We are now midway between the 2011 census and the next one in 2021, and that has significant implications for our understanding of even basic trends in population and households.

The UK census has been undertaken every ten years since 1801 (with the exception of 1941 and Ireland in 1921). It is a valuable resource and provides a unique opportunity to analyse and understand the housing market. However, as we get further from the previous census, we become more dependent on other surveys and data for up-to-date analysis. And, although the census may not be perfect, these alternative resources are inevitably constrained due to their smaller scope and sample. Therefore our understanding of how the country and the housing market is changing becomes increasingly uncertain as we get closer to the next census.

This uncertainty appears in even the most basic of variables. There is currently considerable debate about the level of net migration into the country and hence there is uncertainty over how many people actually live here. Even if we ignore migration data issues, it is still difficult calculating how many households there are. The 2008 projections had over-estimated the number of households in 2011 despite the population turning out higher than expected. We are now seeing a similar divergence between the 2012 household projections and the Labour Force Survey. Meanwhile the English Housing Survey provides a more volatile measure.

The main reason for the difference in household numbers is due to household representative rates or, more simply, average household sizes. The 2008 household projections had expected the average household size to continue the decline seen during the 20th century. Instead, the average size of households in 2011 was found to be the same as in 2001. The latest 2012 household projections for England have interpolated the change between 2001-11 and projected the decline in household size to resume, albeit at a slower rate. However, both the Labour Force Survey and English Housing Survey suggest that the average household size has actually increased since 2011.

These surveys have previous form in picking up unanticipated trends in household formation, with results in the mid 2000s suggesting that the then projected rise in one-person households was not occurring at the expected rate. Recent survey results indicate that the number of households in England has been growing by less than the ~220,000 per annum predicted by the 2012 projections. Household formation looks to have been constrained by a number of factors including affordability and the lack of new homes. This suggests that the backlog of housing need has continued to rise. Unfortunately we will probably have to wait until the 2021 census to find out whether this trend in household size is correct or not. Meanwhile, this uncertainty creates a number of issues including knowing how the unknown total number of households is split between housing tenures.

Idle Speculation

I’ve looked at what has happened (or not) to household size and possible reasons why in previous notes* but there is still uncertainty around the lack of change between 2001 and 2011. The 2012 household projections just interpolated the small change in household size between the two censuses but it seems unlikely that this is correct. Identifying if, why and when there were changes to the average household size will be essential in creating more robust household projections. Meanwhile, I can speculate on a possible trajectory for household numbers and the average size.

Historic household formation rates have tended to be constrained by the net change in dwellings as shown by Fig 3. Therefore I have used the change in dwellings between 2001 and 2011 as a constraint on household formation and as a guide to the distribution of change within the period (Fig 4). This approach suggests that the decline in household size continued between 2001 and 2009, albeit at a slower rate than originally projected (Fig 5). Household sizes then began to increase as the recession hit and new housing delivery fell. It remains to be seen whether the most recent decrease in the Labour Force Survey’s average household size shows an actual turning point or just volatility in the survey. This year’s Labour Force Survey will help but we will still have to wait for the 2021 census to find out what has actually happened to total household numbers.

Households can only form if there’s somewhere for them to move into. Changes of use, conversions and bringing empty homes back into use have all helped increase the housing stock by more than just new housebuilding. However, these solutions won’t work forever and so more homes are needed to cope with a growing population.

Although I’ve used the dwelling stock figures as a guide, there is no guarantee that this data is correct either. The dwelling stock figures were revised upwards following the 2011 census and the post 2011 figures could also be undercounting current changes in dwelling stock.

Although the Labour Force Survey shows a decline in average household size for 2015, there are a couple of issues that suggest this may not be the case: the survey appears to assume a large rise in people living in communal establishments and so not counted within the household resident population. It also uses projections for the 2015 population and these may be revised.

Taken from
Housing Market Notes: