A research paradigm for the uncertain world (Kupiszewski and Kupiszewska)

The theoretical underpinning of this volume is the analysis of changes in migration patterns over time, discovering regularities and formulating a theory describing the observed migration processes. Marek Okólski offered such a theory in chapter 1. Migration forecasting allows for the extension, albeit with a degree of uncertainty, of migration patterns into the future. In an ideal and reasonably predictable world, it would be possible to use forecasts to identify the moment of transition from an emigration to an immigration country for those countries that are in the transition phase. To do so, the IDEA project team (see note 9 in Introduction) tried to introduce some quantitative measure of the stage in the migration cycle and to develop some kind of ‘trajectory’ on which each country could have been located, depending on its migration system characteristics. However, operationalising the theory turned out to be an impossible task due to its longterm perspective. Also posing a challenge were the irregularity of the migration processes involved, complexity and inherent differences in the historical changes in migration observed in various countries (see chapter 2) and uncertainty embedded in forecasting (see later in this section). Substantial impact of historical and political processes (for example, the demise of communism in Central Europe and the 2010-2011 revolts sweeping Arab World countries), which have profound impact on migration flow counts, is impossible to account for in a theoretical framework, although it is very apparent in statistics. This failure is not a surprise: Sture Öberg and Babette Wils (1998) and Marek Kupiszewski (2002) noted that theories of migration are always difficult, and very often impossible, to operationalise and use for forecasting purposes.

Apart from purely theoretical considerations, migration forecasting serves utilitarian purposes by supporting the formulation of migration policies. The authors of this chapter feel strongly that migration policies are often formulated without a firm base of factual support and are rather predicated on qualitative information only. While chapters 3 through 8 present data on past trends, this forecasting exercise is done in an effort to provide some quantitative insight into the future and thus provide a better factual basis for policymaking. The Bayesian forecasting method, implemented in the IDEA project, allows for combining qualitative information with statistical data. Moreover, it has an added bonus: very clearly showing how much uncertainty is involved in the forecasts.