To help countries and researchers meet the complex challenges of measuring people on the move, UNECE today published a study on using new data sources to measure international migration, ahead of World Migration Day (18 December).
Migration is an emotional topic that occupies a central place in many political arenas. Policy decisions about migration management rely on statistical evidence, but producing that evidence is notoriously difficult. By definition, migrants move around and are therefore difficult to find and count; especially those who migrate without legal documentation, who may intentionally or unintentionally avoid being recorded in standard data sources.
Global initiatives, including the 2030 Agenda for Sustainable Development and the 2018 Global Compact for Safe, Orderly and Regular Migration, highlight the need for data to help the world tackle the many challenges posed by large-scale migration and other forms of cross-border mobility . These global agendas, combined with the growing demand for detailed and timely migration data from national decision-makers, have prompted the official statistical community to seek new ways to meet data needs. UNECE’s new study responds to calls from official statisticians to examine the state of the art in the use of these new approaches.
As stated in the study, the measurement of migration includes the measurement of the movement of people in and out of countries, and knowledge of the numbers and characteristics of these migrants. Conventional data sources include censuses, household surveys, and administrative records (data sources collected by governments for purposes other than statistics, such as health, education, and tax records), each with their own advantages and disadvantages. A census, for example, is only held every ten years in most countries, so while it might provide a rich and comprehensive picture of where and how migrant communities live, it cannot be relied upon for up-to-date information on large and sudden changes such as movement refugees as a response to the crisis. Administrative sources are much better at capturing changes in real time, but risk missing those who fall under the radar of a country’s administrative systems, which could lead to systematic bias as these are often the poorest, youngest and most vulnerable migrants.
Integrating information gathered from these sources is often suggested as a way to overcome such problems, but it is not a perfect solution. In emergency situations, such as disasters, conflicts and health crises, new and different sources are needed to provide data to understand rapid changes. The sudden and widespread restrictions on international travel that occurred at the start of the Covid-19 pandemic highlighted the limitations of existing migration data sources and methods.
This is where unconventional sources come into play. Data collected using mobile phones, credit cards and social networks – generally known as big data – could be useful for producing migration statistics when used in combination with conventional sources. There are many obstacles that lie in the way: affordability, accuracy and access to these new sources. But in the last few years, some examples have emerged that highlight their potential.
Examining the use of new sources
With the international official statistics community more eager than ever to explore the potential offered by these new sources and techniques, UNECE has launched a study of the current and planned uses of new data sources for the production of official migration statistics – a world first on the subject.
The study, conducted between 2019 and 2021 by an international working group of experts, surveyed participating countries to find out which data sources they use and how; the challenges they faced and what they learned; and the reasons behind their decisions to continue or stop working with these new sources.
This revealed several examples of successful testing or use of new data sources to produce migration statistics; either to supplement traditional data sources or, in some cases, to replace them, for example when exceptional circumstances have affected the availability of conventional data sources. For example, when Puerto Rico was hit by Hurricane Maria in 2017, a devastating storm that caused extensive damage and loss of life, and prompted a mass exodus from Puerto Rico to the United States mainland. The usual method for measuring migration – using annual household surveys – was not adequate because these surveys could not capture sudden and large movements of people, and in the chaotic aftermath of the hurricane, the possibility of even conducting such surveys was hampered. To make up the shortfall, the US Census Bureau combined survey data with data on passenger flights between the mainland United States and Puerto Rico, produced by the US Bureau of Transportation Statistics, to produce ‘mixed’ estimates of migration.
Addressing obstacles to progress
Although the potential of big data sources is considerable, the study found that in reality around three-quarters of the participating countries are not currently using them to measure migration because the barriers to effective use are still too high. Countries cited privacy concerns and legal restrictions that limit the statistical use of the data. However, a third of countries reported using new sources to produce statistics on topics other than migration. This suggests that such sources may still have a role for migration statistics in the future, if these obstacles can be overcome.
Overall, the study concludes, statistical offices see continued potential in the search for new data sources to address urgent and emerging migration data needs. They could help bridge the gap between the need for rapid responses to current migration issues, on the one hand, and the often long delays between data collection and publication of figures using traditional techniques, on the other.
Many countries are transitioning to integrated statistical systems, and new data sources could be an important component of these, potentially serving as benchmarks against which to check the quality of other information.
One thing the study identified as a key missing piece in moving to better exploit these new sources is the sharing of information about their tools, methods and country learning. With this in mind, in collaboration with the study, UNECE has developed an online repository of innovations in migration statistics, bringing together scholarly articles and case studies that could help official statisticians and researchers.
Covering a range of migration topics including international migration, internal migration, human mobility, population displacement and population distribution, the tool will grow as countries’ experiences become broader and more complex, with the aim of becoming a key reference supporting the use of new data sources and related methods for creating migration statistics of the future.