Asylum seekers: In the search of lost well-being definitivo.pdf · Asylum seekers: In the search of lost well-being Jordi Paniagua 1 Jesús Peiró-Palomino 2,3 Andrés J. Picazo-Tadeo

Asylum seekers:In the search of lost well-being

Jordi Paniagua1 Jesús Peiró-Palomino2,3 Andrés J. Picazo-Tadeo1,3

1University of Valencia

2University Jaume I

3INTECO

XXI Applied Economics Meeting

Alcalá de Henares, June 7, 2018

PPP (UV & UJI) In the search of lost well-being XXI EEA, Alcalá 2018 1 / 24

Motivation Stylized Reggae Music

Exodus: Movement of Jah people!

.5

1

1.5

2

2.5

3

Ass

ylum

see

kers

(m

illio

ns)

2000 2005 2010 2015

year



Exodus: Movement of Jah people!



We know where we're going, uh!









We know where we're from.



We know where we're from.



Open your eyes and look within:



Open your eyes and look within:



Are you satised (with the life you're living)? Uh!

0

.2

.4

.6

.8

1

Wel

l−be

ing

indi

cato

r

Tur

key

Mex

ico

Pol

and

Slo

vak

Rep

ublic

Gre

ece

Est

onia

Chi

leH

unga

ryC

zech

Rep

ublic

Por

tuga

lS

love

nia

Kor

eaS

pain

Isra

elN

ew Z

eala

ndF

inla

ndO

EC

DIta

lyIr

elan

dN

orw

ayF

ranc

eA

ustr

iaU

nite

d K

ingd

omA

ustr

alia

Japa

nG

erm

any

Icel

and

Sw

eden

Can

ada

Den

mar

kLu

xem

bour

gN

ethe

rland

sB

elgi

umU

nite

d S

tate

sS

witz

erla

nd

201520132014



Are you satised (with the life you're living)? Uh!

0

100

200

300

400

Assy

lum

see

kers

(tho

usan

ds)

Turk

eyM

exic

oPo

land

Slov

ak R

epub

licG

reec

eEs

toni

aC

hile

Hun

gary

Cze

ch R

epub

licPo

rtuga

lSl

oven

iaKo

rea

Spai

nIs

rael

New

Zea

land

Finl

and

OEC

DIta

lyIre

land

Nor

way

Fran

ceAu

stria

Uni

ted

King

dom

Aust

ralia

Japa

nG

erm

any

Icel

and

Swed

enC

anad

aD

enm

ark

Luxe

mbo

urg

Net

herla

nds

Belg

ium

Uni

ted

Stat

esSw

itzer

land

201520132014



We're leaving Babylon

AUS

AUT

BELCAN

CHE

CHL

CZEDEU

DNK

ESPEST

FIN

FRA

GBRGRC

HUN

IRL

ISRITA

JPN

KOR

MEX

NLD

NOR

NZL

POL

PRT

SVK

SVNSWE

TUR

USA

0

5

10

15

Asy

lum

(lo

g)

0 .2 .4 .6 .8 1

Well−being indicator



We're going to our Father land

AUS

AUTBEL

CANCHE

CZE

DEU

DNKESP

EST

FIN

FRAGBR

GRC

HUN

IRL

ISL

ITA

JPNKOR

LUX

NLDNOR

NZL

POL

PRT

SVK SVN

SWETUR USA

0

5

10

15

Asy

lum

(lo

g)

0 .2 .4 .6 .8 1

Well−being indicator


Outline

1 MotivationStylized Reggae MusicContributions

2 The model

3 Data & EmpiricsEmpiricsData

4 ResultsBaselinePull & Push

5 ConclusionsLessons learned

Motivation Contributions

Background

Why do individuals abandon their home and cross international

borders to seek asylum in a foreign destination?

Standard migration models (Anderson, 2011; Beine et al., 2011;Grogger and Hanson, 2011) present some shortcomings when it comesto explain forced migration (i.e., refugees and asylum seekers), whichelude wage considerations.

distress-driven migration (Missirian and Schlenker, 2017a)Temperature, oods or earthquakes (Feng et al., 2010; Gray andMueller, 2012; Yang, 2008; Missirian and Schlenker, 2017b)politics, oppression and violence in source countries (Davenport et al.,2003; Hatton, 2009; Moore and Shellman, 2004; Missirian andSchlenker, 2017b; Neumayer, 2005b,a; Schmeidl, 1997)Policies on host countries (Holzer et al., 2000; Neumayer, 2004;Thielemann, 2004, 2006; Vink and Meijerink, 2003).



Background







Background







Background







Background







Background







Determinants of asylum seekers

well-being determinants of asylum seekers using the gravity equation(Hatton, 2009 EJ, 2016 AER):

Origin: terror scale, political rights, civil liberties, wars, incomeDestination: unemployment, recognition, migration, welfareBilateral: distance

Issues:A formal model to explain and derive a gravity equation for asylumows

Empirical bias: Multilateral resistance (time-varying third countryeects)

Country-specic crude indicators of welfare (Hatton, 2009 EJ, p.211)

Decision making: dierentials rather than in absolute levels (Ariely,2009)











































































Contributions

1 First, we develop a model which:1 incorporates well-being explicitly in the migration decision and reduces

the uncertainty of the idiosyncratic migration component.2 introduces multilateral resistance in bilateral asylum ows.

2 Second, we construct multi-dimensional home and host well-being

indices

3 Third, and estimate the eects of well-being on asylum ows

controlling for unobserved bilateral heterogeneity, multilateral

resistance terms, zero asylum ows and heteroskedastic residuals



Contributions




indices






Contributions




indices






Contributions




indices






Contributions




indices





The model

The setup

The prospect asylum seeker faces a discrete menu of host locations

with dierent wages (wj ); each with a common bilateral cost to all

migrants in the country pair, which are modeled with iceberg cost

τij > 1.

Wages are bargained at country level and are known to all prospective

migrants. In addition to the nominal wage, an i-country asylum seeker

assesses the well-being of location. The well-being gain ϕj > 0 enters

multiplicative in the model and enhances or deters nominal wages.

We allow for a constant preference between monetary income and well

being of the form wσj ϕ

1−σ

j


The model

The setup




τij > 1.







1−σ

j


The model

The setup




τij > 1.







1−σ

j


The model

The decision and the pie

In line with the relative decision making theory, a rational individual

decides to seek asylum in country j if:

(wσj ϕ

1−σ

j /τij)εijz > wσi ϕ

1−σ

i . (1)

The asylum seeker has a logarithm utility and the observable

component of migrant utility is then:

uij = σ lnwij + (1−σ) lnϕij − lnτij , (2)

the probability that a random migrant select a particular destination is

given by the multinomial logit form. The aggregate probability is the

proportion of identical migrants from i (except for the values of εijz)

that choose j. The predicted aggregate ow of asylum seekers from i

to j is:

Aij = G (uij)Ni =wσij ϕ

1−σ

ij /τij

∑k wσ

ikϕ1−σ

ik /τik

Ni . (3)


The model




(wσj ϕ

1−σ


1−σ

i . (1)








to j is:


1−σ

ij /τij

∑k wσ

ikϕ1−σ

ik /τik

Ni . (3)


The model




(wσj ϕ

1−σ


1−σ

i . (1)








to j is:


1−σ

ij /τij

∑k wσ

ikϕ1−σ

ik /τik

Ni . (3)


The model

A structural gravity equation for asylum seekers

Aij =SjNi

N︸︷︷︸Frictionless asylum

wσij ϕ

1−σ

ij /τij

ΩjLi︸︷︷︸Asylum frictions

. (4)

The second term represents frictions that impede or enhance asylum

ows. In a simple two country setup, bilateral migration ows will ow

towards destinations with higher wage and well-being dierentials with

lower travel cost.

However, multiple alternative destination inuence the migration

decision. This fact is captured by Ωj and Li , whose interpretation is

analogous to the multilateral resistance terms in a gravity model of

trade (Anderson & van Wincoop, 2003).


The model

A structural gravity equation for asylum seekers

Aij =SjNi

N︸︷︷︸Frictionless asylum

wσij ϕ

1−σ

ij /τij

ΩjLi︸︷︷︸Asylum frictions

. (4)

The second term represents frictions that impede or enhance asylum

ows. In a simple two country setup, bilateral migration ows will ow

towards destinations with higher wage and well-being dierentials with

lower travel cost.

However, multiple alternative destination inuence the migration

decision. This fact is captured by Ωj and Li , whose interpretation is

analogous to the multilateral resistance terms in a gravity model of

trade (Anderson & van Wincoop, 2003).


Data & Empirics Empirics

Estimation

We use the the Pseudo-Poisson Maximum likelihood (PPML) estimator

proposed by Silva and Tenreyro (2006) using Larch's et al. (2017)

procedure:

Aijt = exp(β1 lnwijt + β2 lnϕijt + λij + λit + λjt) + εijt .

Data

OECD: Asylum seekers

Better Life Index the BLI following the recent guidelines by the Commission

on the Measurement of Economic Performance and Social Progress

(CMEPSP), based on three domains: material conditions, quality of life

and sustainability.



Estimation



procedure:


Data





and sustainability.



Estimation



procedure:


Data





and sustainability.


Data & Empirics Data

Better Life Index



A composite indicator for better life

We elaborate a global well-being indicator comparable acrosseconomies.

Data Envelopment Analysis (DEA) & Multi-Criteria-Decision-Making(MCDM) techniques (Peiró & Picazo, 2018).

endogenous weights that maximise each country's well-being relative tothe well-being of all other countries in the sample assessed with thesame set of weights

Composite indicator dimension d∗c ′ =Minimiseλc ,S

+ic ′

1

1+ 1

I ∑Ii=1

S+ic ′

indicator ic ′

Subject to:

xc ′ ≥34

∑c=1

λcxc

Indicator ic ′ =34

∑c=1

λc indicator ic −S+ic ′ i = 1, ...I

S+ic ′ ≥ 0 i = 1, ...I

λc ≥ 0 c = 1, ...34

(5)PPP (UV & UJI) In the search of lost well-being XXI EEA, Alcalá 2018 17 / 24







+ic ′

1

1+ 1

I ∑Ii=1

S+ic ′

indicator ic ′

Subject to:

xc ′ ≥34

∑c=1

λcxc


∑c=1


S+ic ′ ≥ 0 i = 1, ...I

λc ≥ 0 c = 1, ...34








+ic ′

1

1+ 1

I ∑Ii=1

S+ic ′

indicator ic ′

Subject to:

xc ′ ≥34

∑c=1

λcxc


∑c=1


S+ic ′ ≥ 0 i = 1, ...I

λc ≥ 0 c = 1, ...34




Issues with DEA:lack of discriminating power (countries vs dimensions)Idiosyncratic weights

Combination of DEA with with Multi-Criteria-Decision-Making

(MCDM) (Despotis, 2002):

Minimisemc ,ωi ,z t1

34

34

∑c=1

mc + (1− t)z

Subject to :

I

∑i=1

ωi indicator ic +mc = composite indicator dimension d∗c c = 1, ...34

(mc − z)≤ 0 c = 1, ...34

mc ≥ 0 c = 1, ...34

ωi ≥ ε i = 1, ...I

z ≥ 0








34

34

∑c=1

mc + (1− t)z

Subject to :

I

∑i=1


(mc − z)≤ 0 c = 1, ...34

mc ≥ 0 c = 1, ...34

ωi ≥ ε i = 1, ...I

z ≥ 0








34

34

∑c=1

mc + (1− t)z

Subject to :

I

∑i=1


(mc − z)≤ 0 c = 1, ...34

mc ≥ 0 c = 1, ...34

ωi ≥ ε i = 1, ...I

z ≥ 0








34

34

∑c=1

mc + (1− t)z

Subject to :

I

∑i=1


(mc − z)≤ 0 c = 1, ...34

mc ≥ 0 c = 1, ...34

ωi ≥ ε i = 1, ...I

z ≥ 0



Asylum vs well-being dimensions destination0

510

15lo

gAss

ylum

0 .2 .4 .6 .8 1(mean) well_being_des

05

1015

logA

ssyl

um

0 .2 .4 .6 .8(mean) housing_des

05

1015

logA

ssyl

um

0 .2 .4 .6 .8(mean) income_des

05

1015

logA

ssyl

um

0 .2 .4 .6 .8(mean) jobs_des

05

1015

logA

ssyl

um

0 .2 .4 .6 .8 1(mean) community_des

05

1015

logA

ssyl

um

0 .2 .4 .6(mean) education_des

05

1015

logA

ssyl

um0 .2 .4 .6 .8 1

(mean) environment_des

05

1015

logA

ssyl

um

0 .1 .2 .3 .4(mean) civic_engagement_des

05

1015

logA

ssyl

um

.2 .4 .6 .8(mean) health_des

05

1015

logA

ssyl

um

0 .2 .4 .6 .8(mean) safety_des

05

1015

logA

ssyl

um

0 .2 .4 .6 .8 1(mean) work_life_balance_des



Asylum vs well-being dimensions origin0

510

15lo

gAss

ylum

0 .2 .4 .6 .8 1(mean) well_being_scr

05

1015

logA

ssyl

um

0 .2 .4 .6 .8(mean) housing_scr

05

1015

logA

ssyl

um

0 .2 .4 .6 .8(mean) income_scr

05

1015

logA

ssyl

um

0 .2 .4 .6(mean) jobs_scr

05

1015

logA

ssyl

um

0 .2 .4 .6 .8 1(mean) community_scr

05

1015

logA

ssyl

um

0 .2 .4 .6(mean) education_scr

05

1015

logA

ssyl

um0 .2 .4 .6 .8 1

(mean) environment_scr

05

1015

logA

ssyl

um

0 .1 .2 .3 .4(mean) civic_engagement_scr

05

1015

logA

ssyl

um

.2 .4 .6 .8(mean) health_scr

05

1015

logA

ssyl

um

0 .2 .4 .6 .8(mean) safety_scr

05

1015

logA

ssyl

um

0 .2 .4 .6 .8 1(mean) work_life_balance_scr


Results Baseline

Asylum ows Migrant ows

(1) (2) (3) (4) (5) (6) (7) (8)

Well-beingij 5.568*** 2.193 12.621∗∗∗ 0.005

(1.79) (2.46) (2.17) (0.12)

Housingij 0.535 0.524 10.989∗∗∗ 0.361

(2.40) (1.01) (3.75) (0.39)

Incomeij 0.188* 0.250 -4.639 -0.005

(0.10) (0.19) (4.09) (0.02)

Jobsij -1.240 -0.197 -1.985 0.368*

(1.47) (0.28) (1.96) (0.19)

Communityij 2.231 -0.474 1.282∗∗∗ -0.121

(2.83) (0.41) (0.47) (0.25)

Educationij -1.080 -0.399 -3.480∗∗∗ -0.399

(2.23) (1.17) (1.11) (0.64)

Environmentij 4.584 3.451∗∗∗ 2.866∗ -1.583

(4.31) (0.97) (1.49) (1.03)

Civicij 2.657** 0.356∗∗ 0.986∗∗∗ 0.173

(1.10) (0.15) (0.38) (0.18)

Healthij -5.863 0.110 -0.620 0.059

(3.83) (0.11) (0.65) (0.09)

Safetyij -0.300* -0.084 -1.447∗ -0.013

(0.17) (0.17) (0.74) (0.06)

Work-life balanceij 4.250*** 0.691 0.558 0.350

(1.15) (0.73) (0.83) (0.31)

Observations 686 686 915 905 3179 3179 3179 3179

R2 0.99 0.99 0.36 0.36 0.054 0.276 0.99 0.99

Country-year FE Yes Yes No No No No Yes Yes

Country-pair FE Yes Yes Yes Yes No No Yes Yes

Notes: Robust standard errors in parentheses, clustered by country pair

Country-year include both home-year and source-year country xed eects.∗ p < 0.10, ∗∗ p < 0.05, ∗∗∗ p < 0.01PPP (UV & UJI) In the search of lost well-being XXI EEA, Alcalá 2018 20 / 24

Results Pull & Push

Source Host Source Host Host

(1a) (1b) (2a) (2b) (3) (4)

Well-being -1.248** 3.937*** 8.392***

(0.57) (0.78) (1.99)

Housing -4.801*** 5.081*** 3.581***

(0.74) (0.34) (0.99)

Income 3.298 5.890 1.462

(2.49) (5.75) (1.85)

Jobs -2.109*** 0.395 -0.311

(0.61) (0.29) (1.10)

Community -0.392** 1.865** 0.431

(0.20) (0.84) (0.40)

Education -2.723 1.732*** 3.215**

(1.73) (0.33) (1.42)

Environment -3.855*** 1.210 -0.469

(0.55) (1.04) (1.41)

Civic -2.898*** -2.699** 0.835

(0.74) (1.08) (1.41)

Health 0.033 -1.065 1.123

(1.37) (1.11) (1.92)

Safety 0.234 -1.153*** -0.161

(0.63) (0.04) (0.33)

Work-life balance -2.869*** 0.246 -0.492

(0.72) (0.66) (1.28)

Observations 915 915 8022 8022

R2 0.20 0.52 0.16 0.18

Country-pair FE Yes Yes Yes Yes

Notes: Robust standard errors in parentheses, clustered by country pair∗ p < 0.10, ∗∗ p < 0.05, ∗∗∗ p < 0.01


Conclusions Lessons learned

Take-away

1 Asylum ows & Wellbeing:1 Theoretical framework2 Composite well-being indicator3 Structural gravity estimation

2 Our ndings give support to the use of this new set of

multidimensional measures of well-being, as the Better Life Index



Take-away






Take-away






Take-away






Take-away






liberté, égalité, fraternité

...¾and well-being?

1 The inspirational foundations of the European project are falling apart

at the seams of the refugee crisis2 Can we do better?

Mind the gap: work-life balance, civic engagement and incomePush factors: enlarging the scope of economic policies at the sourcewith a wider range targets and political and civil actors.Pull factors, a better design of the refugee quota system, which takesinto account not only the population and growth of the host country,but also housing and educational variables.



liberté, égalité, fraternité ...¾and well-being?























The end

Documents

Asylum seekers: In the search of lost well-being definitivo.pdf · Asylum seekers: In the search of lost well-being Jordi Paniagua 1 Jesús Peiró-Palomino 2,3 Andrés J. Picazo-Tadeo