• Sonuç bulunamadı

Understanding the linkage between objective and subjective well-being in Turkey

N/A
N/A
Protected

Academic year: 2021

Share "Understanding the linkage between objective and subjective well-being in Turkey"

Copied!
113
0
0

Yükleniyor.... (view fulltext now)

Tam metin

(1)

Department of Economics M.A. Economics Programme

ISTANBUL TECHNICAL UNIVERSITY  INSTITUTE OF SOCIAL SCIENCES

M.A. THESIS

JUNE 2017

UNDERSTANDING THE LINKAGE BETWEEN OBJECTIVE AND SUBJECTIVE WELL-BEING IN TURKEY

(2)
(3)

M.A. THESIS

Department of Economics M.A. Economics Programme

Thesis Advisor: Assist. Prof. Dr. Quentin STOEFFLER

JUNE 2017

ISTANBUL TECHNICAL UNIVERSITY  INSTITUTE OF SOCIAL SCIENCES

UNDERSTANDING THE LINKAGE BETWEEN OBJECTIVE AND SUBJECTIVE WELL-BEING IN TURKEY

Seda ÇALIŞIR 412151013

(4)
(5)

Ekonomi Anabilim Dalı Ekonomi Yüksek Lisans Programı

İSTANBUL TEKNİK ÜNİVERSİTESİ  SOSYAL BİLİMLER ENSTİTÜSÜ

TÜRKİYE’DE OBJEKTİF VE ÖZNEL İYİ OLUŞ İLİŞKİSİNİ ANLAMAK

YÜKSEK LİSANS TEZİ Seda ÇALIŞIR

412151013

Tez Danışmanı: Yrd. Doç. Dr. Quentin STOEFFLER

(6)
(7)
(8)
(9)
(10)
(11)

vii

,

(12)
(13)

ix FOREWORD

Since my childhood, I try to do my best in each place I have been to (e.g. classes, student clubs, interships, professional work experience, etc.). All of them provided valuable contribution to my personality to some extent. However, the turning point of my life is the decision I made when I was working in a multinational company 2 years ago.

The life offers you many things; prestige, success, money, happiness, sadness, etc. Actually, it depends on your point of view; which one(s) is (are) your purpose of living, surviving. I have chosen to be the person who will never leave to learn for herself and use the things learnt in favor of science. I strongly believe that economics is one of the crucial social sciences although there is an ongoing debate about the status of economics as a science in the literature.

I am extremly grateful to my advisor Dr. Quentin STOEFFLER for his support during my M.A. thesis study. I am also grateful for Assist. Prof. Dr. AyĢegül KAYAOĞLU-YILMAZ and Prof. Dr. Fuat ERDAL for their valuable comments and suggestions for the improvements of thesis. I would like to thank to my family; my grandmother Zeliha, my mother Nurkan, my father Erdal, my aunt Mine, my dear Montana and my love Emre for their valuable support during this process.

May 2017 Seda ÇALIġIR

(14)
(15)

xi TABLE OF CONTENTS Page FOREWORD ... ix TABLE OF CONTENTS ... xi ABBREVIATIONS ... xiii SYMBOLS ... xv

LIST OF TABLES ... xvii

LIST OF FIGURES ... xix

SUMMARY ... xxi

ÖZET………..…… .... ………xxiii

1. INTRODUCTION ... 1

2. OBJECTIVE WELL-BEING ... 3

2.1 Defining Objective Well-Being ... 3

2.2 Measuring Objective Well-Being ... 4

2.3 Multidimensional Deprivation ... 6

2.4 Objective Well-Being Literature in Turkey ... 8

2.4.1 Static approach to objective well-being ... 9

2.4.2 Multidimensional approach to poverty ... 11

2.4.3 Policy suggestions and transition in objective well-being ... 12

3. SUBJECTIVE WELL-BEING ... 15

3.1 Defining Subjective Well-Being ... 15

3.2 Subjective Well-Being Literature ... 16

3.3 Subjective Well-Being Literature in Turkey ... 19

4. DATA and METHODOLOGY ... 21

4.1 Micro Data Source ... 21

4.1.1 The descriptive statistics of well-being variables ... 22

4.1.1.1 Household equivalent per capita income ... 22

4.1.1.2 Deprivation indicators ... 24

4.1.1.3 Subjective well-being ... 27

4.1.1.4 Pairwise tabulations of well-being variables ... 30

4.1.2 Descriptive statistics for other variables ... 32

4.1.3 Pairwise correlation of variables with subjective well-being ... 34

4.2 Methodology ... 35

4.2.1 Cardinal approach modeling ... 37

4.2.2 Ordinal approach modelling ... 38

4.2.3 Diagnostics in ordinal approach ... 39

4.2.4 Generalized ordered logitstic regression (go-logit) ... 39

4.3 Limitations ... 40

5. EMPIRICAL RESULTS AND DISCUSSION ... 43

5.1 OLS and Ordered Logistic Regression Results ... 43

5.2 Generalized Ordered Logistic Regression Results ... 45

6. CONCLUSION and POLICY RECOMMENDATIONS ... 59

(16)

xii

APPENDICES ... 73 CURRICULUM VITAE ... 83

(17)

xiii ABBREVIATIONS

CDF : Cumulative Distribution Function EQLS: : European Quality of Life Surveys

EU : European Union

EUROSTAT : Statistical Office of the European Communities GDP : Gross Domestic Products

Go-Logit : Generalized Ordered Logistic Regression HBS : Household Budget Surveys

HCS : Household Consumption Surveys HIDS : Household Income Distribution Surveys HLS : Household Labor Survey

HPCI : Household Per Capita Income MDP : Multidimensional Poverty MOECD : Modified OECD Scale

MPI : Multidimensional Poverty Index

NUTS : Nomenclature of Territorial Units for Statistics

OECD : Organization for Economic Co-operation and Development O-Logit : Ordered Logistic Regression

OLS : Ordinary Least Squares

OPHI : Oxford Poverty and Human Development Initiative OWB : Objective Well-Being

RR : Enumerated Random Method

SILC : Statistics on Income and Living Conditions SWB : Subjective Well-Being

TURKSTAT : Turkish National Statistics Institute UN : United Nations

WB : World Bank

WDI : Weighted Deprivation Index WVS : World Value Surveys

(18)
(19)

xv SYMBOLS

cn : Sum of Weighted Number of Deprivation Scores d : Deprivation Indicators

Dqi : Household Deprivation Scores ek, vi, zi, i : Error Terms

fa,b : Functional Form of Utility

H : Headcount Ratio

Ha, Hchildren : Number of Household Members

HmodOECD : Number of household members based on scales Id : Deprivation Score

l : Subjective Well-Being Cut-Off k : Deprivation Cut-Off

n : Total Number of Deprivation Indicators

p : Total Sample

q : Number of Households Multidimensionally Deprived SWBi : Subjective Well-Being Score of Households

ui : Utility of Households

wd : Weight Assigned to Indicators Xi : Household Spesific External Factors z: : Deprivation Dimensions

Z: : Total Number of Deprivation Dimensions β0,c,i,j,l,r : Regression Coefficients

δr : Region Spesific Form

2

(20)
(21)

xvii LIST OF TABLES

Page

Table 2.1 Poverty classifications. ... 5

Table 4.1 Sample distribution across household per capita (HPC). ... 24

Table 4.2 Deprivation dimension, indicators, cut-offs and weights. ... 26

Table 4.3 Households deprived by indicators (%). ... 27

Table 4.4 Summary statistics of the variables. ... 28

Table 4.5 Dimensions of objective and subjective well-being. ... 30

Table 4.6 Income poverty vs. SWB. ... 31

Table 4.7 Multidimensional deprivation vs. SWB. ... 32

Table 4.8 Summary statistics for dummy variables. ... 33

Table 4.9 Pairwise correlation with life satisfaction and happiness. ... 35

Table 5.1 OLS vs. O-logit results. ... 44

(22)
(23)

xix LIST OF FIGURES

Page Figure 4.1 Distribution of SWB responses on a 1-10 scale. ... 30

(24)
(25)

xxi

UNDERSTANDING THE LINKAGE BETWEEN OBJECTIVE AND SUBJECTIVE WELL-BEING IN TURKEY

SUMMARY

The well-being is a widely examined topic in economics literature in recent years. The definition of well-being has a broad sense; then it differs from study to study or country to country. In this thesis, we concentrate on two aspects of well-being: objective and subjective well-being. Then the motivation behind this focus is to show how and to which degree objective and subjective well-being indicators coincide. In other words, if an improvement occurs in objective aspect of it, then we investigate is it sufficient to increase subjective well-being as well?

The main contributions of this study as follows. First, we do not find any other study focusing on the linkage between well-being aspects of Turkey. Then our study will be first to contribute this literature. In fact, there are the studies focusing on the association of objective well-being with its determinants or the same for subjective well-being. However, they do not consider these two aspects of well-being together. Moreover, we employ objective well-being with not only with money-metric measure, but also with non-monetary deprivations. Then for subjective well-being measure, we conclude that the proposed SWB terms in the literature could be used for Turkey interchangeably since we assert the pairwise correlations of both life satisfaction and happiness yield roughly the same results. A last contribution is that we propose policy recommendations based on our findings inasmuch as the results obtained are highly policy-relevant.

The definitions of well-being variables and previous literature review for both objective and subjective well-being is presented before analyzing our micro data. We use cross-sectional micro data provided by European Foundation for the Improvement of Living and Working Conditions (Eurofound) with the 3rd European Quality of Life Survey (EQLS). Our well-being variables are introduced first and some modification, weighting or other data preparations procedures are conducted. The first objective well-being variable is money-metric: household equivalent per capita income. This could yield a static measure like poverty rate. Thus, it won‟t enough to capture objective well-being, as a whole. Then, we employ non-monetary well-being variable, namely deprivations. At this point, considering Turkey„s socio-demographic structure, we identified 5 deprivation dimension (education, health, housing condition, economic capability and local environment). Assigning 13 indicators at total (less than 5 years schooling; having chronic health problem; shortage of space, rot in windows, doors or floors, damp or leaks in walls or roof; keeping home adequately warm, replacing any worn-out furniture, a meal with meat every second day, buying new, rather than second-hand, clothes; noise, air quality, quality of drinking water, litter or rubbish on the street), we equally weighted them and calculated the headcount ratios for each of deprivation indicators. According to that, the deepest deprivation is resulted in economic capability and education.

(26)

xxii

Subjective well-being, even its definition or scope could change depending on the country chosen, it is generally understood as life satisfaction, happiness or quality of life in the literature. After conducting pairwise correlation analysis, we conclude that life satisfaction and happiness could be used interchangeably in this thesis.

The empirical approach used in this study is ordinal approach modelling based on the data‟s nature and previous literature. Although the frequently preferred ordinal model is order logistic regression, we propose to use generalized ordered logistic regression for estimation. The main reason for this, the rejection of the parallel lines hypothesis after conducting ordered logit and generalized logit models. In other words, across different response categories of SWB variable, the coefficient does change. However, the interpretations are done with respect to calculated marginal effect. According to our findings, household eqauivalent per capita income, as an objective well-being varibale, has a highly significant and positive marginal impact on life satisfaction. This finding is consistent with most of the studies in the literature. Yet, objective well-being includes another component, named deprivations. We consider overall deprivation score and deprivation in domains of life. While the overall deprivation score has a marginal impact which is statistically significant at 5% level, the breakdown of the deprivations in dimensions is still crucial to extract the policy-relevant linkage among the well-being aspects. Economic capability and local environment have the deepest marginal impacts on life satisfaction, respectively. In terms of deprivation indicators, the density of deprivation can be ranked as follows: replacing any worn-out furniture, keeping home adequately warm, noise, air quality. Furthermore, we include various socio-demographic features to be sure about the robustness of result as we conclude that if an adequate objective well-being occurs, then this latter will improve the subjective well-being in the first models. Consistent with most of the studies in the literature, being married and employed and living in urban areas have positive and significant marginal impacts on SWB. However, family size (e.g number of adults and children in the households) yields different results in the literature. We find evidence that the marginal impact of increase in number of children is negative and statistically significant on SWB. Yet, in crowded families, who live generally in rural areas, the marginal impact of income increase is positive on SWB. An interesting finding is for educational attendance. It is found that the life satisfaction decreases with higher education.

The recommendations based on the findings of this study are as follow. An extension of the study could be done by using time-series and panel data for better capture the linkage between different well-being aspects. Another suggestion is creating a national index and publishing it regularly based on not only objective aspect of well-being, but also including the association of objective and subjective measures at the same time. A last recommendation is again about extension of this study by including some institutional variables (e.g. governance, transparency, freedoms, democracy, etc.).

(27)

xxiii

TÜRKİYE’DE OBJEKTİF VE ÖZNEL İYİ OLUŞ İLİŞKİSİNİ ANLAMAK ÖZET

Son yıllarda ekonomi literatüründe iyi oluĢ konusunu içeren çok sayıda çalıĢma yapılmaktadır. Ġyi oluĢ kavramı kendi içinde birden fazla anlamı barındıran ve tanımlanması da çalıĢmadan çalıĢmaya, ülkeden ülkeye değiĢiklik gösteren bir olgudur. Bizim bu çalıĢmada takip ettiğimiz ve arasındaki iliĢkinin derecesini anlamaya çalıĢtığımız iyi oluĢ kavramları, objektif ve öznel iyi oluĢtur. ÇalıĢmanın motivasyonunu oluĢturan temel etken, iyi oluĢtaki değiĢimlerin birbiri ile ilgileĢim halinde olup olmadığının ortaya koymaktır. Bir baĢka deyiĢle, objektif iyi oluĢtaki bir iyileĢmenin, öznel iyi oluĢa etki edip etmediği ya da ne ölçüde etki ettiğinin belirlenmesidir.

Bu çalıĢmanın literatüre baĢlıca katkıları arasında birincisi, Türkiye için objektif ve öznel iyi oluĢ kriterlerini bir arada irdeleyen ilk çalıĢma olmasıdır. Bizim çalıĢmamız bu açıdan Türkiye‟deki yoksunluk boyutlarının belirlenmesi açısından da gelecek çalıĢmalara yol gösterici olmayı amaçlamaktadır.

ÇalıĢmada öncelikle, tez konusunun odak noktasını oluĢturan iyi oluĢ değiĢkenleri tanımlandıktan sonra, kullanılan veri setinde bağımlı değiĢken olarak belirlediğimiz öznel iyi oluĢ kavramının hangi değiĢken(ler)le ifade edileceğini belirlemek amacıyla ilgileĢim analizleri yapılmıĢtır. Bu çalıĢma, Avrupa YaĢam ve ÇalıĢma KoĢullarını ĠyileĢtirme Kurumu (Eurofound) tarafından en son kamuya açık Ģekilde yayınlanan üçüncü Avrupa YaĢam Kalitesi Anketi (EQLS) veri setinden alınan mikro veri ile yapılmıĢtır. Birinci objektif iyi oluĢ değiĢkenimiz parasal bir gösterge olan hanehalkı geliridir. Ancak bu gösterge, bize tek baĢına çok mana elde edemeyeceğimiz yoksulluk oranı gibi statik bir değer verecektir. Oysa objektif iyi oluĢun bir diğer alanı da parasal olmayan göstergelerle temsil edilen yoksunluklardır. Türkiye için sosyo-demografik yapısı göz önünde bulundurulduğunda 5 yoksunluk boyutu (eğitim, sağlık, konut durumu, ekonomik yapabilirlik, yerel çevresel faktörler) ve bu boyutların alt baĢlığında onları oluĢturan 13 yoksunluk göstergesi belirlenmiĢtir (5 seneden daha az eğitim görme; kalıtsal sağlık sorunu; konutta yer sıkıntısı, pencerelerde kırıklar ve çatı akması; konutu yeterince ısıtabilmek, eski mobilyaları yenileyebilmek, istendiğinde et yiyebilmek, ikinci el yerine yeni kıyafetler alabilmek). Her bir boyutta eĢit ağırlıklandırma yöntemi kullanılarak yoksunluk dereceleri hesaplanmıĢtır. Buna göre yoksunluk sıralamasında en derin yoksunluk ekonomik yetebilirlik ve eğitim boyutlarında gözlenmiĢtir.

Öznel iyi oluş, literatürde çalıĢmadan çalıĢmaya tanımı ve kapsamı değiĢen bir

değiĢken olmakla birlikte genel olarak yaĢam memnuniyeti, mutluluk ya da yaĢam kalitesi Ģeklinde tanımlanabilir. Yapılan ikili ilgileĢim analizleri sonucunda Türkiye için yapılan bu çalıĢmada, bahsi geçen öznel iyi oluĢ terimlerinin birbiri ile aynı anlama gelecek Ģekilde değiĢmeli kullanılmasına karar verilmiĢtir.

Kullanılan ampirik tahminleme yöntemi, veri setinin özellikleri de göz önüne alındığında sıralı modelleme yaklaĢımı tahmin yöntemi belirlenmiĢtir. Literatürde

(28)

xxiv

genelde sıralı lojistik tahmin yöntemi tercih edilse de, hem önceki çalıĢmalar hem de veri setinin özellikleri göz önünde bulundurularak bazı hipotez testleri gerçekleĢtirilmesi gerekmiĢ ve sıfır hipotezi sınamasına istinaden ayrık sıralı veri setimizde tahmin katsayılarının en az bir değiĢken için her kategoride aynı kalmadığı tespit edilmiĢtir. Bu sonuçtan hareketle model tahminleri, GenelleĢtirilmiĢ Sıralı Lojistik Regresyon (go-logit) ile gerçekleĢtirilmiĢtir. Sıralı tahmin yöntemlerinde yorumlanan değerler katsayılar değil, katsayıların marjinal etkileri, ve bu etkilerin yönü ile anlamlı olup olmadığıdır.

Ekonometrik modelin sonuçlarına göre, kiĢi baĢına düĢen hane halkı geliri ile öznel iyi oluĢ arasında istatistiki olarak pozitif ve anlamlı bir iliĢki olduğu belirlenmiĢtir. Bu bulgu geçmiĢ literatür verileri ile de uyuĢmaktadır. Ancak gelir, tek baĢına, objektif iyi oluĢu temsil edemeyeceğinden, diğer objektif iyi oluĢ değiĢkenleri olan toplam yoksunluk puanı ve yoksunluk göstergeleri ile farklı model tahminleri yapılmıĢtır. Buna göre, yoksunluk puanı, %5 düzeyinde negatif anlamlılık göstermiĢtir. Bu çıktı bize gösterir ki, hane haklı gelirindeki artıĢ öznel iyi oluĢu nasıl

ki pozitif yönde etkiliyorsa; yoksunluk skorundaki

artıĢ da öznel iyi oluĢu negatif olarak etkiler. Öznel iyi oluĢu açıklayan ekonomik yetebilirlik yoksunluk boyutunun en derin negatif ve anlamlı iliĢkiyi kurduğu; sırasıyla, eski mobilyaları yenileyebilmek ve konutu yeterince ısıtabilmek göstergelerinin en anlamlı marginal etkileri sağladığıdır. Ġkinci anlamlı yoksunluk boyutu ise, yerel çevresel faktörler olarak belirlenmiĢtir.

Bu bulgulara ek olarak, bazı kontrol değiĢkenleri de dâhil ederek yeni model tahmini yapılmıĢtır. Anket katılımcılarının yaĢ, cinsiyet, medeni durum, eğitim seviyesi, yaĢanılan coğrafi bölge, yaĢanılan bölgenin geliĢmiĢliği, psikolojik iyi oluĢ gibi sosyo-demografik özellikler için de model tahmini yapılmıĢtır. Kullanılan kontrol değiĢkenler arasından, evli ve çalıĢıyor olmak ve Ģehirde yaĢamak, bizim çalıĢmamızda olduğu gibi literatürdeki çoğu çalıĢmada da öznel iyi oluĢa pozitif yönde etki eden ve anlamlı bulunan değiĢkenlerdir. Ancak aile büyüklüğüne dair kullanılan evde yaĢayan yetiĢkin ve çocuk sayısı değiĢkenleri, farklı çalıĢmalarda farklı yönde sonuçlar doğurmaktadır. Bizim çalıĢmamızda çocuk sayısındaki artıĢın, öznel iyi oluĢa negatif yönlü ve istatistiki olarak anlamlı etki yarattığı; geniĢ ailelerin genelde kırsal alanda yaĢadığı ve gelir artıĢının bu durumda öznel iyi oluĢa marjinal etkisinin pozitif yönlü olduğu saptanmıĢtır. Bulgular arasında eğitim seviyesine dair yaptığımız model tahminleri, eğitim seviyesi arttıkça, öznel iyi oluĢtaki marjinal etkinin de negatif yönde arttığını göstermektedir.

Bundan sonraki iyi oluĢ çalıĢmaları için kullanılacak veri setinin panel veri tercih edilmesi ve yatay kesitte gözlenemeyen kısımların da eklenerek çalıĢmanın geniĢletilmesi önerilebilir. Bir baĢka öneri de diğer geliĢmiĢ ülke ekonomilerinde bir gösterge olarak kullanılmaya baĢlanan mutluluk indeksi ve insani geliĢmiĢlik indeksinin, Türkiye‟nin kendi içinde düzenli olarak hesapladığı ölçümler haline getirilmesidir. Son önerimiz de, kurumlar düzeyinde (e.g. yönetiĢim, demokrasi, Ģeffaflık, özgürlükler, vb.) veri setinin geniĢletilerek çalıĢmanın geniĢletilmesidir.

(29)
(30)
(31)

1 1. INTRODUCTION

In recent years, the measures of well-being are widely examined by different domains (e.g. sociology, psychology, and economics) in literature. In fact, well-being could be understood in a broad sense since it includes many aspects in it. In this thesis, we concentrate on two aspects of being: objective and subjective well-being. The objective well-being literature employs money-metric measures (e.g. income, consumption, etc.) as indispensable variables in the studies. Yet, this kind of analysis could be considered to some extent as static measures since they usually reflect some traditional results such as poverty rate. Beyond the money-metric measures, non-monetary measures of well-being is assessed by material deprivation or capabilities. In recent years, some index-based approaches (e.g. Multidimensional Poverty Index (MPI), Human Development Index (HDI), etc.) have become applicable when assessing the well-being in developing countries.

In this study, our motivation is to show how and to which degree objective and subjective well-being indicators coincide. In other words, improvements in money and non-monetary metric measures for higher well-being are desired, but not sufficient if people‟s SWB remain unchanged. According to Human Development Report (2016), Turkey is 64th among 188 economies in Gross National Income (GNI) ranking; however is 71st in Human Development Index (HDI) ranking. The dissonance between these two ranks is questionable for analyzing the well-being. Moreover, World Happiness Report (WHR) 2017 asserts that Turkey is ranked at 69th position among 155 countries. Although these rankings seem close to each other, it is not evident for a pure linkage between objective and subjective well-being. They should be supported with empirical findings in order to reach beyond these rankings for further acknowledgement.

Thus, this study diagnoses the determinants of subjective well-being in Turkey by employing monetary, non-monetary factors and several others specific to socio-economic structure of it. The empirical evidence is obtained based on cross-sectional micro data provided by the 3rd European Quality of Life Survey (EQLS) which is

(32)

2

processed by the European Foundation for the Improvement of Living and Working Conditions (Eurofound) in 2011 and 2012. The approach to data analysis is crucial since the dependent variable has an ordered discrete nature, empirical model should built with respect to it. Following the hypothesis testing procedure, we estimate the marginal effects with generalized ordered logistic regression model.

By addressing the research question, the contribution of this study to literature is as follows. First, we do not find any other study focusing on the linkage between well-being aspects of Turkey. Then our study will be first to contribute this literature. In fact, there are the studies focusing on the association of objective well-being with its determinants or the same for subjective well-being. However, they do not consider these two aspects of well-being together. Moreover, we employ objective well-being with not only with money-metric measure, but also with non-monetary deprivations. Then for subjective well-being measure, we conclude that the proposed SWB terms in the literature could be used for Turkey interchangeably since we assert the pairwise correlations of both life satisfaction and happiness yield roughly the same results. A last contribution is that we propose policy recommendations based on our findings inasmuch as the results obtained are highly policy-relevant.

The remainder of this study is organized as follows. We stress on two broad literatures: objective well-being and subjective well-being. In the following section, named section 2, we present the objective well-being with the literature review and definitions. The section 3 follows the same pattern for subjective well-being. In the section 4, we introduce our data with descriptive statistics, pairwise tabulations for well-being variables and correlation analysis. Also we present our methodology after discussing the relevant modelling approaches. In the following section, namely section 5, we construct our econometric model to answer the research question and discuss the finding with comparison to previous literature. Section 6 is the final part of this thesis, in which we summarize the findings and propose our policy recommendations.

(33)

3 2. OBJECTIVE WELL-BEING

Well-being is a complex phenomenon affected by psychological, social and economic dimensions of an individuals' life; therefore well-being is composed of more than one dimension (e.g. objective and subjective). In order to measuring the objective well-being of people, some traditional money-metric measures (e.g. income, consumption, etc.) are often used as proxies. These measures lead us to some conventional well-being indicators like poverty rate. Essentially, if a person‟s well-being is low, this person can be considered as a poor person. Moreover, while assessing the poverty, consumption is conventionally viewed as the preferred indicator, for practical reasons of reliability and because consumption is thought to better captures long-run welfare levels than current income (World Bank, 2001). In this section, the objective well-being is assessed in terms of definition and measures. Furthermore, objective well-being literature in Turkey will be summarized.

2.1 Defining Objective Well-Being

At first glance, objective well-being refers to money-metric measures and an indicator which represent the proportion of poor in the society, named poverty rate. Townsend (1979) provides a broad definition of being poor. The agreed definition of poor by EU is that the individuals could be classified as poor if they are suffering from a minimum decent type of living conditions because of a lack of resources (EU Council of Ministers, 1985).

Halleröd (1995) questions the distinction between direct and indirect definition of poverty. The latter is defined as the measures based on income. The indirect definition refers to the poor in the society without being deprived (Halleröd, 1995). On the other hand, direct approach developed firstly by Mack and Lansley (1985), is based on the output, in order words, focuses on the current living conditions in the society.

(34)

4

Although both approaches provide at least an intuition about the poverty level of household, they have some shortcomings at some level. Firstly, measuring income poverty, as a whole, does not yield adequate inference about the disposable income of household due to the fact that some income originated from different groups, as an illustration, the self-employed or shadow economy or other type of resources (e.g. non-cash transfers, assets) couldn‟t be calculated accurately (Nolan and Whelan, 1996; Whelan et al., 2001; Eurostat, 2009). Secondly, indirect approach of objective well-being assumes a straightforward relationship between economic capability and living conditions; yet, the preferences of households could differ in terms of consumption (Halleröd, 1995).

Nevertheless, the direct approach to poverty also can‟t evade from some shortcomings. Initially, the societal norms have to some extent regulative and shortening impact on how to live in the society (Smith, 1776). Then, when it comes to defining the poor, a direct poverty line could not define the poor adequately; but rather more dogmatically (Halleröd, 1995).

Moreover, objective well-being indicators are not independent from each other since being poor could cause being deprived at some cases due to the fact that economic capability hardens the reduced well-being (Mack and Lansley, 1985; Halleröd, 1995). Ringen (1988) maintains that poverty should be defined as a low standard of living in the sense of a deprived way of life due to insufficient resources to avoid such deprivation.

Kakwani and Son (2006) point out that the wellbeing or standard of living is not about the possession of commodities, but it is about living. Sen‟s (2000) ideas of “functionings and capabilities” meet this premise as „functioning‟ is considered as an achievement, and a „capability‟ as the ability to achieve. The capability of acquiring the commodities is not the final goal; but getting a certain level of utility is the threshold as Kakwani and Son (2006) declare.

2.2 Measuring Objective Well-Being

Poverty measurements could have one dimensional or multi-dimensional aspect according to different situations or countries. More commonly used way to distinguish poor and non-poor in a society depends on identifying a threshold level

(35)

5

and constructing an index that expresses how far from the threshold the poor is. This separation among poor and non-poor can be generated using not only monetary, but also non-monetary components. As we discussed in poverty definition part, another aspect could refer to input or output approach. In Table 2.1, we present these classifications. Most poverty measures are monetary and input based, where inputs are the resources required to achieve well-being (Boarini and d'Ercole, 2006). For this group of poverty measure, the determination of threshold could be either absolute or relative. In absolute threshold of being poor, generally severe deprivation indicators (e.g. food, safe drinking water, sanitation facilities, etc) play crucial role (UN, 1995).

Table 2.1 Poverty classifications. Source: Boarini and D'Ercole (2006). Input-based methods

(indirect measures)

Outcome-based methods (direct measures) Money metric measures Income measures Basic needs indicators Non-monetary measures Access to employment,

public services

Material deprivation (capability indicators, literacy,

chronic illness, etc.)

Chambers (2002) distinguishes the poverty measures into five clusters. The income-poverty is the most common one to assess income-poverty as it includes a more easily measurable proxy, income. The second cluster refers to the material lack such as clothing, personal means of transport, furniture, etc. The third cluster is derived from Sen‟s (1979) capability deprivation and includes human capabilities such as skills and physical abilities. The fourth cluster is a form of multidimensional deprivation. The final cluster emerges as a reflection of four clusters which extent to emphasize their meanings by multiplying. In other word, the well-being or deprivation become a multidimensional issue as different combination of them can occur. As an illustration, one person could be deprived in health but not in housing facilities. Actually, this situation is a kind of local aspect of deprivation. The degree of deprivation could differ from one country to another; therefore the intersection of well-being and deprivation measures could be subject to change. Chambers (2002) calls this concept as the language of well-being. Sen (2000) argues that using an integrated approach in measuring objective well-being will help including people‟s perceptions and capabilities in the process.

(36)

6

In order to assess poverty in a comprehensive way, a joint work of the money-metric and non-monetary measures is highly recommended (Ringen, 1988; Halleröd, 1995; Nolan and Whelan, 1996; Boarini and d'Ercole, 2006; Bellani and D‟Ambrosio, 2010). Moreover, the relative deprivation refers to a subjective feeling (Stouffer 1949; Runciman 1966). Runciman (1966) explains that there exist four preconditions for relative deprivation: (1) there is a person “A” who does not have object X, (2) person “A” knows that person “B” has X, (3) person “A” wants to have X, (4) person “A” finds that having X is reasonable. He implies that the people incline to compare themselves to other people in terms of not only income, but also other material possessions. Furthermore, the deprivation of the person who does not have X is an increasing function of the number of the person having X (Yitzhaki, 1979). By assessing Gini coefficient in order to test this affirmation, Yitzhaki (1979) concludes that this coefficient is a well quantitative representative of relative deprivation with an aggregate index.

Although the income and deprivation measure different phenomena, by measuring both of them, the other characteristics, such as subjective characteristics of household become more likely to be understood (Whelan et al., 2001). Nolan and Whelan (2009) assert the use of non-monetary indicators could help to capture the multidimensional deprivation and social exclusion. For doing this, various control variables should be included (e.g. age, consumption, living area, etc.).

2.3 Multidimensional Deprivation

The multi-dimensional structure of deprivation is questioned in early studies by constructing some index-based approach. Townsend (1979) employs twelve indicators using the survey which includes sixty indicators in his pioneering work on deprivation of British people. However, Townsend‟s approach has been criticized in many ways by other economists. In the first contrast, the discernment between the preferences and the financial constraints as the cause of the lack of a good is emphasized. Mack and Lansley (1985) made this point clearer as they emphasized that deprivation is not composed only by the lack of a necessity, but also enforced lack of necessities and proposed some questions in the survey to differentiate the preferences and constraints of people. Another criticism has been made by Ringen

(37)

7

(1988) toward to Townsend‟s work as he tries to reveal a direct measure of poverty by proposing an income threshold which is an indirect measurement of poverty. Whelan et al. (2001) examine the relationship between deprivation and income. They propose a life-style deprivation index based on 24 household items. The first approach that they employed is constructing a summary index; however based on Irish data evidence, summary index does not serve well enough as to assess the different dimensions of deprivations (Nolan and Whelan, 1996). The aim is to combine them into a summary index of deprivation by capturing different dimensions consist of indicators. In order to construct the index, the simplest approach would be assigning each item a value of 1 where the household is deprived; otherwise zero (Nolan and Whelan, 2009). However, looking for the range of deprivation items in order to check whether they could be separated into different sub-groups is crucial when an index is constructed. Whelan et al. (2001), in their life-style derivation index, use five categories: basic, secondary, housing facilities, housing conditions and environmental factors. They consider the affordability of household or absences of the items as the item choose criteria. In determination of clustering the items into groups, they employ confirmatory factor analysis. This method is preferable especially when a study is looking for cross-country analysis. Halleröd et al. (2006) applied the method which has been proposed by Mack and Lansley (1985) in which they employ consensual poverty approach by looking at the failure of consuming socially preferred essentials due to budget constraints. If an item is preferred by at least 50% of interviewees, then this item is a socially preferred essential therefore this approach could be considered as a consensual definition of deprivation.

According to Kakwani and Son (2006), the determination of deprivation items is an issue of value judgment depending on the societies‟ priorities (e.g. socio-economic structure); therefore, no clear-cut formula will be adequate for determining basic capabilities. Halleröd et al. (2006), for the sake of associating deprivation measure to the current lifestyle, weighted the consumption items with respect to proportion of population who is not deprived in it with a Weighted Deprivation Index (WDI) measure. Guio et al. (2009) propose to use the threshold for each dimension as setting at two or more/or three or more enforced lack (out of 9 items) in the total of

(38)

8

the indicators. It is essential to emphasize that this kind of calculation stresses on the total number of deprivations at the individual level.

Guio et al. (2009) assert different approached proposed in determination of deprivation items and the concerns developed against them. In Mack and Lansley (1985) study, the idea was collecting the views of the general public about which items they consider necessary to have a decent standard of living. For instance, the needs of individuals could be interpreted by the experts from different point of view. Alkire and Santos (2010) propose a flexible approach in determining the describing poor and the list of deprivation items due to the fact that there is no consensus on the selection of deprivation items for every country. Then the selection procedure could be based on country-specific conditions, previous experiences or empirical evidences.

As we mentioned above with reference to relevant literature, the definition of objective well-being and its measure have several approaches. The multidimensionality of deprivation provides the flexibility to researchers in terms of indicator choice or weighting. In the following sub-section, we will present the objective well-being literature in Turkey.

2.4 Objective Well-Being Literature in Turkey

History of poverty studies in Turkey is not erstwhile due to several reasons. Initially, the official statistical institute of Turkey has declared poverty definition and poverty line for Turkey for the first time in early 2000s. Secondly, the studies concerning to poverty issues mostly started after liberalization of Turkish economy in 1980s. By the socio-economic structure of the society differs from the past, the poverty, as a notion, emerged in Turkish economy and needed to calculate. However, the poverty studies employed for Turkey has a static aspect until 2000s. Most of the studies tried to determine the poverty line for Turkish economy mainly based on the data of consumption expenditures. Therefore it would be reasonable to summarize Turkish objective well-being literature by distinguishing in four strands.

In Turkey, first strand of the poverty literature shows that the studies haven‟t gone beyond the static measures of poverty until late 2000s, because of the lack of adequate data on poverty measure. In other words, the main motivation was

(39)

9

calculating absolute deprivation line, definition of poor in the society etc. By the data became more available and the need for understanding the determinants of well-being of the society increased, other strands were born. Secondly, we see the studies stressed on multidimensionality of poverty. These are the studies mostly focusing on dynamic aspects of the poverty by employing such index (e.g. MPI, MDP, etc.) to assess the determinants of well-being. It was begun to be used econometric analysis on poverty issues about a decade ago since both cross-sectional and panel data have become available. Furthermore, the longitudinal has initiated creating another strand of literature which is poverty transition analysis including both descriptive and econometric analysis. In the following sub-sections, we summarize these four strands for Turkish objective well-being literature. summarize these four strands for Turkish objective well-being literature.

2.4.1 Static approach to objective well-being

In static approach, we see the usage of money metric measures are widely preferred. As money-metric measures, consumption expenditure is primarily assessed with the data provided by TURKSTAT in HIDS 1987 and 1994. The regional comparisons, basic characteristics of the poor were identifies in most of the studies.

In his study, Dumanlı (1996) calculates the poverty line for the cities of Turkey according to the daily calorie requirement of an adult which was 2450. Dansuk (1997) mainly analyzes the measurement of poverty by including socio-economic structures. He employed consumption pattern approach in order to compute the poverty line including all consumption items. The rural areas were more suffering from poverty. Alıcı (2002) uses consumption expenditure as the indicator of welfare in her study and concluded that parallel to the finding of previous studies, the type of employment of the breadwinner of house is the key determinant of family‟s poverty state. For instance, if the breadwinner is seasonal worker, then the risk of facing poverty is three times higher than the average of the society.

Gürsel et al. (2000) investigate the dimension of poverty by calculating both absolute and relative poverty rates for Turkey and looked for the effectiveness of social transfers over the poverty rates. In order to provide the classification of poor, they calculate the portion of poor in earning type, education level, employment statue, magnitude of household, and regions. According to the results, the south-east

(40)

10

Anatolian of Turkey suffers from the poverty at most. Moreover, the households with many children and whose income is originated from agricultural activities, stipend, pensions and transfers rather than interest earnings, suffered more from poverty. Pamuk (2002) studies the relative poverty in rural areas of Turkey. The results mostly agree with the previous studies that the education level is a major determinant of being poor, especially in rural areas. Taking into consideration the household‟s facilities such as having TV, dishwasher etc., and the author concluded the quality of life in rural areas was lower than those having living facilities at home. Parallel to the findings of Alıcı (2002), having a permanent income is a significant factor for being non-poor. The sector construction and agriculture are disadvantageous if the bread winner works in one of them.

Erdoğan (2002) calculates the poverty level for Turkish economy by employing both the daily calorie requirement of an adult and the general deprivation of life facilities. Her study included policy suggestions concerning to relative poverty measurements and using different index in order to calculate comparable poverty rates. ġengül and Tuncer (2005) assess food poverty with urban/rural separation by assessing Linearly Approximated Almost Ideal Demand System (LA/AIDS) in Turkey by using the HCS 1994. They conclude that the sensibility of extremely poor households is more than those of poor households on alternation of both price and income. The severity of poverty was more evident if there have already existed persistent inequality. Yükseler and Türkan (2008) examine the dynamics of households in Turkey based on expenditure, income, and poverty for the period 2002-2006 by assessing HLS provided by TURKSTAT. Their principal remark about poverty studies was rather than using consumption expenditure in order to calculate poverty line, using income as a money-metric measure would be more reasonable. Aran et al. (2010) examine the poverty rates in Turkey between 2003-2006 and they result that the poverty rates has declined in urban areas whereas the rural areas still suffer from poverty almost at the same level at the end of the period. Therefore the risk of being poor has steadily increased for workers in agriculture sector, people living in rural areas and crowded household. Moreover, their study results another significant point that the child poverty (0-19 ages) becomes chronic and the risk of being poor for children (0-19 ages) has steadily increased in the same period.

(41)

11

Çağlayan and Dayıoğlu (2011) employ poverty probabilities of households in Turkey using econometric models. Parametric and semi-parametric logit models that they implemented for Turkish data provided by TURKSTAT in HBS 2008, result that income and number of household workers were found non-parametric and the other factors were better to explain poverty if they were considered semi-parametrically. Moreover, consistent with the previous studies they concluded that the rural areas suffer more from poverty. Kızılgöl and Demir (2010) by using money metric measures, consumption expenditure and income, states the determinants of poverty in Turkey over 2002 and 2006 with logit model. The risk of being poor declined by the breadwinner‟s age and education level increased; however the poverty was more severe in rural areas. Çağlayan et al. (2012) discuss the determinants of household poverty for Turkey by assessing Household Budget Survey 2009. Using ordered logit model, they asserted that the gap between the upper and the middle class has been widened whereas the middle class approached to the poor class. Aydın and Güloğlu (2013) discuss mostly the factors that cause poverty and the dimension of relative poverty by assessing HBS and HCS between 2003 and 2006 by providing statistical comparisons and variance analysis. Their results were similar to the findings of the previous studies concerning to education and employment characteristics; a discrepancy was that they also provided evidence that the people over 60 ages were subject to be poor more frequently than other people.

2.4.2 Multidimensional approach to poverty

Akder (2000) employs human development approach in order to project the dimensions of rural poverty in Turkey with the projection of the works done previously. According to this study, among the 7 regions of Turkey, the more a region included rural cities, the more its risk of having low human development index increased. Dağdemir (1999) focuses on the dimensions of the poverty and its effects on the redistribution of income between 1987-1994 timespan where the drawbacks of economic instability were visible after Gulf War in Turkish economy. Similar to the previous static approach studies, the poverty was a more cognizant issue in rural areas. In addition, by the migration towards urban areas increased, the poverty also migrated to urban areas.

(42)

12

KabaĢ (2010) searches answers for the characteristics of the people who has the lowest 20 percent of income distribution. According to the major results, human capital is the key determinant of improving human development index of poor people and it is considered as the pro-poor growth strategy in many economies. Moreover, this study has a proposal about using non-monetary metrics in measuring poverty as these metrics, such as literacy and enrolment rates, infant mortality rate, life expectancy rate, reflects better the characteristics of a society. Yıldız (2011) employs income based poverty measures and multidimensional poverty measures in Turkey by employing 2003 HBS provided by TURKSTAT. The findings show that, South Eastern Anatolia is the poorest region and Istanbul has lowest poverty. With multidimensional approach, non-monetary dimensions (education, health insurance, assets) are included in poverty measurement, and it shows that for all regions, severity and depth of poverty is higher rather than income based poverty.

Acar (2014) implies the dynamics of multidimensional poverty in Turkey by assessing the SILC panel data for years 2007 and 2010 with random effect probit model. According to that study, the more people educated, owned a house or gained rental/asset income, the more the probability of being multidimensionally poor declines. On the other hand, family size, working in agricultural sector or taking social assistance cause the probability of classified as multidimensionally poor. Acar concludes that EU severe material deprivation overestimates the poverty line in Turkey. Karadağ and Saraçoğlu (2015) calculated two different multidimensional poverty measures over 2006-2012. They assessed not only monetary indicators, but also severe material deprivation, work intensity, health, education and environmental indicators while constructing multidimensional poverty measures. The major results that they found are follows: the poverty level was still higher in Turkey than other European countries. Turkey‟s results are close only those of Greece and Portugal.

2.4.3 Policy suggestions and transition in objective well-being

Buğra and Sığmazdemir (2004), recommended cash transfer program for people who live under poverty line. Their works debated the arguments against cash transfer program, especially based on insufficiency of resources. Gündoğan (2008) discusses the direct and indirect ways of combat against poverty from world and Turkish literature; and proposes policies in order to be successful in the purpose of decreasing

(43)

13

poverty. Ġncedağ (2013) questioned existing poverty alleviation policy and programs by handling dimensions of poverty in Turkey and provided certain proposals against it.

ġeker (2011) is seen as the first study on poverty dynamics in Turkey in which the transition in and out of poverty, the state dependency and social assistance programs as a remedy are examined using SILC 2006-2007. She resulted that change in earnings was the major reason as a determinant of poverty transitions; if one person was low-pay initially and it continued for the following periods, it was more likely for him to remain poor in the future. Social assistance programs have a limited influence on reducing poverty.

Acar and BaĢlevent (2014) study transitions of households into and out of poverty in Turkey for 2007-2010 periods. Their approach depended on using binary choice model which takes into account the different characteristics of the poor separately. This data was provided by SILC data from Turkstat. They concluded that the employment status and schooling of the household head and household size were closely associated with transition of household out of or from poverty. Additionally, some social transfers, including cash transfers, child benefits etc. had a significant effect on increasing the probability of exiting poverty.

Dalgıç et al. (2015) examine the determinants of household transition out of and from poverty by assessing 2005-2011 SILC data by TURKSTAT with probit model application. According to the major results from this work, although the absolute poverty declined during the period assessed; the relative poverty still persisted at a high level. The risk of being poor increases if the household was deprived from social security and the gender of the breadwinner of the household (if woman), as well as the education level played a crucial role in decreasing the poverty risk. Limanlı (2015) employs SILC 2006-2009 conducted by TURKSTAT in order to measure the intertemporal poverty in Turkey. He used both intertemporal poverty index which allowed for evaluating the differences in being poor and its dissemination among individuals and probit regression models in order to provide estimation results. Of the numerous results, perhaps the most attractive ones were, if the breadwinner of the household were female, the household was more likely to be poor that asserted a contrast to the finding of Dalgıç et al. (2015), the number of the

(44)

14

non-worker members in the household should be compensated with social welfare transfers.

ġeker and Dayıoğlu (2015) look at the poverty dynamics by assessing poverty persistence, time spent in poverty, re-entry rates, trigger event which cause entering into poverty again over the period of 2005-2008. Besides, their study stated the characteristics of the poor and the persistent poor for the examined periods in which they concluded that 8 percent of the population was persistent poor in Turkey. Labor market dynamics played a key role in poverty persistence.

(45)

15 3. SUBJECTIVE WELL-BEING

Subjective well-being literature does not stress on the proposition that: „money does buy happiness‟, nor do we. The crucial aim is assessing it to estimate the overall quality of life or the satisfaction in different domains of life within a country or a specific group. Then, the assessment of SWB requires the collection of relevant data for that purpose; therefore the survey design for SWB studies has a major role in better understanding life satisfaction. Yet, before analyzing SWB, it is crucial to define it; because, SWB is conceptualized with different terms (e.g. life satisfaction, happiness, quality of life, etc.). In this section, we will summarize its definitions from literature and the selected preeminent works both from world and Turkish literature.

This latter includes a broader framework as the dimensions are multiple and reflect different domains of life (e.g. literacy, health, living conditions). Nevertheless, it isn‟t a priori clear which aspects of well-being should or not be considered in a standardized measure, nor this measure is applicable to all societies in terms of social or cultural norms, etc. Yet, human being is more than a consumer as an economic agent. Thus, it can be stated that well-being is a complex phenomenon affected by psychological, social and economic dimensions of an individuals' life. Furthermore, the heterogeneity of personal perceptions plays a crucial role in well-being measure inasmuch as perceptions refer to subjectivity.

3.1 Defining Subjective Well-Being

The decomposition of SWB is not easy as it totally reflects its own meaning,

subjectivity. It generally refers to life satisfaction, happiness or quality of life. The

focus of SWB literature is also on those concepts. However, there is not any consensus on whether these terms are the same or they include distinctions.

OECD (2013) considers SWB like the expression of the good mental well-being which consist life assessment, affect and eudaimonia which is defined as the aim of life, or good mental health. Easterlin (2001) states that life satisfaction and happiness

(46)

16

can be assessed interchangeably whereas Diener (1984) and Diener and Tay (2015) suggest a distinction among them as the former refers to an cognitive element, on the other hand the latter includes momentary moods inasmuch as it is an affective element. Tsou and Liu (2001) propose to consider the two concepts separately although they show to some extent the correlation. Kahneman and Deaton (2010) suggest that SWB can be considered with a distinction of emotional well-being and evaluation of life where emotional well-being is about emotional quality of people (e.g. how frequently one feels stressful or anger); on the other hand, the evaluation of life is about people‟s declaration about this frequencies. This latter refers to the cognitive component of SWB which is able to show the quality of life as a whole; on the other hand, the affective part should be experimented in a laboratory environment or using other techniques (e.g. Day Construction Method, Experience Sampling Method). Household surveys are appropriate and widely used for this purpose.

Happiness is the degree to which an individual judges the overall quality of his or her life as favorable; in other words, happiness is a summary statistics of subjective well-being (Veenhoven 1991, 1993). Rojas (2008) defines SWB as not experiencing poverty; in other words, a human being is in experienced poverty if he/she has low life satisfaction. Diener et al. (2009) explains SWB as the achievement of a level of material abundance and health which allow people to seek for other aspects of quality of life rather than economic prosperity. Therefore, quality of life refers to the overall well-being of individuals in a broad and multidimensional sense. In other words, there exists a confrontation of objective living conditions and their subjective assessment by satisfaction scales.

Hence, the decomposition of SWB is not straightforward; its composition differs from country to country based on cultural factors as well as other dimensions.

3.2 Subjective Well-Being Literature

Subjective well-being has been studied in sociology and psychology domains before in economics. The reason behind this is at least twofold: Firstly, the subjectivity is naturally a different notion than objectivity and this is explicable with psychological theories. People‟s perceptions and life purposes differ from one person to another. Therefore, SWB could be linked to various psychological processes. Inglehart (1990)

(47)

17

defines happiness as a short-term result of becoming richer in term of income. Secondly, subjectivity is affected by socioeconomic structure of societies.

In economics, the studies concerning to SWB assessments has increased in the last decades. Among these studies, life satisfaction measurements based on the individual declaration on their utility is widely assessed. The main reason for this as follows: most generally, the concept of wellbeing has been defined as material wellbeing measured by traditional money-metric measures such as income or consumption. The basic understanding relies on the assumption that increased income triggers more expenditure and this latter causes to obtain more utility (Conceição and Bandura, 2010).

Nevertheless, SWB literature prefers to go beyond these basic assumptions. Initially, Allardt (1976) constructed a well-being components concept which includes “having, loving and being” components in analyzing citizens‟ personal declarations in the Scandinavian Welfare Survey. In his study, “having” is related to the material and impersonal resources; “loving” expresses the relations which determines social identities with other people; “being” refers to how people integrate themselves in the society. Townsend (1979) in his pioneering work which he employs twelve indicators using the survey which includes sixty indicators on deprivation of British people, differentiates deprivation as objective and subjective. For the latter, the answers for the question “to what extent did people feel deprived?” constitute the subjective deprivation.

The studies stressed on micro data usually investigate the relationship between income and SWB and mostly agree that relative income decreases SWB (Clark and Oswald, 1996; Ferrer-i- Carbonell, 2005).. On the other hand, the people generally incline to compare themselves with other people in term of income rather than other aspects such as happiness or life satisfaction (Blasquez and Budria, 2014).

The early studies focused on cross-country comparisons, include relationship between marginal utility and the SWB. There was no causal path from income to SWB in European countries; although the economics was in a growth era (Clark and Oswald, 1994; Oswald, 1997). Various socioeconomic variables (e.g. age, gender, income, etc.) are shown that they are related to SWB; therefore happiness is not a definable notion based on a single component. Moreover, demographic variables

(48)

18

show different consequences in different cultures. In other words, cultural norms matter. Diener et al. (2000), result that based on marriage variable, living with a partner without officially marrying could change the happiness degree if it is not socially accepted.

Although there is a statistically significant relevance between income and the life satisfaction, the coefficient results small relevancy. The correlation between income and SWB is less as it is the case for poor countries (Easterlin, 1974, 2001; Argyle, 2001; Veenhoven, 1997). Raising the incomes of all does not increase the happiness of all (Easterlin, 1974). This is called Easterlin paradox in the happiness literature. Moreover, in Japan which is a developed country, Easterlin (1995) finds that the average SWB did not increase between 1958 and 1987, although real income increased fivefold. Income has a limited capacity to capture the association of both economic and life satisfaction (Rojas, 2008). Blanchflower and Oswald (2004) provide similar results for US and UK.

In their study conducted for Taiwan, Tsou and Liu (2001) find out that the SWB is associated with a higher income. Clark et al. (2008) provide an extensive overview of the recent findings with respect to the correlation between income and happiness, both at the country-level, in terms of GDP, and the individual-level. At the individual level, subjective well-being is found to rise with income in both cross-section and panel data. Tsou and Liu (2001) also assert that the personal characteristics play major role as determinant in domain satisfaction. On the other hand, although socio-demographic and economic factors such as being married, educated, and healthy (Argyle 2001; Diener et al. 1999; Myers 2000), and a higher income and stable employment status (Clark and Oswald 1994; Diener et al. 1999; Di Tella et al. 2001) are positively associated with happiness, they fail to account for important differences in individual happiness (Diener et al. 1999). The discrepancy between individuals and also countries could be better referenced to other institutional differences (e.g. governance, democracy, transparency, etc.) (Veenhoven, 2000; Frey and Stutzer, 2000; Di Tella et al. 2003; Helliwell 2003).

In the following sub-section, we will summarize the studies about SWB in Turkish literature.

(49)

19 3.3 Subjective Well-Being Literature in Turkey

There are few studies in SWB literature for Turkey; during the last decade it became an emerging literature due to the availability of representative survey datas. These studies are mainly concentrating on the association between SWB and the socio-demographic features of the respondents. We will summarize some of them in this section (for a detailed list of SWB literature in Turkey, please see Table B.2 in Appendix B).

The first study that we find in the literature is from Gitmez and Morçöl (1994) in which they investigate the relationship between life satisfaction and socio-economic status based on a survey implemented by them in Ankara, the capital of Turkey, in 1990. Their main findings are the result of the socio-economic composite index calculated with survey data and imply that socio-economic status is a crucial determinant for the satisfaction. They control for economic well-being with asset and income variables. After this preeminent work, we meet this literature with the studies conducted by 2000s. Selim (2008) finds similar results for Turkey as the world SWB literature suggest for positive influences of income and health status whereas negative effects of unemployment and age on life satisfaction. Moreover, education level shows different pattern; being female with a secondary level education causes to be less satisfied. Moreover, she results that the number of children on happiness and life satisfaction is negatively associated.

Bülbül and Giray (2011) employ non-linear canonical correlation analysis in order to analyze the relationship among the socio-demographic features and happiness in Turkey. Married, homemaker and young women are found to be more satisfied. In addition, the men, who live in urban areas, have low education level and whose average earning is around minimum wage, are also found to be more satisfied with life. This latter group is the most satisfied with the family life in terms of domains of life which is consistent with our findings. Akın and ġentürk (2012) employ the previous wave of our dataset, second EQLS, for Turkey and result that people whose marital status is married, gender is male and employment status is retired or student are the happiest. The happiness level increases with the increase in income. Caner (2015) employs both TURKSTAT and WVS datasets in order to assess the SWB determinants for Turkey. She concludes that the results could differ based on the

(50)

20

dataset employed. Nevertheless, some of the common findings among the datasets are as follows: that the being unemployed is a cause for being less happy; while being married has a positive impact on happiness. Ekici and Köydemir (2014) differs from the studies done before them as they include for the first time the social capital indicators and investigate their linkage with SWB by employing European Values Survey for 2011. The main findings are as follows: the trust in institutions, the satisfaction in government and democracy show a positive correlation with SWB. Eren and AĢıcı (2016) investigate the determinants of happiness based on city-level data and result that being employed and married increase the happiness; however the education level has an increasing effect on happiness only if it helps to increase income.

Dumludağ (2013), Dumludağ et al. (2015) and Dumludağ et al. (2016) focus on the association between SWB and household consumption, internal and external comparisons. In the former, Dumludağ (2013) concludes that being married is not significant and has a negative effect on life satisfaction. Men are found to be much satisfied than women. Life satisfaction is inclined to increase with age, education level, consumption and health status, and to decline with negative effect of unemployment. Dumludağ et al. (2015) concentrate on the life satisfaction of Turkish immigrants living in Netherlands based on the fieldwork data collected by the authors and differ in the literature by assessing the SWB of the immigrant living abroad with respect to two reference groups: the ethnic reference groups and life-domain reference groups. The former results that perceived importance of income comparison with Dutch natives is significantly and positively correlated to life satisfaction, while the latter also shows a similar result that the people are impressed by comparing their income with relatives in the Netherlands in terms of subjective well-being in negative way. The association between relative income and SWB is consistent with Easterlin Paradox.

Dumludağ et al. (2016) explore the effect of income comparisons on life satisfaction in a collectivistic society. The main findings of this paper as follows: an increase in income, being a homemaker or retired and living in rural areas have positive effects on the reported life satisfaction; age variable presents a U-shaped relationship.

Referanslar

Benzer Belgeler

It was determined that trailer parks that would be developed based on the criteria determined in the present study would encourage visitors to become an important part of the

In a community-based program implemented in India over four years, it was reported that, after the intervention, the number of individuals who had a 10% or higher FRS risk

Öldükten sonra su .çensıne atılan veya su içerisinde fakat suda boğulmanın dışında başka bir nedenle ölen ve burada bir süre kalan cesetlerin

Sabahattin Eyuboğlu’nun yurt ve uygarlık yorumu, Anadolu bileşiminin bütün kaynaklarını kapsar; Olympos’u da, Sultanahmet se­ bilini de, Halk Ana’nın bütün

Babüssaade önünde çocuk hü­ kümdara bi’at edilmiş ve kalabalık halinde Sultan İbrahimin yanına varılmıştı. Ben her biri­ nize ihsanlar etmedim mi, şimdi

Birinci bölümün son alt başlığı Deniz Harbine Hazır Olma (119-124) herhangi bir deniz harbi öncesinde donanmaya üs olarak hizmet vere cek olan ikmal noktalarının

Koruluklar/baltalıklar özellikle doğu köknarından (Picea tarihinde arkadaşları M. Ancketil ve Dr. Neroutsos ile birlikte ikinci Anadolu seyahatine çıkan

İnfaz koruma memurlarının cezaevi içerisindeki gündelik hayat, özlük hakları, mahpus, meslektaş ve amirleri ile olan ilişkilerinde, taleplerinde ve çalıştıkları