Methodology for ‘To build, or not to build: that is the question’

To arrive at the figures in our analysis requires a number of bits of information:

1) The estimated number of pensioner and non-pensioner benefit units in the UK in 2060.

2) The stock of social housing in 2060 and the stock required to keep the current ratio (0.88) of social and affordable rented dwellings to benefit units in the social rented sector constant.

3) The share of pensioners and non-pensioners that will rent, rather than own their own home, in retirement in 2060 and how many of these rent socially.

4) The amount spent on housing benefit for the pensioner population and housing benefit saving from housing someone in the social rented sector.

1)  The first of these can be estimated using the ONS’ households and partnership projections which provide estimates of the number of benefit units in England out to 2041. These are supplemented with household projections in Scotland and population projections in Wales and Northern Ireland (household projections are not available for Wales and Northern Ireland). Using the benefit unit projection for England we can estimate the number of benefit units in the other countries of the UK using the ratio between England’s household projections and Scotland’s and between England’s population projections and Wales’ and Northern Ireland’s. Once we have UK benefit unit projections out to 2041 we roll these forward to 2060 using UK population projections. This gives us 41 million benefit units in 2060 (compared to 34.4 million now), 13 million of which are pensioners (8.4 million now) and 28 million non-pensioner families (26 million now).

2) To estimate the stock of social housing in 2060 we assume that the building rates achieved between 2012 and 2017 continue to 2060. This increases the stock by 13,000 a year, meaning that by 2060 we have 5.55 million social and affordable rented dwellings (up from 5 million now). To continue to house 16 per cent of the pensioner population in social housing we will need to build an additional 9,000 a year because of population growth. Depending on what proportion of pensioners rent in retirement, and what proportion of these rent socially, we will need to build more homes to keep the current ratio of benefit units to social rented homes constant at 0.88. Although this analysis is primarily about pensioners we need to estimate total social housing stock, not just stock used by pensioner benefit units, because we do not have figures for building for different age groups and obviously stock is fungible across age groups.

3) To estimate the share of pensioners that will rent in retirement we use the estimates given in our previous research. To isolate just the impact on pensioner housing benefit spend and shifts in pensioner tenure we assume that the tenure patterns of non-pensioners remains the same as now. To estimate how many pensioners rent socially we either assume that the current percentage point share (16 per cent) rent socially in retirement or that the share of renting pensioners that rent socially remains the same as it is now (76 per cent). Under the former scenario 16 per cent of pensioners rent socially in 2060 under both our optimistic and pessimistic scenarios whereas under the latter 20 per cent of pensioners rent socially under the optimistic scenario and 26 per cent under the pessimistic scenario.

4) Our estimates of housing benefit spending is the same as in our previous research where we take the current figures for the average amount of housing benefit awarded to retirees renting privately and socially (£94.71 and £84.99). We make no assumptions about spending on housing benefit for non-pensioners and spending on this group is not included in our estimates of the possible increase in spending.