The Future of Estimating Housing Requirements. How Can Demographic Forecasting Help Plan for Housing?

I would like to outline four themes that may be helpful to promote future work on housing requirements in the UK. First, what is the role of forecasting within planning? I think a two part formulation is a useful guide to what planners need. Forecasters should predict that part of the future which planners don’t control, so that they can concentrate on what they can control; and forecasters should predict the consequences of a plan. In our context, the ‘trend-based’ official projections encapsulate business as usual without the impact of future-changing plans, and in particular deal with fertility, mortality and international migration which are considered outside the planners’ control. Dwelling-led and jobs-led projections explore the consequences of a plan. The distinction, made a long time ago by RJS Baker (1972: 119-126) seems a useful one to me, not only because it makes sense of the variety of scenarios that planners deal with, but also because it highlights the need for demographic forecasters and planners to collaborate, with neither assuming a determining role.

Second, how should uncertainty in forecasting best be dealt with? The first two decades of the 21st  century are a time of great uncertainty in housing requirement. Norms in UK experience of international and internal migration are recognised as difficult to establish, and at the same time the supposed steady decrease in household size has been disturbed and although we can identify the reasons we do not know which of them are lasting factors and which are temporary. Uncertainty is always considerable in forecasting, but the current recognition of it allows us to consider how it should best be dealt with in the planning process. To do so requires considering the costs and the risks of under-providing and over-providing; this is a familiar choice for educational planners who balance the costs of permanent school buildings and temporary accommodation against the risks of over-provision and crowded sub-standard classrooms. In land-use planning would it sometimes be wise to provide less than the principal projection of housing required, in order to review new information a few years down the line, only releasing further land if that already released has been taken up? Or would it be advantageous – to provision of housing suited to those who require it – to release land further towards the upper end of the range of uncertainty?

Inspectors at examinations of draft Local Plans are confronted with a multitude of future scenarios which represent uncertainty, or at least lack of agreement, among forecasters about the future. Often the finer points of the construction of scenarios give significantly different indications of housing requirement, making planning decisions dependent on areas of technical understanding which are not well shared. There is therefore a need to set out the definitions of an acceptable set of scenarios that reflect different common assumptions about the future. As yet neither the Inspectors nor the producers of forecast scenarios have found a way of meeting this need.

Third, how can the capacity of public research to meet the forecasting needs of planners be repaired? The leading role of the Department of Communities and Local Government in reducing public expenditure after the banking crisis of 2008 coincides with the lack of co-ordination of that department with the Office for National Statistics that has resulted in incomplete analysis of the Census (completion promised by CLG (2015: 4) but not yet delivered), reduced household type outputs from household projections, and a lack of annual household estimates needed to monitor change between censuses. At the same time more devastating cuts in capacity have been forced on local government where many of the most experienced research staff have been paid to leave public service before their career was complete, and others have added demographic work to their existing loads. Most of the academic contributors to the interface between demography and planning are in this room and reaching the end of their official working lives. Though I am sure many will continue Alan Holman’s example and contribute for many more years, it is worrying that my recent search of UK textbooks on planning found scant mention of housing requirement and no mention of demographic modelling to support planning. What institutions will make the effort to strengthen capacity and to capture the understanding of existing and recently retired staff for the generations who will undoubtedly follow?

Fourth and finally, isn’t the notion of ‘predict and provide’ which informs planning in England, failing those who need housing? While ‘predict and provide’ is a crude characterisation, it does match fairly well the government’s guidance to start with projections of households to guide the assessment of housing need, and to provide land for at least that level of need. However, as the Planning Advisory Service clarifies (2014), the projections are of ‘effective demand’ by those that have the resources to buy or rent housing in the current market. Unless we also measure unmet need, suppressed in the market for many reasons discussed elsewhere (eg McDonald and Whitehead, 2015), the current housing crisis will be converted into a new normality as the lower household formation rates become the basis for future projections. When methodology of household projections is reviewed it must find ways of including concealed families and indicators of other suppressed housing need, while standards are needed for measuring backlog of unmet housing need and the housing need implied on average by households with complex  adult structures.

At present the second part of ‘predict and provide’ is not working well either. Even though the ‘predict’ part of the equation is an underestimate of need, the ‘provide’ part has not nearly matched it (McDonald and Whitehead, 2015). Of the solutions tried and proposed, none have worked except at the edges and insufficiently. If the provision of housing is to meet the need for it, ways must be found of helping the government out of their entrenched position against the social housing which clearly addresses the failure of market solutions.

References:

Baker RJS. (1972) Administrative theory and public administration. London: Hutchinson University Library.

CLG. (2015) Household Projections 2012-based: Methodological Report. London: Department of Communities and Local Government.

McDonald N and Whitehead C. (2015) New estimates of housing requirements in england, 2012 to 2037. Town and Country PlanningTomorrow Series Paper 17.


This entry was posted in Environment, Welfare by Ludi Simpson. Bookmark the permalink.
Ludi Simpson

About Ludi Simpson

Ludi Simpson has worked for local and national government, in university posts, and independently. He is honorary Professor of Population Studies at the University of Manchester, designer of the demographic software POPGROUP, was President of the British Society for Population Studies 2011-2013, and chair of the local authorities advisory group on population statistics to the Office for National Statistics 1993-2003. He writes and teaches on methods and issues in demographic forecasting, and contributes to current projects in the UK, Latin America and Australia, focused on sub-national demography and local planning. He co-authored Statistics in Society, Sleep-walking to segregation?, Making local population statistics and Ethnic identity and inequalities in Britain.

2 thoughts on “The Future of Estimating Housing Requirements. How Can Demographic Forecasting Help Plan for Housing?

  1. Ludi, I understand where you are coming from in distinguishing between trend and jobs/housing-led projections when thinking about national populations. My concern arises once if we extrapolate this to smaller geographies (regions, cities, parts of cities etc), where internal migration becomes relatively more important.
    We need to remember that past trends are simply the observed outcomes of past plans and past patterns of infrastructure provision (especially housing and transport). Therefore, when we project based on trend internal migration patterns and propensities, we are implicitly assuming that either:
    - past patterns of planning and infrastructure decisions will be repeated, or
    - that planning and infrastructure has no significant impact on where people live.

    Either of these assumptions would be subjective and brave. I worry that characterising trend based projections as showing that part of the future that is outside of planners’ control implies a degree of objectivity that is not justified. Instead, trend based projections show what would happen if planners did again, what they did over the period used to estimate the trend. Therefore, my preference would be to see more demographers tackle the subjectivity implicit in trend assumptions head on.

    regards

    Adam

    • Thanks Adam, that’s helpful. As you say, the official projections are ‘business as usual’, and represent one future based on the impact of past policies and circumstances continuing to have the same demographic impact as they did in the recent past. Planners and politicians then set to work and can affect population change, especially they can affect migration at local level. Demographers help by modelling the impact of each of the many plans. I think this is a helpful two-part description to the relationship between demographers and planners. I’m not sure it is helpful to call any of it subjective; we should clarify the basis of assumptions in each projection, and their approximation to any real future. I think that at present the planning regulations make a rock out of the official projections, insisting that planners should provide for that single trend-based official projected population, or more.

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>