Spatial modelling fuel poverty in Northern Ireland
In the following guest article, Ryan Walker outlines how area-based modelling could improve the process of identifying and targeting fuel poor households.
Current fuel poverty initiatives cannot identify households in fuel poverty. In the search for improved targeting, there is an increasing interest on delivering fuel poverty policy from area-based platforms. This paper outlines a case study from Northern Ireland of an evidence-based approach to area-level targeting of fuel poverty and its application in a recent energy efficiency intervention.
Recent audits (e.g. NIAO, 2008) show that anti-fuel poverty initiatives in Northern Ireland are poorly targeted. This is due to a variety of factors including definition of fuel poverty status, eligibility criteria, and referral systems (Liddell et al., 2011). As a result, those in greatest need of remedial measures are often missed. Finding and treating the most severely fuel poor households is a priority for policy in a region like Northern Ireland where fuel poverty is particularly widespread and severe: 42% of households were classified as fuel poor in 2011 (NIHE, 2013). The main fuel poverty initiative in Northern Ireland is the Warm Homes scheme, which installs energy efficiency measures in eligible households.
Area-based approaches offer several opportunities for improved targeting. Social disadvantage and poor housing conditions are known to overlap in certain geographic areas, producing identifiable pockets of fuel poverty. Economies of scale can be gained by targeting these areas, capturing a greater number of fuel poor households per unit cost, compared to existing self-referral methods (Elbers et al., 2007). Further, a focus on improving dwellings has potential for local area regeneration. Proactive, local targeting can also encourage uptake of measures (Sefton, 2004). Providing ‘something for everyone’ in a neighbourhood or street can remove a ‘poverty stigma’ that can be perceived with fuel poverty measures.
Developing an area-based model
Implementing an area-based policy requires knowing which areas should be targeted. Geographic Information Systems (GIS) techniques were used to construct a multi-factorial model, using data on incomes and social welfare benefits, fuel types and energy costs and dwelling energy efficiency. Analysis took place using administrative areas of approximately 125 households (census output areas or ‘COAs’) (N=5022). This high resolution model permits high need areas to be identified and pinpointed to quite small localities, on an evidence-based and objective basis (see Walker et al., 2012). Area-level income and welfare statistics are available from the Northern Ireland Neighbourhood Information Service (online). Energy costs were modelled using data from oil and natural gas suppliers. Energy demand was modelled using temperature data from the Met Office (places experiencing colder temperatures require more heating). Data on the age, size, type and value of housing was obtained from Northern Ireland’s national mapping agency (Land & Property Services) under the Northern Ireland Mapping Agreement. The model is summarized in Figure 1 below.
This model has been applied within a recent pilot project which set out to explore new area-based approaches to targeting. This project is known as the Affordable Warmth Pilot (AWP) and set out to identify households in severe fuel poverty and target them with remedial measures under the existing Warm Homes scheme. The highest scoring areas in the model are those where risk factors overlap and where severe fuel poverty is likely to exist. The model’s incorporation of social welfare data also accounts for the eligibility criteria for Warm Homes (see Boardman, 2010, p.63-65). Using additional GIS techniques, areas of predominantly social housing tenure (which are ineligible for the Warm Homes scheme) can be removed. The model should therefore enhance economies of scale by selecting areas where households are at risk of fuel poverty and who are also eligible for current schemes.
Local authorities (Councils) acted as the main facilitators for AWP. The model was used to map fuel poverty across each Council area to identify COAs in greatest need of targeting (Figure 2). After some discussion, four ‘high need’ COAs were selected for targeting in each Council area. Target COAs were mapped in fine detail and postal addresses were sourced, generating a list of around 400 target households per Council, of which 125 households were required to be surveyed. The main purpose of the survey was to assess the household’s eligibility for the WH scheme. If found to be eligible, they were referred onto the existing WH system (with the householder’s consent), through which they could receive practical heating and/or insulation improvements.
How well does the model work?
The survey also collected a wide range of additional information on household and dwelling characteristics and represents a rich, large-scale (N=2147 households) source of real-world data on the demography of households at risk of severe fuel poverty. In target areas, there is a predominance of low-income households (45% earning less than £12,000/€14,000 per year) who are living in dwellings with out-dated and inefficient heating and insulation kit (e.g. 39% of households were using a boiler over 15 years old).
This empirical data indicates that severe fuel poverty is likely to be a reality for many households in areas identified by the model. Indeed, anecdotal observations from Council staff indicate that very few of the households they visited were not in severe fuel poverty. By reflecting the real world situation, the GIS model could be argued to be a valid targeting tool for finding areas of severe fuel poverty.
A large proportion of households in areas identified by the model were eligible for the Warm Homes scheme: 56% of targeted households qualified for practical heating and/or insulation improvements. This figure well exceeds the 10% target of schemes like Warm Zones in GB, which use similar delivery methods. These initiatives target areas based on indicators of multiple deprivation which may not have the highest rates of fuel poverty or eligibility (see Boardman, 2010). Our model is more attuned to fuel poverty and identifies areas of severe fuel poverty (where households are in greater need) as well as areas with high eligibility rates that, if targeted, will yield greater returns on investment.
This case study shows how existing fuel poverty/energy efficiency initiatives can be delivered in a local setting, from an area-based platform, facilitated by an objective, evidence-based decision-support tool (the GIS model). It should be noted that the availability of up-to-date, high resolution data was key to developing this model. Empirical data highlights the validity of the model in finding areas of severe fuel poverty, where most residents are in need of (and eligible for) remedial measures. Using the model as an identification tool offers real potential for improving the identification and targeting of fuel poor households.
Boardman, B. (2010). Fixing Fuel Poverty: Challenges and Solutions.London: Earthscan.
Elbers, C., Tomoki, F., Lanjouw, P.F., et al. (2007). Poverty alleviation through geographic targeting: how much does disaggregation help? Journal of Development Economics 83 (1): 198-213.
Liddell, C., Morris, C., McKenzie, P., Rae, G. (2011a). Defining Fuel Poverty in Northern Ireland: A Preliminary Review. Coleraine: University of Ulster.
NIAO (Northern Ireland Audit Office) (2008). Warm Homes: Tackling fuel poverty. Belfast: TSO.
NIHE (Northern Ireland Housing Executive) (2013). Northern Ireland House Condition Survey 2011. Belfast: NIHE.
Sefton, T. (2004). Aiming High – An Evaluation of the Potential Contribution of Warm Front towards Meeting the Government’s Fuel Poverty Target in England. London: CASE.
Walker, R., McKenzie, P., Liddell, C., Morris, C. (2012). Tackling fuel poverty in Northern Ireland: An evidence-based approach to targeting. Applied Geography 34: 639-649.