bilat_mig_type.csv (25.65 MB)

Bilateral international migration flow estimates by type of move

Download (25.65 MB)
Version 4 2022-10-28, 03:32
Version 3 2022-04-03, 03:23
Version 2 2022-02-24, 02:32
Version 1 2021-07-16, 06:19
posted on 2022-10-28, 03:32 authored by Guy AbelGuy Abel

Estimates of bilateral international migration flows by move type between 1990-1995 and 2015-2020 based on methods presented in Abel & Cohen (2019). In the paper, estimates by move type were not discussed in any detail, but are by-products from the three demographic accounting methods, that might be of potential interest. If used, please cite the Abel and Cohen (2019) paper.

Part of collection of files related to bilateral international migration flow estimates. See the Figshare collection for other files, including bilateral migration flow estimates from six different methods.

You might also be interested in files for sex-specific bilateral international migration flow estimates, provided in another Figshare collection


Use the download button above for the CSV file or run the code below to directly bring into R:


f <- read_csv("")

Version Details

Estimates of international migration flows by move type. Move types are a dis-aggregation of the total estimates of bilateral migration flows. The flow estimation methods use the most recent published UN DESA International Migrant Stock (IMS2020) and World Population Prospects (WPP2022) data inputs. 

Move types are divided into outward migration, return migration or transit migration, following Azose and Raftery (2019). Outward migrants move away from their country of birth. Return migrants move to their country of birth. Transit migrants move from and to countries, neither of which is their country of birth. 

Estimates are only available for demographic accounting methods, which utilize the framing of bilateral migrant stock data as marginal sums of birthplace x origin x destination migration flow tables introduced by Abel (2013)

A description of the changes between each version of the estimates (of total flows, not by type) can be found here.

Data Details

Row for each migration corridor - period - move type combination (225 origins x 225 destinations x 3 periods x 3 move types + 226 origins x 226 destinations x 1 period x 3 move types + 229 origins x 229 destinations x 2 periods x 3 move types = 923,499).

Four columns for each of the data dimensions:

year0 First year of five-year period
orig Origin ISO three letter country code
dest                         Destination ISO three letter country
type Move type (one of outward, return or transit)

See the county list file for country names corresponding to the ISO three letter country codes.

Three columns for flow estimates from each method:

da_min_open Demographic accounting, minimize migrant transitions (maximize stayers) - open accounting system
da_min_closed                             Demographic accounting, minimize migrant transitions (maximize stayers) - closed accounting system
da_pb_closed Demographic accounting, pseudo-Bayesian average of migrant transitions and stayers - closed accounting system

See the paper for full details on each of the methods and the usage notes.


National Natural Science Foundation of China, General Program, Grant/Award Number: 41871142