PREDICT: Migration Algorithms PREDICT

The project aims to create three calculation models (algorithms) for Migration divided into inflow, through flow and outflow.

The calculation models will respond to the needs that the Swedish Migration Agency and its stakeholders have for better forward planning in all parts of the flow. These models will be able to predict inflows, calculate delivery times and illustrate the effects on Swedish society in the range of 3-6 months in the future. The three models differ mainly in terms of access to data, statistical calculation methods and target groups.

Project ID: FFF2016/000055
From 2016-05-01 to 2017-12-31


Coordinator: Migrationsverket , Migrationsverket

  • 5asy

    Asylum and forced displacement

    Asylum is a fundamental human right guaranteed universally by the 1951 Geneva Convention. The European Union’s Common European Asylum System (CEAS) set up common procedures to ensure fair...

    View Asylum and forced displacement


The Swedish Migration Agency process 350,000 applications annually from third country citizens who, for various reasons, want to immigrate to Sweden. For a number of years the Agency has worked with forecasts, particularly for asylum immigration and associated family reunion immigration to Sweden. The forecasts can be said to consist of three parts: business intelligence, business analysis and economic analysis of the effects of the Swedish Migration Agency's activities over the next three years. This creates the basis for a dialogue with the Ministry of Justice regarding the financing the Agency will need in the coming years. Forecasts are submitted five times a year and therefore a proposed financing plan is revised continuously each time a forecast is submitted.

There are a number of different reasons why the Swedish Migration Agency needs to develop forecasts:

1) In recent years, the need to be able to determine a more operational production planning and monitoring has been argued. Although the forecast will be operationalised, it will primarily be a matter of recruitment needs to meet an influx which is to be regulated with estimated production needs so that the closing balance is acceptable at the end of the year. The forecast model does not account for variations that arise during the year with the precision that operational activities need in their daily lives. The recent large influx to Europe and later Sweden has exposed this operational requirement further. Although good efforts with forecasts for the next two years have been made, what is missing today is a good model and procedures for how forecasts can cover the short-term planning, i.e., daily, weekly, monthly and quarterly for up to 3-6 months.

2) The handling of cases is done according to several factors that together can be expressed as being the organisation's behaviour. These factors include process classification, skill levels or decisions on priorities. Ultimately, cases are allocated according to principles which are based on organisational divisions or factors that are not obvious when seen from a statistical perspective.

The categorisation of cases submitted to the Agency for handling is flawed and is mainly based on groupings of entire populations (e.g. asylum seekers), the fragmentation of application reasons linked to a specific activity (cf. permit workers, athletes, students), categorisation by section of the law (cf. established versus so-called newly established relationships, rough perceptions of severity (final decision, further investigation cases and investigation cases) or part of flow (e.g. first-time application, extension application or return cases). The groups are not based on statistical data modelling. Similar cases should be handled equally, but similarities are not the same thing as an individual's nationality, the organisation's divisions and rougher divisions of the flow meaning that we treat similar cases. We therefore cannot answer what "delivery timeframe" a case will, or should, have because we do not categorise the cases in queues or indicate the estimated timeframes in which the Swedish Migration Agency will deliver decisions.

3) Following on from point 2, the Swedish Migration Agency is unable to provide information externally about the groups and, above all, when and where they are expected to arrive in the municipalities. This applies to both primary and secondary flows. Studies conducted by the Swedish Migration Agency for the years 2006-2010 show that asylum seekers, for example, move largely to areas in Sweden where there are jobs and housing. Consequential immigration therefore has a double effect as both the applicant and their family later arrive in a municipality other than the municipality that the asylum seeker or the family ties applicants first settle in. For our external contacts, the need for good data and information about migration is a prerequisite for their social planning.

The target groups will therefore greatly benefit from advance planning, both internally within the Swedish Migration Agency and externally among other community stakeholders, if the Swedish Migration Agency can provide a picture of how the near future will look in terms of all parts of the migration flow. This ensures better social planning that primarily satisfies e.g. municipalities' use and planning of their resources, which will benefit future citizens. In concrete terms it is about the design of social resources, housing and institutions so that reception can be managed in the most optimal manner. Ultimately, it means a better start to the future establishment of new arrivals, where the same need for good advance planning that the Swedish Migration Agency has in its own planning is made available to actors further down the migrant flow.

Additional Info

  • Forecasting
Funded / Co-funded AMIF - Asylum Action Grants 2014

Please be aware that the information and links provided in the metadata above are maintained in distributed and heterogeneous information systems. Although we strive to maintain and keep links and information updated, this may not always be possible because of changes that are not registered and updated in the relevant information systems. Please, help us to maintain the system updated by indicating broken links or any other outdated information by contacting the relevant contact point. You can also inform us using the "Contact" link of this page.