Did you know that over 80 per cent of immigrants leave Switzerland again within 17 years of their arrival? Or that in In Italy, 15 per cent of immigrants from China move at least twice within the country in the five years following their entry?   

These are some of the examples of use of longitudinal data contained in the new Guidance published today by UNECE that will help answer questions about the effects of international migration long after people migrate.

Statistical work on international migration usually focuses on how to measure the numbers of people entering and leaving a country, as well as the characteristics of those people—are they male or female, young or old, what is their religion, their socioeconomic status, ethnic and national origin? This may sound simple but it brings many challenges, as migrants can be hard to identify and even harder to survey, while non-survey sources of data rarely include all the kinds of information that policymakers need. 

Challenging as this basic accounting of migrants is, though, to paint a complete picture the statistics on international migration need to go further. Policymakers need to know whether and how the characteristics of migrants change after they arrive at their destination, and how they integrate into and impact on societies.    

This means measuring change—change in residence, change in legal status, and change in socio-economic outcomes. This is best accomplished using longitudinal data; that is, data where the same individuals are followed over time. We could compare information about people who migrated at different times, say, those that migrated one year ago, two years ago, and so on, but this makes it hard to separate out the effects of the passage of time from the different circumstances when people migrated. Longitudinal data get around this problem by tracking the same individuals, so we know that any changes we see are real-life changes experienced by those individuals. 

The Global Compact for Safe, Orderly and Regular Migration underscores that for countries to monitor and evaluate how well they are fulfilling their commitments, they depend on quality data that permit tracking of change over time. 

With such data it would be possible to measure how long immigrants stay in their destination country and how gaps between immigrants and the non-immigrant population widen or narrow over time. Do immigrants’ wages come to resemble those of the non-immigrant population more closely as time passes? What about their unemployment rates? Longitudinal data can shed light on many aspects of migrants’ lives after migration, from health outcomes to childbearing to where they make their homes and what jobs they hold. 

Obtaining longitudinal data is not as simple, though, as just conducting a survey with the same respondents at different points in time: such surveys are expensive and it can be hard to maintain contact with respondents. With the increase in availability of administrative data, techniques such as data integration can be better leveraged to develop longitudinal datasets built from joining together various administrative sources. Many national statistical offices have sources that could potentially fulfil this need, but the opportunities are so far underused.  

The new Guidance, developed for UNECE’s Conference of European Statisticians by a task force of experts led by Statistics Canada, aims to help countries tap the potential of the data they already have. It consolidates the best available expertise on how to develop data sources and how to compile and disseminate key indicators and findings. The Guidance contains examples of the creation and use of longitudinal data sets from 11 countries from across the UNECE region. 

For example, Switzerland combines data from its register of foreign nationals and its population statistics files to create the Swiss Longitudinal Demographic Database.  In Italy, statistical linkage of records of residence permits enables analysts to see the connections between international migration and internal mobility. 

In Canada, the Longitudinal Immigration Database integrates immigration data and annual tax files. It shows, for example, that over several decades, less than a quarter of temporary foreign workers became permanent residents; but in recent years this has become more likely and occurs within a few years of obtaining the first work permit in Canada.  

The new Guidance is a crucial step in developing statistics on migration that truly recognize that migration and settlement are processes, not states, and that outcomes can be short-term, long-term and dynamic.