Most of the presentations from this session have been made available for download and can be accessed from the hyperlink in the title line of the abstract. Copyright in all presentations remains with the author(s), to whom permissions requests should be made. The views expressed are those of the author(s).
Demographic potential and component drivers for ethnic group populations in the UK, 2001-2051
Philip Rees, Pia Wohland, Paul Norman, Peter Boden, University of Leeds
The first question that users of population projections ask is what is driving the results presented. Part of the answer lies in the assumption made for future mortality rates, future fertility rates and future migration rates and flows. But future populations also depend on the age-sex structure of the starting populations: it takes 70 to 80 years for the effect of the starting structure to work itself out. The contribution of the existing age structure to the future population is termed the demographic potential. We adapt and extend a methodology used by Bongaarts and Bullateo (1999) to projections in which both international and internal migration components are influential for population sub-groups whose starting age structures are very diverse. The paper reports on the methods, assumptions and results of running a systematic series of scenario projections which enable us work out the drivers of future ethnic population change for the UK's local areas over the time horizon 2001 to 2051.
Email: p.h.rees@leeds.ac.uk
Projections of population, health, labour force and households in localities in Northern England
Philip Rees, University of Leeds; Chengchao Zuo, University of Leeds; Pia Wohland, University of Leeds; Carol Jagger, University of Newcastle; Paul Norman, University of Leeds; Peter Boden, University of Leeds & Edge Analytics Ltd; and Martyna Jasinska, University of Leeds & Edge Analytics Ltd
Northern England faces important demographic challenges over the next 25 years: population ageing, a more ethnically diverse society, moderate demographic growth but with large differences between most and least successful localities. We examine past and future demographic changes in Northern England and their effect on population health, labour force size and household numbers, drawing out implications for public policy. We use new projections for England's local authority (LA) populations by ethnicity from 2001 to 2036. The population of Northern England will grow between 8 and 12%, depending on projection scenario. So, aggregate demand for goods and services in Northern England will continue to increase. The population will experience, however, falling old age support ratios. To maintain current ratios, pension entitlement age will need to increase to 70 by 2036 rather than the 67 planned in the 2007 Pension Act. Population ageing will impact numbers with limiting long term illness, which will increase more than the population as a whole because the age structure will become older. A decrease in the labour force will occur as the baby boomers retire, if labour force participation rates remain constant. Action needs to be taken to encourage employers to introduce flexible transitions to retirement that reward older workers who wish to continue in work. We also project the numbers of households, which increase in line with population as we judged that that the long term reduction in household size would come to an end.
Email: p.h.rees@leeds.ac.uk
Estimating the impact of demographic change on housing, health and social care in the North of England
Lisa Buckner, University of Leeds; Gary Fry, University of Leeds; Katen Croucher, University of York; Martyna Jasynska, University of Leeds
The population in the UK is ageing at an unprecedented rate with more people over state pension age than under 16. In addition the population is also becoming more diverse. These demographic changes have a number of housing, health and social care implications and present key challenges for public policy and infrastructure planning. . Results from a recent research project which sought to quantify the size and impact of these demographic changes in the North of England between 2011-2036 with respect to housing, health and social care are presented and the policy implications discussed. Although these issues are not unique to the North of England, this region is experiencing population ageing together with pockets of high levels of deprivation and poverty, contrasting urban and rural areas and communities disproportionally affected by recession with some being more resilient than others. Local authority level geographical differences are also presented with respect to these area types: deprivation;. rurality and resilience.
Email: l.j.buckner@leeds.ac.uk
Does the period life table provide more valid estimates of length of wait than the census- or event-based methods of estimation?
Paul W. Armstrong, University of East London
Aim: To discover whether the period lifetable provides more valid estimates of length of wait in prospect than are obtained using the lengths either of (current) waits captured at the time of the mid-period census or of the (prior) waits of those admitted over a specified period.
Method: We determined whether there was a surplus (or a deficiency) of admissions within the cross-tabulation of cohort and waiting time category which straddled each census. We used census-, event-, and lifetable- based methods to produce three period-specific estimates of the percentage of waits of 0-2 months, and we determined whether length of wait grew shorter (or longer) from one period to the next. We used Lambda B to indicate the extent to which we were able to predict the direction of change in length of wait once we knew the direction of change in size of list.
Results: We found a direct correlation between change in length of wait and change in size of list, as expected under the stock-flow model, when length of wait was estimated using the lifetable for the period (LB = 58.33, 95% CI = 29-88), but we obtained a null correlation when we used census-based estimates (LB = 6.45) and we obtained an inverse correlation when we used event-based estimates (LB = 57.14, 95% CI = 31-83).
Discussion: The period lifetable appears to provide more valid estimates of length of wait and should therefore be substituted for census- and event- based methods of estimation, wherever possible.
Email: p.w.armstrong@uel.ac.uk
Ageing in place or transitory populations? Scottish population change by ten year period – cohorts, 1991 to 2001
John Paul McCarthy, University of Leeds
The development of the Scottish Neighbourhood Statistics (SNS) programme has enabled improvements to be made to the availability, consistency and sophistication of Scottish 'small area' statistics. This in turn has presented the opportunity to add depth to the understanding of sub-national population change and the components of population change for new spatial units. The research presented here utilises vital statistics, population estimates, and interim life tables to compute and analyse rates of population change, net migration, and natural increase at Council Area and Data Zone levels between 1991 and 2008. These rates have been disaggregated by age, enabling the analysis to be conducted for separate period-age groups and period-cohorts. The picture which emerges for Scotland is one of a complicated 'mosaic' of sub-national population change over time, with the populations of some areas experiencing 'ageing in place' and other areas possessing more transitory populations. A comprehension of the distribution of these 6,505 small areas will be beneficial for public and private service allocation.
Email: geo1jpm@leeds.ac.uk
Improving the United Kingdom's immigration estimates
James Raymer, University of Southampton; Sarah Crofts, Office for National Statistics
In this presentation, we first identify the current issues concerning the supply of migration data required by Eurostat. We then introduce several estimation techniques and strategies that can be used to improve the quality of data that come primarily from the International Passenger Survey (IPS). For illustration, we focus on immigration by age, sex and citizenship. We present a methodology that allows for smoothing, repairing and augmenting the IPS data. The results represent synthetic data benchmarked to IPS marginal totals that are deemed reliable. Furthermore, we show how the methods can be extended to estimate other multidimensional flow tables, including those for emigration.
Email: j.raymer@soton.ac.uk
Integrated modelling of European migration: A first look at the final results
Arkadiusz Wisniowski, Jonathan J. Forster, James Raymer, Peter WF Smith, Jakub Bijak, University of Southampton
In this presentation, we present the final results from the Integrated Modelling of European Migration (IMEM) project. This two year project is funded by NORFACE and brings together three international teams specialising in migration data, uncertainty and modelling. We apply a Bayesian model to harmonise and correct for the inadequacies in the available data and to estimate the completely missing flows. Recent international migration flows amongst countries in the European Union (EU) and European Free Trade Association (EFTA) are estimated by using data provided by Eurostat, covariate information and expert judgements. The result is a synthetic data base with measures of uncertainty for international migration flows and other model parameters.
Email: a.wisniowski@soton.ac.uk
Inter-regional migration in Europe: a spatial interaction modelling perspective
Adam Dennett, Kimberley Claydon, Pablo Mateos, University College London
The study of migration – both internal and international – is of perennial interest to scholars of population. This interest, it could be argued, is at least in part due to the conspicuous lack of comprehensive and reliable data at scales from the global to the local. A paucity of data for many years, combined with a continued need for knowledge of migration flows, has fuelled a significant body of research focusing on the underlying influences acting upon migrants; and key to this focus in many cases has been the specification, development and testing of migration models. These models have tended to fall into two main camps; those based on fixed-rate demographic methods in the Markov tradition which rely heavily on prior information to produce estimates; and those based upon Newton's laws of gravity which concentrate on the influence of spatial structure. In the latter of these two camps a bifurcation can be observed with some choosing to develop mathematical programmes based on entropy maximising/information minimising principles to derive estimates of migration flows, and others choosing log-linear/Poisson regression statistical derivations of the same models to achieve the same goal. A number of pieces of research have demonstrated the theoretical equivalence of the two spatial interaction approaches, but there has been little empirical comparison of the techniques. In this paper we evaluate the two approaches using internal (inter-NUTS2-regional) migration data for 21 EU countries. In doing so we also present a framework for modelling a full intra-Europe matrix of NUTS2 level migration flows.
Email: a.dennett@ucl.ac.uk
Modelling fertility schedules with a skew-symmetric distribution
Stefano Mazzuco, Bruno Scarpa, University of Padova
New fertility models had been recently proposed, trying to describe properly a fertility schedule – emerging in some European countries – with a pronounced additional hump at young ages. These models are mostly mixture models (e.g. Hadwiger mixture model or Normal mixture model). In this paper we take a different approach basing on known results on skew-symmetric distributions family, which is a generalization of the skew-normal distribution. Its pdf is particularly flexible and can have a bimodal shape. Such a fertility model shows a good fit both with fertility schedules with an additional 'hump' at young ages (typical of many English-speaking countries) and with 'traditional' fertility schedules (i.e. with no additional hump, typical of Mediterranean countries). Fertility models based on mixture distributions show a good fit with English-speaking countries fertility patterns but sometimes with traditional ones. Moreover, we looked at trends over time of parameters of these models and found that parameters of the new model are more stable while mixture models parameters are much less stable over time, suggesting that these models can be overidentified, especially when there is no additional hump in the fertility curve. In addition, we empirically studied the value of parameters in relation to the shape of fertility pattern, in order to give a proper demographic interpretation to them. We therefore conclude that a model based on skew-symmetric distributions might be a useful tool for studying old and modern fertility patterns and forecasting the future trends.
Email: mazzuco@stat.unipd.it
Developing a relational model of disability
Alan Marshall, University of Leeds; Paul Norman, University of Leeds; Ian Plewis, University of Manchester; Ludi Simpson, University of Manchester
Local estimates and projections of population with disabilities distinguishing disability type and severity are important for planning purposes to inform the provision of specialist services, equipment and support (Siegel 2002). Projections of disability, which require local estimates as their base, quantify the impact of trends in population and age-specific disability rates on future disabled populations, enabling practitioners to plan for future needs. This paper extends relational models beyond their original application, the estimation of mortality schedules (a schedule is a set of age specific rates), to the estimation of disability schedules Relational models are developed that accurately quantify the relationship between national schedules of limiting long term illness (LLTI) (census) and particular disability types (Health Survey for England). The specification of the relational model depends upon the complexity of the age-specific relationship between LLTI and particular disability types. For some disability types a simple relational model developed by Brass (1973) with two parameters provides a good fit. For other disability types a modified version of a relational model originally proposed by Ewbank et. al (1983) with three parameters is required. A procedure to select the best fitting and statistically robust model is developed. The relational approach to the estimation of disability schedules is particularly useful where local schedules of disability rates are either unreliable or absent. In such scenarios LLTI data (census) provides the reliable standard schedule from which local disability schedules then can be derived. An evaluation of the use of relational models for this purpose is undertaken. Brass, W. (1971). Biological aspects of demography. On the scale of mortality. Brass, W. London, Taylor Francis: 69-110. Ewbank, D. C., Gomez de Leon, J. and Stoto, M. (1983). "A reducible fourparameter system of model life tables." Population Studies 37(1): 105-127. Siegel, J. (2002). Applied Demography. London, Academic Press.
Email: a.d.marshall@leeds.ac.uk
Methodological triangulation is applicable to health and mortality research in demography!
Christian Wegner, Priska Flandorfer, Vienna Institute of Demography
Our paper contributes to the discussion about the use and successful adaption of methodological triangulation as an innovative new methodological approach in demography. The triangulation combines qualitative and quantitative methods and complements those results to a comprehensive picture of the phenomenon under study. The approach has been used in some health studies, but has only scarcely been applied in demographic research. We evaluated this approach by considering a know phenomenon - sex differences in smoking - in order to illustrate how this method could be applied in health and mortality research. The results present that the triangulation replicates determinants which were examined in previous quantitative analysis and expands it by new factors, especially due to the qualitative analysis. An important new factor however is the change in male health behaviour which can be characterised as the adoption of feminine attributes. Due to the positive evaluation of the triangulation method, we discuss relevant issues and describe why we consider this methodological approach as highly beneficial not only for health and mortality research but also for studying other demographic phenomena.
Email: christian.wegner@oeaw.ac.at
An alternative measure of mortality: the mortality synthetic indicator
Michela C. Pellicani, University of Bari; Alessandro Polli, University of Rome 'La Sapienza'
Aims Main aims are the estimation of a non weighted indicator of mortality in the Italian population and its temporal comparison with the life expectancy. The construction of a non weighted (synthetic) indicator of mortality has not been matter of detailed discussion in the scientific literature and, as a matter of fact, there's only a very limited number of publications on this subject (Sardon, 1993 and 1998; Termote, 1998; Barbi, Bongaarts, Vaupel, 2008). We would like focus on non weighted indicators for their interesting features. First, they are additive in the sense that also the application of a non weighted sum procedure allows the local decomposition of the index. Moreover, non weighted indicators are sensitive to the level of disaggregation by age, unlike the weighted ones. Data The data used are those contained in the cohort life tables calculated for Italy for the cohorts 1872-1917. Results Being, above all, a theoretical study, the main result of the present work is to demonstrate that the synthetic indicator of mortality has a key advantage over similar indicators based on a weighted sum, because in space/temporal comparison, it can neutralize the bias resulting from dynamic factors, like age-specific lifestyles resulting in different mortality rates which, as emphasized in the literature, change over time and space. The adoption of this procedures can neutralize the homogenization effect, a central feature of all weighted sum procedures.
Email: mc.pellicani@scienzepolitiche.uniba.it