AQMeN Annual Lecture
16th December 2009
The Royal Society of Edinburgh, 22-26
George Street , Edinburgh
Patrick Sturgis: A National Strategy for Research and Capacity Building in Research Methods: the National Centre for Research Methods
Professor Patrick Sturgis, Director of NCRM, opened the event with an introduction to the National Centre for Research Methods, NCRM.
Phil Rees: The Dynamics of Population Large and Small
:Processes, Models and Futures, with Some Illustrations for the Scottish Population
Emeritus Professor Phil Rees, from the School of Geography, University of Leeds presented on the the processes of population change, the models we use to simulate the processes and what we can say about the future.
We have always been interested in change. Recent observations of background radiation suggest that the universe has existed for 13.73 (±0.12) billion years. The universe’s fate could be “big freeze”, “big rip”, “big “crunch” and many others, each with its own set of predictive equations and insufficient data to test them. This will sound familiar to demographers wrestling with demands of policy makers to inform them about the future of national and sub-national populations and their productive capability/destructive potential. This Lecture reviews the processes of population change, the models we use to simulate the processes and what we can say about the future. Key processes influencing most national populations are ageing driven by low or declining fertility, rising life expectancy and international migration gains and losses, creating older and more diverse populations. Within national borders internal migration redistributes our changing populations. But many policy questions need more than demography: we need to understand changes in educational attainments, participation in the labour force, family and household formation (many demands are household generated), changes in retirement dependant on state support and savings, and on health/illness outcomes.
The Lecture will illustrate the processes and recent experience, drawing on evidence from Scotland. A variety of strategies are available to model our changing populations. Sometimes you have to model stocks as a time series (when a consultancy project needs swift execution). However, it is better to study and model transitions between states. The classic cohort-component model is still at the heart of the projection of large and small populations, in many forms as single region, bi-regional, multi-region models, whether implemented using macro-populations experiencing average group intensities or using micro-populations of individuals sampling their transitions from a macrodistribution.
What is vital in implementing such models is that the input intensities match model intensities in definition. Even the best demographers get this wrong. Sometimes our input data are inadequate and we must resort to other modelling strategies. To solve the student migration issue (poor data on the migration of students on graduation), recent work has introduced agent based models into micro-simulation, for example. In other situations, no direct estimates of inputs to our projection models are available and the needed inputs must be modelled from proxy variables, survey data or administrative sources, as has been done recently for ethnic group mortality, fertility and immigration in the UK.
The Lecture will report on some recent population projections of Scotland’s
population, distinguished by ethnicity. To drive our projection models we need to make assumptions and to deal with the uncertainty that they carry. Traditionally high and low variants are used, but will be replaced in the future by stochastic projections. These use time series models or historical analysis or expert opinion to produce error distributions of the main projection drivers for sampling. We also need to think out of the box and develop scenarios that establish “what might happen if …”. What if the conventional view of continuous life expectancy improvement was replaced by predictions of higher mortality from the obesity epidemic? What if climate change crises caused a rise in the migration of environmental migrants? What if resource depletion (the end of oil) reduced mobility and migration in Europe? These are scenarios which the European Commission has charged a network of European demographers to explore. Here we are into not only “known unknowns” but “unknown unknowns”. The lecture will review what assumptions might be made about the future and how these might play out in Scotland.
