Open Access

Identifying Syrian refugees in Turkish microdata

Luis Pinedo-Caro
Luis Pinedo-Caro
   | Oct 02, 2020
IZA Journal of Development and Migration's Cover Image
About this article
Previous Article
Next Article
Cite
Published Online: Oct 02, 2020
Page range: -
DOI: https://doi.org/10.2478/izajodm-2020-0017
Keywords
© 2020 Luis Pinedo-Caro, published by Sciendo
This work is licensed under the Creative Commons Attribution 4.0 International License.

Figure 1

Number of foreign-born individuals in Turkey by the year of arrival, 2004–2017.Sources: Household Labour Force Survey 2017 and author’s own calculations.Notes: The figure shows the number of foreign-born observations living in Turkey. “Syrian refugees” are obtained by subtracting the average number of foreign-born individuals during the 2004–2010 period (the so-called ex-ante standard migrants) from the total number of observations in each of the years between 2011 and 2017 (ex-post migrants).
Number of foreign-born individuals in Turkey by the year of arrival, 2004–2017.Sources: Household Labour Force Survey 2017 and author’s own calculations.Notes: The figure shows the number of foreign-born observations living in Turkey. “Syrian refugees” are obtained by subtracting the average number of foreign-born individuals during the 2004–2010 period (the so-called ex-ante standard migrants) from the total number of observations in each of the years between 2011 and 2017 (ex-post migrants).

Figure 2

Propensity scores, before and after matching: (a) before matching and (b) after matching.Sources: Household Labour Force Survey 2017 and author’s own calculations.Notes: The box plots show propensity scores distributions. The three horizontal lines of the blue boxes denote the third quantile, the median, and the first quantile, from top to bottom. (a) shows the distribution before matching for foreign-born families who arrived, respectively, between 2004 and 2010 and between 2011 and 2017. (b) splits the scores of 2011–2017 migrant families between those families matched (ex-post “standard” migrant families) and those unmatched (Syrian families).
Propensity scores, before and after matching: (a) before matching and (b) after matching.Sources: Household Labour Force Survey 2017 and author’s own calculations.Notes: The box plots show propensity scores distributions. The three horizontal lines of the blue boxes denote the third quantile, the median, and the first quantile, from top to bottom. (a) shows the distribution before matching for foreign-born families who arrived, respectively, between 2004 and 2010 and between 2011 and 2017. (b) splits the scores of 2011–2017 migrant families between those families matched (ex-post “standard” migrant families) and those unmatched (Syrian families).

Figure 3

Syrian refugees’ population pyramid, HLFS vs. DGMM.Sources: Household Labour Force Survey 2017, Directorate General of Migration Management, and author’s own calculations.Notes: The figure shows the proportion of Syrian refugees in a number of age groups from two sources: the HLFS 2017 and the DGMM. A chi-square test cannot reject that the HLFS and DGMM data are drawn from the same distribution. It also shows the proportion of ex-post “standard” migrants for comparison.
Syrian refugees’ population pyramid, HLFS vs. DGMM.Sources: Household Labour Force Survey 2017, Directorate General of Migration Management, and author’s own calculations.Notes: The figure shows the proportion of Syrian refugees in a number of age groups from two sources: the HLFS 2017 and the DGMM. A chi-square test cannot reject that the HLFS and DGMM data are drawn from the same distribution. It also shows the proportion of ex-post “standard” migrants for comparison.

Figure A1

Born abroad by year of arrival to Turkey.
Born abroad by year of arrival to Turkey.

Figure C1

Number of Syrian refugees under temporary protection by subregion, DGMM July 2017.
Number of Syrian refugees under temporary protection by subregion, DGMM July 2017.

Figure C2

Number of Syrian refugees under temporary protection by subregion, HLFS 2017.
Number of Syrian refugees under temporary protection by subregion, HLFS 2017.

Figure C3

Difference in Syrian refugees under temporary protection by subregion, HLFS vs. DGMM.
Difference in Syrian refugees under temporary protection by subregion, HLFS vs. DGMM.

Summary statistics at the family level: before matching

VariableMigrant families(2004–2010)Migrants families(2011–2017)Ratio(ex-post/ex-ante)
Family size1.683.342.00
Proportion of 0–140.090.202.23
Proportion of 15–240.160.241.50
Proportion of 15+ women0.750.630.83
Existence of a widow0.030.082.37
Existence tertiary educ.0.410.230.55
Existence of 15–24 students0.150.100.65
Existence of 15+ female workers0.270.130.50
Proportion of 15+ NEETs0.500.591.18
Existence of workers0.420.551.31
Number of informal workers0.180.633.50
Existence of male garment workers0.020.116.61
Existence of non-migrants0.570.270.47

Construction of the variable ‘year of arrival to Turkey’.

Year of arrivalVariableSubpopulationObserved HLFS 2017
buil yildogum yer=Abroad8,310
buil yasama=No*
onceki ikamet=Abroad
tr yildogum yer=Abroad1,377
buil yasama=No
onceki ikamet=Turkey
yurtdisi durum=Yes

Syrian refugees under temporary protection in Turkey by subregion (NUTS-2), 2017

Syrian refugees in Turkey
SubregionProvincesOfficialHLFS estimatedDifference
İstanbulİstanbul497,1351,148,199651,064
BursaBursa, Eskişehir, Bilecik113,989260,277146,288
KonyaKonya, Karaman76,744198,198121,454
KırıkkaleKırıkkale, Aksaray15,05580,72465,669
Niğde, Nevşehir, Kırşehir
AdanaAdana, Mersin308,641358,83050,189
AntalyaAntalya, Isparta, Burdur15,43860,27044,832
AydınAydın, Denizli, Mŭgla26,48369,10042,617
ManisaManisa, Afyon, Kütahya, Uşak13,55454,50440,950
AnkaraAnkara75,881111,67235,791
SamsunSamsun, Tokat, Ç orum, Amasya7,57929,32821,749
ErzurumErzurum, Erzincan, Bayburt9509,4838,533
KastamonuKastamonu, Ç ankırı, Sinop1,4498,7487,299
BalıkesirBalıkesir, Ç anakkale6,22211,7605,538
KocaeliKocaeli, Sakarya, Düzce, Bolu, Yalova46,53351,3364,803
AğrıAğrı, Kars, Iğdır, Ardahan1,3333,6962,363
ZonguldakZonguldak, Karabük, Bartın8832,8922,009
TekirdağTekirdağ, Edirne, Kırklareli15,71916,7301,011
VanVan, Muş, Bitlis, Hakkâri4,6625,642980
TrabzonTrabzon, Ordu, Giresun Rize, Artvin, Gümüşhane4,0274,352325
İzmirİzmir112,881106,764−6,117
MalatyaMalatya, Elazığ, Bingöl, Tunceli29,6897,965−21,724
KayseriKayseri, Sivas, Yozgat67,20742,480−24,727
MardinMardin, Batman, Şırnak, Siirt136,67312,672−124,001
ŞanlıurfaŞanlıurfa, Diyarbakır466,811180,728−286,083
GaziantepGaziantep, Adıyaman, Kilis497,371208,257−289,114
HatayHatay, Kahramanmaraş, Osmaniye536,986143,244−393,742

Variable description

MnemonicShort descriptionFull description
kidsProp. 0–14Proportion of people aged 0–14 in the family.
youngProp. 15–24Proportion of people aged 15–24 in the family.
womenProp. 15+ womenProportion of women among 15+ family members.
widowExist widowsExistence of at least one widow in the family.
universityExist 15+ tertiary educ.Existence of at least one 15+ university graduate.
studentExist 15–24 studentsExistence of at least one student aged 15–24.
fem workExist 15+ female workersExistence of at least one working women.
NEETProp. 15+ NEETsProportion of NEETs among 15+ family members.
workersExist workersExistence of at least one 15+ worker in the family.
informalNumber informal workersNumber of 15+ informal workers in the family.
garmentExist male garment workerExistence of at least one 15+ male garment worker.
turkishExist nonmigrantsExistence of at least one nonmigrant in the family.

Estimates of the probability of being a “standard” migrant family, marginal effects

VariableMarginal effectVariableMarginal effect
Prop. 0–14−0.05Exist 15+ female workers0.10***
Prop. 15–24−0.19***Prop. 15+ NEETs−0.15***
Prop. 15+ women0.02Exist 15+ workers−0.09**
Exist widows−0.04Number of informal workers−0.10***
Exist 15+ tertiary education0.05**Exist male garment worker−0.07
Exist 15–24 students0.12**Exist nonmigrants0.12***

Re-weighting by subregion (NUTS-2), 2017 HLFS

TurkishSyrian refugees
SubregionProvincesAdj. factorInv. prob.Adj. weight
İstanbulİstanbul1.02033381,511
TekirdağTekirdağ, Edirne, Kırklareli1.0024109485
BalıkesirBalıkesir, Ç anakkale1.002096428
İzmirİzmir1.0076196876
AydınAydın, Denizli, Mŭgla1.0079157700
ManisaManisa, Afyon, Kütahya, Uşak1.0041171765
BursaBursa, Eskişehir, Bilecik1.0207195871
KocaeliKocaeli, Sakarya, Düzce, Bolu, Yalova1.0032188838
AnkaraAnkara1.0077192856
KonyaKonya, Karaman1.042983370
AntalyaAntalya, Isparta, Burdur1.0081166743
AdanaAdana, Mersin1.0268183817
HatayHatay, Kahramanmaraş, Osmaniye1.0143187837
KırıkkaleKırıkkale, Aksaray1.026385378
Niğde, Nevşehir, Kırşehir
KayseriKayseri, Sivas, Yozgat1.0075160716
ZonguldakZonguldak, Karabük, Bartın1.0008110490
KastamonuKastamonu, Ç ankırı, Sinop1.002955246
SamsunSamsun, Tokat, Ç orum, Amasya1.0026139620
TrabzonTrabzon, Ordu, Giresun Rize, Artvin, Gümüşhane1.0003124552
ErzurumErzurum, Erzincan, Bayburt1.001675335
AğrıAğrı, Kars, Iğdır, Ardahan1.000977343
MalatyaMalatya, Elazığ, Bingöl, Tunceli1.0008120537
VanVan, Muş, Bitlis, Hakkâri1.000492411
GaziantepGaziantep, Adıyaman, Kilis1.0255144642
ŞanlıurfaŞanlıurfa, Diyarbakır1.0161177791
MardinMardin, Batman, Şırnak, Siirt1.0019159711

Summary statistics at the family level: after matching

Variable“Standard” migrantfamilies (2004–2010)“Standard” migrantfamilies (2011–2017)Syrianfamilies
Family size1.671.913.88
Proportion of 0–140.090.090.25
Proportion of 15–240.160.150.27
Proportion of 15+ women0.750.770.57
Existence of a widow0.030.040.09
Existence tertiary educ.0.410.380.17
Existence of 15–24 students0.140.140.07
Existence of 15+ female workers0.260.250.07
Proportion of 15+ NEETs0.500.510.62
Existence of workers0.420.390.61
Number of informal workers0.180.160.81
Existence of male garment workers0.020.020.14
Existence of non-migrants0.570.550.16
eISSN:
2520-1786
Language:
English
Publication timeframe:
Volume Open
Journal Subjects:
Business and Economics, Political Economics, Microeconomics, Macroecomics, Economic Policy, Mathematics and Statistics for Economists, Econometrics
Journal RSS Feed