2

The numbers
Understanding who moves,
where to, and why
Key messages
• Migration is a mechanism used by people in responding to long-term global imbalances, such as differ-
ences in welfare, and in adapting to shocks, such as conflicts.
• Some 184 million people live outside of their country of nationality, about 20 percent of whom are
refugees. Patterns of movement differ based on migrants’ motives (figure 2.1).
• Migrants and refugees live in countries in all income groups—43 percent in low- and middle-income
countries; 40 percent in high-income member countries of the Organisation for Economic Co-operation
and Development (OECD); and 17 percent in member countries of the Gulf Cooperation Council (GCC).
• There is no simple dichotomy between migrants’ countries of origin and countries of destination.
Many countries at all levels of income are both, at the same time.

Figure 2.1 Patterns of movements reflect distinct matches and motives

Undocumented Bangladeshi
Albanian migrant workers in the
caring for an elderly United Arab
Italian person Emirates
Indian engineers
in Silicon Valley

Stronger Syrian
match entrepreneur
refugees in
Many Refugees with Türkiye
economic skills in demand
Benefits exceed costs migrants at destination

MATCH
Unaccompanied
refugee minors in
Costs exceed benefits Ethiopia
Distressed
migrants,
mainly Many refugees
irregular Rohingya
Weaker (Forcibly Displaced
match Myanmar Nationals),
who are not allowed to
Some poorly work in Bangladesh
skilled migrants Opportunity Fear
at the US MOTIVE
at destination at origin
southern border
Choice whether Obligation
to accept to host
2.1

Source: WDR 2023 team.

Note: Match refers to the degree to which a migrant’s skills and related attributes meet the demand in the destination
country. Motive refers to the circumstances under which a person moves—whether in search of opportunity or because of a
“well-founded fear” of persecution, armed conflict, or violence in their origin country.

43
Current trends
This Report focuses on people who live outside their country of nationality, whether they moved in
search of better economic opportunities or were displaced by conflict or persecution. There are approx-
imately 184 million such people worldwide, of whom 37 million are refugees, and they constitute about
2.3 percent of the global population. They live in countries in all income groups (figure 2.2):

• Low- and middle-income countries. Of Figure 2.2 A large share of migrants and
the approximately 79 million migrants refugees live in low- and middle-income
and refugees who live in these coun- countries
tries, some moved for job opportunities,
family, or other reasons under a variety s, 17%
ant
migr Ec
on
of statuses, including undocumented m
ic om
o i
(detailed data are lacking, however,

on

c
GCC

m
Ec
in most countries). They also include

igr
countries,

an
about 27 million refugees. Although
1

ts,
17%

28%
migrants represent a relatively small Low- and
proportion of the population in most middle-income
low- and middle-income countries, countries,
there are exceptions such as ­Colombia, 43%
High-income
Ec on

Côte d’Ivoire, Djibouti, Gabon, Jordan,

countries,
and Lebanon. Some migrants eventu-
om

40%

5%
ic

ally acquire citizenship in the country

,1
mi

es
gr

of destination. ts

ge
an

,3 fu
• High-income OECD countries. The 5% Re
,
approximately 74 million migrants ees
Refug
5%
and refugees who live in these coun-
tries include both high- and low-skilled
Source: WDR2023 Migration Database, World Bank, Washing-
workers, migrants on student visas, ton, DC, https://www.worldbank.org/wdr2023/data.
as well as undocumented migrants. Note: High-income countries exclude Gulf Cooperation Council
Family reunification with spouses, (GCC) countries.
parents, or children accounts for a
large share of regular migrants—about 35 percent in the European Union (EU).2 Among these
migrants are people who have extensive residency rights, such as the 11 million EU nationals
who live in another EU country and the 13.6 million green card holders in the United States. Also
among them are about 10 million refugees, who are receiving international protection. Some
migrants to the OECD high-income countries move temporarily, while others intend to settle
there. Many are eventually naturalized—about 62 million naturalized citizens are spread across
OECD high-income countries (and are not considered to be migrants in this Report).
• GCC countries. Of the roughly 31 million migrants living in GCC countries, nearly all have a
temporary status, t­ ypically a multiyear work visa that can be renewed. They are both high- and
low-skilled. Only high-skilled migrants can be accompanied by their families. GCC countries do
not host large numbers of refugees. Overall, migrants constitute about half of the GCC popula-
tion—about 79 percent if Saudi Arabia is excluded (box 2.1).

44 | WORLD DEVELOPMENT REPORT 2023

 There is no sharp distinction between countries of origin and countries of destination for migrants.
In fact, most countries are both—at the same time. For example, the United Kingdom is home to about
3.5 million immigrants,3 but it is also the origin of 4.7 million emigrants. At a lower level of income,
Nigeria is home to almost 1.3 million immigrants and is the origin for 1.7 million emigrants. Türkiye
has a large diaspora of economic migrants in Europe, but also hosts 3.5 million Syrian refugees and over
2 million other migrants. Each society needs a combination of policies to best address the situation of
both the people who enter and those who leave.

Box 2.1 Migration data in this Report

Unless otherwise indicated, the data and figures in this Report are based on the WDR2023 Migration
Database.a The database is constructed from the bilateral immigration data produced by the censuses
of individual destination countries. Because of COVID-19 restrictions, only a handful of high-income
­countries managed to conduct the decadal censuses or nationwide surveys in 2020.b Data for member
countries of the European Union, as well as for Norway, Switzerland, and the United Kingdom, are based
on the European Union Labour Force Survey.c Data for all other countries are from the International
Migrant Stock estimates of the Population Division of the United Nations Department of Economic and
Social Affairs (UN DESA).d Most of these data are based on a definition of migrant as a person who lives
in a country that is different from the country of birth. For the purposes of this Report, the data have
been adjusted with citizenship data obtained from a variety of sources or estimations.e Data on refugees
are based on the Refugee Population Statistics Database of the United Nations High Commissioner for
Refugees (UNHCR) and include refugees and asylum-seekers and other people in need of international
protection as determined by UNHCR as of mid-year 2022.f

a. WDR2023 Migration Database, World Bank, Washington, DC, https://www.worldbank.org/wdr2023/data.

b. They include Australia, Canada, Chile, and the United States.
c. “European Union Labour Force Survey (EU-LFS),” Eurostat, European Commission, Luxembourg, https://ec.europa.eu
/eurostat/web/microdata/european-union-labour-force-survey.
d. International Migrant Stock (dashboard), Population Division, Department of Economic and Social Affairs, United Nations,
New York, https://www.un.org/development/desa/pd/content/international-migrant-stock.
e. WDR2023 Migration Database, World Bank, Washington, DC, https://www.worldbank.org/wdr2023/data.
f. Refugee Population Statistics Database, United Nations High Commissioner for Refugees, Geneva, https://www.unhcr
.org/refugee-statistics/.

Changes over time

Historical data on migration patterns are not available. What is known is that the share of foreign-
born people (both migrants and naturalized citizens) has fluctuated between 2.7 percent and 3.5 per-
cent of the world population since 1960.4 However, the apparent stability of this number is somewhat
misleading because, worldwide, demographic growth has been very uneven since 1960. The world pop-
ulation increased by about 156 percent between 1960 and 2020, but high-income countries grew by only
58 percent, while middle-income countries grew by 177 percent and low-income countries by 383 per-
cent.5 As a result, migration trends vary considerably across country income groups (figures 2.3 and 2.4).

THE NUMBERS: UNDERSTANDING WHO MOVES, WHERE TO, AND WHY | 45

 Figure 2.3 Since 1960, the share of Figure 2.4 Since 1960, the share of
emigrants in low-income countries' immigrants and naturalized citizens in high-
population has almost doubled income countries' population has tripled
Emigrants as share of population, by country Foreign-born as share of population, by country
income group, 1960, 1990, 2020 income group, 1960, 1990, 2020
6 15

Immigrant share of population (%)

Emigrant share of population (%)

4 10

2 5

0 0
Low-income Middle-income High-income Low-income Middle-income High-income
countries countries countries countries countries countries
1960 1990 2020 1960 1990 2020

Sources: WDR2023 Migration Database, World Bank, Sources: WDR2023 Migration Database, World Bank,
Washington, DC, https://www.worldbank.org/wdr2023 Washington, DC, https://www.worldbank.org/wdr2023
/data; population, 1960–2020: Population Estimates and /data; population, 1960–2020: Population Estimates and
Projections (database), World Bank, Washington, DC, Projections (database), World Bank, Washington, DC,
https://datacatalog.worldbank.org/dataset/population https://datacatalog.worldbank.org/dataset/population
-estimates-and-projections. -estimates-and-projections.
Note: The 2020 World Bank income level groups are used for Note: The 2020 World Bank income level groups are used for
the 1960 and 1990 data (Serajuddin and Hamadeh 2020). the 1960 and 1990 data (Serajuddin and Hamadeh 2020).

As income levels and demographic trajectories change over time, the directions of migration flows
change as well. Some countries or regions emerge as important origins or destinations, while others
disappear. For example, the large movements from Europe to Latin America a century ago are no longer
happening today. Migration to the GCC countries was almost nonexistent 60 years ago, and yet today
these countries are the destination for some of the largest migration corridors. Meanwhile, Ireland and
Italy, once countries of origin, have become countries of destination.
Cross-border movements are increasingly distributed across a substantial number of corridors.
In 1970, just 150 corridors—out of more than 40,000 possible pairs of origin and destination
countries—accounted for 65 percent of the world’s migration. By 2020, that share had declined to
50 percent. Today’s main corridors include Mexico to the United States; India to the United Arab Emir-
ates and Saudi Arabia; India and China to the United States; Kazakhstan to the Russian Federation
and the Russian Federation to Kazakhstan; Bangladesh to India; and the Philippines to the United
States. Additional large corridors are associated with the main forced displacement situations, such as
between the Syrian Arab Republic and Türkiye, República Bolivariana de Venezuela and Colombia, and
Ukraine and Poland.

Origin countries
The largest share of emigrants6 are from middle-income countries. They are typically among neither the
poorest nor the wealthiest in their country of origin; they can afford to move, and they have an incentive
to do so. Even in situations of conflict and persecution, those who have more means tend to leave first,
although there are exceptions, such as when an entire group is targeted for violence.

46 | WORLD DEVELOPMENT REPORT 2023

Map 2.1 In most countries, only a small share of the population has emigrated to another
country
Ratio of persons living abroad to origin country population, 2020

Ratio (%)
100
30
20
10
5
1
0
No data
IBRD 47145 |
MARCH 2023

Source: WDR2023 Migration Database, World Bank, Washington, DC, https://www.worldbank.org/wdr2023/data.

Emigrants constitute a significant share of the population of some origin countries. Many Small
Island Developing States have emigration rates well above 25 percent of their population. A number
of Central and Eastern European countries also have relatively large emigration rates, typically above
15 percent (their citizens have easier access to Western European countries). The share of refugees to
the total population of origin countries is also high in Afghanistan, the Central African Republic,
Somalia, South Sudan, Syria, Ukraine, and República Bolivariana de Venezuela (map 2.1). Across all
countries, the median emigration rate stands at 7 percent of the population.

Destination countries
Immigrants7 are spread across the world in countries at all levels of income. The main destination
countries (in numbers of migrants) include the United States, Saudi Arabia, the United Arab Emirates,
Germany, and France. Other countries, such as Australia, Canada, and the United Kingdom, have natu-
ralized large numbers of migrants over time.
The share of immigrants in the population of destination countries varies (map 2.2). It is largest in
the GCC countries—up to 88 percent in the United Arab Emirates. It is also significant in a number of
high-income OECD countries, typically 5–15 percent. Part of migration is intraregional, directed to
countries that are relatively better-off than their neighbors, such as Costa Rica, Côte d’Ivoire, Gabon,
Kazakhstan, Malaysia, or Singapore. The share of immigrants is also large in some countries with a
small population, such as Belize, Djibouti, and the Seychelles. Finally, although the share of refugees
in the host population is typically small—below 1 percent—there are exceptions. For example, as of
mid-2022, one person in six was a refugee in Lebanon; one in 16 in Jordan; and one in 21 in Colombia.8

THE NUMBERS: UNDERSTANDING WHO MOVES, WHERE TO, AND WHY | 47

Map 2.2 Immigrants are spread across the world in countries at all levels of income
Ratio of immigrants to destination population, 2020

Ratio (%)
100
30
20
10
5
1
0
No data
IBRD 47144 |
MARCH 2023

Source: WDR2023 Migration Database, World Bank, Washington, DC, https://www.worldbank.org/wdr2023/data.

Regional perspectives
Patterns of movements vary widely across regions (figure 2.5):
• In the East Asia and Pacific region, immigration remains limited, except for high-income coun-
tries such as Australia and New Zealand. Emigration from the region is directed to a range of
destinations, both within the region and outside, such as North America and the GCC countries.
• High-income European countries are home to about 43 million migrants, including 8 million
refugees. They are predominantly from other European countries (56 percent) and to a lesser
extent from other regions, mainly the Middle East and North Africa (13 percent), Latin America
(9 percent), and Sub-Saharan Africa (8 percent). Emigration is mainly directed to other high-
income European countries and North America.
• In other European countries and Central Asian countries, the movements mainly take place
within the region, totaling about 14 million immigrants. These movements are centered on a few
corridors, including between countries of the former Soviet Union. Some people from the region
have also migrated to high-income European countries (about 11 million).
• In Latin America and the Caribbean, two main trends are evident. First, relatively large move-
ments occur within the region (about 10.7 million people), including the 4.4 million people who
left República Bolivariana de Venezuela. Second, a large number of people originating from the
region have emigrated, mainly to North America (about 60 percent) and to a much lesser extent
to the European Union (about 10 percent).
• In the Middle East and North Africa, there are three distinct patterns. First, GCC countries
receive large numbers of immigrants, mainly but not only from South Asia (60 percent). Second,
the rest of the region is the origin of relatively large emigration flows mainly toward high-income
European countries (8 million) and to the GCC countries (6 million). And, third, the Syrian crisis
and the ongoing insecurity in Iraq have also produced a large number of refugees who are hosted
in the region (about 3.5 million).

48 | WORLD DEVELOPMENT REPORT 2023

 • In North America, the number of immigrants is about 6 times larger than the number of
­emigrants. Many of those who come to the region are from Latin America and the Caribbean
(about 43 percent). Other large regions of origin are East Asia and Pacific (21 percent), Europe
and Central Asia (16 percent), and South Asia (9 percent), with relatively smaller numbers from
the Middle East and North Africa. Many of the immigrants to the United States and Canada are
eventually naturalized.
• In South Asia, migration is relatively limited considering the demographic size of the region.
Three main trends are at play. First, about 19 million people have emigrated to GCC coun-
tries. Second, an additional 15 million have emigrated to other regions, mainly North America
and high-income European countries. And, third, forced displacement—from Afghanistan to
Pakistan and from Myanmar to Bangladesh—accounts for additional movements.
• In Sub-Saharan Africa, most movements take place within the region. Of the approximately
22 million people who live outside of their country, about 35 percent are refugees. These move-
ments are particularly intense along some corridors such as from Burkina Faso to Côte d’Ivoire
or to regional economic poles such as N
­ igeria or South Africa. There are also large refugee move-
ments—for example, out of South Sudan, the Democratic Republic of Congo, Sudan, Somalia,
and the Central African Republic. Emigration to other countries is mainly directed to EU coun-
tries, the United Kingdom, and the United States—about 10.3 million people.

Motives and patterns

People move for a variety of reasons. Their motives partly determine the socioeconomic outcomes of
their migration and their need for international protection. Patterns of movements differ between those
who seek economic opportunities in the destination countries and those who move because of a “well-
founded fear” of persecution and conflict. In some situations, however, the line is blurred because some
people are looking for both opportunities and safety.
The decision to migrate is a complex one, forcing people to weigh their options: staying, moving
within their own country, or migrating to a foreign destination. Some migrants decide to move on their
own, whereas others do so at the behest and with the support of an entire group—their family or their
community. Many factors come into play in making such a decision, including both economic and per-
sonal considerations. Economic theory suggests that potential migrants compare their expected welfare
in various situations and the corresponding costs of moving—financial and nonfinancial. They eventu-
ally settle for the option most likely to let them achieve their objectives in terms of economic prospects,
social and psychological well-being, or safety.

Aspirations for a better life

The vast majority of migrants—over 80 percent—move in search of opportunities in the destination
country. Their movements are often gradual, with predictable trends that reflect medium-term eco-
nomic and demographic patterns. These migrants are mainly driven by the potential for higher wages
and for access to better services (map 2.3).9 Although they are not refugees, some people also move to find
improved personal safety, a stronger rule of law, and more personal freedoms.
In 2020, the vast majority—about 84 percent—of migrants (and naturalized citizens) lived in a coun-
try with an income higher than that of their country of origin. But migration levels are not the highest
for corridors where welfare disparities are the largest. Where people are coming from largely deter-
mines where they are moving to. Movements are mainly determined by the demands for skills in the
labor markets of destination countries, historical and geographic links, and the costs of migration.
Most migrants from low-income countries migrate to other low-income countries or to middle-income
countries, often because the cost of migration to higher-income destinations is prohibitive. Migrants

THE NUMBERS: UNDERSTANDING WHO MOVES, WHERE TO, AND WHY | 49

Figure 2.5 Cross-border movements vary greatly by region

10 million including
coming from 0.2 million Up to
SA other countries refugees 6 million NA
EAP (other) moving to ECA (HICs)
ECA (HICs) other countries
EAP (other)
Other High-income
Other
EAP (HICs)
Up to EAP (HICs)
35 million
East Asia moving to
9 million including GCC
coming from 0.3 million and Pacific other countries
NA
other countries refugees
SA Other
EAP (HICs) EAP (HICs)
Other ECA (HICs)
EAP (other) SA
Other
EAP (other)

43 million including Other

coming from 8 million Up to
37 million NA
other countries refugees
LAC moving to EAP (HICs)
SA other countries LAC
EAP (other) ECA (other)
MENA (other) MENA (other)
Other High-income
SSA
ECA (other) ECA (HICs)

ECA (HICs)
Europe
and Up to
37 million
14 million including Central moving to
coming from 5.7 million
other countries refugees Asia other countries
NA
Other
EAP (other) ECA (HICs)
MENA (other) Other
MENA (other)
ECA (HICs) Other
ECA (other)
ECA (other)

Up to
41 million
moving to
14 million including
other countries NA
coming from 6.8 million
other countries refugees
ECA (HICs)
NA Latin EAP (HICs)
EAP (other) America ECA (HICs)
Other and the Other
LAC Caribbean LAC

(Figure continues next page)

50 | WORLD DEVELOPMENT REPORT 2023

Figure 2.5 Cross-border movements vary greatly by region (continued)

31 million including
coming from <0.1 million
other countries refugees
Up to
SA 1 million
moving to
EAP (other) GCC other countries
All other
MENA (other)
SSA
Middle GCC
Up to
Other East 28 million
GCC 12 million including and North moving to
coming from 3.5 million other countries
other countries refugees
Africa
GCC
SA
Other NA
SSA
ECA (HICs)
Other
ECA (other)
MENA (other) Other
MENA (other)

24 million including
coming from 1.9 million
other countries refugees
LAC Up to
4 million ECA (HICs)
SA
moving to
EAP (other) LAC
other countries
MENA (other) North Other
EAP (HICs)
America NA
ECA (HICs)
SSA
ECA (other)
GCC
NA
Up to
42 million
moving to
6 million including other countries
coming from 2.8 million GCC
other countries refugees
South
EAP (other) NA
Asia
Other EAP (HICs)
ECA (HICs)
SA
MENA (other)
EAP (other)
Other

SA
Up to
30 million GCC
22 million including
moving to NA
coming from 7.7 million
other countries refugees Sub- other countries
ECA (HICs)
All other Saharan MENA (other)
Africa
SSA SSA

Source: WDR2023 Migration Database, World Bank, Washington, DC, https://www.worldbank.org/wdr2023/data.

Note: Due to limitations on available data, immigration numbers for each region include all of foreign nationals; emigration
numbers include foreign-born people (including naturalized). EAP = East Asia and Pacific; ECA = Europe and Central Asia
(including Western Europe); GCC = Gulf Cooperation Council (countries); HICs = high-income countries; LAC = Latin America
and the Caribbean; MENA = Middle East and North Africa; NA = North America; SA = South Asia; SSA = Sub-Saharan Africa.

THE NUMBERS: UNDERSTANDING WHO MOVES, WHERE TO, AND WHY | 51

Map 2.3 Some of the global imbalances that drive migration movements are reflected in the
Human Development Index

Human Development
Index
Highest
Very high
High
Medium
Low
No data
IBRD 47141 |
MARCH 2023

Source: Heat map based on 2021 data, Human Development Insights (table), United Nations Development Programme
(UNDP), New York, https://hdr.undp.org/data-center/country-insights#/ranks.
Note: According to UNDP, the Human Development Index is “a summary measure of average achievement in key dimensions
of human development: a long and healthy life, being knowledgeable and having a decent standard of living.” It ranges
between 0 and 1. The categories in the map are defined as follows: low (below 0.55), medium (0.55–0.70), high (0.70–0.80),
very high (0.80–0.90), and highest (above 0.90).

from middle-income countries often move to high-income countries. Likewise, many emigrants from
high-income countries move to other high-income countries (figure 2.6).
Overall, migration is constrained by challenges and barriers that most migrants face:

• Uncertainty. Migration is inherently risky. It involves dealing with unexpected and uncertain
­outcomes, including the possibility of unemployment, social isolation, psychological stress, or
even injuries and death while in transit. People who migrate in search of opportunities tend to
be more willing to take risks than others in their communities of origin. They also tend to be
more adaptable to new environments and situations, regardless of their skill level or socio­
economic background.10
• Unfamiliarity. Moving to unfamiliar settings entails costs, both monetary and nonmonetary. To
succeed, migrants must familiarize themselves with the language, social norms, and culture of
their destination society.11 This may be difficult, and it takes time, although for some people, the
social and cultural differences between origin and destination societies are precisely what moti-
vates their movement. Examples are some women and members of ethnic, sexual, and political
minorities.12 The internet and new technologies have enhanced access to information and created
both new aspirations and a better awareness of the potential risks and benefits of migration.13
• Job search. Finding a job in a new country can be challenging. Skills, credentials, or diplo-
mas acquired in one country may not transfer easily to another country. Many migrants end
up “downgrading” to a lower-skill occupation, leading to “brain waste.”14 Some migrants rely
on information passed through informal networks of friends and family to find decent work.

52 | WORLD DEVELOPMENT REPORT 2023

Figure 2.6 Where migrants go to largely depends on where they come from

Low-income countries Low-income countries

Middle-income countries Middle-income countries

High-income countries High-income countries

Source: WDR2023 Migration Database, World Bank, Washington, DC, https://www.worldbank.org/wdr2023/data.

Note: The thickness of the arrows reflects the size of the corresponding movements.

They tend to move to areas that are already home to other migrants of their nationality. Other
migrants use formal intermediaries—typical for migrants moving from South Asia to the GCC
countries—although such intermediation has a high cost.
• Financing. The upfront costs of migration can be substantial. They typically include travel and
relocation, visas, and processing, as well as payments to intermediaries to find a job or arrange
a move. The costs vary widely across corridors. For example, for low-skilled migrants the
cost of moving from Central America to Mexico can be as low as US$100, whereas it can reach
more than US$4,000 for a move from Pakistan to Saudi Arabia.15 Irregular migration often
requires paying high smugglers’ fees. For example, the cost of crossing the southern border
of the United States irregularly has been estimated at US$2,000–$10,000, depending on the
migrant’s point of origin.16
These constraints apply differently, depending on skill levels, which largely explains why low- and
high-skilled migrants aim for different destinations. A large share of low-skilled migrants move within
their region: in 2020, about half of all low-skilled migrants were in a neighboring country.17 When they
go farther, they tend to aim for more familiar places—that is, where they speak the dominant language
or where they have access to social networks based on their ethnicity, community, or nationality.18 They
therefore avoid the higher barriers in farther or less familiar destinations—such as higher financing
constraints or greater difficulties in finding a job. By contrast, high-skilled individuals are more likely
to migrate to high-income countries, and this trend has intensified over time. High-skilled individu-
als often benefit from a stronger demand for their skills and more-welcoming migration regulations.19
In some countries, they also have easier access to pathways to residency and citizenship.
Against this backdrop, migration can be both permanent or temporary. Some people, such as many
high-skilled migrants in Australia, move with the intention of living permanently in the destination
country. Some move with their families, whereas some plan to bring them later. But for others, migra-
tion is only temporary. They move for a fixed period of time, to study or to work, with the intention of
returning home afterward. This strategy accounts for most movements to GCC countries, the Repub-
lic of Korea, and Malaysia. Yet the distinction between temporary and permanent migrants is blurred
because many people who initially intended to move for only a few years extend their stay for decades,

THE NUMBERS: UNDERSTANDING WHO MOVES, WHERE TO, AND WHY | 53

 and sometimes for a lifetime. The impacts of temporary and permanent migration—the benefits they
yield, the challenges they pose, and the policy responses they require—are markedly distinct.

Fear in the country of origin

The patterns of forced displacement differ from those of economic migration in terms of the concen­
tration of movements, the vulnerability of those who move, the destinations they choose, and the
suddenness and rapid pace under which their movements occur.
Unlike economic migrants who move from a wide range of countries, most refugees come from a
limited number of countries of origin—and increasingly so (figure 2.7). Although there are refugees
from almost all countries in the world, crises in six countries account for 76 percent of all people in
need of international protection: Ukraine (8 million as of February 2023), Syria (6.8 million), República
Bolivariana de Venezuela (5.6 million), Afghanistan (2.8 million), South Sudan (2.4 million), and Myan-
mar (1.2 million Rohingya).20
Refugee flows include large numbers of vulnerable people—those a family or a community wants to
see out of harm’s way—unlike economic migrants, who are primarily working-age adults. In fact, chil-
dren account for 41 percent of refugees,21 and some are unaccompanied. For example, as of 2023 more
than 70,000 unaccompanied or separated children were in Uganda.22 Regardless of their situation at
origin, many refugees reach their destination in a state of destitution, having left behind their assets
and arriving with little or no savings.23 Some have undergone traumatic ordeals that can make their
economic and social inclusion challenging.24
Refugees and economic migrants choose their destinations differently. Economic migrants typically
move to a place where they believe there is demand for their skills, regardless of the distance from their
country of origin.25 By contrast, refugees prioritize safety and security over labor market considerations,
and therefore they tend to move to a safe country that borders their country of origin.
Overall, more than half of all refugees are hosted in just 10 countries26 that typically border the origin
countries (maps 2.4 and 2.5).27 As of mid-2022, 99 percent of refugees from South Sudan were hosted in
neighboring countries, as were 86 percent of refugees from Myanmar; 78 percent of refugees from Syria;

Figure 2.7 Most refugees come from a limited number of countries of origin—and increasingly so
12 100
90
10
80
Share of refugees (%)
Number of countries

70
8
60
6 50
40
4
30
20
2
10
0 0
05

15
07

17
01

09

11

19

21
03
00

10

13

20
02

08

12

18

22
06

16
04

14
20
20

20

20

20
20
20

20

20

20

20

20
20
20

20

20

20
20

20
20

20
20

20

Origin countries with more than 500,000 refugees

Share of refugees from large crises (more than 500,000 displaced)

Source: Midyear 2022 data, Refugee Population Statistics Database, United Nations High Commissioner for Refugees,
Geneva, https://www.unhcr.org/refugee-statistics/.

54 | WORLD DEVELOPMENT REPORT 2023

 Map 2.4 Most refugees flee to neighboring countries

! !! !
!
!
!
!

!
!
!!
!
!

!
!!
! !

! !
!
!
!

!
!
!!
!

!
!
!

!!
! !
!
!

! !
!!
!

!!
!

!!!!!! !
!
!
!

!
!
!

! !
!

!
!
!
!

!
!

!
!
!
!

Refugee Origin
!
!
!

!
!
!
! ! !
!

!
! !
!

!!!
!

!
!!
!

! !
!
!
!
!
!
!

! !
!
!
!
!

!
!
!
!

!
!

!
!

!
!

!
!

!
!
!

Countries
!
!
!
!

!
!

!!
!
! !

! !
! !
!
!

!
!!

!
!

! !!
!
!

!
!
!
!

!
!
!

! !
!
!
!

!!
!
!
!

! !!
!

! !

!
!
!
!!
!

!
! !
!
! !

!
!
! !

!
!
!
!

!
!
! !
!
!
!
!
!
!
!

Afghanistan
!
! !
! !

! !

!
!

!
!
!
!
!

!
!
!

!
!
!

!!
!
!

Refugee origin
!
!

!
!
!
!
!
!

! !
!

!
!

!
! !
!

!
!
!

country Congo, Dem. Rep.

!
! !

!
! !
!

! !
!
!
! !

!
!
!

!
!

!
!
!
!
!
!

!
!

!
Afghanistan
!
!
!

!
!
!
!

!
!

Myanmar
! !
!
!

!
!
!
!
!

!!
!
!
!
!

!
!

!
!
!

!
!

Congo, Dem. Rep. !

!

!
!

South Sudan
! !
!

!
!

!
!
!
!
! !
!
!

!
!

!
!

!!

!!
!! !

!
!
!
!

!
!
!
!

!
!
!

Myanmar
!
!
!

!
!!

!
Sudan
!
!
!
!!

!
! !

!
!
!
!
!
!
!
!
!
!

!
!
!

!
!
!
!

!
!
!

!
!

!
!
!

!
!
!
!
!
!
!

South Sudan
Syrian Arab Rep.
!

!
! !

!!
!
!

!
!

!
!
! !
!
!

!!
!!

!
!
!

!
!
!

!
!
!
!
!

Sudan
Ukraine !
!
!
! !
!

!
!

!
!
!
!

!
!

Syrian Arab Republic ! !

!
!

!
!
!

Venezuela, R.B. !
!
!

! !
!
! !
!

Ukraine
!

! !

!
!

Number of Refugees
!
!

Venezuela, RB ! !
!
!
!

!
!
!
!

1,000,000

!!
! !
!
!

500,000 !
!
!
!

Number of refugees
!
!
!

100,000
1,000,000
500,000
!
!
!

100,000
Source: UNHCR Refugee Population Statistics Database, 2022. IBRD 47151 |
MARCH 2023
IBRD 47151 |
Note: Larger bubbles represent a higher number of refugees. The category of “refugees” includes refugees, asylum-seekers, MARCH 2023
Source: Midyear
and other people 2022 data,
in need of international protectionRefugee
as of mid-yearPopulation Statistics
2022. Bubbles are placed at the center Database,
of each country United Nations High Commissioner for Refugees,
and not where refugees are hosted.
Geneva, https://www.unhcr.org/refugee-statistics/.
Note: The bubbles indicating the number of refugees are placed at the center of each host country and not over the specific
subregions in which refugees are hosted.

Map 2.5 Ten countries host more than half of all refugees

! Russian
Federation

!! Poland

! !!
Germany
Türkiye

Pakistan

Bangladesh

Sudan
!
Countries
country
Hosting
Refugee host

Refugees
Number of refugees
!
! Peru
Colombia
Uganda
!
2,000,000
1,000,000

Source: UNHCR Refugee Population Statistics Database, 2022. IBRD 47161 |

MARCH 2023
IBRD 47161 |
Note: Larger bubbles represent a higher number of refugees. The category of “refugees” includes refugees, asylum-seekers, MARCH 2023
andSource:
other people inMidyear 2022protection
need of international data,asRefugee Population
of mid-year 2022. Bubbles are placedStatistics
at the center of Database,
each country United Nations High Commissioner for Refugees,
andGeneva,
not where refugees are hosted.
https://www.unhcr.org/refugee-statistics/.
Note: The bubbles indicating the number of refugees are placed at the center of each host country and not over the specific
subregions where refugees are hosted.

THE NUMBERS: UNDERSTANDING WHO MOVES, WHERE TO, AND WHY | 55

and 77 percent of refugees from Afghanistan. Figure 2.8 Refugee flows spike after a crisis
In some cases, refugees have moved to other and then slow over time
countries within their region—such as those
60
leaving Ukraine and República Bolivariana
de Venezuela—but they remain very concen- 50

Share of refugees (%)

trated. In total, about 74 percent of refugees
live in low- and middle-income countries, 40
and 26 percent are in high-income countries,
mainly in OECD high-income countries.28 30
Refugees who move beyond neighboring
countries are increasingly traveling farther 20
and to a larger number of destinations.29 They
10
typically have better incomes, assets, educa-
tion, skills, and access to migration networks
than refugees who flee to neighboring coun- 0 1 2 3 4 5 6
tries.30 Their movements are often influenced Number of years since onset of crisis
by opportunities for better economic benefits,
Source: WDR 2023 team calculations, based on October 2022
family ties, and political freedoms.31 data from Forced Displacement Flow Dataset (dashboard), Refu-
Refugee movements are also often charac- gee Data Finder, Statistics and Demographics Section, Global
terized by their suddenness and rapid pace. Data Service, United Nations High Commissioner for Refugees
(UNHCR), Copenhagen, https://www.unhcr.org/refugee-statistics
Some refugee crises build up gradually, giving /insights/explainers/forcibly-displaced-flow-data.html.
destination countries and the international Note: The figure plots the average flow for refugee “situations”
community some time to prepare. But many commencing between 1991 and 2017. A “situation” begins when
occur suddenly,32 adding to the challenges of the refugee flow exceeds 25,000. The shaded area indicates
the 95 percent confidence interval. The category “refugees”
providing adequate assistance to the forcibly includes refugees, asylum-seekers, and other people in need of
displaced and their hosting communities. The international protection as determined by UNHCR.
number of refugees fleeing a particular conflict
may fluctuate as the intensity and geographic Figure 2.9 Refugees are increasingly originating
spread of violence evolve, triggering successive from middle-income countries
waves of movements. On average, however,
70
over 40 percent of refugees in a given situa-
Share of refugees coming from

tion flee in the first year after violence breaks

middle-income countries (%)

60
out (figure 2.8).33 When the numbers are large,
hosting countries undergo a significant shock. 50
These spikes often dominate the policy debate
40
and the news headlines because of their inten-
sity and human toll, even though refugees rep- 30
resent a small share of all migrants.
Over the last decade, the nature of refugee 20
movements has begun to change, although it 10
is difficult to assess whether this is part of a
longer-term trend. In particular, the countries 0
00
02
04
06
08
10
12
14
16
18
20
22

of origin have changed. Although they were

20
20
20
20
20
20
20
20
20
20
20
20

predominantly low-income and low-capacity

Source: Midyear 2022 data, Refugee Population Statistics Data-
countries until 2014, they are now increas- base, United Nations High Commissioner for Refugees, Geneva,
ingly middle-income countries (figure 2.9). https://www.unhcr.org/refugee-statistics/.

56 | WORLD DEVELOPMENT REPORT 2023

This shift is changing some key characteristics of refugee flows. Refugees originating from middle-
income countries typically have higher levels of skills than those leaving low-income countries, and
their skills and attributes can make for a stronger match with the needs of the destination countries.

A continuum of protection needs

Complex patterns of movements have emerged that blur the distinction between refugees and eco-
nomic migrants. Some routes, especially to high-income countries, are used by both economic migrants
and refugees. Although they travel alongside each other, they still have distinct motivations, pros-
pects, and needs for protection. Yet for destination countries, such “mixed movements” pose partic-
ular challenges, requiring significant resources to determine who should be granted entry and under
what status. Even on an individual level, movements may be determined by a combination of security
threats and other considerations. Although not necessarily direct causes of refugee movements, climate
change, environmental degradation, and natural disasters are increasingly interacting with the drivers
of such movements.34
There is growing recognition that there is a continuum of international protection needs between
those who move to seek opportunities and can enjoy the continued protection of their country of cit-
izenship and those who are recognized as refugees under international law. Some people risk harm if
they return to their country of origin, even if that risk does not meet the threshold necessary for being
granted refugee status—for example, when their country is in deep political crisis or subject to wide-
spread criminal violence. The challenge for both destination countries and the international commu-
nity is to ensure that all those who need international protection can access it and that the international
framework evolves in line with the emergence of new protection needs.

Notes
1. In this Report, unless otherwise stated, the term refu- 9. Beine, Machado, and Ruyssen (2020); Czaika and
gees refers to refugees and other people in need of Reinprecht (2020).
international protection. 10. Bütikofer and Peri (2021); Gibson and McKenzie (2012);
2. Data cover all valid permits. They represent all Euro- Jaeger et al. (2010).
pean Union countries for which data were available in 11. For a recent review of migration barriers with parti-
2021. See Eurostat (2022). cular emphasis on familiarity barriers, see McKenzie
3. Sturge (2022, 25). (2022).
4. WDR2023 Migration Database, World Bank, Washing- 12. For example, in destination countries higher female
ton, DC, https://www.worldbank.org/wdr2023/data; empowerment—as measured by the political empower-
World Development Indicators (dashboard), World ment subindex of The Global Gender Gap Report 2020
Bank, Washington, DC, https://datatopics.worldbank (WEF 2019)—is associated with a higher share of
.org/world-development-indicators/. females among migrants. Conversely, higher female
5. World Development Indicators (dashboard), World empowerment in origin countries is associated with a
Bank, Washington, DC, https://datatopics.worldbank lower share of females among migrants. The associa-
.org/world-development-indicators/. tion is much stronger among migrants from low-
6. Emigrants are defined as people who have departed income countries. In the United States, between 1.2
from their country of origin to become migrants. and 1.7 percent of applicants who applied for asylum
7. Immigrants are defined as migrants who have arrived in because of fear of persecution mentioned gender iden-
a destination country. tities in their asylum interviews (Shaw et al. 2021).
8. These figures, based on UNHCR (2022b), do not 13. Bah et al. (2022).
include Palestinian refugees. The government esti- 14. Mattoo, Neagu, and Özden (2008).
mate of the number of refugees in Lebanon is more 15. For the microdata sets, see KNOMAD and ILO (2021a,
than 1.5 million, including Palestinian refugees. The 2021b). For South Asian migrants to the GCC countries
government estimate of the number of refugees in and Malaysia, migration costs ranged from US$600 to
Jordan is more than 2 million, including Palestinian US$4,400, which amounted to 2–10 months of migrant
refugees. earnings. A large share of the migration cost in this

THE NUMBERS: UNDERSTANDING WHO MOVES, WHERE TO, AND WHY | 57

 corridor can be attributed to high intermediation fees 25. Moore and Shellman (2007).
(Bossavie 2023). 26. Devictor, Do, and Levchenko (2021); UNHCR (2022b).
16. See Migrant Smuggling (dashboard), United Nations 27. Devictor, Do, and Levchenko (2021); UNHCR (2022b).
Office on Drugs and Crime, Vienna, https://www.unodc 28. The numbers on foreign citizens and foreign-born
.org/unodc/en/human-trafficking/migrant-smuggling populations are based on 2020 data of WDR2023
/migrant-smuggling.html. Migration Database, World Bank, Washington, DC,
17. World Bank (2018). https://www.worldbank.org/wdr2023/data. However,
18. McKenzie and Rapoport (2010). data on refugees are from 2022 and include the move-
19. Clemens (2013); de Haas, Natter, and Vezzoli (2016). ment of refugees stemming from the Russian invasion
20. UNHCR (2022b). of Ukraine.
21. Data on refugees are based on Refugee Population 29. Devictor, Do, and Levchenko (2021).
Statistics Database, United Nations High Commis- 30. Aksoy and Poutvaara (2021).
sioner for Refugees, Geneva, https://www.unhcr.org 31. Moore and Shellman (2007); Neumayer (2004).
/refugee-statistics/. 32. UNHCR (2022a).
22. Refugee Population Statistics Database, United 33. This result is consistent with the finding of Melander
Nations High Commissioner for Refugees, Geneva, and Öberg (2006) that the rate of forced migration
https://www.unhcr.org/refugee-statistics/. (including both refugees and internally displaced per-
23. World Bank (2017, 80–81). sons) abates rather than escalates over time.
24. Fazel, Wheeler, and Danesh (2005); Porter and Haslam 34. United Nations (2018, 4).
(2005); Steel et al. (2009).

References
Aksoy, Cevat Giray, and Panu Poutvaara. 2021. “Refugees’ Eurostat (Statistical Office of the European Communi-
and Irregular Migrants’ Self-Selection into Europe.” ties). 2022. “Migration and Migrant Population Statis-
Journal of Development Economics 152 (September): tics.” Version November 21, 2022. Eurostat Explained,
102681. Eurostat, European Commission, Luxembourg.
Bah, Tijan L., Catia Batista, Flore Gubert, and David J. https://ec.europa.eu/eurostat/statistics-explained
McKenzie. 2022. “Can Information and Alternatives to /index.php?title=Migration_and_migrant_population_
Irregular Migration Reduce ‘Backway’ Migration from statistics.
the Gambia?” Policy Research Working Paper 10146, Fazel, Mina, Jeremy Wheeler, and John Danesh. 2005.
World Bank, Washington, DC. “Prevalence of Serious Mental Disorder in 7000 Ref-
Beine, Michel, Joël Machado, and Ilse Ruyssen. 2020. “Do ugees Resettled in Western Countries: A Systematic
Potential Migrants Internalize Migrant Rights in OECD Review.” Lancet 365 (9467): 1309–14.
Host Societies?” Canadian Journal of Economics 53 (4): Gibson, John, and David J. McKenzie. 2012. “The Economic
1429–56. Consequences of ‘Brain Drain’ of the Best and Bright-
Bossavie, Laurent. 2023. “Low-Skilled Temporary Migration est: Microeconomic Evidence from Five Countries.”
Policy: The Case of Bangladesh.” Background paper pre- Economic Journal 122 (560): 339–75.
pared for World Development Report 2023, World Bank, Jaeger, David A., Thomas Dohmen, Armin Falk, David
Washington, DC. Huffman, Uwe Sunde, and Holger Bonin. 2010. “Direct
Bütikofer, Aline, and Giovanni Peri. 2021. “How Cognitive Evidence on Risk Attitudes and Migration.” Review of
Ability and Personality Traits Affect Geographic Mobil- Economics and Statistics 92 (3): 684–89.
ity.” Journal of Labor Economics 39 (2): 559–95. KNOMAD (Global Knowledge Partnership on Migration and
Clemens, Michael A. 2013. “Why Do Programmers Earn Development) and ILO (International Labour Organiza-
More in Houston than Hyderabad? Evidence from Ran- tion). 2021a. “KNOMAD-ILO Migration Costs Surveys
domized Processing of US Visas.” American Economic 2015: El Salvador, Ethiopia, Guatemala, Honduras, India,
Review 103 (3): 198–202. Nepal, Pakistan, Philippines, Vietnam, 2015–2016.”
Czaika, Mathias, and Constantin Reinprecht. 2020. “Drivers Microdata set, version May 24, 2021, World Bank, Wash-
of Migration: A Synthesis of Knowledge.” IMI Working ington, DC. https://microdata.worldbank.org/index.php
Paper 163 (April), International Migration Institute, /catalog/2938.
Amsterdam Institute for Social Science Research, KNOMAD (Global Knowledge Partnership on Migration and
University of Amsterdam, Amsterdam. Development) and ILO (International Labour Organiza-
de Haas, Hein, Katharina Natter, and Simona Vezzoli. 2016. tion). 2021b. “KNOMAD-ILO Migration Costs Surveys
“Growing Restrictiveness or Changing Selection? The 2016: Benin, Burkina Faso, Cabo Verde, Gambia, Ghana,
Nature and Evolution of Migration Policies.” Interna- Guinea, Guinea-Bissau, India, Kyrgyz Republic, Liberia,
tional Migration Review 52 (2): 324–67. Mali, Mauritania, Nepal, Niger, Nigeria . . . , 2016–2017.”
Devictor, Xavier, Quy-Toan Do, and Andrei A. Levchenko. Microdata set, version May 24, 2021, World Bank, Wash-
2021. “The Globalization of Refugee Flows.” Journal of ington, DC. https://microdata.worldbank.org/index.php
Development Economics 150 (May): 102605. /catalog/2944.

58 | WORLD DEVELOPMENT REPORT 2023

Mattoo, Aaditya, Ileana Cristina Neagu, and Çağlar Özden. UCLA School of Law, University of California–Los
2008. “Brain Waste? Educated Immigrants in the US Angeles, Los Angeles.
Labor Market.” Journal of Development Economics 87 Steel, Zachary, Tien Chey, Derrick Silove, Claire Marnane,
(2): 255–69. Richard A. Bryant, and Mark van Ommeren. 2009.
McKenzie, David J. 2022. “Fears and Tears: Should More “Association of Torture and Other Potentially Traumatic
People Be Moving within and from Developing Coun- Events with Mental Health Outcomes among Popu­
tries, and What Stops This Movement?” Policy Research lations Exposed to Mass Conflict and Displacement:
Working Paper 10128, World Bank, Washington, DC. A Systematic Review and Meta-Analysis.” JAMA 302
McKenzie, David J., and Hillel Rapoport. 2010. “Self- (5): 537–49.
Selection Patterns in Mexico-U.S. Migration: The Role Sturge, Georgina. 2022. “Migration Statistics.” Research
of Migration Networks.” Review of Economics and Briefing (November 24), House of Commons Library,
Statistics 92 (4): 811–21. London.
Melander, Erik, and Magnus Öberg. 2006. “Time to Go? UNHCR (United Nations High Commissioner for Refugees).
Duration Dependence in Forced Migration.” Interna- 2022a. “Global Trends: Forced Displacement in 2021.”
tional Interactions 32 (2): 129–52. June 16, Statistics and Demographics Section, Global
Moore, Will H., and Stephen M. Shellman. 2007. “Whither Data Service, UNHCR, Copenhagen. https://www
Will They Go? A Global Study of Refugees’ Destinations, .unhcr.org/en-us/publications/brochures/62a9d1494
1965–1995.” International Studies Quarterly 51 (4): /global-trends-report-2021.html.
811–34. UNHCR (United Nations High Commissioner for Refugees).
Neumayer, Eric. 2004. “Asylum Destination Choice: What 2022b. “Mid-Year Trends 2022.” October 27, Statis-
Makes Some West European Countries More Attractive tics and Demographics Section, Global Data Service,
Than Others?” European Union Politics 5 (2): 155–80. UNHCR, Copenhagen. https://www.unhcr.org/en-us
Porter, Matthew, and Nick Haslam. 2005. “Predisplacement /mid-year-trends.html.
and Postdisplacement Factors Associated with Mental United Nations. 2018. “Global Compact on Refugees.”
Health of Refugees and Internally Displaced Persons: United Nations, New York.
A Meta-Analysis.” JAMA 294 (5): 602–12. WEF (World Economic Forum). 2019. Insight Report: The
Serajuddin, Umar, and Nada Hamadeh. 2020. “New World Global Gender Gap Report 2020. Geneva: WEF.
Bank Country Classifications by Income Level: 2020– World Bank. 2017. Forcibly Displaced: Toward a Develop-
2021.” Data Blog, July 1, 2020. https://blogs.worldbank ment Approach Supporting Refugees, the Internally Dis-
.org/opendata/new-world-bank-country-classifications placed, and Their Hosts. Washington, DC: World Bank.
-income-level-2020-2021. World Bank. 2018. Moving for Prosperity: Global Migration
Shaw, Ari, Winston Luhur, Ingrid Eagly, and Kerith J. Con- and Labor Markets. Policy Research Report. Washing-
ron. 2021. “LGBT Asylum Claims in the United States.” ton, DC: World Bank.
Research That Matters (March), Williams Institute,

THE NUMBERS: UNDERSTANDING WHO MOVES, WHERE TO, AND WHY | 59