Using Census data to derive a new area-based measure of deprivation - Section 5: Housing tenure

Updated 19 March 2024

Section 5: Should housing tenure form part of a composite measure based on the Census?

As we discussed earlier, the relationship between income and housing generally indicates that it is low levels of income that drive housing outcomes, rather than vice-versa. This is also noted in a report by the Joseph Rowntree Foundation (2015). However, one of the key features of the UK housing market over recent decades has been the growing proportion of households in the private rented sector. The continued rise in rental costs combined with stagnating wages has meant that housing arrangements can lead to more people entering poverty once living costs are taken into account, with this particularly being the case for those in (private) rental accommodation.

We therefore carried out a sensitivity analysis to see how the inclusion of a variable relating to housing tenure in our composite measure altered our findings. Hence, we began by creating a field that indicated the proportion of households in an output area that were either privately or socially renting their accommodation. A principal component analysis was then carried out using this variable, alongside our two indicators on education and occupation. The correlation between a composite index including housing tenure and one created using only education/occupation (i.e. SEISA) was 0.97. Typically, when attempting to determine which areas to target, organisations will tend to concentrate first on those in the bottom quintile. We therefore also assessed how the bottom quintile of our composite measure inclusive of housing tenure differed to one based on only education and occupation. A two-by-two matrix of the findings is illustrated in the table below

Table 2: Cross-tabulation of SEISA against a measure using occupation, education and household tenure. Figures represent output areas.

However, when using MSOA income data for England and Wales to evaluate the incomes of those in the bottom quintile of the two composite measures, we found median incomes in the bottom quintiles to be equivalent. Furthermore, the average incomes in output areas that emerge in the bottom quintile of SEISA, but a higher quintile when including housing tenure, are slightly lower (at £380) when compared with output areas that are in the bottom quintile of a measure that uses housing tenure, but not SEISA (£390). There is therefore no strong theoretical or empirical evidence to suggest that including housing tenure can materially improve the measure.

Table 3: Cross-tabulation of SEISA against a measure using occupation, education and household tenure. Figures represent median net equivalised weekly household income estimates after housing costs in England and Wales (£)

Next: Section 6: How does our measure compare to the Indices of Deprivation?