Ready to Help?

Chapter 3. Anticipating, monitoring and reacting to inflows of refugees and other vulnerable migrants

Abstract

Shortcomings in collection, circulation and use of information hinder public policy action in dealing with inflows of refugees and other vulnerable migrants and supporting their integration outcomes. This chapter examines how systems can better anticipate demand and how outcomes can be monitored over time.

Can early warning mechanisms help prevent a crisis in the face of future large-scale inflows?

The world was caught largely off guard by the mass migration movements from the Middle East and Africa in 2015‑16, even if there were ample early warning signs (OECD, 2018[1]). Similarly, increases in asylum seeking in other regions of the world have often occurred following early warning signs that were not adequately captured. Putting together an early warning system requires investment and collaboration. A number of considerations can be taken into account in developing a platform (OECD, 2018[1]).

First, early warning and alert systems based on monitoring flows in real time require significant resources and information sharing among countries, as well as updated intelligence on the functioning and the evolution of smuggling networks.

Second, early warning systems require a good understanding of trigger points to minimise the risks of ignoring relevant signals (false negatives) and of overstating irrelevant signals (false positives).

By linking to specific analysis of diasporas and social networks, early warning and alert systems may pick up the scope of looming migration surges and their destinations.

New data allow for expanding the range of signals to consider. “Big data” – high-volume, velocity and variety data – have allowed fresh analysis to be used in developing early warning systems. Communication-based information sources such as social media, Internet searches, smartphone apps, the IP addresses of website logins and emails, and call detail records all represent grist for analysis. Geolocation further increases the potential of these methods, but also the risks connected with privacy and confidentiality (Box 3.1).

Box 3.1. Big data use in early warning systems to detect risk of forced migration

Two examples of using big data to provide early warning are the Gdelt project and Google Analytics. Gdelt identifies global trends and emerging social, political and economic risks worldwide. Monitoring news in over 100 languages, Gdelt can track events, locations, organisations, people, etc. Google Analytics’ search term frequency reveals terms that have been searched in great numbers, as well as the language in and location from which they were searched. (Böhme, Gröger and Stöhr, 2017[2]; Connor, 2017[3]) have determined that under certain conditions, search query data can allow forecasts of migration flows.

Source: (OECD, 2018[1]).

Foresight methods focused on specific migration categories or corridors may inform policy if the time horizon is not too distant (less than 10‑15 years). Longer time horizons may limit their relevance for policy makers. However, these longer-range exercises can help build consensus around long-term challenges and objectives, and help future-proof policies.

Due to data limitations and uncertainties, tools to model flows of forced migration are resource-intensive and relatively fragile, notably in the context of major external shocks. Nevertheless, in cases where the investment can be mutualised and the results are fully integrated into a multi-dimensional/inter-ministerial response system, it may well be worth investing in such tools.

More could be done to strengthen co-ordination efforts with regard to the way data are gathered for forecasting and the sharing of data within and beyond the European Union, with transit countries and with other OECD countries.

Whatever information system is adopted, policy makers must trust that information so that they act promptly upon it with a full understanding of the uncertainties – because foresight will indeed contain uncertainties.

A “post mortem” exercise should also be undertaken systematically following all major crises and shocks at national and regional levels, in order to improve policy response and preparedness. One recent key lesson is that although information systems were in place prior to 2015/2016, policy makers did not always act on the indications it provided.

What information needs to be improved to better monitor integration outcomes and inform integration policy?

While information on flows and stocks has steadily improved, much remains to be done to improve monitoring of the integration outcomes of refugees and other vulnerable migrants, and linking of this monitoring to the evaluation and development of integration policy. A recent international statistics forum held at the OECD pointed to shortcomings in statistics on refugees and other migrants in vulnerable situations (Box 3.2).

The measurement of well-being outcomes by migrant status – through for example identifying refugees separately from other categories of migrants – is very challenging for those gathering official statistics. While administrative data sources often include information on migration status, household surveys are the most appropriate vehicle for measuring well-being outcomes across a range of dimensions – but then, household surveys rarely identify migrants by category of entry. Further, sample design that is appropriate for the overall population may not be sufficient to capture information about migrant groups. Since migrants both tend to account for a relatively small share of the population in OECD countries and tend to live in geographically segregated areas of the country, sample sizes may be too small.

Censuses or administrative records that contain the most detailed information on migrants in terms of provenance, reason for migrating and key demographic variables only rarely include information on integration outcomes beyond income, labour market status and education. Some OECD countries, such as Australia and Canada, are making use of integrated data sets that link administrative data with censuses or other surveys.

The need for more detailed and granular data on migrant outcomes than what is found in surveys requires a sufficiently large sample and the inclusion of additional questions to identify different sub-groups. Adapting the methodology of existing surveys, such as by boosting sample sizes, will improve the representativeness of the migrant sample obtained (Šteinbuka, 2009[4]).

Improving survey design to reduce non-response rates will also need to be considered. The European Union Labour Force Survey (EU LFS) and the European Union Statistics on Income and Living Conditions (EU SILC) have both included special ad hoc modules on migrants’ outcomes in recent years; these experiences can inform improvements to the measurement of migrant outcomes in other surveys and countries.

Meeting the need for more detailed and granular data on migrant outcomes will also require inclusion of additional survey questions that allow identifying different sub-groups. In addition to the important demographic and socio-economic variables that are usually included in household surveys (e.g. age, gender, educational attainment), some migrant-specific questions should be considered. These include country of birth, duration of stay and reasons for migrating. The experiences of countries that are already using such variables – for example, from 2017 the German Labour Force Survey will include a question on reasons for migrating – could provide useful lessons for others.

In cases where it would be too difficult to modify the methodology of an existing survey, and where resources allow, developing a special, targeted survey of migrant outcomes could be considered.

Special efforts are needed to ensure that data collections cover and enable identification of the most vulnerable migrants, especially those who are unlikely to be reached through standard household surveys. Some countries have made advances in targeting specific migrant groups who may be at greater risk of well-being deprivations; this is the case with Australia’s Building a New Life in Australia survey, which focuses on the experiences of recently arrived humanitarian migrants.

More longitudinal data are needed to understand the evolution of different well-being outcomes for individual migrants over time. More national longitudinal surveys of migrant outcomes should be carried out where possible, and efforts to harmonise surveys across countries could help to facilitate international comparisons.

Box 3.2. The International Forum on Migration Statistics

To address the gaps in knowledge and statistics regarding migration, the first-ever International Forum on Migration Statistics was held on 15‑16 January 2018 at the OECD (www.oecd.org/migration/forum-migration-statistics). Much of the discussion at the forum focused on refugees and migrants in vulnerable situations. Over 500 people from almost all countries of the OECD and many non-members participated in this event, which featured 240 speakers and presenters. Some of the high-level keynote speakers were from partner organisations (IOM and the UN) and from national statistical offices (NSOs. Conclusions were published by the OECD in a joint brief with the International Organization for Migration and the United Nations Department of Economic and Social Affairs (2018[5]).

In addition, a Call to Action – “Protecting children on the move starts with better data” – was jointly published in February 2018 by UNICEF, UNHCR, IOM, Eurostat and the OECD (2018[6]). The publication was prompted by the observation that data are not collected, or are too poor, to provide information about children on the move: their age and gender; where they come from; where they are going; whether they move with their families or alone; how they fare along the way; and what their vulnerabilities are. These important gaps in child-specific information need to be addressed.

References

[2] Böhme, M., A. Gröger and T. Stöhr (2017), Searching for a Better Life: Now-casting International Migration with Online Search Keywords, https://www.wider.unu.edu/sites/default/files/STOEHR%2C%20Tobias_paper.pdf (accessed on 06 September 2018).

[3] Connor, P. (2017), Can Google Trends Forecast Forced Migration Flows? Perhaps, but Under Certain Conditions, Pew Research Center, Washington.

[5] IOM/OECD/UN DESA (2018), What were the key messages of the International Forum on Migration Statistics, OECD Publishing.

[1] OECD (2018), Can we anticipate future migration flows? Migration Policy Debate n°16, OECD, Paris, http://www.oecd.org/migration/mig/migration-policy-debate-16.pdf (accessed on 06 September 2018).

[4] Šteinbuka, I. (2009), How to improve social surveys to provide better statistics on migrants, http://ec.europa.eu/eurostat/documents/1001617/4339944/Improving-survey-dataI-Steinbuka.pdf/e90f6527-af4d-4585-8f3d-d7dde093b148 (accessed on 15 February 2018).

[6] UNICEF et al. (2018), A call to action: Protecting children on the move starts with better data.