ESTP Course International Migration Statistics: Estimation and Reconciliation Methods when Using Multiple Sources HCSO, Budapest, 3-5 May 2017 Item 3.8 Using migration data reported by sending and receiving countries. Other applications Adam Dickmann Hungarian Central Statistical Office THE CONTRACTOR IS ACTING UNDER A FRAMEWORK CONTRACT CONCLUDED WITH THE COMMISSION
CONTENT 1. MIMOSA 2. IMEM 3. Circular Plots 2
MIMOSA o MIgration MOdelling for Statistical Analysis o Development and application of statistical modelling techniques for the estimation of missing data on migration flows and foreign population stocks oimmigration and emigration flows by previous / next country of residence (32 by 32 matrix) o James Raymer and Guy Abel 3
MIMOSA Methodology Harmonise and clean available data Identify base data (DK, FI, SE and NO) Calculate adjustment ratios and adjust reported data by using iterative procedure Estimate missing data Immigration and emigration totals for European matrix Associations between origins and destinations By age, sex and citizenship 4
MIMOSA Methodology Destination Origin BE CZ DK DE EE GR ES FR IE IT BE I 80 587 4291 3037 1959 E CZ I 232 9258 388 915 E 78 47 950 2 66 70 283 31 197 DK I 65 2693 764 281 E 511 180 2540 133 229 1720 1333 264 782 DE I 1228 3221 13746 12902 E 4623 8909 2712 597 18106 16236 19060 2415 33802 EE I 4 169 947 60 103 E GR I 57 278 12959 273 638 E ES I 103 1665 14647 2051 E 647 34 130 2109 4 38 2474 487 801 FR I 462 1488 18133 8847 4647 E IE I 45 306 2046 1649 292 E IT I 274 895 23702 5796 E 1414 20 155 9778 1 211 895 2933 130 5
IMEM Integrated Modelling of European Migration tables of migration flows among 31 countries in the EU and EFTA, and to/from world. The estimated migration flows relate to the flows of longterm migrants (UN def.) Estimated flows can be presented by country of origin, country of destination, age and sex, for years 2002 to 2008. They also include measures of uncertainty. New Opportunities for Research Funding Agency Cooperation in Europe (NORFACE Migration) http://www.imem.cpc.ac.uk/ 6
IMEM: Origin-destination (OD) model Expert opinion Expert opinion Source: James Raymer 7
IMEM: Delphi survey Accuracy: distinguish three types of systems: (1) interlinked population registers in the Nordic countries (2) other good-quality registers (NL,DE,AT,BE,CH,SE) (3) less reliable registers and survey-based systems Duration: reflects the difference between the duration of stay criterion adopted by the country and 12 month criterion if a country uses a 3-month criterion, the number of true migrants should be smaller than the reported number of migrants. in practice the duration is intended or planned rather than actual. Undercount: fraction of the true flow that is captured by the data collection system in a given country Intra EU / Extra-Eu High /Low Coverage: missing subgroups (eg. international students, refugees) 8
IMEM The model combines the official migration flow data with the qualitative and subjective judgements of the experts. It also incorporates auxiliary covariate information on the associations between origins and destinations of migration (e.g., language, borders and distance) to estimate missing patterns. The model permits disaggregation of the estimated flows by age and sex. Single resource for policy making and research Reference for data validation Platform for sharing information, harmonising definitions and removing inconsistencies 9
, MIMOSA, IMEM 10
Estimation method for annual migration flows by Guy Abel Origin/destination international migration tables Input data = migrants stocks (country of birth): usually more reliable than migration flows data R package developed by the author: https://gjabel.wordpress.com/2013/10/23/globalmigration-flow-table-estimates/ 2 phases: Data preparation & Estimation 11
Estimation method for annual migration flows by Guy Abel http://www.demographic-research.org/volumes/vol28/18/28-18.pdf 12
Estimation method for annual migration flows by Guy Abel Data preparation Starting data are 2 square matrices 'country of birth' x 'country of residence' (at two different times) Using a demographic balance approach, the numbers of births, deaths and migration to/from the region are progressively removed Totals by row (i.e., persons born in country X and residing in one of the countries considered) are then equal in the two moments Each couple of rows (t1 and t2) becomes respectively the row - column marginal totals in an (empty) origin - destination migration table 13
Estimation method for annual migration flows by Guy Abel Estimation The cells of the each matrix are filled in by an algorithm remove the stayers, thus assumption of maximum number of stayers Data from each matrix are then compiled up to the overall origin-destination matrix Using population projection models: calculates with birth and death probability as well 14
Circular Plots for Directional Bilateral Migration Data Guy Abel https://gjabel.wordpress.co m/category/research/interna tional-migration-estimation/ 15
Sources: MIMOSA project: http://mimosa.cytise.be/ The Integrated Modelling of European Migration (IMEM) project: http://www.cpc.ac.uk/research_programme/?link=imem_project.php Estimating global migration flow tables using place of birth data: http://www.demographic-research.org/volumes/vol28/18/28-18.pdf Updated Circular Plots for Directional Bilateral Migration Data: https://gjabel.wordpress.com/category/research/international-migrationestimation/ 16