The ONS have produced a detailed set of sensitivity analysis on the differences between the 2014 and 2016 based household projections which will be bound to enflame debate about whether they should be used for strategic planning purposes.
Lets out one thing straight. The issue is not whether the 2016 figures are more accurate but that the dramatic drop from the 2014 based projections broke the system, requiring an emergency reset and reconsideration.
A key methodological change was the 2016 based projections only using 2001 and 2011 census data, whereas previous projections spread the data points for projections over 4 decades. What this did was instead of averaging out changes over the business cycles it straddled the abnormality of the Great Recession, where household formation was suppressed for economic reasons. Whilst this made the projections ‘more accurate’ in the short term it considerably reduced there utility for strategic planning purposes, where the long term assumption was that if you built homes to meet projections over the long term then households would form to fill them.
This assumption has come under increasing pressure over the years as a shortfall of homes has not only suppressed household formation but also birth rates.
Demographic drivers are clear in the analysis with subnational population projection changes rather than changes in HRR rates the main driver in many areas; which in itself is not necessarily an expression of less need because of the feedback effect of lack of housing on fertility and emigration.
We have known for years that the ONS emigration model was broken and overestimated student overstayers. It is no longer even used by the ONS as an official statistic.This is seems distorted figures for Oxford and Cambridge. However the underlying economic realities are the same. If an international student doesn’t overstay to fill a high skill job then it is more likely to be filled by an international graduate. This indeed is the story of the latest ONS adjustments to international migration figures with EU migrants made up for by non eu migrants. These international workers not necessarily staying in Oxford or Cambridge due to a lack of housing for them to fill. The figures may have overestimates students at CMOX but not necessarily people with a desire to live/work there.
It is clear then that the medium – long term future of strategic planning has to be based on a more integrated model of job led migration and commutes, using demography as an input to economic modelling rather than a substitute.