POLICY DIFFUSION BETWEEN THE WORLD BANK AND
TURKEY: THE SOCIAL RISK MITIGATION PROJECT
A Master Thesis
by
ARAM YUNUS GEBOLOĞLU
Department of Political Science and Public Administration
İhsan Doğramacı Bilkent University
Ankara
December 2020
ARAM YU NU S GE B O LO ĞL U P OL ICY D IFF USIO N BETW EEN TH E W ORLD B AN K A ND TU R KE Y: B il ke nt Univer sit y 2020 TH E S OCIA L RIS K MI TI GA TI ON P R OJECTPOLICY DIFFUSION BETWEEN THE WORLD
BANK AND TURKEY: THE SOCIAL RISK
MITIGATION PROJECT
The Graduate School of Economics and Social Sciences of İhsan Doğramacı Bilkent University
by
ARAM YUNUS GEBOLOĞLU
In Partial Fulfillment of the Requirements for the Degree of
MASTER OF ARTS IN POLITICAL SCIENCE AND
PUBLIC ADMINISTRATION
THE DEPARTMENT OF
POLITICAL SCIENCE AND PUBLIC
ADMINISTRATION
İHSAN DOĞRAMACI BİLKENT UNIVERSITY
ANKARA
i
ABSTRACT
POLICY DIFFUSION BETWEEN THE WORLD BANK AND
TURKEY: THE SOCIAL RISK MITIGATION PROJECT
Geboloğlu, Aram Yunus
M.A., Department of Political Science and Public Administration Supervisor: Asst. Prof. Dr. Hasan Tolga Bölükbaşı
December 2020
Turkey adopted the Social Risk Mitigation Project (the SRMP) in 2001 with the incentives of the World Bank (the WB) to recover the effects of the 1999 earthquake and 2001 financial crisis on the poor. This thesis aims to indicate how conditionality, learning, and emulation as policy diffusion mechanisms operated together in the adoption process of the SRMP in Turkey. Policy diffusion literature generally tests these mechanisms individually to explore whether or not they cause policy
adoptions. However, this thesis argues that diffusion mechanisms might be
interrelated and work in complementary ways in policy adoptions. The Turkish case supports this argument by relying on process tracing methodology with
semi-ii
structured interview data and official documents. I argue that, during the adoption process of the SRMP, conditionality serves as a facilitator to adopt the project due to the financing by the WB. The WB, in the meanwhile, serves as a link establisher between Turkey and Latin American countries in the learning process regarding the possible outcomes of the project in addition to financing it. Emulation operates together with conditionality and learning with the emphasis on the prestige of the project among the international community due to the modernization of the social assistance institutions of Turkey. Both Turkish officials and the WB consider the adoption of the project as a prestigious and appropriate attitude.
iii
ÖZET
DÜNYA BANKASI VE TÜRKİYE ARASINDA POLİTİKA
DAĞILIMI: SOSYAL RİSKİ AZALTMA PROJESİ
Geboloğlu, Aram Yunus
Yüksek Lisans, Siyaset Bilimi ve Kamu Yönetimi Tez Danışmanı: Dr. Öğr. Üyesi Hasan Tolga Bölükbaşı
Aralık 2020
Türkiye, Dünya Bankası’nın da inisiyatifleriyle 1999 depremi ve 2001 krizinin yoksul insanlar üzerindeki etkisini azaltmak için Sosyal Riski Azaltma Projesi’ni (SRAP) edinmiştir. Bu çalışma, politika dağılımı mekanizmalarından Şartlılık,
Öğrenme ve Öykünme’ nin SRAP’ın edinilme sürecinde birlikte nasıl bir etkide
bulunduğunu göstermeyi amaçlamaktadır. Politika dağılımı literatürü genel olarak bu mekanizmaların bireysel olarak politika edinimine yol açıp açmadığını
incelemektedir. Fakat bu tez, bahsi geçen mekanizmaların politika edinim süreçlerinde birbirleri ile tamamlayıcı ve ilişkili olabileceğini savunmaktadır. Çalışmanın temelini Türkiye üzerine yapılan bir vaka incelemesi oluşturmaktadır. Vaka incelemesi, mülakat verileri ve resmi dokümanlardan faydalanarak süreç takibi
iv
metodu üzerinden yapılmaktadır. Bu tez temel olarak şu sonuca varmıştır: SRAP’ın edinim sürecinde Dünya Bankası’nın projeyi finanse etmesinden dolayı Şartlılık mekanizması Türkiye’nin projeyi kabul edip uygulamasında kolaylaştırıcı bir etkide bulunmuştur. Projeyi finanse etmesine ek olarak Dünya Bankası aynı süreç içerisinde Türkiye ve Latin Amerika ülkeleri arasında bağlantı kurma vasfıyla uygulanacak olan projenin muhtemel sonuçlarına dair Öğrenme mekanizmasını tetiklemiştir.
Öykünme mekanizması ise, Şartlılık ve Öğrenme ’ye ek olarak, uygulanan projenin
uluslararası toplumda Türkiye’ye itibar kazandırma durumuna işaret etmektedir. Hem Dünya Bankası hem de Türk yetkililer, SRAP’ın Türkiye’nin sosyal yardım kurumlarını modernize edecek olmasından dolayı projenin uluslararası toplumda bir itibar kaynağı olacağına inanmaktadır.
v
ACKNOWLEDGEMENTS
I would like to, first, thank my mother and father who always supported me in my entire life in both material and nonmaterial ways. Special thanks to my father, Alper Tunga GEBOLOĞLU, who has respected and supported all my choices in education life both materially and emotionally. Special thanks to my mother, Fatma AKTAŞ, who has always intimately and devotedly supported me and has, always been tolerant of me throughout my entire life. Without her supports in every way, life would undoubtedly be much more difficult for me, especially in this ongoing process of the Covid-19 pandemic. Special thanks to Kurnia AÇIKÖZ who has not given up encouraging and supporting me during the dissertation process and who has always trusted my skills.
I would like to thank all department staff for the great experience during six years both at the undergraduate and graduate level in Political Science at Bilkent
University. I would like to thank all the interviewees who accept to contribute to my data gathering for the dissertation; their efforts are really appreciated.
I would like to thank my supervisor, Dr. H. Tolga BÖLÜKBAŞI, who has supported me with his academic feedback in the whole dissertation process and who has also established connections for the interviews with the key policymakers within the SRMP.
vi
TABLE OF CONTENTS
ABSTRACT ……….………... i ÖZET ……….………... iii ACKNOWLEDGEMENTS ……….…... v TABLE OF CONTENTS ……….. vi CHAPTER 1: INTRODUCTION ………...1CHAPTER 2: RESEARCH DESIGN AND METHODS IN STUDYING POLICY DIFFUSION ……… 9
2.1 Introduction ………9
2.2 From Quantitative to Qualitative ……….10
2.3 Case Study ………...……12
2.4 Process Tracing ………... 14
2.5 Interview Data ………. 16
2.6 Conclusion ………...19
CHAPTER 3: THEORETICAL FRAMEWORK ………..21
3.1 Introduction ………...21
vii
3.3 Diffusion Mechanisms as Explanatory Tools ………... 23
3.3.1 Conditionality ……….. 24
3.3.2 Learning ………....25
3.3.3 Emulation ………...28
3.3.4 Competition ……….. 29
3.4 Conclusion ………...32
CHAPTER 4: THE EMPIRICAL LITERATURE ………...34
4.1 Introduction ………...34
4.2. Conditionality ………...36
4.3 Learning ………...41
4.4 Emulation ……….47
4.5 External Influence on Turkish Policymaking ………...50
4.6 Conclusion ………...53
CHAPTER 5: WORLD BANK’S APPROACH TO SOCIAL PROTECTION AND THE SRMP IN TURKEY ...56
5.1 Introduction ………...56
5.2 The WB’s Conventional Wisdom to Social Protection ………..….57
5.3 Social Risk Management as A New Concept ………...58
viii
5.5 Adjustment Portion ………...60
5.5.1 Rapid Response ……….60
5.6 Investment Portion ………...61
5.6.1 Institutional Development ……….61
5.6.2 Conditional Cash Transfer ………61
5.6.3 Local Incentives ………62
5.7 Conclusion ………...63
CHAPTER 6: THE ADOPTION PROCESS OF THE SRMP IN TURKEY……….65
6.1 Introduction ………...65
6.2 Main Drivers for the SRMP ………...66
6.3 Earthquake, Trust Building, and Negotiation ………..67
6.4 Conditionality: Facilitation rather than Coercion ………70
6.5 Learning: Looking at ‘the Other’ ……….73
6.6 Emulation: Being ‘Prestigious’ ………77
6.7 Conclusion ………...80
CHAPTER 7: CONCLUSION ………...82
REFERENCES ………..87
1
CHAPTER 1: INTRODUCTION
This thesis explores how policy diffusion mechanisms -conditionality, learning, and emulation- operate in a complementary way in the adoption of the Social Risk
Mitigation Project (the SRMP) to reduce poverty in Turkey. Adopted from the World Bank (the WB) repertoire in 2001, the SRMP is one of Turkey’s largest and
sustainable projects in the area of social policy. The project led to the long-term programming for Conditional Cash Transfer (CCT), an exceptionally sustained project to help the poor. This thesis aims to reveal how the negotiation process between the WB and Turkey shapes the process of policy diffusion through these mechanisms. To indicate that these mechanisms are complementary in the adoption of the SRMP, I will conduct a within-case analysis using archival research among official documents and interview data from three key policy players who build the project for Turkey. Before going into details of those issues in the following parts of my analysis, I will briefly explain the process of policy adoption through policy diffusion.
Adopting a new policy contains various internal and external dynamics in its process. Policy adoptions generally have their main drivers that necessitate reforms; domestic dynamics matter in terms of politics and power relations specific to the states or countries. However, even though the main drivers might be internal or structural problems such as economic crisis, aging, or political tensions among the parties, policy adoptions might have external dimensions as well. In a globalizing world,
2
countries are getting closer to each other thanks to developments in communication and technology. This situation has political consequences such as policy diffusion, and one can observe the spread of the same or similar policies across the world due to the interdependence among the political units. In that sense, policy diffusion means a process in which a political unit’s decision is affected by the other political units’ decisions (Magetti & Gilardi, 2015: 3). Diffusion studies first identify whether there is policy diffusion from one political setting to another and then, aim to explain the causes of the policy adoptions. The literature presents four main diffusion
mechanisms as explanations for diffusion: conditionality, learning, emulation, and competition. Conditionality refers to policy adoption in exchange for financial aid from an International Organization (IO) or for accession to a supranational
organization such as the European Union (the EU) (Gilardi, 2012: 461); Learning relies on a cost and benefit calculation in light of the experience of previous adopters and requires a rational assessment on the possible outcomes if a policy is adopted, i.e., whether the policy will produce the expected outcomes or not (Weyland, 2005: 271); Emulation emphasizes the logic of appropriateness, prestige, and legitimacy in international community gained in return for policy adoption regardless of the
outcome of the policy (Dobbin, Simmons & Garrett, 2007: 452); Finally, competition concerns adopting similar policies among economically competitor countries to obtain financial resources from the international economy such as Foreign Direct Investment (FDI) (Simmons and Elkins, 2004: 182).
In light of diffusion literature, this study aims to explore how diffusion mechanisms operate together for the policy diffusion from the WB to Turkey, specific to the SRMP in 2001, a project aimed at overcoming the economic effects of the 1999 earthquake and the 2001 financial crisis in Turkey (Prime Ministry, 2007: 13).
3
Turkey experienced a catastrophic earthquake in 1999 and then a serious financial crisis in 2001, both of which had negative impacts on the poor. To reform its social policy programs and institutions and, to reduce the effects of the crisis on the poor, Turkey negotiated with the WB to adopt the SRMP with the financial support of USD 500 million. This study aims to show the diffusion mechanisms through which Turkey adopted the SRMP from the WB.
Some scholars who study the SRMP heavily address the implementation of the project regarding CCT and the impact of the project on women’s socio-economic status. These scholars underline the logic for implementation of the project due to the concerns for a social explosion as a result of the 2001 financial crisis (Zabcı, 2009: 124; 2003: 232). They draw attention to the WB’s social risk management concept for the design of the project aiming to mitigate poverty and transforming the poor into risk-taking individuals to cope with future crises (2003: 215). The
implementation and outcomes of the project are controversial issues among scholars. In these accounts, poverty is seen as a structural problem that cannot be solved through social assistance projects rather, accession to social services and employment are rights that should be provided by the state for every individual (2009: 124; 2003:237).
In contrast with these arguments, others stress the importance of CCT, as a component of the SRMP, to reduce child poverty (Karaaslan, 2015: 26).
Furthermore, others argue that CCT contributes to the education and empowerment of poor women (Öcal, 2012: 31). However, some scholars find no remarkable
influence of CCT on women’s socio-economic status even though it aims to improve women’s position in society. Women utilizing CCT are not better-off in comparison with women without CCT and, it contributes to the practical needs of women rather
4
than their socio-economic status in the society (Emir, Erbaydar & Yüksel, 2013: 120). Furthermore, women entrepreneurship and employment do not develop through projects providing micro-credits, such as the SRMP with Local Incentives
component, due to cultural barriers, prejudices, and discrimination (Sallan Gül & Altındal, 2016: 1361). Demircioğlu supports this argument by indicating that the income and social status of women are consolidated through those projects being not interested in the maintenance of women entrepreneurship and, considering women as homogenous (2016: 163). All these studies analyze the outcomes and
implementation of the SRMP. Even though Zabcı indicates the effect of the WB by financing and consultancy in the adoption of the project, she does not specifically address diffusion mechanisms (2003: 232). However, this thesis aims to identify the adoption process of the project through diffusion mechanisms in a complementary way.
The diffusion literature tends to assess the causal effects of these mechanisms individually with statistical pattern-finding analysis. This means the literature predominantly takes for granted that one mechanism is more significant than the others to produce policy adoption. I will show, however, that we gain analytical leverage by showing how a set of mechanisms may be complementary for producing the outcome. I will do so by focusing on a case study of Turkey’s adoption process of the SRMP where a specific combination of several mechanisms – conditionality, learning, and emulation - produces the policy adoption. The competition mechanism is not covered in this process because there is no expectation or regulation to gain more financial resources in Turkey. First, the SRMP does not foresee any financial regulation to obtain economic in-flows from the international economy. Second, the WB, as an IO, does not stand as a competitor of Turkey with an ambition to gain
5
financial resources. This thesis aims to uncover a policy diffusion process from an IO to a state rather than among competitor countries. One key contribution this study hopes to make, therefore, is to show that diffusion mechanisms may operate in complementary, rather than competing, terms in specific contexts. In that sense, this thesis shows how the diffusion mechanisms may in fact be interrelated and
complementary (Graham, Shiphan & Volden, 2012: 695; Marsh & Sharman, 2009:273). These scholars argue that separating and measuring the diffusion mechanisms is difficult and those mechanisms often operate together in leading policy adoptions. This thesis argues that the Turkish case of the adoption of the SRMP provides strong evidence to show how the co-existence and co-operation of conditionality, learning, and emulation work in a complementary way. In that sense, I argue that all these mechanisms are necessary individually but not sufficient to lead the adoption of the SRMP alone. I try to indicate how they operate individually and how their combination produces the outcome. Thus, this thesis aims to contribute to the literature on Turkey with fresh empirical evidence from this case.
To support this argument, this study relies on qualitative within case studies with a process-tracing approach to uncover the causal process of policy adoption. To open the black-box of the causal chain, three interviews were conducted with the key policymakers from both the WB side and the former Turkish officials who took part in the negotiation process of the SRMP in 2001. This study aims to indicate,
ultimately, how the combination of conditionality, learning, and emulation operate in the adoption process of the SRMP in which all those mechanisms are necessary individually, but not sufficient enough for producing the policy adoption. In that sense, the adoption of the SRMP by Turkey indicated a successful case of learning and emulation, not much conditionality because it did not operate in a carrot
6
and stick model. The operation of those mechanisms occurred in the Turkish context with specific conditions. The main drivers for the adoption of the SRMP stemmed from structural and internal dynamics including the 1999 earthquake and the 2001 financial crisis. The cooperation among Turkey and the WB for the Earthquake Emergency Recovery Loan (the EERL) in the 1999 earthquake helped for trust-building among the partners, which also enabled the adoption of the SRMP. These conditions necessitated the adoption of the project to develop social assistance institutions and to poverty reduction in Turkey.
Conditionality facilitated the adoption of this project with financial support without any coercive attributes by the WB because the content of the agreement was set mutually among the partners rather than the imposition of the WB. Turkey actively participated in the negotiation process for the SRMP and asked for conditions in addition to the financing of the project such as; prevention of a 40% budget cut of Social Assistance and Solidarity Associations; a law called General Directorate of Social Assistance and Solidarity to ensure the future implementation of the project; Rapid Response for women and children after the 2001 crisis; and Institutional Development to gather data regarding the poor. For all that, the WB desired to extract a success story regarding the project in Turkey so that it could present it to the rest of the developing countries. Furthermore, the WB recommended CCT to Turkish policymakers to implement for poverty reduction. All those findings indicate that conditionality did not operate as an anti-liberal carrot and stick model or
unilateral imposition by the WB rather, it served as a facilitator by financing and enabling the adoption of the project.
When conditionality opened a channel to adopt the SRMP, in the meanwhile,
7
by the WB with the Latin American countries, especially Mexico regarding CCT. The WB performed as a link establisher in the adoption process of the project. Apart from financing the project, Turkish policymakers were also concerned with the consequences of the project that to what extent the project would prevent poverty. This indicates that conditionality and learning operate together in the adoption process of the SRMP. This corresponds to the motivation of Turkish policymakers for the possible outcomes of the project with cost and benefit analysis along with the concerns for financing the project. However, the learning mechanism among Turkey and Latin American countries did not operate consistent with geographical proximity and ideological similarity as the literature addressed. Rather, the interests, aims, and concerns of the policymakers were at stake in the learning mechanism for the Turkish case.
In addition to that, both the WB and Turkish officials considered the adoption and implementation of this project as a matter of prestige that indicated Turkey’s compliance with modernity and international norms. Turkish policymakers, in addition to learning from previous experience as a rational problem-solving attempt, considered the modernization of social assistance institutions as a representation of modernity and a source of prestige. The WB also attributed prestige to the adoption and successful implementation of the project by Turkey so that it could present Turkey to the rest of developing countries as a modern member of the international community. This indicates that learning and emulation operate together along with conditionality in the adoption process of the SRMP. Emulation did not operate in merely imitating the WB’s recommendations or Latin American countries without considering the policy outcomes. From all of these findings, the Turkish case provides a successful policy diffusion by indicating the operation of conditionality,
8
learning, and emulation in a complementary way for the adoption process of the SRMP, which supports the arguments constituted by Graham, Shiphan, Volden, Marsh, and Sharman.
The literature on the external influence on Turkish policymaking lacks a specific study that directly addresses the diffusion mechanisms for policy adoptions from the EU, the IMF, the ILO, and the WB. This thesis also aims to fill this gap by directly addressing diffusion mechanisms in policy adoption processes.
The organization of the study is as follows: first I will justify the qualitative case study with a process-tracing approach relying on semi-structured interview data to indicate how diffusion mechanisms will operate together for the adoption of the SRMP. Second, I will present the theoretical background of diffusion mechanisms and explain them in a detailed way. Then, I will critically approach the empirical diffusion literature that explains policy adoptions in the light of a single mechanism. In that sense, I will argue that various mechanisms might operate for a policy
adoption depending on the context. After that, I will briefly identify the
transformation of the WB’s approach to social protection and the concept of Social Risk Management (SRM) which is the conceptual framework of the SRMP. Then, I will explain the content of the project in Turkey. Finally, I will indicate, with smoking-gun evidence in light of the negotiation process among policymakers, how the combination and the co-existence of conditionality, learning, and emulation lead to the adoption of the SRMP by Turkey.
9
CHAPTER 2: RESEARCH DESIGN AND METHODS ON
STUDYING DIFFUSION
2.1 Introduction
In this chapter, I will present the methodological choices of this study to show how diffusion mechanisms operate together in the adoption process of the SRMP. First, I will make a distinction between quantitative and qualitative ways of studying policy diffusion. Quantitative studies mostly rely on a large-N statistical analysis of
diffusion mechanisms. The effects of those mechanisms on policy adoption are considered individually to identify and explain policy adoptions. However, I will indicate the views of the scholars that statistical analysis is not enough to explore the causal process of those mechanisms in policy adoption. Qualitative studies, on the other hand, are argued to be more suitable for the causal explanation with the investigation on the interactions of policymakers via in-debt analysis of within-case studies. Thus, I will explain the logic of case studies to study policy diffusion either for the outcome or the explanations of an outcome. I will underline the case selection of Turkey, as a causal case study, based on my argument that diffusion mechanisms are interrelated and their combination produces the outcome. The outcome here is the adoption of the SRMP and I will argue that the case study on Turkey for policy diffusion supports my approach. To construct my argument in a within-case study, I will emphasize the role of process tracing methodology in policy diffusion studies to
10
open and explore the black-box. In other words, I will offer an explanation and identification of the factors or causes of a specific outcome with smoking-gun evidence. To accomplish this, this study will conduct both deductive process-tracing and inductive process tracing. Deductive process tracing relies on testing the validity of diffusion mechanisms in the adoption of the SRMP. Inductive process tracing relies on providing strong evidence to show how the combination of those
mechanisms produces the outcome. Finally, to conduct process tracing, I will explain the data gathering process from semi-structured interviews with key policymakers participating in the adoption process of the project in Turkey.
2.2 From Quantitative to Qualitative
Policy diffusion literature heavily relies on large-N statistical analysis across cases with various operationalization of diffusion mechanisms. Scholars draw hypotheses from diffusion mechanisms, theories that explain casual processes of policy
diffusion, and individually test them to indicate whether there is a channel for diffusion or not (Starke, 2013: 566). Weyland clarifies the very nature of these quantitative studies by emphasizing an S-shaped pattern that takes place in temporal and spatial dimensions (2005: 265). In that sense, in the quantitative term, diffusion is indicated by exploring the frequency of policy adoptions throughout time in different political settings, called the temporal dimension (Weyland, 2005: 265). Policy diffusion is identified with the rise of policy adoption in a specific region and then spread to the other regions in the world (Weyland, 2005: 265). That S-shaped pattern-finding diffusion is identified by testing hypothesis derived from the individual diffusion mechanisms, called conditionality, learning, emulation, and competition. Thus, pattern-finding studies find individual effects of those mechanisms on policy adoptions across cases/governments or states. However,
11
quantitative studies might have limitations to indicate the causal process of these mechanisms. Qualitative studies might fill these causal gaps.
Graham, Shiphan and Volden argue that quantitative studies are limited to answer the "what" question, which identifies the existence of diffusion in policy adoptions, rather than the "why" question, which explains the causal process behind the policy adoptions (2012: 695). They further point out that diffusion studies generally ignore the roles of policymakers and actors involved in the process of policy diffusion, which constitutes the internal limits of these studies. The goals, preferences, and objectives of those actors are important to understand how policies are formulated and prioritized and, to explain why the governments adopt them (Graham, Shiphan & Volden, 2012: 685). Thus, to understand and explain the process of diffusion,
qualitative studies constitute a complementary way of study in addition to statistical and pattern finding ones. Starke argues that qualitative approaches, methodologically speaking, are more suitable to explore and uncover the causal mechanisms in terms of empirical process rather than the identification of the effects of specific causes such as diffusion mechanisms (2013: 566). In that sense, interactions of
policymakers among political units are needed to be analyzed in detail via qualitative studies (Graham, Shiphan and Volden, 2012: 695). Thus, in-debt case studies based on interview and fieldwork data are needed to fill the gap left by statistical studies and to uncover the causal processes of diffusion mechanisms (Marsh and Sharman, 2009: 273).
In light of these methodological arguments, this study aims to uncover the causal process of diffusion mechanisms via the case study of Turkey. It is a case of policy diffusion, due to its relations with the WB in the adoption process of the SRMP in 2001. The design of this research is constructed by a single case study with a
12
process-tracing methodology. Semi-structured and purposive interview data gathered from key policymakers of the SRMP both in Turkey and the WB will be analyzed via process-tracing. Hence, the purpose of the case study is to explore the negotiation process of the SRMP between the WB and Turkish policymakers in light of diffusion mechanisms. The methodological preferences will be identified and justified in the following sections accordingly.
2.3 Case Study
Case studies hold an important position within the qualitative tradition of social sciences due to their contribution with detailed explanations and investigations regarding a social phenomenon. Case studies are detailed investigations of causes, mechanisms, and processes of specific outcomes (Hancke, 2009: 61). However, cases need to be relevant to the other cases in terms of either their outcomes that they produce or the theories that explain the outcomes. " A case study, for present
purposes, is an intensive study of a single case or a small number of cases that promises to shed light on a larger population of cases" (Gerring & Cojocaru, 2016: 394). The connection of the case with other cases within the universe is an important factor for case selector to investigate. In the qualitative tradition, one way of case selection relies on the outcome that the researchers are interested in to explain (Mahoney & Goertz, 2006: 239). The reason for this is that explanation of the outcome will serve for the broader theory or arguments in terms of its relevance (Hancke, 2009: 62). It should have a different value from other cases either because of the outcome, otherwise known as the theory (Hancke, 2009: 62).
Turkey is a case of policy diffusion due to the adoption of the SRMP, a policy from the WB, which indicates diffusion or policy transfer. It is important to note that the
13
SRMP is a case of the WB’s social policy and also a case of Turkish social policy as well. Thus, this project relies on a case study of diffusion mechanisms in social policy in a middle-income country, Turkey. However, the reason for my case selection does not stem from the outcome/policy adoption but the explanation of the outcome. I argue that the causal process of policy adoption binds together the
intersection of various diffusion mechanisms. Each mechanism represents a different explanation for policy adoption and scholars generally rely on one or two
mechanisms to explain policy diffusion in quantitative studies. Though diffusion mechanisms are often evaluated individually, which makes them compete to explain policy adoptions, they are generally interrelated (Graham, Shiphan & Volden, 2012: 695). They might lead to policy adoption only when they are at stake together in a specific context (Marsh & Sharman, 2009: 273). I aim to explore the process of policy adoption from the WB via the combination of different diffusion mechanisms rather than individually separating them. Thus, I argue that the WB as an external actor influences Turkish policymakers in the adoption of the SRMP in addition to the internal dynamics and context of Turkey. The combination of conditionality,
learning, and emulation constitute my argument in; how they operate together in producing the adoption of the SRMP.
To show how Turkey adopts the SRMP from the WB, this study adopts a causal case study approach. A causal case study generally investigates whether a change in X leads to a change in Y, however, X might also refer to the different factors that lead causation for Y (Gerring & Cojocaru, 2016: 397). In that sense, a causal case study identifies the causes of the outcome which are relevant to the theory and provides explanatory arguments (Mahoney & Goertz, 2006: 230). Diffusion mechanisms are explanatory tools and provide theoretical backgrounds to explain policy adoptions
14
across political settings. This study investigates the roles of conditionality, learning, and emulation together whether they produce the outcome (in this case, the policy adoption). To put it simply, these mechanisms are causal explanations for the
adoption of the SRMP in Turkey and, their combination constitutes the arguments for the exploration of this process. For this purpose, the policy adoption process will be analyzed through causal process observation. It aims to gather pieces of information in within-case analysis to evaluate whether the causal factors are consistent with the theory explaining the outcome (Goertz & Mahoney, 2012: 90). This approach is strongly related to the process-tracing methodology which this study relies on to uncover the causal process of the adoption of the SRMP. The following section identifies the relevance of within-case process-tracing to this study.
2.4 Process-Tracing
Process-tracing stands as an important method for qualitative case studies to identify causal relations between factors, causes, and outcomes in specific contexts. Process-tracing, initially a method to analyze the decision-making processes, aims to identify causal mechanisms, and, in doing so, overcome the weakness of statistical analysis (Trampusch & Palier, 2016: 441). It is considered as an efficient solution for the mechanismic problem of causation presented by purely quantitative studies (Trampusch & Palier, 2016: 441). The motivation behind process-tracing is to understand the causal links between factors and outcomes. Thus, it aims to open the black box of causal chains/mechanisms producing a specific outcome (Trampusch & Palier, 2016: 442). In that sense, by opening the black box, process-tracing turns implicit theoretical mechanisms explicit through a detailed examination of the relevant cases (Hancke, 2009: 68). To open the black-box or indicate theories explicitly, process-tracing requires an investigation by combining the factors to
15
identify specific processes considered as necessary to produce the outcome (Goertz & Mahoney, 2012: 108).
I try to identify the black boxes of conditionality, learning, and emulation in which their combination leads to the adoption of the SRMP in Turkey. Even though they individually constitute factors to produce policy adoption, I argue that none of them is individually sufficient to identify the causal process of policy adoption. However, they are necessary for the production of the outcome, thus their combination makes sense. In that sense, Starke recommends that identification of smoking-gun evidence of these mechanisms/factors is a proper way of investigation through process tracing in policy diffusion studies within the qualitative tradition (2013: 572). Smoking gun tests provide strong evidence for sufficient conditions or a combination of conditions for a theory, which makes a theory valid in a case study (Starke, 2013: 572; Goertz & Mahoney, 2012: 95). Furthermore, within process-tracing, hoop tests are
recommended to identify the necessary conditions that arguably produce the outcome when they engage each other (Goertz & Mahoney, 2012: 92). In sum, I will use process-tracing to conduct smoking gun and hoop tests and identify necessary and sufficient factors within diffusion mechanisms for the adoption of the SRMP. However, this approach to study policy diffusion arguably contains both inductive and deductive incidents of process tracing. Thus, I will test the theories
(conditionality, learning and emulation) using deductive process tracing and determine their validity in the Turkish case. In this way, I aim to contribute to the literature based on inductive process-tracing that proves or disproves my hypothesis that the combination of the mechanisms leads to policy adoption in a specific context.
16
Deductive process-tracing relies on the comparison of the evidence gathered from the case with the theoretical expectations, which means testing the validity of the
theories/diffusion mechanisms. Inductive process-tracing, on the other hand, relies on the detailed observation of a case by uncovering the causal mechanisms to support and contribute to an existing theory with new evidence (Trampusch & Palier, 2016: 445). In that sense, in light of official documents and interview data gathered from key policy players, I will look for evidence that verifies conditionality, learning, and emulation. I will then aim to contribute to the literature that evaluates diffusion mechanisms as complementary and interrelated (Graham, Shiphan & Volden, 2012: 695; Marsh & Sharman, 2009: 273).
2.5 Interview Data
An interview is one of the fundamental ways of data collection within the qualitative tradition, especially within case studies. "Interviews are important, and often an essential tool for making sense of political phenomenon" (Mosley, 2013: 2). Interviews provide a way for researchers to uncover the causal mechanisms regarding the beliefs, motivations, and behaviors of the actors (Mosley, 2013: p2). Thus, they enable researchers to test theories that provide meaning for certain actions or to construct theories explaining these actions (Mosley, 2013: 2). In other words, they help researchers to give meaning to the opinions and mindset of the people under investigation within a case study (Lynch, 2013: 34). Following these arguments, interview data in that study aims to uncover the mindset of key policymakers regarding diffusion mechanisms in the SRMP.
This study sets out to accomplish two goals: test the validity of the diffusion mechanisms for the adoption of a policy and propose a model combining
17
mechanisms to explain the outcome in a specific context. For this reason, semi-structured interviews in this study are conducted for testing the theories and proposing a model combining diffusion mechanisms to explain the outcome.
Interviewees are asked to identify: whether or not the loan that the WB provided for Turkey facilitated the adoption, a process of learning from other experiences
provided positive feedback, and whether or not the adoption of the policy would make Turkey gain more legitimacy and prestige in the international community. The answers to such issues provide smoking-gun evidence for the validity of theories, which corresponds to testing the theories. However, these interviews are essentially conducted with semi-structured and open-ended questions for interviewees so that their way of understanding the project and the motivation might also be covered. This is mainly to indicate whether the combination of mechanisms produces the outcome/ the adoption of the SRMP. This side refers to proposing a model combining diffusion mechanisms.
To test the validity of diffusion mechanisms and to propose a model underlining the combination of those mechanisms, the sampling of the interviewees stands as a crucial issue to obtain validity and reliability of the data for the whole population. While validity concerns with asking the right questions to the right people and properly asking them to obtain truthful answers; reliability concerns the accuracy of the data whether it changes or not if the same interviews are conducted in a different time and context (Mosley, 2013: 25). Thus, validity and reliability would be obtained with a proper sample for an interview to gather truthful and unchangeable data throughout the time. "Sampling involves selecting a subset of elements from the universe of population of all such elements" (Lynch, 2013: 38). The decision for sampling relies on the purpose of the study, i.e. whether the purpose is theory-testing
18
or theory building (Mosley, 2013, p. 19). To discuss the issue of sampling in this study, I take the key policymakers within the SRMP as the unit of analysis. The whole population of the sample refers to all actors within the project. Because I try to explore the adoption process of the SRMP by Turkey from the WB through diffusion mechanisms, sampling needs to be purposive following the research goals. Purposive sampling refers to the selection of the elements following their certain qualities which represent the characteristics of the whole population (Lynch, 2013: 41). In that sense, I identified three key policymakers in the SRMP from both Turkey and the WB who built the project and took part in the agenda-setting, decision making, policy-making, and negotiation processes. These key policymakers are: a former member of the cabinet, Turkey, a former consultant in the SRMP in the WB, and the former policy expert in the SRMP in Turkey. Their key roles in the formulation of the project, eliminate any possible issue regarding representativeness of the sample. These individuals, from many angles, were the most important actors in the adoption process.
Validity is obtained by asking relevant questions with the objectives of the study to the right people. Reliability is not an issue due to the position of the interviewees. They took part in a completed project and all of their occupational position has changed since, thus, they do not have a positive or negative interest to change the answers when the same interviews are conducted again because of political or occupational concerns.
Ultimately, non-random, purposive sampling helps to uncover causal explanations following theories (Mosley, 2013: 19). Furthermore, it might facilitate causal process observations to test the mechanisms or the arguments derived from the theories
19
(Lynch, 2013: 40). It is strongly consistent with the process-tracing methodology to open the black-boxes. Interviews in this study provide primary data for exploring the causal processes and the identification of the combination of the diffusion
mechanisms, which will explain the adoption of the SRMP.
2.6 Conclusion
In this chapter, I presented the methodological preferences of this study to show how diffusion mechanisms operated together in the adoption process of the SRMP. First, I discussed the distinction between quantitative and qualitative ways of studying policy diffusion, in that quantitative studies mostly relied on a large-N statistical analysis of diffusion mechanisms. The effects of these mechanisms on policy adoption are considered individually to identify and explain policy adoptions.
However, I indicated the views of scholars that statistical analysis was not enough to explore the causal process of those mechanisms in policy adoption. Qualitative studies, on the other hand, were argued to be more suitable to indicate the causal explanation with the investigation of the interactions of policymakers via in-depth analysis of within-case studies. Then, I clarified the logic of case studies to study policy diffusion either for the outcome or for the explanations of an outcome. I selected Turkey as a causal case study based on my argument that diffusion mechanisms were interrelated and their combination produces outcome/policy adoption. This was an explanation for an outcome. The outcome here was the SRMP and I argued that the Turkish case of policy diffusion provided strong evidence for the model combining diffusion mechanisms in a complementary way in policy adoption. To construct my argument in a within-case study, I emphasized the role of process tracing methodology in policy diffusion studies to explore and open the black-box. Through this methodological approach, the explanation and identification
20
of the factors or causes of a specific outcome would be identified with smoking-gun evidence. In that sense, I underlined that this study would conduct both deductive process-tracing and inductive process tracing. Deductive process tracing relied on testing the validity of diffusion mechanisms in the adoption of the SRMP. Inductive process tracing enabled to provide a model combining diffusion mechanisms
producing the outcome. Finally, to conduct process tracing, I indicated the way of data gathering from semi-structured interviews with three key policymakers participating in the adoption process of the project.
21
CHAPTER 3: THEORETICAL FRAMEWORK
3.1 Introduction
In this chapter, I will start with the explanation of diffusion that occurs when a decision made by a political unit is under the influence of a decision of another political unit. To put it simply, I will explain policy diffusion as a sort of external influence on decision making among political units. After presenting the definition of diffusion, I will explain diffusion mechanisms as explanatory tools for similar policy adoptions among political units. For the clarification of these mechanisms, I will first start with conditionality. Conditionality takes place in policy adoptions via
implementing a policy, following certain criteria set by the donor political unit in exchange for financial support from a country or an IO. This incidence of
conditionality corresponds to a vertical way of policy adoption. However, it also comes about via fulfilling the requirements to gain accession to a supranational organization such as the EU. Second, I will explain learning, the process by which adopting a policy requires observations of the others’ experience regarding a specific policy. It mainly aims to foresee whether the policy will produce the expected
outcome or not. Then, I will emphasize the logic of emulation, which operates with the motivation of appropriateness, prestige, or being modern in the international community without rational calculations regarding policy adoption. Finally, even though this study does not focus on this factor, I will explain the competition
22
mechanism for policy diffusion that economically competitor countries tend to adopt similar policies to gain more financial resources in the global economy such as tax reductions for foreign investments.
3.2 Policy Diffusion as an External Effect on Policymaking
The relations among the states, governments, and international organizations are studied widely by social scientists to identify in which ways these units affect each other on decision making as external players. Political units, in general terms, influence others’ policymaking in various ways, which is a complex phenomenon in a globalizing world. This is what the policy diffusion literature aims to uncover. In the broadest definition, policy diffusion is defined as a process in which a political unit’s decision is affected by other political units’ decisions (Magetti & Gilardi, 2015: 3). By this definition, policy diffusion might take place not only among the international community but also among domestic political entities in a given state. However, policy diffusion literature is more interested in inter-state relations. It is a matter of whether a government is affected by the decisions of other governments to adopt a policy or not (Graham, Shiphan & Volden, 2012: 675). This process gains importance to understand to what extent a government is open to external effects to adopt a policy or not. Gilardi clarifies this process with these words; "how
policymakers in one country are influenced by prior choices in other countries or other words, how policies diffuse from one country to another” (2010: 650). To put it simply, policy diffusion, in the international sense, stands as a process in which external political powers have a say on one government’s decisions and
23
These ways might differ based on the context that the actors share. Weyland
underlines the external pressure as a vertical way to cause a rapid adoption of similar policies in different political entities (2005: 269). Weyland underlines the
observation that powerful external organizations such as IMF, the EU, or the WB might impose reforms by using the carrot and stick to make a government cooperate with them. Thus, policy diffusion occurs as an external effect on a given
government’s decision making through pressure. However, pressure does not have to be a force by an external political entity to influence others’ policymaking.
Innovations spreading from one government to another might also create pressure to reform (Shiphan & Volden, 2008: 842), but it is not as the pressure of IOs’ in a vertical way. This way of diffusion is similar to the observation of a government by another government based on a specific policy as an external effect on policy adoption in a horizontal way (Meseguer, 2005: 72). Successful implementation of a policy by a government following its policy objectives turns out to be a policy innovation and it spreads to other governments through observation and learning among political units.
Policy diffusion as an external effect on policymaking might take place with different processes (or mechanisms) to explain and understand the ways that they occur. The literature highlights four different mechanisms that may explain why diffusion occurs: learning, competition, conditionality, and emulation. These are theoretical frameworks to understand why a government is under the influence of external actors to adopt a policy and in what ways this influence takes place in a process that leads to policy diffusion.
24 3.3.1 Conditionality
Conditionality corresponds to a process in which mutual agreements among parties constitute a vertical imposition of policies and, the relations of the International Organizations (IOs) with partner countries enable conditionality to operate
(Weyland, 2005: 269). In that sense, by means of financial assistance or sanctions, if the agreements are not fulfilled by the partners, conditionality performs as an
insurance for the agreements to be fulfilled (Weyland, 2005: 269). Given the states are sovereign political entities and driven by national interests in the international system, the imposition of a policy by an IO over a state is not a zero-sum game for the political unit imposed. Rather, the logic behind conditionality relies on the financial help, benefits, and assistance from IOs such as IMF and the WB to the states that need to implement certain policies to gain economic resources in return (Gilardi, 2012: 461).
The vertical spectrum of coercion, industrialized countries- IOs -developing
countries, is framed as a top-down mechanism that IOs introduce in a coercive way (Meseguer, 2005: 72). This is because, the adoption of a policy as a result of a condition to provide financial aid might lead to involuntary regulations by the states with their various agendas (Meseguer, 2005: 72). The coercive nature of
conditionality is due to asymmetrical power relations among the parties, especially in an economic sense. Thus, the restriction of the states’ autonomy in return for
financial help make conditionality an anti-liberal mechanism that shapes the policies of countries that strongly rely on foreign direct investments, aid, grants, or loans (Dobbin, Simmons & Garrett, 2007: 454).
25
Apart from IOs’ imposition for reforms, policies, and projects consistent with their economic and normative power in the global economy, conditionality has an
important role when it comes to relations with a supranational institution such as the European Union (the EU). Policy adoptions become a necessity to gain accession for a supranational organization and countries are required to fulfill certain conditions by the organization (Gilardi, 2012: 461). The EU forms a broad framework for the governance of political, social, and economic affairs and asks for a wide range of regulations from human rights to tariffs for both the member countries and the candidate ones. In that sense, the EU as a supranational organization and influential actor in these countries constitutes another part of conditionality that seeks for broader compliance with the adoption of different kinds of reforms in return for economic benefits.
3.3.2 Learning
Learning is a complex social phenomenon that relies on human interactions and communications in different settings, which enables people to accumulate knowledge and ideational sharing. However, as a policy diffusion mechanism, learning is a social process that causes policy adoptions among political units through policy networks and observation of the implemented policies. Dunlop and Radaelli explain the learning process by emphasizing the accumulation of policy-related evidence within public organizations, changes of mindsets within policy networks and advocacy coalitions among decision-makers and, the extraction of policy results of other governments to re-formulate the policies (2017: 307). In that sense, learning differs from conditionality due to the emphasis on the efforts of policymakers to seek policy solutions through interaction rather than imposition. Policy learning is an outcome of the socialization of individuals in a political setting and is investigated
26
via tracing the micro-to-micro relations which aim to understand how policymakers behave when they interact in a group (Dunlop and Radaelli, 2017: 307).
To be more concrete, learning is a horizontal mechanism among policymakers that enables them to deal with the difficulties of policy decisions through observing the prior adoptions of the related policies and their outcomes (Mesequer, 2005: 72). When policies are adopted through such kind of investigation to reach a "good policy" and there is an expectation to eventually match the observed policy objectives and the outcomes, it is no doubt that policy learning is at stake. Gilardi underlines the expectation of policy adopters with their efforts to estimate the possible consequences of policy adoption in light of other adopters’ experience (2012: 463). This effort requires a voluntary action for policy adoption because it indicates that policymakers are aware of what they need and are motivated to overcome the deficiencies or incapability of the current policy by observing others. Learning mechanism in policy diffusion, in that sense, brings about an ideational change and policy adoption through cause-and-effect calculation on the relationship between observed practices of other governments and possible outcomes in a different political unit (Dobbin, Simmons & Garrett, 2007: 460).
Rational learning, in the context of cause-effect calculation, is referred to by scholars as highlighting the assumptions that consider policymakers self-interested,
autonomous, and goal-oriented for policy adoptions via observation. It assumes that policymakers, irrespective of their belief and ideology, are capable of seeking
solutions for a problem (Weyland, 2005: 271). They are assumed to conduct cost and benefit analysis on whether or not a new policy adoption provides expected
outcomes in comparison with the current policy (Weyland, 2005: 271). Thus, following this assumption, learning is considered as purposive with the logic of
27
rationality, which ensures the detection of the problems and the solutions being sought for a new policy adopted in light of its possible outcomes (Meseguer, 2005, p. 74). Although that theoretical approach, known also as Bayesian rationality, relies on policymakers’ logic, the effect of their prior beliefs, experience, and knowledge on the policy is also taken into account for the accumulation of knowledge process from the other adopters. In that sense, policymakers would re-consider their previous knowledge as a result of new information they gathered via observation and
interaction in the learning process (Gilardi, 2012: 464). Policy decision making through the learning process, in Bayesian logic, occurs as a result of this
reconsideration expected to reinforce the current beliefs to adopt a policy or refuse it by ending with high costs (Dobbin, Simmons & Garrett, 2007: 460).
The learning process of policy diffusion not only consists of Bayesian rationality for policy adoptions but also bounded rationality. It relies on the idea that policymakers’ analytical skills are limited due to their cognitive biases (Meseguer, 2005: 74). Dobbin, Simmons, and Garrett explain that policymakers tend to pay more attention to the successful cases or outcomes of a specific policy rather than giving the effort to assess all available cases which Bayesian logic heavily insists on (2007: 461). This means that policymakers are selective to observe other implementations of the policies and assess their outcomes. This might be delusive to adopt the "good policy" because it creates prejudices and leads to ignoring the possible risks. Gilardi frames this bounded rationality which might lead to inappropriate policy adoption, in cognitive psychology terms as cognitive shortcuts such as "Availability" and
"Representativeness" (2012: 465). Availability refers to the selectivity of observation of a policy in compliance with successful cases as I have just mentioned regarding Dobbin, Simmons, and Garrett. Representativeness simply indicates the motivation
28
to infer from limited evidence of a policy and generalizes it for all outcomes (Gilardi, 2012: 465). Inference from the initial success of the outcome is considered by the policymakers in that the initial success represents the whole success of the policy.
3.3.3 Emulation
Adopting a policy is not necessarily driven from the motivation for seeking solutions for the deficiency of a current policy, or driven from the imposition of a powerful actor in the international community. Emulation as a policy diffusion mechanism has a different kind of rationale from two other mechanisms aforementioned. It relies on the motivation of social appropriateness to adopt a policy rather than cost and benefit analysis or obligation for an agreement made with the partners of a government (Gilardi, 2012: 467). Before the clarification of social appropriateness, it is important to notice that scholars who explain policy diffusion with emulation question how policies are adopted socially (Dobbin, Simmons & Garrett, 2007: 452). Weyland, with a normative framework approach, clearly states that "decision-makers attempt to gain international legitimacy by importing advanced innovations and thus demonstrating the emulating countries’ modernity and compliance with the
international norms" (2005: 270). In short, policy adoption occurs in a way that a government engages in social interaction with another government or an IO perceived prestigious in the international community. Then, it imitates the norms, values, and policies to be accepted as an appropriate member.
The logic of emulation for policy adoption, a certain way of policy diffusion, relies on the need for legitimacy of a political unit for its existence, actions, and
interactions in the international community. Adoption of a policy is, theoretically, independent from the solutions of a problem. Weyland frames this logic as "properly
29
adopting innovations regardless of functional needs" (2005: 270). Functional needs here refer to the logic of learning insisting on the relations between problems and solutions in policymaking via social interaction and observation among
administrative elites. Thus, a policy adopted with emulation mechanism “has no causal paths lending from policies to outcomes" (Meseguer, 2005: 73). It means that the reasons for adopting a policy are independent of its consequences, effects, and outcomes. Because emulation is not a goal-oriented diffusion process, attractiveness in the international community is much more important than problem-solving (Weyland, 2005: 279).
The process of imitation is not simply imitating a modern government. IOs such as the IMF and the WB also shape the minds of policymakers in partner countries through policy experts and specialists. These experts constitute and rationalize the arguments of policies and announce that the compliance of these policies is appropriate behavior (Dobbin, Simmons & Garrett, 2007: 452). Thus, the logic of appropriateness is not only the motivation of the governments adopting policies but also policy innovators who aim to facilitate the spread of norms via emulation in their bilateral agreements with partners. To make it more concrete, IOs’ provisions might be perceived as an innovation that turns into an attractive symbol of modernity or a normatively appropriate model. It might lead many countries to improve their legitimate position via the imports of the new policy agendas. In that sense,
emulation works as a process that IOs’ construct legitimate, appropriate, modern, and contemporary norms in social interactions with the partners whose motivation is to obtain approval in the international community.
30
Even though this study does not aim to scrutinize competition, it is worth
clarification as a mechanism widely accepted in the literature. The last mechanism of policy diffusion differs from the previous mechanisms with a strong reliance on the economic interests of policy adopters and the race among them to gain more
financial resources from the global economy. Gilardi defines competition as “the process whereby policymakers anticipate or react to the behavior of other countries to attract or retain economic resources" (2012: 462). This motivation operates the behaviors of policymakers in several economic regulations such as tax competition or economic globalization. Tax competition would drive countries to adopt
regulations with lower taxes so that they attract more investments. Policymakers might adopt a set of standards that reduce the benefits of the social protection system on behalf of deregulation and marketization expected to bring additional resources. However, this might cause the race to the bottom despite the expectation to be wealthier in a national sense (Gilardi, 2012: 462).
Although countries adopt liberalizing reforms, the adoptions are not necessarily caused by policy diffusion, competition. Countries might adopt similar policies for different reasons. Because policy diffusion is subject to the interdependence of at least two political settings, the adoption of a policy due to competition must be identified. Diffusion literature identifies competition via investigating that whether or not a government tends to adopt an economic policy when its competitors do the same (Simmons & Elkins, 2004: 182). For clarification, diffusion literature
investigates similar countries in terms of economic structure and, to determine if they compete with each other to gain more resources and attract FDI (Gilardi, 2012: 462). This scrutiny might be traced via the indicators of the economic openness of
31
barriers on imports and exports are also subject to the identification of competition. Theoretically speaking, competition as a mechanism assumes that "governments know who their competitors are and can connect policy choices to competitive advantages" (Dobbin, Simmons & Garrett, 2007: 458). Thus the actions of competitors are interdependent. In product competition, if country A reduces the tariffs to export to the market of country C to gain more resources than the country B, B will react in the same direction by reducing the trade barriers for export to access the market of C as well as A (Dobbin, Simmons & Garrett, 2007: 458). The entrance to a market is not only a matter of motivation on competition among similar countries but also the specific conditions of the market influence the decisions of policymakers of those countries. This process is called the "California Effect", which refers to the strict regulations regarding carbon emission for
automobile producers after California inserted conditions in the 1970s and 1990s (Gilardi, 2012: 463). The California Effect demonstrates for competing countries that how export markets are able to formulate conditions for access, and thus force
countries to meet certain environmental standards, human rights, or certain working conditions (Gilardi, 2012: 463). While these countries compete with each other via tightening their regulations and sometimes lowering their standards for labor, putting conditions for the improvement of the standards indicates a dynamic process of competition. The California Effect is significant in the sense that, through
competition, not only economic policies diffuse but also different kinds of norms and regulations might be at stake, which indicates the complex nature of policy diffusion. Because this study investigates policy diffusion from the WB to Turkey, competition is irrelevant to this case study. The WB and Turkey do not compete for more
32
rather than economic policy. Also, competition generally occurs among the states and, this study analyzes the relations between an IO and a nation-state. Despite the irrelevance to this study, competition is worth explaining as one of the theoretical diffusion mechanisms. However, the literature review of this study does not include competition. The main purpose of this study is to indicate that diffusion mechanisms might work in a complementary way to cause a single outcome/policy adoption. Due to the irrelevance of competition in the WB-Turkey relations but the very presence of the other mechanisms, the literature review covers only conditionality, learning, and emulation to indicate that they might be complementary. Still, in some cases,
competition might co-exist with one or more mechanisms depending on the contexts.
3.4 Conclusion
In this chapter, I started with the explanation of diffusion that occurs when a decision made by a political unit was under the influence of a decision of another political unit. To put it simply, I clarified policy diffusion as a sort of external influence on decision-making among political units be it among countries or between countries and IOs. After defining diffusion, I explained diffusion mechanisms as explanatory tools for similar policy adoptions among political units. For the clarification of those mechanisms, I first started with conditionality, which took place in policy adoptions via implementing a policy and, following certain criteria set by the donor political unit, in exchange for financial support from a country or an IO. This could
correspond to a vertical way of policy adoption or coercion. However, it included via fulfilling the requirements to gain accession to a supranational organization such as the EU. Second, I clarified learning as adopting a policy based on observations of others’ experiences regarding a specific policy. It mainly aimed to foresee whether the policy would produce the expected outcome or not. Then, I emphasized the logic
33
of emulation, which operated with the motivation of appropriateness, prestige, or being modern in the international community without rational calculations regarding policy adoption. Finally, even though this study did not cover it, I clarified the competition mechanism for policy diffusion, in which economically competitor countries tend to adopt similar policies to gain more financial resources in the global economy such as tax reductions for foreign investments.
34
CHAPTER 4: THE EMPIRICAL LITERATURE
4.1 Introduction
In this chapter, I will critically analyze the empirical literature on policy diffusion. This study aims to indicate that multiple diffusion mechanisms might operate together to cause policy adoption in a specific context. Thus, I will approach critically the diffusion literature which explains policy adoptions mainly with one mechanism or arguments that one diffusion mechanism is more important than others to produce an outcome/policy adoption. Scholars emphasize conditionality, learning, and emulation independently for different contexts. However, they miss the possible intersection of them that might indicate a more comprehensive explanation for policy adoption. This study aims to underline the intersection of conditionality, learning, and emulation. Thus the literature on competition will not be covered consciously but it does not mean that competition cannot intersect with other mechanisms in specific contexts. I will discuss the studies that analyze policy adoptions through these mechanisms that had individual effects on the causal chain of policy adoptions. Then, I will re-evaluate them for the possible intersection of diffusion mechanisms including competition, conditionality, learning, and emulation.
In terms of conditionality, I will argue that, depending on the political contexts, policy adoptions of Sri Lanka, Papua New Guinea and African countries cooperating with the IMF and the WB might be also explained with a combination of learning
35
and emulation. For learning, the epistemic communities and expertise of those organizations might be influential for policy adoption. For emulation, the prestigious position of those organizations in the international community might matter for policy adoption. Furthermore, when it comes to the EU, policy adoptions of the candidate, member, or trade partner countries might be explained with a combination of learning and emulation in addition to conditionality due to the technical expertise of policymaking and normative stance of the EU.
In terms of learning, scholars generally underline the importance of geographical proximity, wherein learning typically comes from a neighbor, and political similarity, wherein learning comes from ideologically similar political actors. However, I will discuss that learning might also be compatible with emulation and competition. A prestigious success in a political unit might lead to policy adoption by another unit through emulation, be it a neighbor, an ideologically similar actor, or neither. Furthermore, countries within the same region might be competitors as well and ideologically similar countries might also compete with each other due to their resemblance to the political and economic structure.
For the last mechanism, emulation, I will discuss how it could co-exist with learning due to the bounded rationality, which operates apart from geographical proximity and ideological similarity. It might be observed with the tendency to follow the leader as indicated in the diffusion of public downsizing policies within the OECD countries following the U.S model. In terms of being prestigious, it might be
observed in the adoption of UNCRC among African countries following Ghana due to its reputation stemming from the first independence movement in the region.
