Labour Protection in the Globalized Workplace
The interrelation between national and international political
initiatives and developments: reinforcing or replacing?
Max Joosten, University of Amsterdam, Amsterdam, The Netherlands
Correspondence: max.joosten@student.uva.nl Master thesis
Research Master Social Sciences
Student ID: 10445684
1st reader(s): Brian Burgoon and Luc Fransen 2nd reader: Sijeong Lim
31 July 2018, Amsterdam Word count: 9293
Abstract
Scholars have long disagreed about whether political initiatives and developments at the international level operate in harmony or tension with initiatives at the national level. Given the increase in policy interventions, primarily at the international level, in labour governance, this paper explores the interrelation between different instruments of labour protection within political economies. Relying on an original dataset, this paper provides insights relevant to both the study of comparative political economy and of global labour governance. It does so by studying national polities that harbour the industries producing labour-intensive
manufacturing goods. The results show that national labour protection is a statistically and substantively significant predictor of international labour protection, and vice versa, while few other conditions are significant factors in shaping national or international protection. Furthermore, statistical analysis suggests that labour protections in both domains are part of the same underlying political and economic processes.
Key words: labour market institutions, institutional complementarity, developing countries, globalization
JEL classification: J80
1. Introduction
On 4 July 2016, a broad coalition of partners signed the Dutch Agreement on Sustainable Garments and Textile. Under the auspices of Dutch Minister for Foreign Trade and Development Cooperation Lilianne Ploumen, the coalition included industry associations, trade unions and non-governmental organizations (NGOs). The agreement aimed to promote safety and equality for workers in producing countries by requiring firms to make public a list of their production sites, as transparency was believed to be a critical factor in making
improvements (IRBC Agreements, 2018). One-and-a-half years later, in December 2017, the agreement has only received six reports of violations in the affiliated factories. Remarkably low, Tara Scally from Clean Clothes Campaign says, indicating that in reality much more violations occur. Certainly in comparison to the thousands of problems found and resolved by the Accord on Fire and Building Safety in Bangladesh (initiated after the Rana Plaza factory collapse in 2013), whilst this only concentrates on one country and one standard. Scally therefore suggests that workers might lack knowledge of the fact that they can make a complaint at all (One World, 2018).
This example illustrates two related areas of interest in the study of labour governance. The first considers substantial improvement for the lives of workers and is involved with studying the consequences of such new initiatives. Although receiving reports of violations might not be the main objective of the above-mentioned agreement (rather changing the attitudes of firms), it indicates that after one-and-a-half years it has not shown promising potential. The second is related to the notion that workers might not know where to go or who to turn to if their rights are violated. Such confusion is imaginable considering that in the wake of globalization labour governance has become increasingly complex: new actors and governance instruments in labour protection emerge (Hassel, 2008). This begs the question of whether and how these actors and instruments interact with each other.
The consequences of – and interplay between – old and new policy instruments is a topic of keen interest to academics. Particularly scholars of comparative political economy (CPE) aim to understand the interactions across different parts and types of policy by analysing the dynamics of inter-institutional links that are key to the formation of national political economic arrangements. A widely debated concept is institutional complementarity: whether different types of policy interventions strengthen or undermine each other (see for example Crouch et al., 2005). A growing body of empirical literature examines this concept covering different substantive (for climate governance see Roger et al., 2017; for labour governance see Burgoon and Fransen, 2017), and geographical areas (for Vietnam see Locke, 2013; for OECD see Kinderman and Lutter, 2018), using different methodological approaches (for case study see Kinderman, 2011; for cross-national study see Burgoon and Fransen 2018). Despite this growing body of empirical literature, the results are ambiguous. Some research finds support for replacing dynamics, for example where corporate social
responsibility (CSR) fills institutional voids left behind by deregulation. Others find support for mutually reinforcing dynamics, where CSR flourishes in countries with already-generous social policies. Still others find something in between.
Institutional dynamics are not only examined in the context of the public-private dichotomy. In the study of ‘social Europe’, scholars have long disagreed about whether political initiatives and developments at the international level operate in harmony or tension with initiatives at the national level. Mau (2005) and Sánchez-Cuenca (2000), for example, suggest undermining dynamics, where strong national social standards decrease support for European integration. On the other hand, Atkinson and Davoudi (2000) suggest mutually reinforcing dynamics flowing from the EU level to the national agenda. Geyer (2000), as a third alternative, argues that EU social policy is too modest to have an effect on national
welfare regimes. This paper chooses to further explore a range of policy interventions that, most crudely, can be positioned under the umbrella of either national or international protection, thereby taking into account various kinds of political protection. We have little leverage to understand whether and how the politics of national and supranational social policy are intertwined (Burgoon, 2009), especially not in a non-European Union (EU) context. I conduct a cross-country analysis of 50 political economies that depend on the export of labour-intensive manufacturing products. Needless to say, studying labour in these countries is an important and necessary exercise. Not only because of terrible incidents that catch the attention of the media every now and then (such as the Rana Plaza incident in 2013), but also because of the much less spectacular tragedies – long hours, low pay, hazardous conditions – that happen in many workplaces around the world. Advancement of the literature could provide essential insights for the improvement of labour conditions for policy-makers, NGOs, firms and so forth. To depict whether and how labour policy efforts across various levels of political-economic governance interact with each other, I address the following research question:
“Are national labour protection and international labour protection in tension or harmony with each other?”
With this research question I directly engage in the above-outlined theoretical and empirical debates in CPE and global labour governance. Labour protection across national and international instruments are argued to be in harmony when they mutually reinforce each other, i.e. where national labour protection flourishes in contexts of already high international labour protection and vice versa. Contrastingly, these domains would be in tension when increased efforts in one domain undermine developments in the other. Finding no systematic pattern would lead to the conclusion that the domains of labour protection have little or no influence on each other.
In examining this research question, I rely on original country-level variables that serve as proxies for international labour protection and compare them to existing national labour protection variables from various datasets. First fitting the data into ordinary least squares (OLS), the paper suggests that national labour protection is a statistically and substantively significant predictor of international labour protection, and vice versa. I find empirical support for the hypothesis that social policy development on one level of political-economic governance strengthens development on the other. Minor differences in estimated
effects suggest that national protection comes before international protection in such recursive harmony. Subsequent seemingly unrelated regressions (SUR) models, however, suggest that national and international labour protection are not fostered by distinct political economic processes, indicating that there is something underlying labour protection that causes developments in both domains. Results show that this leads to a group of countries that maximizes pro-labour policies in both domains on the one hand, and a group of countries that puts less efforts in both the domains on the other hand. Furthermore, the models indicate that there are few other conditions that are significantly shaping national or international
protection.
To systematically come to my conclusion, the paper is organized as follows. First the paper is positioned in CPE and global labour governance literature and hypotheses will be formulated. This is followed by the data and methods section that discusses the variables and the estimation approach. Then the empirical analysis and concluded with a discussion of the results.
2. Theoretical framework: Interaction between labour governance instruments
Traditionally, labour governance was characterized by hard-law, centralization and
compliance by government. The protection of labour and social standards was taken care of by tax code, social services and transfers, and employment regulations. This system, however, came under pressure in the wake of an increasingly globalized economy wherein production chains internationalized. Today, as a result, traditional labour protection instruments are supplemented by buyer-driven and trade-related policy initiatives that should be better capable of dealing with internationalized labour relations through incorporating strategies such as soft-law, decentralization and/or compliance by business (Hassel, 2008). Buyer-driven regulation, on the one hand, focuses on how buyers can improve social and labour standards by adopting codes of conduct, joining social certification schemes, participating in social audit processes, financing safety improvements and making purchasing commitments (Koenig-Archibugi, 2017, p. 1). Ethical consumer movements often support such efforts (Burgoon and Fransen, 2017, p. 42). Trade-related interventions, on the other hand, such as social clauses in preferential trade agreements (Raess and Sari, 2017) or bilateral investment treaties (Brown, 2016) are steadily gaining momentum as well. Lim et al. (2015) find that when in developing countries foreign aid levels are low, bilateral trade-based pressures are associated with improved labour rights (the effect loses significance as aid levels rise). Greenhill et al. (2009) find that labour laws and actual labour practices in developing
countries are influenced by the labour standards of their trading partners. Suggesting that using international labour protection instruments, which are often a requirement for trade, would directly affect national labour protection. Other trade-related research shows that during economic downturns in importing countries, through cost-cutting efforts, workplace safety deteriorates in exporting countries (Lim and Prakash, 2017).
What are the consequences of these increases in labour protection instruments for workers? This paper is first and foremost interested in studying the interrelation among different instruments of labour protection. Studying the interrelation among different institutions in CPE often occurs through an analysis of institutional complementarities. Basically, institutional complementarities suggest that “a particular type of coordination in one sphere of the economy that should tend to develop complementary practices in other spheres as well” (Hall and Sockice, 2001, pp. 17-20). This is because policies are embedded in institutional settings that are interlinked to each other and therefore developments in the one will lead to developments in the same direction in the other – say for example
liberalisation in any institution is going to lead to consequences that will drive the system towards more liberalisation in another institution (Crouch et al., 2005). The following section will introduce the hypotheses that are based on this notion of institutional complementarities.
Hypotheses
Pearson and Seyfang argue that the development of voluntary codes of conducts reflects the political and economic forces behind the demise of strong institutional statutory standards (2001, pp. 52-53). In their argument, private initiatives are driving the dissolution of public standards. In line with suggesting that different policy instruments replace each other – however indicating that private initiatives are a reaction to (not a driver of) the demise of public standards – Bernstein and Cashore argue that the global social policy void left behind by state regulation has been filled by “corporate social responsibility, industry self-regulation, political consumerism, voluntary instruments, and public-private partnerships” (2007, p. 348). These suggested dynamics are supported by empirical evidence provided by for example Jackson and Apostolakou (2010), who find that CSR is higher in more liberal market economies (to be found in Anglo-Saxon countries) than in more social coordinated market economies (to be found in Continental Europe) and by Lim and Prakash (2017), who find empirical support for the hypothesis that private regulatory activities positively improve workplace standards where public regulation of such standards is weakest.
Such dynamics found in the public-private divide can be tested under a crowding-out
hypothesis. In the context of EU integration (and therefor closer to my analysis of different
levels of political-economic governance), Mau (2005) finds that citizens in democratic welfare states are less supportive of EU sharing decision-making over health and social welfare. Sánchez-Cuenca (2000) finds that countries with more generous welfare provisions tend to be more sceptical about European integration, including social policy integration. Citizens, Sánchez-Cuenca argues, make a decision based on their view towards the different institutions: if they conceive supranational institutions to function well and national
institutions not, their support for integration is greater. Furthermore, in his analysis of the rise and stall of labour linkage in globalization politics, Burgoon (2004) argues that, through political learning experiences, interest groups interact over trade-linkage politics. He, for example, finds that unions in Europe consider that “scarce political resources are better spent on welfare compensation than labour linkage or other trade-protectionism” (Burgoon, 2004, p. 198). This portraits a dynamic, as is the case in Sánchez-Cuenca’s piece, wherein political actors have to choose the most effective policy instrument (in their specific context) given their scarce resources, indicating that choosing for one means that they cannot choose for the other. These trade-offs should then lead to clustering of different types labour regimes in national polities. Anner (2015) portrays such clustering, depicting a world wherein workers, dependent on the labour control regime that they are in, choose to push for different national or international strategies to maximize pro-labour policies in a specific domain. Effects can be indirect: the existence of significant provision of protection at one level can diminish or moderate negatively the effect of such insecurity on support at another level. If this is accurate, I should find support for the following hypothesis:
H1: Social policy development at the international level undermines policy efforts at the national level and/or vice versa.
In contrast to the first hypothesis, we can also expect that institutional efforts to maximize pro-labour policies on the national and international level mutually reinforce each other. In this case, we expect the political economies that are socially embedded on the national level to be the most ardent supporters of international instruments (and vice versa). This is what Middtun et al. (2006, p. 370), in terms of welfare state policies and CSR, label the systematic embdeddedness hypothesis where “social policy efforts are mirrored over different institutional settings”. Building on this, Gjølberg (2009, p. 611) finds that the main
institutional mechanism for better CSR performance is a strong social embedding of the economy. This is furthermore supported by Burgoon and Fransen (2017), who, in analysing European public opinion, argue that red consumerism (i.e. paying more for ethically produced goods) is more popular in settings with already-generous public protection (although
suggesting that support for welfare state-redistribution might be eroded in exactly these settings). Moving to the EU context, Atkinson and Davoudi (2000) suggest that discussion of the importance of problems such as social exclusion and poverty on the EU level play a positive role in keeping such issues on the policy agenda of national polities as well. Again, the effect can be indirect: background conditions like economic insecurity or low-income can be expected to have stronger implications for political support for both domains. Following this line of thought, I should find empirical support for this hypothesis:
H2: Social policy development at the international level strengthens rather than dampens policy efforts at the national level and/or vice versa.
Where the first two hypotheses represent examples of systematic divergence among countries, I include the suggestion that there is no pattern to be found in my analysis. Analyses flowing from stagist or modernization theories of the 1960s argue that post-industrial societies could experience convergence, thereby inferring that lesser developed countries are marked by “extreme divergence” (Rudra, 2007, p. 378). Such a mechanism indicates that the domestic and the international should be conceived as separate spheres that have no influence on each other. In the case of social policy in the EU, Geyer (2000) argues that EU social policy is too modest, lacking vital financial and political muscle to
significantly affect national welfare regimes (p. 245). The notion that the two levels do not much affect each other, existing in different spheres, led to the last hypothesis:
H3: The national and international domains have separate politics with little influence on each other.
3. Methods and Data: Do international labour efforts empirically affect national social policy and vice versa?
To analyse the above-mentioned debates, I analyse cross-national data on a range of approaches to labour protection across the national and international levels of political-economic governance. Cross-national comparison is useful to get an initial indication of the
existence of complementarity between particular institutions (Höpner, 2005, p. 334). The instruments serve as proxies for either the national dimension of labour protection or the international dimension of labour protection. This analysis thereby addresses various kinds of political protection that represent the most essential dichotomies in labour protection today: soft law and hard law; regulations and benefits; public and private; and welfare state and CSR. By focussing on the national-international dichotomy of social policy, I will be able to provide original insights in the interactions and consequences of labour protection.
The selection of country-cases is based on the requirement that the country is active in exporting labour intensive manufacturing goods, which are all low(er) income countries according to the World Bank. Data availability for these countries allows me to move away from the dominant focus on Western and industrialized countries and study the dynamics in the countries in which workers more often experience a vulnerable positions due to relatively low wage levels and abundant non-skilled labour supply (Fransen, 2010, p.3). Existing CPE literature, as Rudra (2007) argues, overlooks the possibility of systematic divergence in the developing world. The list of countries is taken from Khondonker and Kalirajan (2012) and includes 65 developing countries from Asia, Africa, Latin America and Europe. As the descriptive Table 1 at the end of this section shows, data is available varying from 56 cases (the two indicators from the Bertelmann Transformation Index dataset) to 64 cases (private standards, labour provisions in PTAs and two of the controls). In the analysis this means that there is full data on 50 cases, thereby excluding 15 countries.
I acknowledge that in some countries the informal economy is significant, however, for many countries statistical data on the informal economy is at best fragmented, cursory and anecdotal (ILO, p. 2). Additionally, I aim to focus on key elements within a polity’s control, rather than outcome variables that are influenced by events beyond a polity’s control. The next section will elaborate on the indices that I developed to measure the character of the co-ordination in national and international labour regulation. Needless to say, these indices are measured in different units; varying from seven point ordinal scales to interval ratio scales. To present a coherent analysis and compare coefficients, the variables were standardized around the mean of the subset of countries.
Domestic Labour Protection
The following variables are parts of national political-economic configurations of labour protection. The national focus on labour conditions can be measured in a range of ways. In picking the variables I aimed to cover a range of these policy ways, while keeping the inward
looking national characteristic constant. This resulted in including: (1) a score of basic administration, as government capability to secure compliance); (2) social safety nets, as a measure of public welfare state efforts; and (3) generosity of private severance payments, measured as mandated severance arrangements. These three measures are also combined into a composite measure.
(a) Basic Administration (BA index)
The Transformation Index of the Bertelsmann Stiftung (BTI) dataset contains a variable that examines the basic civil functions of a state apparatus in terms of regulation, administration and implementation. Although the variable does not exclusively analyses labour law
enforcement or social security systems (e.g. also police efficiency and tax service
effectiveness), it does depict important domestic orientation of a polity. The variable most importantly examines whether central governments can ensure that national policies are correctly implemented at lower administrative levels. This variable ranges from 1 to 10 (from limited to differentiated administrative structures), with a sample mean of 5.57 and a standard deviation of 1.55. Data is from 2015.
(b) Social Safety Nets (SSN index)
Also from the BTI database, this variable refers to social safety nets that compensate for the social risks of the capitalist economic system. It depicts the existence of a welfare regime, independent whether it is public, private or both. Scores indicate to what extent social risks, such as unemployment or poverty, are tackled in terms of coverage, quality and efficiency of programs. This variable is national because it first and foremost focuses on state-funded welfare systems, but if these are not present it may also comprise family, clan or village structures if these provide a viable compensation on a broad scale for risks. This variable ranges from 1-7, with a mean of 4.02 and a standard deviation of 1.41. Data is from 2015.
(c) Severance Pay Programs (SP index)
How generous are firms in protecting their labour as an alternative for welfare state provisions? Holzmann et al. (2011) developed a measure that captures the generosity of severance pay programs in countries. This variable refers to the average severance payments in weeks of salary for employees with 1, 5 and 10 years of service. It depicts efforts to cover risks of unemployment by private domestic actors. Severance pay mandates require that firms make payments to workers upon separation from the firm. Although the legal source of these
programs can be voluntary – if firms themselves decide whether to make such payments or not, this information is due to lack of data availability not included in this measure. What is included is firstly mandated schemes, of which there is almost complete data, and quasi-mandated schemes (collective bargaining provision of severance benefits) of which data is available for approximately half of the countries in their dataset (Holzmann et al., 2011, p. 13). Mandated programs originate in labour codes or similar legislation and the
quasi-mandated programs originate in collective bargaining. The legal sources of the quasi-mandated and quasi-mandated programs are limited to the national scope and therefor this variable fits as a domestic-national proxy. This variable ranges from 0-10.5 with a mean of 3.47 and a standard deviation of 2.52.
FIGURE 1
National labour protection components: BA index; SSN index; and SP index
−3 −2 −1 0 1 − 2 − 1 0 1 2 Basic administration, 2015 Social saf ety nets , 2015 Albania Angola Bangladesh Bhutan Bolivia Cambodia Cameroon
Central African Rep. Chad
China
Congo, Dem. Rep.
Cote d'Ivoire Ecuador Egypt El Salvador Eritrea Ethiopia Georgia Ghana Guatemala Guinea Honduras India Iran Jordan Kenya Kyrgyzstan
Madagascar MalawiMali Mauritania Moldova Mongolia Morocco Mozambique Nepal Nicaragua Nigeria Pakistan Papua New Guinea
Paraguay Philippines Rwanda Senegal Sierra Leone Sri Lanka Sudan Syria Tanzania ThailandTunisia Uganda Ukraine Uzbekistan Vietnam Zambia −3 −2 −1 0 1 − 1 0 1 2 Basic administration, 2015 Inde x of se ver ance pa y progr
ams (Holzmann et al. 2011)
Albania Angola Bangladesh Bhutan Cambodia Cameroon Central African Rep.
Chad China Cote d'Ivoire Ecuador Egypt El Salvador Eritrea Ethiopia Georgia Ghana Guatemala Guinea Honduras India Iran Jordan Kenya Kyrgyzstan Madagascar Malawi Mali Mauritania Moldova Mongolia Morocco Mozambique Nepal Nicaragua Nigeria Pakistan
Papua New Guinea Paraguay Philippines Rwanda Senegal Sierra Leone Sri Lanka Sudan Syria Tanzania Thailand Tunisia Uganda Ukraine Uzbekistan Vietnam Zambia −2 −1 0 1 2 − 1 0 1 2
Social safety nets, 2015
Inde
x of se
ver
ance pa
y progr
ams (Holzmann et al. 2011)
Albania Angola Bangladesh Bhutan Cambodia Cameroon Central African Rep.
Chad China Cote d'Ivoire Ecuador Egypt El Salvador Eritrea Ethiopia Georgia Ghana Guatemala Guinea Honduras India Iran Jordan Kenya Kyrgyzstan Madagascar Malawi Mali Mauritania Moldova Mongolia Morocco Mozambique Nepal Nicaragua Nigeria Pakistan
Papua New Guinea Paraguay Philippines Rwanda Senegal Sierra Leone Sri Lanka Sudan Syria Tanzania Thailand Tunisia Uganda Ukraine Uzbekistan Vietnam Zambia
(d) National composite
Figure 1 provides a first impression of the distribution of my national measures. Because I theoretically expect these measures to represent different, not necessarily overlapping, parts of national labour protection, it is not expected that they have very high correlations.
Nevertheless, it is constructive that the individual measures positively correlate with each other: basic administration and social safety nets 0.77; basic administration and severance pay programs 0.13; and social safety nets and severance pay programs 0.15. The three measures have a Cronbach’s alpha of 0.62, which convention in cluster analysis suggests is (just) short of the threshold to combine them into a scale. However, as mentioned, there are theoretical grounds mentioned before to combine the measures. Figure A1 in the Appendix shows the sample distribution of the national composite score. The composite score excludes 10 out of the 64 cases because they lack data on one of the three proxies. The country cases that are excluded from the analysis are Tonga, Cape Verde, Guinea-Bissau, Swaziland, Comoros, Solomon Islands, Vanuatu, the Gambia, the Democratic Republic of Congo and Bolivia. Sri Lanka is the most nationally oriented polity (score 5.83) and Syria is the least nationally oriented polity (score -6.48).
International Labour Protection
To reflect relevant variation in the international aspects of labour protection, I created three original country-level measures that cover a range of possible global activities with relation to labour and social interventions: private standards, labour provisions in preferential trade agreements, and ILO conventions ratified. Both labour provisions in PTAs and ILO conventions are more focussed on labour rights, whereas private standards focus more on labour standards (i.e. workplace safety, worker compensation, etc.). Additionally, and valuable for contrasting to the majority of public-private literature, these three proxies
compose a mix of public and private interventions. This, of course, is not an exhaustive list of labour initiatives from outside a country to another country to improve labour conditions, but these allow me to judge how international interventions interrelate with national activity. These three measures are also combined into a composite measure.
(a) Private Standards (PS index)
This variable depicts the intensity of private standard-setting organizations’ activity in a country. As these private standards emerge in high-income industrialized countries, we can safely assume that for the countries in the analysis these are international interventions.
The variable measures the presence of private regulatory organizations on three levels of maturity of monitoring efforts as an integrated approach to advancing worker rights in a country: standardization, benchmarking and cross-auditing. The first level is lowest level is private standard-setting and is measured by the presence of private standards in the regulation of labour rights, including Better Cotton Initiative, Business Social Compliance Initiative, Electronic Industry Coalition, Ethical Trading Initiative(s), Fair Labor Association, Fair-trade
Labeling Organization (FLO)1, Fair Wear Foundation, Flower Label Programme,
Goodweave, Initiative Clause Sociale, the International Council for the Toy Industries CARE Process, Rainforest Alliance, Social Accountability International’s Corporate Involvement Programme and Utz Certified. This list of standards is taken from Burgoon and Fransen (2017). Secondly, I included private organizations involved in benchmarking. These are coded as one step up in the maturity of monitoring efforts in a country because they aims to harmonize existing efforts by building consensus and strengthen collaboration among level 1 private standards. In social standard setting there are three relevant organizations, including ISEAL Alliance, the Global Social Compliance Programme (GSCP) and Global Sustainable Tourism. Third, SEDEX membership is included an additional level of integration of
monitoring efforts in a country because it allows participants in private standard schemes to mutually review their audit reports so that private labour protection efforts are no longer isolated per firm or per standard organizations.
Data is gathered from StandarsMap.org. Standards Map provides information on standards, codes of conduct and audit protocols addressing sustainability hotspots in global supply chains. This is an initiative by the Trade for Sustainable Development programme from the International Trade Centre. Data is gathered in 2018 and should represent the
presence of standards in 2018. The presence of each of the 14 standard-setting organization is coded as 1; the presence of each of the three benchmarkers is coded as 2; and the presence of SEDEX is coded as 3. This adds up to a total, which is used as the proxy for international private regulatory activity.
(b) Labour Provisions in Preferential Trade Agreements (LP in PTA index)
As a second proxy for international labour protection, I created the variable Labour Provisions (LPs) in Preferential Trade Agreements (PTAs) as an individual country measure. Trade agreements are one of the most obvious ways to participate in the international economy and
through the rise of labour provisions therein a relevant proxy for international labour protection. Trade relations can serve to transmit higher labour standards across national borders (Lim et al., 2015, p. 295). Labour provisions indicate that through meeting core standards a country can get preferential access to markets
This indicator is based on the scoring of labour provisions in preferential trade agreements as provided by Raess and Sari (2017). They have created a dataset covering 487 PTAs from 1990 to 2015 coded against 140 distinct items pertaining to six main categories, namely “preambular provisions on aspirational labor commitment, substantive commitments on LPs, obligations in relation to LPs, enforceability of the substantive LPs, cooperation over labour issues, and the institutions set up to implement the labour-related commitments” (Raess and Sari, 2017, 1). Clearly, a score 1 (preambular provisions) is less intensive than a score 6 (institutions). The authors set the lower temporal boundary to 1990 because they consider that to be the year that LPs in PTAs emerged – considering the North American Free Trade Agreement to be the first PTA with comprehensive LPs (Raess and Sari, 2017;
Burgoon, 2004).
This dataset contains information of each PTA, however my analysis requires a score per each individual country. I recoded the data so that each country involved in a PTA gets the LP score of that particular PTA, this included both bilateral agreements, such as that between Afghanistan and India in 2003, but also large trade frameworks such as the Cotonou Agreement that covers 103 European, African and Caribbean countries.
(c) Conventions from the International Labour Organization ratified (ILO index)
By ratifying conventions from the International Labour Organization (ILO), a country affirms its formal commitment to apply that international instrument. While the ILO employs a tripartite structure, where government, employers’, and workers’ delegates and advisers take seat in the conferences and meetings, a country’s Minister of Labour/Social Affairs (or likewise) in the end signs the conventions, whereby committing themselves to applying the convention in national law and practice. The ILO is one of the most traditional and
institutionalized international instruments.
There are three different types of conventions: Fundamental, Governance (Priority) and Technical. The coding of this proxy is done along the lines of the priority the ILO sets, meaning that ratification of each one of eight fundamental conventions is appointed a score 3, of each one of four governance conventions a score 2, and each of 187 technical conventions a score 1. Data is available at the website of the ILO and is up to date as of May 2018.
FIGURE 2
International labour protection components: ILO index; LP in PTA index; and PS index.
(d) International composite
The scatterplots in Figure 2 provide a first impression of the distribution with respect to our three international measures. The analysis of the international domain incorporates the same logic about different, not necessarily overlapping nature of the composite measure. Also in this case, however, it is good to see that the individual measures are positively correlated: ILO conventions and labour provisions in PTAs 0.2; ILO conventions and private standards 0.23; and labour provisions in PTAs and private standards 0.08. Figure A2 in the Appendix shows the sample distribution of the composite measure of the three proxies for international labour protection. Unsurprisingly, the Cronbach’s Alpha score of 0.38 does not provide empirical
−1 0 1 2 − 1 0 1 2
Ratified ILO conventions, 2018
Labour pro visions in PT As , 1990 − 2015 a ver age Albania Angola Bangladesh Bolivia Cambodia Cameroon Cape Verde
Central African Rep. Chad
China
Comoros Congo, Dem. Rep.
Cote d'Ivoire Ecuador Egypt El Salvador Eritrea Ethiopia Gambia, The Georgia Ghana Guatemala Guinea Guinea−Bissau Honduras India Iran Jordan Kenya Kyrgyzstan Madagascar Malawi Mali Mauritania Moldova Mongolia Morocco Mozambique Nepal Nicaragua Nigeria Pakistan Papua New Guinea
Paraguay Philippines Rwanda Senegal Sierra Leone Solomon Islands Sri Lanka Sudan Swaziland Syria Tanzania Thailand Tunisia Uganda Ukraine Uzbekistan Vanuatu Vietnam Zambia −1 0 1 2 − 1 0 1 2
Ratified ILO conventions, 2018
Pr iv ate standards , 2018 Albania Angola Bangladesh Bolivia Cambodia Cameroon Cape Verde
Central African Rep. Chad
China
Comoros Congo, Dem. Rep.
Cote d'Ivoire Ecuador Egypt El Salvador Eritrea Ethiopia Gambia, The Georgia Ghana Guatemala Guinea Guinea−Bissau Honduras India Iran Jordan Kenya Kyrgyzstan Madagascar Malawi Mali Mauritania Moldova Mongolia Morocco Mozambique Nepal Nicaragua Nigeria Pakistan
Papua New Guinea Paraguay Philippines Rwanda Senegal Sierra Leone Solomon Islands Sri Lanka Sudan Swaziland Syria Tanzania Thailand Tunisia Uganda Ukraine Uzbekistan Vanuatu Vietnam Zambia −1 0 1 2 − 1 0 1 2
Labour provisions in PTAs, 1990−2015 average
Pr iv ate standards , 2018 Albania Angola Bangladesh Bhutan Bolivia Cambodia Cameroon Cape Verde Central African Rep.Chad
China
Comoros Congo, Dem. Rep. Cote d'Ivoire Ecuador Egypt El Salvador Eritrea Ethiopia Gambia, The Georgia Ghana Guatemala Guinea Guinea−Bissau Honduras India Iran Jordan Kenya Kyrgyzstan Madagascar Malawi Mali Mauritania Moldova Mongolia Morocco Mozambique Nepal Nicaragua Nigeria Pakistan
Papua New Guinea Paraguay Philippines Rwanda Senegal Sierra Leone Solomon Islands Sri Lanka Sudan Swaziland Syria Tanzania Thailand Tonga Tunisia Uganda Ukraine Uzbekistan Vanuatu Vietnam Zambia
support for combining the three measures. But, just as for the national composite, there is theoretical support to combine the three measures. Guatemala stands out as one of the most internationally oriented polities (score of 5.56), which is first and foremost because of the country’s high scores on ILO conventions and LP in PTAs. On the other side of the figure Iran is the polity with the least international-oriented interventions (score of -4.03), which is caused by its’ low scores on all three proxies.
Controls
To check if the relationship between the domestic-national and the global-international is robust, in the models I added control variables. GDP divided by midyear population (data are in constant 2010 US dollars). Economic globalization ranges from 24.7 to 83 and
measures actual flows of trade and investments, and restrictions on trade and capital such as tariff rates. This control should evaluate whether the use of international policy instruments of a political economy is not just a function of how internationally oriented a polity is.
Associational and organizational rights evaluates the freedom of assembly, demonstrations
and open public discussion; the freedom for NGOs, trade unions and other organizations. Scores vary from 0-11. This variable is not included in the domestic-national dimension because the proxies should ipso facto be independent and in the ILO conventions these associational rights are warranted. Furthermore, I controlled for rural population as a percentage of the total population (varying from 16.6-87%) and official development assistance (ODA). All variables included in the analysis are presented in Table 1.
Estimation approach
Key questions of interest in my analysis are whether the national protection composite is a significant predictor of the international protection composite, and vice versa. These questions are represented in my hypotheses, which will be tested fitting the data into several models. In the analysis I will report on the most important models. As baseline models, the analysis starts with an ordinary least squares (OLS) regression. To furthermore check if the classical OLS assumptions are met, particularly addressing the endogeneity problem (independent error terms), I continue to fit a generalized least squares (GLS) model. In this case the GLS estimator is generally referred to as a seemingly unrelated regressions (SUR) estimator (Zellner, 1962). SUR basically examines whether the (same) panel of independent variables play out the same for the national as for the international labour protection composite variable. This would indicate if the same underlying political and economic processes can
explain scores on national and international labour protection: i.e. do they have similar background politics or economics?
TABLE 1
Summary statistics for analysis of national and international labour protection
Source Mean SD Range Year N
International proxies
ILO index International
Labour Organization 53.34 18.03 21-94 2018 62 PS index International Trade Centre 12.93 3.36 8.6-22 2018 64
LP in PTA index Raess and Sari
2017 8.05 6.11 0-24 1990-2011 64 International composite 0.08 1.99 -3.94-5.62 62 National proxies BA index Bertelmann Transformation Index 5.57 1.55 1-8 2015 56 SSN index Bertelmann Transformation Index 4.02 1.41 1-7 2015 56 SP index Holzmann et al. 2011 3.47 2.52 0-10.5 2011 61 National composite 0.06 2.21 -3.94-5.62 54 Country controls GDP World Bank 2108 1538 302-6109 2017 63 Economic globalization KOF Index of Globalization 52.79 12.74 24.72-83.01 2014 59 Associational and organizational rights Freedom House 5.77 3.07 0-11 2014 64
Rural population World Bank 57.39 16.95 16.55-87.02 2014 64
Official development assistance (in million $)
World Bank 1041.92 1361.44 0.00*-8869.64 2016 64
*China is a net provider of ODA, for the analysis this country is recoded to 0. Indicating that the country does not receive ODA.
4. Analysis
In this section I will empirically test the relationship between international and national labour protection in both ways of causality. As explained in the previous section, the data will first
be modelled using OLS. The robustness of the results for both dependent variables will be tested for the same set of country-controls, thereby allowing the controls to wipe out
relationships where national conditions are the outcome variables and not when international conditions are the outcome variables (and vice versa). The question of whether the predictors operate differently in correlation with national protection than in correlation with international protection – suggesting that we are talking about distinct political and economic processes underlying the different types of protection, will further be analysed using SUR.
But before moving to the regression models, it is helpful to look at the more disaggregate story. Table 2 and 3 show the correlation matrix of the three domestic and the three
international components. All individual correlations are positive. The correlation matrix in Table 2 shows that ensuring that national policies are correctly implemented at lower administrative levels (BA index) strongly correlates (0.77) with the coverage, quality and efficiency of programs tackling social risks (SSN index). In making sense of this relation, one could argue that although a strong basic administration does not inherently indicate a strong welfare state, conversely coverage, quality and efficiency of social programs might require a strong basic administration. Lower positive correlations appear between the generosity of severance pay programs (SP index) and basic administration (0.13) and social safety nets (0.15).
TABLE 2
National labour protection correlation matrix
BA index SSN index SP index
BA index 1
SSN index 0.77 1
SP index 0.13 0.15 1
Within the international composite, shown in Table 3, ratifying ILO conventions (ILO index) is roughly equally correlated to the intensity of labour provisions in trade agreements (LP in PTA index) as to the intensity of private standard-setting organizations’ activity (PS index) in a country. With 0.08, the lowest correlation is between labour provisions in PTAs and private standards. Considering that concluding trade agreements and ratifying ILO conventions are primarily part of the public hard-law domain and private standards are more in the private soft-law sphere, these correlations do not suggest that these two realms are necessarily cluster together within the international composite.
TABLE 3
International labour protection correlation matrix
ILO index LP in PTA index PS index
ILO index 1
LP in PTA index 0.20 1
PS index 0.23 0.08 1
Figure 3 shows that the composite labour protection variables are positive predictors for each other with a correlation of 0.39. The following section will subject this relation to various controls in different models in order to find out whether these two domains actually mutually reinforce each other or whether there is an endogeneity problem where both domains are part of the same underlying political and economic processes.
FIGURE 3
Composite international labour protection and national labour protection
Models (1) – (4): Dependent variable: national composite
Table 4 shows models 1-4 and reveals the effects of the individual and composite proxies for international labour protection. PS index shows up as the most important predictor: this variable has the largest estimate (0.73) of all predictors in the model and is significant using a confidence interval of 5 percent (with an N of 51). Both ILO index and LP in PTA index are not statistically significant, but have positive and substantial estimated effects (respectively 0.44 and 0.25). GDP appears to be the only significant other factor in predicting the national composite, indicating that wealthy countries are significantly more protective of their labour
−6 −4 −2 0 2 4 6 − 4 − 2 0 2 4
Index total score 'national'
Inde
x total score 'inter
national'
Syria
Sudan Eritrea Central African Rep.
Chad Papua New Guinea Angola Nepal Guinea Madagascar Cambodia Nigeria Pakistan Mali Uganda Cameroon Malawi Ethiopia Mauritania Jordan Kenya Sierra Leone Morocco Tanzania Cote d'Ivoire Senegal Bangladesh Guatemala Georgia Mongolia Honduras Iran Paraguay Egypt Nicaragua India Uzbekistan Kyrgyzstan Tunisia Mozambique Ukraine Rwanda Philippines Vietnam El Salvador Moldova Albania Zambia China Ghana Ecuador Thailand Sri Lanka
force through national instruments. Most interestingly, however, is that the international composite score has an estimated effect of 0.31 significant at the 0.05 level, indicating that there is a substantial association between the two composites. Table A1 in the Appendix shows the effect of international labour protection on national labour protection while gradually adding the control variables. Naturally, the explained variance in the national composite increases by adding variables (from 15% in the bivariate model up to 45% in model 6) and the estimated size of the effect decreases from 0.44 to 0.31, which is still substantial.
TABLE 4
National labour protection and the direct effects of various forms of international labour
protection (1) (2) (3) (4) GDP 0.00 *** 0.00 * 0.00 *** 0.00 *** (0.00) (0.00) (0.00) (0.00) Economic globalization 0.04 0.04 0.03 0.03 (0.02) (0.02) (0.02) (0.02)
Associational and organizational rights 0.08 0.06 0.15 0.09
(0.10) (0.09) (0.09) (0.09)
Rural population 0.04 0.01 0.03 0.03
(0.02) (0.02) (0.02) (0.02)
Official development assistance 0.00 0.00 0.00 0.00
(0.00) (0.00) (0.00) (0.00) ILO index 0.44 (0.28) PS index 0.73 * (0.28) LP in PTA index 0.25 (0.24)
Composite score international 0.31 *
(0.13)
N 50 51 51 50
R2 0.41 0.45 0.38 0.45
Notes: Dependent variable: National labour protection composite score. Robust standard errors in parentheses.
Models (5) – (8): Dependent variable: international composite
We observe more or less the same pattern when we look at the predictors for international labour protections in models 5-8 in Table 5. BA index is positively and statistically
significantly correlated with the international composite score with a large estimated effect of 0.80. Having an effective state apparatus is an important component in explaining
international labour protection. It would make sense if this particularly applied to concluding trade agreements and ratifying international conventions, as these would naturally require coordinated action. However, Appendix Table A3 suggests that BA index has the largest estimated and only significant effect on PS index, running against narratives of private institutions ‘filling the gap’ of weakening public institutions. Both other individual proxies are not statistically significant and while SSN index has a large positive estimated effect (0.64), the estimated effect of SP index on international labour protection is fairly low at 0.17. Strikingly and in contrast to the role of GDP in Table 4, none of the controls has a significant effect on the outcome variables. The national composite for labour protection is significant at the 0.05 level and has an estimated effect of 0.39: for each increase of 1 in the composite for international protection, the national protection composite increases on average with 0.39. Table A2 in the Appendix shows the gradual changes to the effects while adding control variables, the estimated effect increases from 0.35 in the bivariate model to 0.39 in the last model and the explained variance increases from 15% to 23%, which is significantly less than
in the reversed causal models.We can conclude, however, that the effect found in Table 4
TABLE 5
International labour protection and the direct effects of various forms of national labour
protection (5) (6) (7) (8) GDP -0.00 -0.00 -0.00 -0.00 (0.00) (0.00) (0.00) (0.00) Economic globalization 0.00 0.00 0.03 0.00 (0.02) (0.02) (0.02) (0.02)
Associational and organizational rights 0.09 0.10 0.06 0.11
(0.10) (0.10) (0.10) (0.10)
Rural population -0.02 -0.01 -0.02 -0.02
(0.02) (0.02) (0.02) (0.02)
Official development assistance 0.00 0.00 0.00 0.00
(0.00) (0.00) (0.00) (0.00) BA index 0.80 * (0.33) SSN index 0.64 (0.37) SP index 0.17 (0.28)
Composite score national 0.39 *
(0.16)
N 52 52 54 50
R2 0.22 0.18 0.12 0.23
Notes: Dependent variable: International labour protection composite score. Robust standard errors in parentheses.
*** p < 0.001; ** p < 0.01; * p < 0.05.
In comparing the two ways of causality, we see that the national composite has implications for the international slightly more than the other way around. This could represent that
national institutions are likely to be more established and ex ante developed than international institutions, suggesting that the national is causally prior to the international. However, this is hard to say because the differences are rather small (0.31 to 0.39) and the variance explained in the models with national labour protection as the dependent variable is almost twice as large as the other models. Both sets of models show one individual proxy that drives most
importantly drives the correlation: the intensity of private standards in explaining national labour protection; and the strength of the basic administration in explaining international labour protection. Looking at the more disaggregate story in Appendix Table A3, PS index and BA index are two of the components that are statistically significant predictors for each other. Furthermore in analysing components that predict components in the ‘other domain’, ILO index and SSN index are statistically significant predictors for each other. In both cases, the national component has a larger estimated effect on the international component.
The main issue for further analysis, however, is whether the controls appear to play out for national labour protection as they do for international labour protection. Comparing the OLS models provides us with a first glimpse. Apart from the negligible differences in estimated effects, the biggest different is the effect of GDP. GDP is statistically significant at the 0.001 level in explaining all national labour protection components, but is not statistically
significant in explaining any of the international labour protection components. Table A3 in the Appendix, on the other hand, shows that GDP is a statistically significant factor in explaining ILO index, PS index and SSN index. The notion that GDP predicts two
components from the international domain and only one from the national domain deviates from these earlier mentioned findings. These contrasting findings create confusion about the political and economic processes underlying national and international labour protection. Using SUR to analyse whether the error terms of the two regressions are correlated, we find that, for the same labour protections, the correlation of the residuals in the national and international equations is 0.3499 and that we can reject the hypothesis that this correlation is zero. This suggests that we are not talking about distinct political and economic processes underlying national protection compared to international protection: there is some underlying factor that explains both national and international labour protection.
5. Discussion
In recent decades global labour governance went through significant changes. In an attempt to facilitate labour protection in the increasingly globalized economy, protecting social and labour standards expanded from essentially domestically oriented hard-law to additional international buyer-driven and trade-related initiatives. Today, a mixture of new and old policy instruments varying from tax code to private standard setting characterizes the field of labour protection. Scholars have long disagreed about whether such political initiatives and developments replace or reinforce each other on different aspects of political-economic
governance (e.g. public versus private or national versus international). Current academic literature, however, has a dominant focus on European and other industrialized contexts. Whereas arguably the most vulnerable positions tend to exist in labour-intensive
manufacturing industries that virtually moved from these Western countries to other places in the world. This paper examines the interrelation between different national and international policy interventions in labour protection in exactly these contexts.
Based on an original dataset covering 50 national polities, this paper finds empirical support for labour protection interventions on the national and international level of political-economic governance to operate in harmony, rather than in tension, with each other. This is in line with hypothesis 2, thereby simultaneously rejecting hypothesis 1 and 3. The analysis suggests that few other conditions are significant in shaping labour protection on both levels. I find that there is a group of countries that maximizes interventions in both domains and a group of countries that minimizes the use of these policy instruments. Econometric models furthermore suggest that national and international protections are not part of distinct political economic processes, but could be explained by some factor underlying both domains of labour protection.
This paper thereby contributes to comparative political economic literature in general and global labour governance literature in particular. In CPE, scholars disagree whether particular types of coordination in one sphere tend to develop complementary practices in other spheres as well. This paper contributes to this debate by adding new empirical-analytical insights to this theoretical debate on institutional complementarities. In global labour
governance, this paper provides insights to the consequences of new types of labour governance. This strand of literature is not yet involved in studying very sophisticated
institutional differences between countries. Conducting cross-national research on 50 national polities provides new insights on the interrelation between different instruments of labour protection and adds to the study of the effectiveness of such instruments.
There are several limitations to this study that could be further investigated by future research. First, the paper ‘only’ provides a cross-national analysis. As everything social takes place in time and considering that we can only understand what we see today if we know how it looked yesterday and where it might be heading (Streeck, 2014, p.xi), studying institutional dynamics and change requires longitudinal analysis. For most of the variables used in this analysis it is possible to collect data over the years, but other proxies might be developed as well. A second important suggestion for future research is a case study that examines the mechanisms that are outlined in this paper. The cross-national analysis is limited in
suggesting the direction of causality from either national to international labour protection or
vice versa. By studying the relevant political actors that are assumed to be at the centre of
labour protection (such as unions, business or government actors), literature should study the causal mechanisms. If the suggested patterns do hold true in future research, I have
successfully shed light on undertheorized interactions of a range of approaches to labour protection across various levels of political-economic governance.
6. Literature
Anner, M. (2015) ‘Labor control regimes and worker resistance in global supply chains’,
Labor History, 56(3), 292-307.
Atkinson, R., and Da Voudi, S. (2000). ‘The concept of social exclusion in the European Union: context, development and possibilities’, Journal of Common Market Studies, 38(3), 427-448.
Bernstein, S., and Cashore, B. (2007). ‘Can non-state global governance be legitimate? An analytical framework’, Regulation & Governance, 1(4), 347-371.
Brown, R. (2016). ‘Promoting labour rights in the global economy: Could the United States’ new model trade and investment frameworks advance international labour standards in Bangladesh?’, International Labour Review, 155(3), 383-406.
Burgoon, B. (2004) ‘The rise and stall of labor linkage in globalization politics’, International
Politics, 41(2), 196-220.
Burgoon, B. (2009). ‘Social nation and social Europe: Support for national and supranational welfare compensation in Europe’, European Union Politics, 10(4), 427-455.
Burgoon, B., and Fransen, L. (2017) ‘Support for Ethical Consumerism and Welfare States in the Global Economy: Complements or Substitutes?’, Global Policy, 8, 42-55.
Burgoon, B., and Fransen, L. (2018) ‘Might Corporate Social Responsibility Hollow Out Support for Public Assistance in Europe?’, British Journal of Industrial Relations, 56(1), 128-163.
Crouch, C., Streeck, W., Boyer, R., Amable, B., Hall, P. A., and Jackson, G. (2005) ‘Dialogue on ‘Institutional complementarity and political economy’’, Socio-Economic Review, 3(2), 359-382.
Fransen, L. W. (2010). Minding their own business? Firms and activists in the making of private labour regulation.
Gjølberg, M. (2009) ‘The origin of corporate social responsibility: global forces or national legacies?’, Socio-economic review, 7(4), 605-637.
Greenhill, B., Mosley, L., and Prakash, A. (2009) ‘Trade-based diffusion of labor rights: A panel study, 1986–2002’, American Political Science Review, 103(4), 669-690.
Hall, P. A. and Soskice, D. (eds) (2001) Varieties of capitalism: The institutional foundations
of comparative advantage, Oxford, Oxford University Press.
Hassel, A. (2008) ‘The evolution of a global labor governance regime’, Governance, 21(2), 231-251.
Höpner, M. (2005) ‘What connects industrial relations and corporate governance? Explaining institutional complementarity’, Socio-Economic Review, 3(2), 331-358.
Holzmann, R., Pouget, Y., Vodopivec, M., & Weber, M. (2011). ‘Severance pay programs around the world: history, rationale, status, and reforms’, World Bank.
ILO (International Labour Organization) ‘Measurement of the Informal Economy – ILO’, accessed at
http://www.ilo.org/wcmsp5/groups/public/---ed_emp/---emp_policy/documents/publication/wcms_210443.pdf on June 18, 2018.
IRBC Agreements (2018) ‘Dutch Agreement on Sustainable Garments and Textile’, accessed at https://www.imvoconvenanten.nl/garments-textile?sc_lang=en, on July 6, 2018. Jackson, G., and Apostolakou, A. (2010). ‘Corporate social responsibility in Western Europe:
an institutional mirror or substitute?’, Journal of Business Ethics, 94(3), 371-394. Khondoker, M., and Kalirajan, K. (2012) Determinants of Labor-Intensive Exports by the
Developing Countries: A Cross Country Analysis, ASARC Working Paper 2012/09,
Kinderman, D. (2011). ‘‘Free us up so we can be responsible!’The co-evolution of Corporate Social Responsibility and neo-liberalism in the UK, 1977–2010’, Socio-Economic
Review, 10(1), 29-57.
Kinderman, D. P., and Lutter, M. (2018) Explaining the growth of CSR within OECD
countries: The role of institutional legitimacy in resolving the institutional mirror vs. substitute debate, MPIfG Discussion Paper 18/2, Cologne, Max Planck Institute for the
Study of Societies
Koenig-Archibugi, M. (2017). ‘Does transnational private governance reduce or displace labor abuses? Addressing sorting dynamics across global supply chains’, Regulation &
Governance, 11(4), 343-352.
Lim, S., Mosley, L., and Prakash, A. (2015). ‘Revenue substitution? how foreign aid inflows moderate the effect of bilateral trade pressures on labor rights’, World Development, 67, 295-309.
Lim, S. and Prakash, A. (2017). ‘Do Economic Problems at Home Undermine Worker Safety Abroad?: A Panel Study, 1980–2009’, World Development, 96, 562–577.
Locke, R. M. (2013) The promise and limits of private power: Promoting labor standards in
a global economy, Cambridge/New York, NY, Cambridge University Press.
Mau, S. (2005) ‘Democratic Demand for a Social Europe? Preferences of the European Citizenry’, International Journal of Social Welfare, 14(2), 76–85
Midttun, A., Gautesen, K., and Gjølberg, M. (2006). ‘The political economy of CSR in Western Europe’, Corporate Governance: The international journal of business in
society, 6(4), 369-385.
One World (2018) ‘Afspraken textielsector: de arbeider merkt er nog maar weinig van’, accessed at https://www.oneworld.nl/fashion/een-jaar-kledingconvenant/ on July 6, 2018.
Pearson, R., and Seyfang, G. (2001). ‘New hope or false dawn? Voluntary codes of conduct, labour regulation and social policy in a globalizing world’, Global Social Policy, 1(1), 48-78.
Raess, D. and Sari, D. (2017). ‘Labor provisions in trade agreements (LABPTA): Introducing a new dataset’, accessed at https://snis.ch/project/social-clause-through-back-door/ on April 10, 2018.
Roger, C., Hale, T., and Andonova, L. (2017) ‘The comparative politics of transnational climate governance’, International Interactions, 43(1), 1-25.
Rudra, N. (2007). ‘Welfare states in developing countries: unique or universal?’, The Journal
of Politics, 69(2), 378-396.
Sánchez-Cuenca, I. (2000) ‘The Political Basis of Support for European Integration’,
European Union Politics, 1(2), 147–71.
Streeck, W. (2014). Buying time: The delayed crisis of democratic capitalism. London: Verso Books.
Zellner, A. (1962). ‘An efficient method of estimating seemingly unrelated regressions and tests for aggregation bias’, Journal of the American statistical Association, 57(298), 348-368.
7. Appendix
FIGURE A1
National labour protection composite
Figure A2
International labour protection composite
Syria Sudan Eritrea Central African Rep. Chad Papua New Guinea Angola Nepal Guinea Madagascar Cambodia Nigeria Pakistan Mali Uganda Cameroon Malawi Ethiopia Mauritania Bhutan Jordan Kenya Sierra LeoneMorocco Tanzania Cote d'Ivoire Senegal Bangladesh Guatemala Georgia Mongolia HondurasIran Paraguay Egypt Nicaragua India Uzbekistan Kyrgyzstan Tunisia MozambiqueUkraine Rwanda Philippines Vietnam El Salvador Moldova Albania Zambia China Ghana Ecuador Thailand Sri Lanka Bolivia Cape Verde Comoros Congo, Dem. Rep. Gambia, The Guinea−Bissau Solomon Islands Swaziland Vanuatu Tonga Congo, Rep. −4 0 4 total_nat cname Domesticness Above Average Below Average
Normalised scores domesticness
Diverging Bars Iran Sudan Vanuatu Uzbekistan Mongolia Nepal Solomon Islands Cape Verde Eritrea Gambia, The Cambodia Georgia Chad Papua New Guinea Syria Guinea−Bissau Comoros Kyrgyzstan Sierra Leone Mauritania Central African Rep. Swaziland Rwanda Paraguay Mozambique Thailand Bangladesh Nigeria Philippines Angola Cameroon Congo, Dem. Rep. Bolivia Pakistan Jordan Ethiopia Mali Guinea Vietnam Senegal Sri Lanka Uganda Albania Zambia Malawi Tanzania Madagascar El Salvador Ghana Moldova Cote d'Ivoire Ecuador Morocco Egypt India Ukraine Tunisia Kenya Honduras China Nicaragua Guatemala −4 −2 0 2 4 6 total_int countr y name Globalness Above Average Below Average
Normalised scores globalness
TABLE A1
National labour protection and the direct effects of international labour protection (OLS)
(1) (2) (3) (4) (5) (6)
International labour protection composite 0.44 ** 0.34 ** 0.33 * 0.31 * 0.32 * 0.31 * (0.15) (0.12) (0.12) (0.13) (0.12) (0.13) GDP 0.00 *** 0.00 *** 0.00 *** 0.00 *** 0.00 *** (0.00) (0.00) (0.00) (0.00) (0.00) Economic globalization 0.03 0.02 0.03 0.03 (0.02) (0.02) (0.02) (0.02)
Associational and organizational rights 0.07 0.08 0.09 (0.09) (0.09) (0.09)
Rural population 0.03 0.03
(0.02) (0.02)
Official development assistance 0.00
(0.00)
Constant -0.08 -1.28** -2.48* -2.63** -5.14* -5.54**
N 53 52 50 50 50 50
R2 0.15 0.38 0.40 0.41 0.44 0.45
Notes: Dependent variable: National labour protection composite score. Robust standard errors in parentheses.
TABLE A2
International labour protection and the direct effects of national labour protection (OLS)
(7) (8) (9) (10) (11) (12)
National labour protection composite 0.35 ** 0.43 ** 0.41 * 0.38 * 0.41 * 0.39 *
(0.11) (0.15) (0.15) (0.16) (0.16) (0.16)
GDP -0.00 -0.00 -0.00 -0.00 -0.00
(0.00) (0.00) (0.00) (0.00) (0.00)
Economic globalization 0.01 0.00 -0.00 0.00
(0.02) (0.02) (0.02) (0.02)
Associational and organizational rights 0.11 0.10 0.11
(0.10) (0.10) (0.10)
Rural population -0.03 -0.02
(0.02) (0.02)
Official development assistance 0.00
(0.00)
Constant 0.31 0.64 0.22 -0.08 2.25 1.31
N 53 52 50 50 50 50
R2 0.15 0.15 0.17 0.19 0.21 0.23
Notes: Dependent variable: International labour protection composite score. Robust standard errors in parentheses.
TABLE A3
OLS regression of individual components
(1) (2) (3) (4) (5) (6)
GDP -0.00 * 0.00 * -0.00 -0.00 0.00 ** 0.00
(0.00) (0.00) (0.00) (0.00) (0.00) (0.00)
Economic globalization -0.02 -0.01 0.03 -0.00 0.02 -0.00
(0.01) (0.01) (0.01) (0.01) (0.01) (0.02) Associational and organizational rights 0.09 0.06 -0.11 0.01 0.01 -0.03
(0.05) (0.05) (0.06) (0.03) (0.03) (0.07)
Rural population -0.03 * 0.03 * -0.01 -0.01 0.01 0.01
(0.01) (0.01) (0.01) (0.01) (0.01) (0.01)
Official development assistance 0.00 0.00 ** 0.00 0.00 0.00 0.00
(0.00) (0.00) (0.00) (0.00) (0.00) (0.00) ILO index 0.17 0.25 -0.15 0.22 * 0.11 (0.14) (0.18) (0.10) (0.10) (0.20) PS index 0.22 0.09 0.31 ** -0.15 0.23 (0.18) (0.21) (0.10) (0.11) (0.22) LP in PTA index 0.18 0.05 0.13 -0.06 -0.10 (0.13) (0.12) (0.09) (0.09) (0.17) BA index -0.37 0.60 ** 0.42 0.63 *** -0.04 (0.25) (0.20) (0.29) (0.13) (0.31) SSN index 0.54 * -0.29 -0.20 0.62 *** 0.04 (0.24) (0.22) (0.29) (0.12) (0.31) SP index 0.07 0.12 -0.09 -0.01 0.01 (0.13) (0.11) (0.15) (0.08) (0.08) N 50 50 50 50 50 50 R2 0.37 0.50 0.28 0.68 0.72 0.12 *** p < 0.001; ** p < 0.01; * p < 0.05.
(1) DV = ILO index; (2) DV = PS index; (3) DV = LP in PTA index; (4) DV = BA index; (5) DV = SSN index; and (6) DV = SP index.
