  26.7
[19.7–33.7]
19.5
[12.3–26.7] 
 21.0
[15.0–26.9]
 18.8
[9.8–27.8]
 33.9
[22.1–45.8]
25.6
[16.1–35.1] 
 19.4
[9.1–29.6]
 12.1
[9.4–14.8]
 Other landlord  s  s  s  s
 Total renting households 100.0  100.0   100.0 100.0   100.0  100.0  100.0  100.0
 Percentage of people in households aged 15 and over
 Labour force status (people aged 15 and over)
 Employed 37.8
[34.0–41.6] 
40.7
[34.2–47.1] 
42.0
[38.3–45.7] 
 49.6
[42.7–56.5]
52.1
[44.8–59.4] 
60.7
[54.7–66.7] 
56.7
[48.8–64.6] 
63.4
[61.9–64.9] 
 Unemployed 7.4
[5.5–9.3] 
9.7
[6.7–12.8] 
 7.3
[5.5–9.1]
 9.3
[5.3–13.2]
 12.0
[8.2–16.9]
9.3
[5.8–12.8] 
 11.0
[6.2–15.8]
3.7
[3.0–4.3] 
 Not in the labour force  54.8
[51.1–58.6]
49.6
[43.9–55.3] 
 50.7
[47.1–54.3]
 41.1
[34.2–48.0]
 35.3
[28.4–42.3]
 30.0
[24.3–35.8]
 32.3
[26.0–38.7]
 32.9
[31.5–34.3]
 Total people aged 15 and over 100.0  100.0  100.0   100.0 100.0   100.0  100.0 100.0 
 Highest formal qualification (people aged 15 and over)
 No qualifications  34.4
[30.8–38.0]
 30.8
[25.9–35.6]
 32.4
[28.8–35.9]
31.7
[25–38.4] 
 33.5
[26.4–40.6]
 23.5
[18.4–28.6]
24.3
[17.6–31.0] 
19.9
[18.7–21.1] 
 School qualification  28.0
[24.3–31.8]
 27.0
[20.6–33.5]
 27.8
[23.7–31.9]
 38.5
[30.8–46.2]
 31.7
[24.8–38.7]
 35.8
[30.2–41.5]
 33.9
[26.0–41.7]
 28.5
[27.0–29.9]
 Vocational  24.5
[21.3–27.6]
 26.2
[21.1–31.3]
25.1
[21.8–28.4] 
 21.6
[16.1–27.1]
 20.7
[16.1–25.3]
 25.2
[21.2–29.3]
 25.2
[19.2–31.2]
 25.4
[24.1–26.8]
 Tertiary  11.7
[8.9–14.5]
 14.1
[9.8–18.4]
12.4
[10.0–14.9] 
6.6
[3.1–10.1] 
12.0
[8.3–15.6] 
 16.6
[10.1–19.2]
14.9
[7.5–22.2] 
 24.7
[23.4–26.1]
 Total people aged 15 and over  100.0  100.0  100.0  100.0  100.0  100.0  100.0  100.0
 Note: Vocational includes level 4–6 qualifications, while tertiary is level 7 and above.
 Percentage of households – distribution by broad region
 Auckland  17.7
[14.1–21.3]
 21.8
[16.7–26.9]
19.7
[16.3–23.1] 
27.3
[17.7–36.8] 
 18.0
[10.5–25.5]
24.4
[16.5–32.3] 
28.3
[16.8–39.8] 
 31.4
[31.1–31.7]
 Wellington  11.7
[9.6–13.8]
 11.2
[8.2–14.2]
 11.4
[9.4–13.3]
 12.0
[7.5–16.5]
 16.9
[11.4–22.5]
 16.1
[11.1–21.0]
 11.8
[6.3–17.3]
11.0
[10.9–11.2] 
 Rest of North Island  39.0
[35.1–43.0]
 37.2
[31.0–43.3]
 39.1
[35.1–43.0]
34.3
[25.1–43.5] 
 44.1
[35.0–53.3]
 33.9
[26.9–41.0]
 39.2
[28.3–50.0]
32.7
[32.4–33.0] 
 Canterbury  13.7
[11.3–16.0]
13.1
[9.8–16.4] 
12.8
[10.5–15.1] 
 15.0
[8.0–22.0]
 11.5
[6.8–16.2]
13.1
[8.4–17.8] 
 8.9
[3.4–14.4]
 13.4
[13.3–13.5]
 Rest of South Island  17.9
[14.4–21.4]
16.7
[12.3–21.2] 
 17.1
[14–20.2]
 11.4
[6.0–16.8]
9.5
[5.3–13.6] 
12.5
[8.1–16.9] 
11.8
[5.7–17.9] 
 11.4
[11.3–11.6]
 Symbol:
s – suppressed due to small numbers and high sample error.

As table 4 shows, there were some differences in socio-economic characteristics by energy hardship indicators. These differences have also emerged in other studies that considered a range of indicators, such as Healy (2004) and Aspitarte et al (2015). Households identified by the 10 percent or more of household income spent on fuel were the most concentrated in the lowest equivalised income quintile.

As in the Australian study (Azpitarte et al, 2015), households identified as fuel poor by cost-to-income indicators tended to be different to households identified as fuel poor by consensual indicators. Households identified as fuel poor by cost-to-income indicators had higher rates of home ownership (though still lower than the national average) and an older age structure. The pattern of household composition was also quite different with around one-third of these being one-person households.

Rates of home ownership were much lower for fuel-poor households identified through consensual indicators – around one-third compared with two-thirds for all households. The results from these consensual indicators are supported by objective measurement in the housing condition survey carried out five-yearly by the Building Research Association of New Zealand (BRANZ). The latest survey in 2015 found that rental properties tended to be in poorer condition and to have greater problems with dampness and mould. One-fifth of rental properties did not have a fixed heating source and relied on portable heating devices, which are more expensive and less efficient.

The age structure of fuel-poor households varied according to indicator, as figure 7 shows.

Figure 7
Graph, Household age structure by selected energy hardship indicator, HES 2015/16.

Children (aged under 15) and young people (aged 15–24) were overrepresented in households identified as fuel poor by consensual indicators. Lone parents and social-housing renters were more likely to be identified as fuel poor regardless of the indicator used. 

Despite some differences according to indicator, fuel-poor households demonstrated greater socio-economic disadvantage when compared with all households.

It is clear that despite the differing characteristics of households when identified by different fuel poverty indicators, each indicator usefully identifies groups of people that are experiencing poor quality housing conditions associated with inadequate heating, high costs for fuel in relation to income, or are at risk of disconnection.

Next we will consider the overlap between these indicators before attempting to create a composite indicator as recommended by Jonathan Healey (2004).

Overlap between indicators

Table 5 shows the percentage of households experiencing one fuel poverty indicator who are also experiencing another fuel poverty indicator. For example, in 2015/16, of the 110,000 households who said heating their dwelling or keeping it warm in winter was a major problem, 76,000 (around 70 percent) also experienced a minor or major problem with dampness and/or mould. Just over one-third (40,000 or 36 percent) of these households experienced both cold and dampness/mould as major problems.

The overlap between subjective measures and cost-to-income measures was lower. For example, around one-third of households that had a major problem with damp were paying two or more times the median on fuel as percentage of expenditure.

Table 5 shows the overlapping between indicators for all households, while table 6 shows overlapping for households in the lowest equivalised income quintile. It is important to remember that households may not heat their dwellings adequately in order to reduce their energy expenditure.

Table 5

 Overlapping between indicators
HES 2015/16
 Indicator

 10% or more of household income spent on fuel

 10% or more of AHC income spent on fuel

 Energy costs twice median as percentage of income

 Energy costs twice median AHC income

 High fuel expenditure as percentage of total expenditure

 Damp and/or mould a major problem

 Heating a major problem

 Bills late more than once 

 Put up with feeling cold a lot 

 Percentage [95% CI]

 10% or more of household income spent on fuel

 ..

 80.2
[74.0–86.3]

 100
[100–100]

 80.2
[74.0–86.3]

 72.5
[64.8–80.1]

 8.6
[3.4–13.8]

 15.4
[9.9–20.9]

 13.7
[9.0–18.4]

 11.6
[7.2–15.9]

 10% or more of AHC income spent on fuel

 47.0
[41.2–52.9]

 ..

 89.4
[85.2–93.7]

 100
[100–100]

 67.4
[61.6–73.3]

 9.3
[6.0–12.7]

 16.1
[11.8–20.4]

 14.6
[10.8–18.4]

 13.5
[10.0–16.9]

 Energy costs twice median as percentage of income

 30.9
[27.0–34.9]

 47.2
[42.2–52.1]

 ..

 83.3
[80.0–86.7]

 70.0
[65.5–74.5]

 6.7
[4.5–8.9]

 11.3
[8.8–13.7]

 11.2
[8.7–13.8]

 10.1
[7.7–12.5]

 Energy costs twice median AHC income

 30.9
[27.0–34.9]

 49.6
[45.0–54.1]

 78.4
[73.7–83.0]

 ..

 62.0
[56.8–67.1]

 7.6
[5.1–10.1]

 12.1
[8.9–15.3]

 10.5
[8.0–13.1]

 12.1
[9.3–14.8]

 High fuel expenditure as percentage of total expenditure

 17.6
[14.5–20.7]

 27.9
[24.1–31.7]

 55.0
[51.1–58.8]

 51.7
[48.1–55.4]

 ..

 6.2
[4.1–8.3]

 10.5
[8.2–12.7]

 8.9
[6.7–11.1]

 0.5
[8.2–12.8]

 Damp and/or mould a major problem  12.2
[4.6–19.7]
22.5
[14.5–30.6] 
30.5
[20.7–40.2] 
 36.6
[25.6–47.6]
 36.1
[24.7–47.4]
 ..  56.9
[45.3–68.5]
15.8
[8.1–23.6] 
29.7
[19.8–39.5] 
 Heating a major problem  14.2
[8.9–19.5]
 25.1
[18.8–31.3]
 33.3
[26.3–40.3]
 37.8
[29.1–46.4]
39.6
[31.9–47.3] 
 36.7
[28.6–44.8]
 .. 17.0
[11.0–23.0] 
 43.1
[35.7–50.5]
 Bills late more than once  18.1
[11.6–24.7]
 32.8
[24.5–41.2]
 47.7
[38.8–56.5]
 47.5
[38.4–56.7]
 48.0
[38.9–57.1]
 14.9
[7.4–22.4]
 24.5
[15.2–33.7]
 .. 22.8
[14.4–31.2] 
 Put up with feeling cold a lot 13.5
[8.3–18.6] 
 26.7
[20.3–33.1]
 37.8
[29.9–45.6]
47.9
[39.0–56.8] 
 49.9
[40–59.9]
 24.7
[16.9–32.5]
56.0
[46.7–65.3] 
 20.1
[13.2–26.9]
 ..

Overlapping between indicators – low-income households only

As we can see from table 5, the overlap was greater for low-income households – here we have looked at households in the lowest equivalised income quintile. For some of these indicators the differences were significant, particularly for subjective measures. Around two-thirds of low-income households that had a major problem with dampness and mould were paying at least twice the median for energy as a proportion of their income and there was little difference regardless of whether we used a before- or after-housing cost measure. Over one-third of households (37.8 percent) where heating the home was a major problem spent twice or more than the median of their AHC income on domestic energy. This proportion rose to almost three-quarters of low-income households (71.5%).

Table 6

 Overlapping between indicators
Low-income households only
HES 2015/16
 Indicator

 10% or more of household income spent on fuel

 10% or more of AHC income spent on fuel

 Energy costs twice median as percentage of income

 Energy costs twice median AHC income

 High fuel expenditure as percentage of total expenditure

 Damp and/or mould a major problem

 Heating a major problem

 Bills late more than once 

 Put up with feeling cold a lot 

 Percentage [95% CI]

 10% or more of household income spent on fuel

.. 

77.4
[70.3–84.6] 

 100.0

77.4 [70.3–84.6] 

71.2
[62.7–79.6] 

 9.8
[3.7–15.8]

 17.6
[11.4–23.8]

 13.9
[8.6–19.1]

12.9
[8.0–17.9] 

 10% or more of AHC income spent on fuel

 55.3
[47.8–62.9]

 ..

92.9
[88.3–96.9] 

 100.0 [100–100]

 68.4
[61.0–75.9]

 11.7
[7.3–16.1]

 18.0
[12.4–23.6]

 16.2
[11.8–20.5]

 16.8
[12.3–21.3]

 Energy costs twice median as percentage of income

 42.1
[36.7–47.6]

 54.6
[48.8–60.5]

.. 

81.9
[77.7–86.1] 

 67.0
[60.7–73.3]

 9.4
[5.9–12.9]

 14.4
[10.7–18.0]

 13.5
[10.2–16.8]

 13.8
[10.2–17.3]

 Energy costs twice median AHC income

33.5
[27.6–39.4] 

 60.6
[54.9–66.3]

 84.1
[78.3–89.9]

 ..

64.4
[58.7–70.2] 

 9.9
[5.8–14.0]

 15.6
[10.4–20.7]

 13.6
[10.1–17.0]

 16.8
[12.8–20.9]

 High fuel expenditure as percentage of total expenditure

34.9
[29.1–40.7] 

 47.0
[40.7–53.3]

78.0
[72.6–83.4] 

 73.0
[68.4–77.6]

 ..

11.0
[6.5–15.5] 

17.4
[12.4–22.4] 

 13.3
[9.2–17.5]

17.3
[12.6–22.0] 

 Damp and/or mould a major problem 28.9
[12.6–45.1]
48.5
[32.6–64.5]
65.2
[47.1–83.3]
66.3
[50.4–82.1]
 66.1
[50.7–81.5]
.. 56.8
[39.4–74.2] 
22.3
[10.3–34.4]
37.5
[20.7–54.4]
 Heating a major problem  35.9
[21.8–50.1]
51.3
[37.6–65.0] 
 68.7
[52.0–85.3]
71.5
[59.7–83.4]
72.2
[62.4–82.0]
39.4
[25.4–53.4] 
.. 24.7
[12.1–37.3]
 58.0
[48.1–68.0]
 Bills late more than once 33.5
[20.3–46.8]
54.8
[41.2–68.3]
 76.2
[64.8–87.5]
74.3
[61.7–86.9] 
65.3
[53.0–77.6]
18.2
[7.9–28.5]
 29.2
[13.5–44.8]
.. 32.3
[17.5–47.2]
 Put up with feeling cold a lot 26.1
[16.2–35.9]
 47.5
[34.4–60.3]
 64.9
[51.7–78.1]
76.8
[67.6–86.1]
 70.7
[59.7–81.7]
 25.9
[11.7–40.1]
59.8
[47.6–72.1]
26.9
[15.3–38.6]
 ..

Combining indicators

Number of households experiencing a fuel poverty indicator

In HES 2015/2016, we also looked at the number of households that had experienced an energy hardship indicator. This approach is used by Thomson (2014, 2016) in her work on fuel poverty in the European Union.

We selected five indicators to use:

  • whether a household paid twice or more the median in their AHC household in fuel costs
  • whether an individual put up with feeling cold a lot
  • whether there was a major problem with heating and/or keeping dwelling warm in winter
  • whether dwelling had a major problem with dampness and/or mould
  • whether they had trouble paying energy bills on time more than once.

Subjective measures give more direct evidence of the vulnerability of a household to fuel poverty. We chose the AHC income measure as relating more closely to hardship. The 10 percent benchmark that is commonly used came from Boardman’s work in the late 1980s. When she developed her work on fuel poverty, 10 percent was twice the median expenditure on domestic energy as a percentage of household income. We believed that paying twice or more the median as a proportion of AHC income more clearly reflected both the cost burden of fuel and related specifically to the situation in New Zealand.

In total, almost three-quarters of all households (70.8 percent) did not experience an energy hardship indicator. Around one-fifth of households (21.1 percent or 356,000) experienced one energy hardship indicator. The proportion of households experiencing three or more energy hardship indicators was much lower (3.0 percent).

The proportions were quite different for low-income households. Only one-third of households in the lowest equivalised household income quintile did not have an energy hardship indicator, while around one-fifth of households had two or more energy hardship indicators, as figure 8 shows.

Figure 8
Graph, Number of energy hardship indicators by equivalised household income quintile, HES 2015/16.

Later in this paper we explore creating a composite indicator as recommended by J Healy (2004) and compare the results. Households with multiple indicators of energy hardship were slightly larger at around 2.8 people per household.

Table 7

 Households and people by number of energy hardship indicators
HES 2015/16

 Number of indicators

 Estimated number of households

 Percentage of total households

 Estimated number of people

 Estimated people per household

 Number of low-income households

 Percentage of low-income households

 0

 1,193,200
[1,163,600–1,222,800]

 70.8%
[69.0–72.5]

 3,276,000
[3,184,000–3,369,000]

 2.7

 111,000
[94,000–128,000]

 33.1%
[29.2–37.0]

 1

 356,100
[330,500–381,600]

 21.1%
[19.6–22.6]

 887,000
[809,000–965,000]

 2.5

 150,000
[131,000–169,000]

 44.6%
[40.1–49.1]

 2

 85,400
[70,700–100,000]

 5.1%
[4.2–5.9]

 237,000
[191,000–283,000]

 2.8

 45,000
[34,000–55,000]

 13.3%
[10.4–16.1]

 3+

 51,200
[39,900–62,600]

 3%
[2.4–3.7]

 144,000
[106,000–181,000]

 2.8

 30,000
[22,000–39,000]

 9.0%
[6.5–11.6]

We observed some differences with age structure. Children under 15 were more likely to be in a household experiencing multiple energy hardship indicators, as figure 9 shows. People aged 65 and over were less likely to be experiencing two or more energy hardship indicators.

Figure 9
Graph, Number of energy hardship indicators by age group, HES 2015/16.

Creating a composite indicator for energy hardship

We have derived a composite indicator using the same five variables as above – one objective and four subjective.

When creating this indicator, we needed to assign a numeric response to each variable. We scored the responses for each variable. For example, where a household gave a response of ‘no problem’ to the question on heating or keeping the dwelling warm, we gave them a score of ‘0’. We gave ‘2’ when the household said it was a minor problem and ‘4’ for a major problem. This scoring was followed for all the subjective indicators. For the final objective indicator, a rating from ‘0’ to ‘4’ was given for each household, with two or more times the median AHC income spent on fuel being given the highest rating.

We then compiled a final score using different weighting scenarios to see whether there were differences if we gave equal weights to each indicator, or higher weights to heating or dampness and mould as being more direct indicators of fuel poverty.

Here we followed some of the methodology used by Healey (2004, see pp 46–49):

a) whether heating or keeping the dwelling warm in winter was a problem
b) whether dampness or mould was a problem
c) whether there was difficulty paying utility bills on time
d) percent of after housing cost income spent on domestic energy
e) whether a person put up with feeling cold.

We tried a range of scenarios.

Scenario 1: Problem with heating is considered a key indicator and given strong preference

We assigned a higher weight to heating, as having a problem with heating the dwelling or keeping it warm in winter relates most strongly to the issue of fuel poverty:

0.5a + 0.125b + 0.125c + 0.1125d + 0.125e

Scenario 2: Key indicator given moderate preference

We assigned a more moderate weighting to heating due to its relationship to fuel poverty:

0.4a + 0.15b + 0.15c + 0.15d + 0.15e

We repeated this scenario but used dampness and mould as a key indicator instead of heating.

We considered the presence of dampness and mould as a significant issue related to fuel poverty because of its relationship to health, so gave this indicator a slightly higher weighting.

0.16675a + 0.33b + 0.16675c + 0.16675d + 0.16675e

Scenario 3: All indicators given equal weight

All five indicators were treated as of equal importance, and equal weights were given to each:

0.2a + 0.2b + 0.2c + 0.2d + 0.2e

We then used the combined scores to assign each household to an energy hardship quintile. The highest quintile has been used to designate which households were experiencing energy hardship. Depending on which weighting scenario was used, between 350,000 and 400,000 households were in the highest energy hardship quintile.

Figure 10 shows the differences in the proportion of households that experienced a problem with heating their house in winter, by the different weighting scenarios.

Figure 10
Graph, Whether heating dwelling in winter was a problem by composite indicator type, HES 2015/16.

The results of sensitivity testing showed that there was very little difference in the characteristics of households experiencing the most severe energy hardship (quintile 5).

In table 8 we compared results for three scenarios, with either equal or moderate rating for heating and dampness. We also compared with the number of indicators experienced. Households that reported two energy hardship indicators were very similar to the highest quintile for the composite indicator. They were significantly different from households with no energy hardship indicators.

To recap, for combined indicators, each household was given a score for indicators of dampness, heating, feeling cold, whether bills were paid late, and percentage of income spent on domestic energy. We used the highest scoring quintile as representing households in energy hardship. We show the results for each weighting scenario. The same indicators were also used to create the ‘number of energy hardship indicators’ measure.

We have included a comparison of responses to the question about whether income was enough for everyday needs. There were clear differences between households that experienced energy hardship indicators and households that did not. For example, only 5.2 percent of households without an energy hardship indicator said that they did not have enough money, compared with 50 percent of households with three or more energy hardship indicators. Around one-quarter of households where indicators were combined said they did not have enough money. These differences were repeated for the other indicators we looked at: life satisfaction, tenure, and when we compared households in different equivalised income quintiles (see Appendix 3: Additional tables).

Table 8

 Comparing combined indicators
Income enough for everyday needs
HES 2015/16
 Composite indicator – highest quintile  Income
 Not enough  Just enough  Enough or more than enough
 Equal weighting scenario  27.6 [23.5–31.8]  42.6 [38.3–46.8]  29.8 [25.5–34.1]
 Heating – moderate weighting scenario  24.6 [21.0–28.3]  39.1 [35.3–42.8]  36.3 [32.3–40.3]
 Dampness – moderate weighting scenario  26.5 [22.4–30.6]  41.7 [37.6–45.8]  31.8 [27.3–36.4]
 Number of indicators
 None  5.2 [4.1–6.3]  20.1 [17.9–22.3]  74.7 [69.3–80.1]
 One  14.5 [11.1–18.0]  37.6 [32.6–42.7]  47.8 [38.1–57.5]
 Two  37.4 [26.5–48.1]  39.5 [28.8–50.1]  23.2 [12.5–33.8]
 Three or more  50.0 [35.2–64.6]  38.5 [23.6–53.5]  11.5 [3.7–19.5]

For composite indicators, very little difference emerges between the different weighting scenarios. Because of its closer relationship with energy hardship, we would recommend the use of the composite indicator with moderate weighting given to the ‘dwelling hard to keep warm’ indicator. As can be seen from figure 11, households with two or more energy hardship indicators corresponded closely in age structure to each of the composite indicators.

Figure 11
Graph, Combined energy hardship indicators by age group, HES 2015/16.

Whether to follow a ‘number of indicators’ or a composite indicator approach

The ‘number of indicators’ approach enables us to look at the depth of energy hardship, with table 8 showing significant differences between households experiencing just one indicator compared with households with two or more indicators. We noted that results are similar whether a number of indicators or a composite indicator approach is used.

We would recommend the ‘number of indicators’ approach due to simplicity and ease of understanding. However, since there is no significant difference between households with two, or three or more energy hardship indicators, this category could be combined for use in HES. Combining two or more indicators would also be advantageous due to the high sample errors for the ’three or more’ category in HES. However, this may not be such an issue when using the General Social Survey (GSS) physical environment supplement, as the GSS has a larger sample size.  

  • Share this page to Facebook
  • Share this page to Twitter
  • Share this page to Google+
Top
  • Share this page to Facebook
  • Share this page to Twitter
  • Share this page to Google+