Using MicMac to project living arrangements: an illustration of biographic projections
Nicole Van der Gaag, Netherlands Interdisciplinary Demographic Institute (NIDI)
Joop De Beer, Netherlands Interdisciplinary Demographic Institute (NIDI)
Peter Ekamper, Netherlands Interdisciplinary Demographic Institute (NIDI)
Frans Willekens, Netherlands Interdisciplinary Demographic Institute (NIDI)
MicMac is a multistate model for projecting both cohort biographies and individual biographies. This paper illustrates which kind of information the projection of individual biographies can add to projections on the cohort level. For this purpose we apply a prototype of the MicMac model to Dutch data on transitions between six categories of living arrangements: ‘in parental home’, ‘alone’, ‘with partner without children’, ‘with partner with children’, ‘lone parent’ and ‘in institution’.
The projections are based on empirical transition probabilities for women aged between 15 and 98 years. By assuming the probabilities to remain constant, the macro projections generate the distribution of the number of women by living arrangement for a synthetic cohort. On the basis of the same transition probabilities and a random mechanism, life courses (i.e. successions of living arrangements) for 10,000 individuals are simulated. Added together, these results are consistent with the projections at the cohort level. The results at the micro level show the diversity of individual life courses. Whereas the macro projections show that at young ages about a quarter of all women are living alone and at older ages less than 40 per cent, the micro projections show that a much larger proportion of all women live alone for some time during their life. Nevertheless only very few women live alone their whole life until they enter into an institution or die.
The results presented in this paper are just a few illustrations of the type of projections that can be made by using MicMac. The overall aim of MicMac is to be an instrument that can be used for the calculation of scenarios and impact assessment. The development of MicMac is as yet only in an early stage. Both the model and the data need further improvement.
Session 16: Bridging the micro-magro gap in forecasting