Beyond the Stereotype: Restating the Relevance of the Dependency Research Programme

This article evaluates the relevance of dependency theory for understand- ing contemporary development challenges, especially in the light of changes in the global economy over the past 50 years. In order to do so, the arti- cle rectiﬁes previous misunderstandings of the scholarship and offers a new deﬁnition of dependency theory as a research programme, rather than a sin- gular theory. Four core tenets of this research programme are identiﬁed: a global historical approach; theorizing of the polarizing tendencies of global capitalism; a focus on structures of production; and a focus on the speciﬁc constraints faced by peripheral economies. While each of these elements can be found in many contemporary theories, what makes dependency theory unique — and a particularly strong research programme — is the combin- ation of these elements. The article demonstrates how this approach provides a deep and broad understanding that is necessary to appreciate the persis- tence of uneven development with reference to two case studies, namely successful industrialization in South Korea, and how the fragmentation of global value chains has impacted industrialization in Indonesia. Finally, the article argues that approaching these kinds of cases through a dependency research programme can contribute to a fruitful renewal of development studies.


INTRODUCTION
In a time when developing countries are increasingly moving into manufacturing production and there has been a degree of rebalancing of global power relations, it has become common to claim that the world has changed so drastically since the 1970s that we need new theories to understand contemporary development challenges. This article intervenes in this debate to demonstrate that dependency theory can offer important insights that remain relevant for understanding the changing development landscape.
Despite its enduring relevance, dependency theory has largely been excluded from mainstream development analysis since the late 1970s (Booth, 1985;Kufakurinani et al., 2017). Since then, the neglect of dependency theory has been so extreme that Margulis (2017) dubbed the phenomenon a 'peripheralization' of centre-periphery analysis. This article argues that one 2 Ingrid Harvold Kvangraven of the reasons dependency theory was so easily dismissed is the common focus on individual theorists who are mistakenly taken as spokespersons for 'the theory' as a whole. Given the misrepresentation of the tradition, this article argues that it is essential to develop a new framework for understanding dependency theory if we are to evaluate its relevance and preserve its strengths. Despite the common ground that can be identified across the competing dependency theories, little effort has been directed towards identifying the common core of the family of theories -which is a gap this article aims to close. The diversity within the tradition calls for a definition that can capture the strengths of a wide range of dependency theories, which is why this article defines it as a research programme.
In developing a coherent definition of the research programme, it becomes clear that dependency theory is not defined by a single element, but rather by the combination of several elements. This combination reflects a deep interdisciplinarity and a systemic approach, combined with attention to the particularities of peripheral economies. While outlining what dependency theory as a research programme entails, this article also unpacks the various approaches within dependency theory, focusing on the essential concepts, particularly in the structuralist and neo-Marxist schools, but also pointing to the way that related concepts and theories have appeared in other theories, from colonial times to modern frameworks that are seemingly far removed from dependency theory. 1 This article starts by laying out the pillars of dependency theory as a Lakatosian research programme, which includes a global historical approach; theorizing about the polarizing tendencies of capitalist development; a focus on structures of production; and particular constraints to peripheral development. While this definition broadens dependency theory as a research programme, it also excludes some approaches that may appear related to dependency theory. In order to demonstrate the distinctiveness of the research programme, the article therefore clarifies how it relates to other development theories, both mainstream and heterodox.
Furthermore, the article engages seriously with the critique of the tradition in order to assess whether the research programme is subject to the common critiques of dependency theory. Particular attention is paid to epistemological critiques such as economic reductionism, the focus on the nation state, and an overly strong focus on external factors. This is followed by an assessment of changes in the global economy that have taken place over the past half century, which some claim render dependency theory 1. It is common in the literature to make a distinction between neo-Marxists and structuralists (e.g. Kay, 1989), and some categorizations will include world systems theory (Arrighi, 2002;Foster-Carter, 1979). See also Cardoso (1973), Lall (1975) and Palma (1978), who categorize dependency theory into three groups -though their categories are slightly different. An in-depth overview of different strands and debates can be found in Kvangraven (2018).
Japanese scholarship on the power relations between centre and periphery (Ohno, 1998), radical African scholarship (Amin, 1974;Offiong, 1982;Rodney, 1972), Soviet development theory (Mark and Feygin, 2020), Canadian 'staple theory', 4 and the Caribbean dependency school. 5 What's more, dependency theories have also evolved over the past half century, addressing different regions of the world, and addressing new challenges as they have presented themselves, such as the increasing dominance of multinational corporations (MNCs), international financialization, centre-periphery relations in Europe, and Chinese investment in Latin America (Amin, 2010;Bruszt and Vukov, 2017;Casanova, 1970;Giraudo, 2020;Higginbottom, 2013). Two prominent early attempts to develop a broad definition of the family of dependency theories were by Cardoso (1977) and Kay (1989). Cardoso (1977) distinguishes the Latin American approach from the neo-Marxist strand popularized by Frank, but does not take stock of dependency theories that emerged elsewhere. Furthermore, his account is biased towards an understanding of dependency theory consistent with associated dependent development -the strand of dependency theory that he himself pioneered with Faletto. Moreover, Cardoso (1977) views dependency theory as a constantly evolving discourse, rather than a formalized theory. Meanwhile, Kay (1989: 18) limits himself to Latin American dependency theories. He identifies two key characterizations of this approach: 1) a concern with the peculiarities and distinctiveness of 'the Latin American economic, social, and political dynamics as compared with those of the developed capitalist countries'; and 2) a preoccupation with uncovering 'the external and internal mechanisms of exploitation and domination'. More recent attempts have focused on evaluating whether specific dimensions of dependency theory are still important, rather than trying to define a broader agenda (e.g. Fischer, 2015;Heller et al., 2009). For example, Heller et al. (2009: 287) specifically revisit Cardoso and Faletto's (1979) approach to dependency, which they see as a set of principles that combine a focus on how economies are inserted into the global economy, 'with a focus on the balance of domestic class forces, the capacity of state institutions, and contingent choices by political actors to explain the contrasting developmental fortunes of countries'. Fischer (2015: 701) revisits dependency theory through the concept of 'peripherality', which he defines as 'an assessment of structural modes of influenced by these Indian writers. In this way, 'the Indian debate on dependency could have had an impact on the Latin American debate through Paul Baran!' (ibid.: 254). 4. Innis (1950) argued that a staple-producing economy such as Canada would never be transformed by industrialization, but would remain a net exporter of resources and importer of manufactured goods, continuing to depend on borrowed capital and technology. 5. The Caribbean dependency school draws on the specific features and historical background of Caribbean countries (small islands, plantation economies, history of African slavery and a British colonial past) to elaborate the dependent features of these economies (Best and Polanyi Levitt, 2009;Girvan, 2005;Oxaal, 1975;Pantin, 1980;Styve, 2017).
The Relevance of the Dependency Research Programme 5 integration into the world economy via the dissemination of technological and industrial development'. He argues that this concept reflects certain common constraints that continue to structure the subordination of peripheral economies. 6 While these are both important contributions to dependency research, Heller's definition misses the radical essence of dependency theory (as we will see in the next section), and Fischer's is more of a review of constraints that remain relevant rather than a broader research programme. It should perhaps not be surprising that there are few systematic and comprehensive attempts to evaluate dependency theory as a research programme across time and space, given the variety of the contributions in terms of discipline, ideology and approach. While Kay and Cardoso focus on a limited sub-section of dependency theories, and Heller et al. and Fischer focus on specific principles of dependency or peripherality, there has not yet been an attempt to identify what dependency theory as a global research programme within development studies would look like, and how such a programme can be defined through building on the tradition's strengths.

Dependency Theory as a Research Programme
Considering the widely different perspectives within dependency theories, this body of scholarship is best thought of as a programme that allows for disagreement, but agrees on fundamental elements of how to approach the study of development -in line with Lakatos' (1978) definition of a research programme. According to Lakatos, research programmes are collections of interrelated theories that have common hypotheses that form a 'hard core'. The hard core also establishes a methodology for scientific investigation. Meanwhile, surrounding the hard core is a 'soft core' or a protective belt, which Lakatos calls the falsifiable auxiliary hypotheses, that handle some of the hanging threads derived from the hypotheses established in the core.
With this model in mind, it is possible to systematically define dependency theory as a (Lakatosian) research programme. The core hypothesis associated with the research programme is the polarizing tendency of capitalist development, related to both structures of production and the common constraints related to peripheral development. Note that this means that the research programme is broader than simply including scholarship that self-identifies as 'dependency theory'. For example, Prebisch and Singer were not originally dependency theorists, but given that centre-periphery dynamics are at the core of their research, and that dependency theorists build on and extend many of their core concepts, they are certainly a part of the dependency research programme as defined here.
Regarding methodology, dependency scholarship always takes a global historical approach to development (as also noted by Samir Amin in Kvangraven, 2017). This leaves us with four essential characteristics of the research programme related to theory and method: 1) the global historical approach; 2) economic theorizing that addresses the polarizing tendencies of capitalist development; 3) the attention to structures of production; and 4) the attention to the constraints that result from being a peripheral economy in the global economic system (see Figure 1). While each of these four elements may be found across the history of social sciences, across disciplines and also in contemporary approaches to development, it is important to note that it is the combination of these multiple elements to form a holistic and comprehensive approach to uneven development that characterizes the research programme.
This combination of factors also points to dependency theory's deeply interdisciplinary nature. 7 It is deep because the research programme is not about 'adding' different disciplines or methods to each other to understand a phenomenon, but rather about approaching the research question in an open and systemic way, addressing how underdevelopment has been historically and structurally produced, and developing explanations that traverse political, sociological, economic and historical boundaries. A dependency 7. As Kay (1989) also identifies in his review of Latin American dependency theory.
The Relevance of the Dependency Research Programme 7 research programme defined in this way makes it possible to address the weaknesses of individual theories, as it permits further research and the discovery of new phenomena, while preserving the strengths that lie at its core. Given this comprehensive approach, it is worth noting that both the strengths and weaknesses of the dependency research programme may lie in 'the scale of its aim and ambition' (Kay, 1989: 194). Consequently, if research within the programme is done well, it is highly insightful and innovative, but if it is done poorly, it runs the risk of being circular and reductionist.
Let us consider first the global historical approach to development. All dependency approaches are, in one way or another, critiques of linear historiography; from the structuralists' analysis of the role of imperialism in shaping peripheral structures of production (e.g. Furtado, 1970) to the neo-Marxists' attention to the extension of capitalism differing across time and geographies (e.g. Amin, 1974;Baran, 1957). Many dependency scholars have gone to great lengths to document the divergence of the periphery from the centre under colonialism (e.g. Frank, 1974) and the origin and persistence of technological dependence that reinforces the development of the centre at the expense of the periphery (Prebisch, 1950;Singer, 1970). This led Cardoso (1977: 14) to emphasize the importance of the 'historicosocial process' in dependency analysis.
In line with this, a core motivation in the neo-Marxist approach to dependency theory was to challenge the stageist view that many Marxist authors held at the time, namely that the developed nation shows the underdeveloped 'the image of its own future' (Marx, 1967: 8-9). It was Frank (1967b) who popularized the neo-Marxist approach to dependency theory, largely building on Baran and Sweezy,8 further laying the foundations for the work by Amin, Emmanuel, and others. This strand of dependency theory can also be considered a part of the vast literature on imperialism, which clearly also has a strong global historical approach (see Amin, 1974Amin, , 1976Patnaik and Patnaik, 2016;Quijano, 1974;Smith, 2016). 9 In terms of method, dependency research's emphasis on global, dynamic and interactive analysis goes against the idea of isolating specific variables for hypothesis testing (Cardoso and Faletto, 1979;Kay, 1989;Vasconi, 1971). As Valenzuela and Valenzuela (1978: 556) put it, dependency research is the study of structural relations which are intimately related, and 8. Baran (1957) is usually considered to be the origin of the Marxist dependency tradition (e.g. by Palma, 1995), although there were several Latin American Marxists that made similar arguments in the 1940s, such as Bagú (1949). The work of Baran and Sweezy (1966) also had important ramifications in the works of Bambirra (1974), Dos Santos (1970), Marini (1978) and Quijano (1974), thus muddling the distinction between Latin American structuralist and American neo-Marxists. 9. Colonial drain theory, which can be seen as a precursor to dependency theory, also exemplified the global historical approach with its focus on the role of the colonial legacy in shaping structures of production, investment and consumption (Dasgupta, 2001;Pasricha, 2008).

8
Ingrid Harvold Kvangraven how they evolve over time. Therefore, it cannot be 'tested' by doing crosscountry regressions at one point in time. However, anyone reviewing dependency research will find a vast array of concrete, empirical, regional and country studies, as well as comparative studies.

Core Hypothesis: Capitalist Development Tends to be Polarizing
At the core of dependency theory as a radical research programme is the attempt to explain why global economic development has a tendency to be polarizing rather than equalizing. Having an economic theory of uneven development is a necessary aspect of the programme, as without it the research programme is reduced to mere descriptions. 10 The vast literature on dependency includes a variety of theoretical propositions regarding what is driving global unevenness. The Latin American structuralist dependency theory, sometimes called dependistas or dependencia, developed in response to the inadequacies of the UN Economic Commission for Latin America's (CEPAL) analysis, including that of Prebisch, Singer and Furtado, as well as the perceived failure of import-substitution industrialization (ISI) to spur sustained industrialization. Prebisch (1950) and Singer (1950) examined the long-term trends in relative prices in goods traded by industrialized and peripheral countries and found that the terms of trade had been moving against the latter. They pointed to several explanations for this trend. One of them was the differences in income elasticities of high value-added goods versus low valueadded goods (often primary commodities). Another was the differences in the market structure between the periphery and the centre, with the industrial markets of the centre countries being more oligopolistic and subject to greater wage rigidity, and the primary goods markets of the periphery countries more competitive and flexible. The workers in the periphery were therefore the ones that tended to absorb most of the global economy's income contraction. 11 10. The definitions proposed by Cardoso and Faletto (1979), Evans (2009) andHeller et al. (2009) are not in line with this way of defining the research programme because they take a more descriptive approach to understanding economic organization and do not explicitly theorize about the polarizing tendencies of capitalism. 11. Contrary to the most common focus on barter terms of trade, it is important to note that Prebisch (1950) and Singer (1950) argued that it was the income terms of trade that were the most important for peripheral economies. Although the thesis was quickly dubbed the Prebisch-Singer thesis, both Prebisch and Singer argue that they had not directly exchanged any ideas at the time when they were both individually working on their papers (Love, 1980). See Fischer (2015) and Ho (2012) for a comprehensive overview of their position and a review of the various explanations for the deteriorating terms of trade they identified. See Brown (1978) on the distinctions between Prebisch and Singer, particularly with respect to the role of demand. The neo-Marxist dependency theorists tend to root their economic theory in the extension of monopoly capitalism. For example, Baran (1957) explains the lack of dynamism in the underdeveloped world as a result of its insertion into the capitalist world economy, which makes it vulnerable to extraction of surplus by the centre. He argues that since foreign capital tends to control domestic markets, the periphery enters straight into the monopolistic phase of capitalist development, which is different from the competitive capitalism that Marx observed. A central distinction between competitive and monopoly capitalism is that while in competitive capitalism surplus arises from the production process, in monopoly capitalism, surplus can also arise due to the possession of monopoly power, and therefore it can arise from the exchange process as well (Sawyer, 1988). There are also important varieties within the tradition, 12 for example regarding the (partial) survival of feudalism in the periphery. 13 Amin (1974) and Emmanuel (1972) popularized the concept of unequal exchange, with modifications from Prebisch's original idea (Brown, 1978;Evans, 1975). Broadly, unequal exchange theories attempt to explain factor price non-equalization in the world economy, where factor price refers to the remuneration to labour or other primary non-produced factors. Versions of unequal exchange originating within the dependency tradition are often based on some concern with monopoly power, surplus extraction and core-periphery trade relations (Amin, 1974;Bambirra, 1978;Frank, 1967a;Marini, 1978). 14 Amin (1976) was among the first to measure unequal exchange empirically, which has later been done by many, including Gibson (1980), Higginbottom (2014), Nakajima and Izumi (1995), Patnaik and Patnaik (2016), Ricci (2018) and Williams (1985). While Polanyi Levitt recently noted that the existence of unequal exchange in the global economy is 'obvious' (Fischer, 2019: 558), this form of theorizing is largely marginalized from mainstream discussions of development. 15

See Kvangraven (2019) for more on the distinctions between Amin, Arrighi, Frank and
Wallerstein, who were known as the 'gang of four'. 13. This debate may be considered an 'auxiliary hypothesis' within the research programme, as the survival or not of feudal structures does not necessarily impact the core hypotheses of the research programme. This is an old debate, with significant interventions made by Baran and Sweezy (1966), Frank (1967b), Luxemburg (1951) and Wallerstein (1976), in addition to less well-known contributions such as Boeke (1953) in his writings on Indonesia, Patnaik and Patnaik (2016) in their writing on imperialism, and James ' (1963/1989) study of the Haitian revolution in which he described the slaves as the most proletarianized workers in existence at the end of the 18th century. 14. There have been other explanations for unequal exchange as well, such as intersectoral wages (Lewis, 1954), intersectoral profit rates (Prebisch, 1950;Singer, 1950), capital composition (classical Marxists such as Bettelheim, 1972;Shaikh, 1980), in addition to wage differences (Emmanuel, 1972) and international profit rates (the neo-Marxists). 15. There is also the interesting spin-off ecological dependency literature that is based on unequal exchange, pioneered by Bunker (1985).
Dependency theories that explain polarizing tendencies in the global economy do not claim that it is impossible for some countries to catch up, but rather that it is unlikely and difficult. While there has been some convergence among clubs of countries since the 1980s, the main consensus is still that there is no strong trend of gradual global convergence in income levels (Goda and García, 2017;Islam, 2003;Johnson and Papageorgiou, forthcoming;Popov and Jomo, 2018;Pritchett, 1997). Furthermore, the fact that between-country inequality is significant for explaining the high global inequality levels (Anand and Segal, 2015;Milanovic, 2015) calls for a research programme that takes these inequalities as a starting point.

Focus on Structures of Production and Constraints to Development
While the polarizing nature of capitalist development is at the core of dependency theory, it is approached with a particular focus on structures of production and peripheral constraints to development. The focus on structures of production to explain uneven development dates back to the 1600s with the work of Antonio Serra and others (Reinert and Patalano, 2016). Structures of production include structures of the labour markets, social relations of production and the characteristics of the goods that are being produced and consumed in an economy.
While Prebisch and Singer are most well known for pointing out that countries in the periphery face unfavourable terms of trade, they argued from the beginning that it was the structure of production that was the 'fundamental economic problem' that impeded autonomous industrialization (Prebisch, 1950: 8). By structure, they meant the technological capacity and organization of labour markets, which led to certain specializations and productivities (Singer, 1992). 16 A core concern for structuralist dependency theorists was the international dualism resulting in uneven technological development in the centre and periphery (Dos Santos, 1970;Prebisch, 1950;Singer, 1953;Sunkel, 1969), as well as the importance of the improvement of human capital as a way to achieve increasing returns to investment (Singer, 1965). 17 The neo-Marxists' grounding in theories of monopoly capitalism naturally also have structures of production at their core.
The final defining feature of the dependency programme is the attention to specific constraints that peripheral economies face and the relationship 16. In this connection, the deterioration of the terms of trade was seen as something that aggravated the development problem, although it wasn't the essence of the problem. For more recent elaborations on this distinction, see Ocampo and Parra-Lancourt (2009). 17. Although Singer here used similar terminology as the endogenous growth literature, he subsequently distanced himself from what he considered to be the mainstream 'human investment model', because it treats human welfare as an instrument and an increase in GDP as the end goal (Singer, 1965). See Ho (2012) for an explanation of how Singer's approach to 'human capital' evolved.
between these constraints and the domestic structures of production. Such constraints include several of the concepts already mentioned, such as technological dependence, unequal exchange and falling terms of trade for exports from the periphery. Financial constraints were also important for many dependency theorists. Within the financial dependency view, it is the lack of finance, in particular foreign finance and the limits imposed by balance-of-payments constraints, that leads to low levels of growth (Vernengo, 2006). Tavares (1985), for example, who was critical of what she considered to be an excessive focus on technological dependency by some of the Latin American structuralists, considered the real obstacle to development to be the financial dependency reflected in the inability of peripheral countries to borrow in their own currency on international markets (see also Suter, 1989), 18 as well as the limits of monetary policy in the periphery. 19 There have been echoes of some of the problems associated with financial dependence by scholars from a variety of traditions, including Wade and Veneroso (1998) and more recently Stiglitz (2017) and Chandrasekhar and Ghosh (2018). Even though it is possible to identify specific constraints to development, an important part of the research programme is to study how these may vary across time, place and context, for example due to geopolitics or other circumstances. Another example of a paradigm that recognizes both polarizing tendencies of capitalism as well as constraints to development is world systems theory (Chase-Dunn, 1982;Goldfrank, 2000;Wallerstein, 1974Wallerstein, , 1976. However, world systems theory has been critiqued for neglecting structures of production (in favour of exchange) and for letting the superstructure determine the base, rather than there being dynamic interactions between the local and global (Pieterse, 1988;Skocpol, 1977). Because of this, work within the world-systems tradition may be a part of the dependency research programme when it recognizes those factors, but it is not necessarily so (see Table 1 below).

The Research Programme's Relation to other Development Theories
To illustrate what is unique about the dependency research programme, it is worth comparing it to other development theories that have similar elements. As Table 1  encompasses more than just the traditional dependency theorists, but it also excludes many related approaches to development, both mainstream and heterodox. For example, consider Ha-Joon Chang's (2002) historical work. Although Chang combines a global historical approach with analysis of economic structures and a recognition of the specific constraints faced by peripheral economies, his approach lacks a theory of uneven development to explain polarizing tendencies in the global economy. Without such a theory, Chang's work at times appears to suggest that every country can simply follow the path of the East Asian 'miracle' countries if given the policy space to do so. Furthermore, there is some overlap between global value chain (GVC) analysis and the dependency research programme, although much of GVC analysis takes a more technical approach to production than a dependency research programme would, as it tends to not consider the political economy constraints peripheral economies face when attempting to move up GVCs. Similarly, theories of economic complexity provide support to the propositions of the structuralist view that states' production sophistication is a central way to overcome underdevelopment (Gala et al., 2018), but the approach rarely goes beyond describing the characteristics of goods produced. Furthermore, it is limited in its ability to explain the origins and persistence of uneven development due to its lack of historical analysis and lack of attention to the external and internal constraints to development. North-South models have a distinct similarity to dependency theory as they are grounded in the structural and institutional characteristics of countries at different 'levels' of development, often producing the unorthodox result that 'the South lags behind the North indefinitely' (Darity and Davis, 2005: 143). 20 These models are good examples of approaches that consider external constraints as well as economic structure with a theory of uneven development, but that still may not be considered to be a part of the core of the research programme because they do not take a global historical approach.
Notably, there are theories one could draw on that are not explicitly 'dependency theory' that also lay out comprehensive explanations for why capitalism tends to be polarizing, such as classical Marxist theory (e.g. Brenner, 1977;Shaikh, 2016;Weeks, 1981) or the more recent theories of subordinate financialization (Bonizzi et al., 2019;Powell, 2013). While classical Marxists tend to be critical of dependency theories because of their reliance on concepts of monopoly capitalism (e.g. Shaikh, 1980) or their allegedly excessive focus on exchange rather than modes of production (e.g. Brenner, 1977), they often do include many pillars of the dependency theory research programme in their approach. However, many strands of classical Marxism do not engage with the particular constraints faced by peripheral economies. The 'subordinate financialization' view is a notable exception to this, as it recognizes that peripheral economies will experience the tendencies of financialization in a distinctive form, which has been shaped by imperial relations in the current world system. 21 Finally, a politically related strand, but one that cannot be a part of the research programme as defined here, is the neostructuralism that emerged out of CEPAL in 1990 (see ECLAC, 1990). This is because neostructuralism breaks with the basic premise of capitalism being a polarizing force and often does not focus on domestic economic structures of production (Leiva, 2008). 22 Table 1 also poses an interesting challenge: there are theories that are a part of the dependency research programme whose proponents may not sympathize with dependency theory as such. However, given that a lot of 20. The models differ in theoretical starting points, from Findlay's (1980) synthesis of neoclassical trade models with elements from dependency theory, Taylor's (1981) attempt to formalize the structuralist tradition, to Dutt's (1989) formalization of some of the arguments by Baran and Emmanuel. 21. In line with the dependency research programme, Powell (2013: 44) argues that the impact must be understood in relation both to changes at the macroeconomic level as well as to the 'sectoral transformations characteristic of financialization'. 22. As Leiva (2008: xvii) puts it: 'by renouncing structuralism's methodological legacynamely a focus on how economic surplus is produced, appropriated and distributed within a single, world capitalism economy -Latin American neostructuralism becomes analytically impotent in adequately explicating the scope of the qualitative transformations experienced by Latin American capitalism over the past decade'. animosity towards dependency theory comes from disagreements with specific theories of dependence, those scholars may not object to being associated with the dependency research programme as defined here. A related challenge is the fact that scholars who self-identify as a part of the dependency school may be excluded from the research programme if they do not fulfil the 'core' criteria outlined. While this may be inconvenient from a history of thought point of view, it is necessary in order to push the research programme forward in a constructive direction -that is, by preserving the tradition's strengths. Nonetheless, the vast majority of dependency scholarship would be included in the dependency research programme.

WHY WAS DEPENDENCY THEORY DISMISSED? ADDRESSING DEPENDENCY THEORY'S CRITICS
By and large, one can point to three major reasons that often appear as explanations for the decline of dependency theory, namely political reasons, empirical changes in the world economy and epistemological critique. Leaving the political reasons for exclusion aside for now, and dealing with the empirical reasons in the next section, this section delves into the epistemological critique of dependency theory. These critiques are vast and come from both inside and outside the tradition. 23 However, much critique of dependency theory is based on an incomplete, superficial, and at times incorrect understanding of what dependency theory is, which led Cardoso (1977: 15) to argue that the common simplification and misunderstanding of dependency theory had made it 'a straw man easy to destroy'. Indeed, critics of dependency theory often assume a theoretical unity that doesn't exist among dependency scholars. 24 It is precisely this stereotyping of dependency theory that this article aims to refute and move beyond. The critiques of dependency theory can be roughly categorized as being about tautology and precision, economic reductionism, and the lack of agency associated with an overly strong focus on the 'external'. 25 23. The 1981 special issue of Latin American Perspectives (Chilcote, 1981) illustrates the heated and polemical nature of the debates within dependency theory, with each side (structuralist, classical Marxist, neo-Marxist) accusing the other of being 'outdated', for using 'unquestioned formulas', 'eclecticism', for acting as 'ideological cops', etc. Ideological attacks are prevalent in debates outside of the special issue as well, for example, Marini (1978) calls Cardoso the shield-bearer of the bourgeoisie because he appears to be disputing the necessity of a socialist revolution, to which Cardoso and Serra (1978) replied that Marini's position is ineffective and dangerous. 24. For example, Lall (1975: 800) accuses the 'dependencia school' of 'internal confusion', as 'dependence' is given different meanings depending on the scholar who is writing, rather than accepting that there is debate about what dependency is. 25. For other attempts to rectify misrepresentations and misplaced critiques of various aspects of dependency theory, see Fischer (2015), Henfrey (1981), O'Brien (1975) and Palma (1978). Lall (1975) argues that the concept of 'dependence' can only serve as a useful analytical tool if it satisfies the following two criteria: 1) it must identify characteristics of peripheral economies that are not found in centre countries; and 2) those characteristics must be shown to adversely affect the pattern of development of the periphery countries. Lall argues that dependency theory does not satisfy these two criteria, but rather selectively misdirects analysis to certain features of the broad phenomenon of capitalist development. 26 In addition to being imprecise, the dependency literature has also been attacked for being circular, as the argument goes that periphery countries are poor because they are dependent, and that any characteristic they display is a sign of dependence (Booth, 1985;Brenner, 1977;Lall, 1975;Robinson, 2011). While this is not true for all dependency scholars, it is clear that the critique is valid for some cases. For example, while Frank (1972) argues that dependence on exports is a characteristic of a dependent economy, export dependency is not particular to developing countries. Similar characterizations without independent explanations can be found in Dos Santos (1970) and even Taylor (2015). 27 Notably, much of the critique of dependency theory for being imprecise and simplistic is directed particularly at Frank (Brenner, 1977;Chilcote, 1974;Laclau, 1971) and Wallerstein (Appleby, 1975;Brenner, 1977;Pieterse, 1988;Robinson, 2011;Skocpol, 1977). However, most dependency theorists do go beyond observations to causal explanations, as they argue that it is the historically produced structure of production that generates dependence, in combination with external constraints. In fact, many dependency scholars would be in agreement with Lall's two criteria, as these aspects have been a part of much of the internal discussions. Furthermore, with economic theorizing about uneven development being placed at the centre of the dependency research programme, it may be able to move beyond the weaknesses of some dependency theories. 26. Similarly, C.  argues that 'dependency' is used as an 'interpretative wild card ', and O'Brien (1975) calls it a 'pseudo-concept' that explains everything in general and hence nothing in particular. Smythe (1981) and Collins (1986) also make similar critiques of different strands of dependency theory. 27. Dos Santos (1970) uses 'dependency' in all the following ways: 'the international situation of dependency', 'dependent structures', 'mercantile dependency', 'financial industrial dependency', 'colonial dependency', 'financial dependency', 'technological industrial dependency', 'the dependent system of production', 'the dependent economic system', and 'the development of dependent capitalism'. Similarly, Taylor (2015: 11) does not explicitly define dependence, but appears to equate it with 'dependency on primary products'.

Economic Reductionism
Dependency theory has been accused of underestimating culture and overemphasizing economics and politics (Grosfoguel, 2000); of failing to bring out the way in which imperialist relationships were internalized in the social relationships within societies (Panitch and Gindin, 2003;Shivji, 2016); of representing a techno-scientific understanding of modernity (Gulalp, 1998); and of explicitly or implicitly maintaining that the dependent countries' social and political structure is determined by, or derivative from, its economy (Leys, 1977;Staniland, 1985). Similarly, critiques have been levelled against dependency theory for ignoring both race (Bonilla and Girling, 1973) and gender (Scott, 1995). While structuralist dependency theory has been critiqued for focusing too much on the nation state (Blaney, 1996;Chase-Dunn, 1982;Evans, 2009;Palma, 2016), it has also been criticized for having a weak analysis of the nature of the state (Ndlovu-Gatsheni, 2013). There are two important responses to this set of critiques that point towards the importance of the research programme. Firstly, one of the strengths of the dependency approach is the focus on production structures and their geographical character, which is an important aspect of development that is often neglected in contemporary analysis. Thus, criticizing the theory for focusing precisely on these structures is not terribly convincing as a fundamental critique. Furthermore, to deny that national and regional units are relevant for understanding development means denying that these units are important for understanding historically produced structures associated with underdevelopment -which persist to this day (Margulis, 2017;UNCTAD, 2016). Moreover, within the dependency research programme it is possible to go beyond the focus on the national, to investigate centreperiphery dynamics within countries as well (Furtado, 1970). Of course, there are also parts of the research programme that adopt a more global approach, with less focus on the state, such as the approaches that consider the global economy as one single system (e.g. Frank, 1978;Sunkel, 1973;Wallerstein, 1974).
Secondly, it is important to note that political and cultural factors were a part of the analysis of dependency theorists from the beginning. Furtado (1970Furtado ( , 1973 was arguably the most influential dependency thinker on issues of colonialism and culture, which are ultimately linked to social relations of production and consumption. 28 Other structuralist work on cultural dependency includes Corradi's (1971) work on the culture and 28. Furtado (1973) defined demand as cultural colonization, by which he meant that the elite and upper middle class avidly buy goods that are consumed by the affluent in the advanced countries (see also Prebisch, 1950;Sunkel, 1969). One can also find traces of cultural dependency in Baran (1957) and Naoroji's (1917) focus on the stagnationary effect of local elites' conspicuous consumption (Pasricha, 2008), and Innis' (1950) and Smythe's (1981) ideology adopted by local elites, Quijano's (1974) study of ways that dependency relates to ideologies such as developmental nationalism and developmentalism and Vasconi's (1977) work on the education system as an ideological apparatus. Regardless, there are important insights to be drawn from studies of coloniality which are related to situations of dependence, although they draw attention to slightly different development problems and aren't necessarily a part of the dependency research programme as such (see Table 1). This does, of course, not rule out important overlaps and synergies between post-colonial literature and a dependency research programme (Kapoor, 2002;Mignolo, 2000;Slater, 2004). In addition, dependency theorists often critiqued CEPAL structuralists for assuming a technocratic developmental state rather than having an explicit theory of the state, which they considered to be central for understanding situations of dependence. Furthermore, while large parts of the dependency literature did not incorporate race and gender, there were also important parts of the literature that did precisely that. For example, part of the literature on internal colonialism, which was part of the dependency debates, refers specifically to racial inequalities (Casanova, 1965;Cotler, 1967;Stavenhagen, 1965). 29 The dependency research programme would benefit from building further on this work to consider how racism and sexism shape the structures of production and how this affects constraints to development in the periphery.
Finally, while the post-development critique might be right to caution against establishing laws of dependency or of exploitation in the social sciences, as it may lead us to lose sight of the historical, political and institutional specificities of uneven development, the strongest strands of dependency theory have been those that have been able to combine economic theorizing with historical, political and institutional analysis of particular situations of dependence.

The External versus the Internal and the Question of Agency
A critique often levelled at dependency theory is the relative absence of agency. 30 These critics claim that the way dependency scholars describe the dynamics of international capitalism leaves little room for action by the state or social groups. This is, however, simplistic if it is meant as a general critique, as it is only applicable to some parts of dependency theory, and is work on the role of popular media in maintaining dependent relations between Canada and the US. 29. See Chapter 3 in Kay (1989) for more on how parts of the Latin American dependency literature incorporate race relations. 30. Different variations of this critique can be found in Angotti (1981), Booth (1985), C. , Mkandawire (2001), Petras and Brill (1985), Pieterse (1988), Sanchez (2003), Skocpol (1981), Smith (1979). most relevant for world systems theory (Kay, 1989). 31 As discussed, dependency theorists placed the relationship between the internal and the external at the centre of their analysis, and they conducted an in-depth discussion of historical and country-specific political factors. This point was also made forcefully by Cardoso (1977), who accuses Frank of failing to draw together the general and the particular, but sees Frank as an exception within the dependency school. 32 Cardoso (1977) argues that the critique of dependency theory for overemphasizing the external stems in part from the misrepresentation of dependency theory in the US. He points out that 'Dependency analyses … were preoccupied much less with external conditioning of the Latin American economies, which was taken for granted, than with the development of a type of analysis that could grasp … the movement of structures within the dependent countries' (ibid.: 12). Similar defences can be mounted for dependency theory outside of Latin America, such as that of Amin (1973), as well as newer strands of the research programme that carefully examine internal as well as global structures in their analyses of dependency (Bonizzi et al., 2019;Patnaik and Patnaik, 2016;Powell, 2013).
These criticisms are wrong to assume that dependency theorists generally believed external factors were the prime determinants of underdevelopment and dependence, as this was a minority view within the tradition (but popularized by widely read scholars such as Frank). Indeed, the core insight from dependency theory is that internal dynamics must be examined in relation to the dynamics of the centre countries (Cardoso, 1977;Dos Santos, 1970;Kay, 1989;Vasconi, 1971). This is essential, because while it is true that context matters, a dependency perspective forces us to see the local context in relation to systemic global processes that 'constrain peripheral countries in similar ways despite their diversity' (Fischer, 2015: 727). Furthermore, how the internal and external relations interact to enforce certain patterns of industrialization was a key issue for dependency theorists across schools.

HOW IS THE DEPENDENCY RESEARCH PROGRAMME RELEVANT TODAY?
The core hypothesis of the dependency theory research programme remains pertinent today given the persistence of a highly uneven distribution of types 31. The lack of space for agency in world systems theories has been called 'structuralist superdeterminism' (D.L. Johnson, 1981: 112) and 'the tyranny of the whole' (Smith, 1979), because the theories fail to account for complex interrelation of international and national forces through history. 32. However, even in some of Frank's work (e.g. 1967b), he analyses the interaction between internal and external power structures in several in-depth cases studies.
of production across the world, along with overall rising global inequality. To demonstrate the relevance of the research programme, this section considers two empirical developments in the global economy that are often invoked as reasons for why dependency theory is no longer relevant, namely the transition from periphery to centre by some traditionally peripheral countries and the development of an integrated global production system. This article argues that these trends can, in fact, be fruitfully understood through a dependency research programme. Most focus is placed on the development of GVCs, given that this has become a central research area within development studies. Furthermore, given that the focus in the mainstream GVC literature tends to be narrowly on industrial upgrading as a matter of firm-level competitiveness (Bair, 2005;Neilson, 2014), the dependency research programme can offer guidance for a renewal of development studies by bringing in broader structural questions of how inequality is produced and reproduced in the global economy.

The Shift from Periphery to Centre: The Case of South Korea
Dependency theory is often criticized for failing to account for the fact that some of the traditional periphery countries have developed over the past half century, suggesting that it is possible to break out from dependence (e.g. Amsden, 2003;Booth, 1985). However, the ability of some countries to transition from the 'periphery' to the 'centre' does not invalidate the core hypothesis of the dependency research programme -that capitalism tends to generate uneven development. To understand how some countries were able to defy this tendency, it is necessary to dig deeper into how the structures of production were transformed and how constraints to development were relaxed in these cases.
Within a dependency research programme, a global historical approach is essential. While there has been much heated debate regarding the success of Korea, most accounts simply start in 1960 and focus on domestic policies, and most mainstream accounts tend to focus solely on the role of institutions and market-based policies. 33 Within the dependency research programme, in contrast, it is imperative to go back to the development of capitalism in Korea. An important element in this development was Japan, Korea's colonizer, which was actively attempting to conquer China (Eckert, 1990;Kohli, 2004). It was beneficial for Japan to integrate the Japanese and Korean economies, which required industrialization in Korea. Faced with opposition from protesting Koreans in 1919, Japan played the capitalist class against the agrarian sector and included Koreans in the industrial commission of 1921, which opened the door for Korean industrial capitalism and the development of a Korean capitalist class with a financial structure intricately linked to the state. Korean businessmen were thus 'not so much subordinated by the political structure as incorporated into it' (Eckert, 1990: 125). This is in stark contrast to how capitalism developed in other peripheral economies, which shaped local industry in a much more exploitative and extractive manner (Amin, 1974;Frank, 1967a;Kohli, 2004;Rodney, 1972).
This historical development of capitalist production structures laid the foundations for the industry that later emerged as a part of the welldocumented developmental state (Amsden, 2001;Chibber, 2003;Kim, 2010;Wade, 1990). During this period, Korea actively managed its trade by using both import substitution and export promotion policies, thus largely following the policy prescriptions of dependency theorists (Amin, 1990;Ho, 2012;Margulis, 2017). In this way, the structures of production were ultimately shaped in a very different way in Korea than in other parts of the periphery at the time. Understanding how the production structures were historically and politically shaped within the global economy leads to a much richer and deeper understanding of Korea's successful industrialization process than approaches that attempt to measure Korea's policies, human capital or institutions at a certain point in time.
Furthermore, the constraints often prevalent in peripheral industrialization that tend to lead to uneven development (e.g. technological dependence, lack of spillover effects, enclave economies, foreign ownership and extraction, financial dependence) were mitigated in the case of Korea in part because of geopolitical factors (Doucette and Müller, 2016;Glassman and Choi, 2017). For example, given that peripheral industrialization has a strong propensity to generate trade deficits (Fischer, 2009(Fischer, , 2018Furtado, 1956Furtado, , 1973Prebisch, 1950), it was important for Korea that it was able to run a large trade deficit for a long period during its industrialization process. Korea's industrial strategy involved substantial imports of foreign licences in a concerted effort to ensure national ownership and break out of technological dependency (Amsden, 1989), and contrary to popular belief, Korea's industrialization process in the 1960s and 1970s was import-rather than export-led (Fischer, 2018). Korea was able to run a deficit during this period because of the ample supplies of external finance it was able to access at that time for geopolitical reasons (Amsden, 1989;Fischer, 2018;Rhee, 1973). While this helped alleviate the challenge of mounting balance-of-payments deficits that many peripheral economies face, it was also essential that Korea was allowed to protect its markets from the global economy and strategically manage them (Wade, 1990).
As pointed out by Heller et al. (2009: 289), Chibber (2003), Evans (1979Evans ( , 1995 and other developmental state scholars owe a direct debt to Cardoso and Faletto's claim 'that under certain circumstances states could become The Relevance of the Dependency Research Programme 21 proactive agents of accelerated development, even in contexts of overall dependency'. While the Korean experience does contradict the caricature version of dependency theory, it does not contradict the dependency research programme. As Evans (1994: 449) put it, if 'we confront the East Asian experience with an historical-structural or Cardosian version of the dependency approach, we find the results for that approach more confirmatory than contradictory'.

Global Production Networks and Industrialization in the Periphery
A new international division of labour has emerged, characterized by the restructuring of global production networks, which has allowed many developing economies to move into manufacturing through participation in far-flung global value chains (Arrighi et al., 2003;ILO, 2015;Milberg and Winkler, 2013). This development has been used as an argument against the relevance of dependency theory, which was developed during a time when global production was less integrated (Gereffi, 1989;Robinson, 2011). The argument goes that the allegedly decentralized character of global commodity chains renders centre-periphery analysis superfluous. This section outlines how a dependency research programme remains helpful to understand global inequalities and how it can fruitfully engage with the rise of GVCs by going beyond the contemporary GVC literature.
Although GVC analysis first emerged as an attempt to account for the structure and dynamics of the global economy within world systems theory (Fröbel et al., 1980;Hopkins and Wallerstein 1986), it has in large part lost its original macrohistorical perspective as it now centres more on organizational analysis of firms and industries, and thereby diverts attention from global patterns of uneven development (Bair and Werner, 2011;Selwyn, 2013;Suwandi, 2019). Indeed, the mainstream GVC literature tends to explain the possibilities for development within GVCs as deceptively straightforward, often primarily drawing on the experiences of the high-performing East Asian economies and assuming replicability (Gereffi, 1999).
While the structural shift in the global economy has involved the peripheral economies shifting into manufacturing activities (UNCTAD, 2016), it is a common misreading of dependency theory that it saw peripheral economies as doomed to primary commodity exports. What scholars in the dependency tradition were concerned with was the technological content of production (e.g. Prebisch, 1950), the ownership of production and how that impacted the domestic economy (e.g. Singer, 1950), potential for spillover effects and increasing returns of domestic production (Baumgartner et al., 1976), and how these structures relate to power, the global economy, and economic development more broadly. Although the related 22 Ingrid Harvold Kvangraven constraints may manifest themselves in different ways today, assessing such constraints in relation to both domestic and global factors is still pertinent. 34 For many developing economies this shift has been associated with a fall in the share of domestic value added in exports (Caraballo and Jiang, 2016). Indeed, recent research shows that production structures across the periphery are characterized by the production of goods with relatively low technological content, including countries in Africa (Hauge, 2019;Morris et al., 2012), South Asia (Abdon and Felipe, 2011), Latin America (Ocampo, 2014) and China (Fischer, 2015). In addition, many contemporary studies have picked up on the terms of trade debate started by Prebisch and Singer, and documented falling terms of trade for periphery exports -despite the fact that they are now often manufacturing exports (Erten, 2011;Lewis, 1978;Ocampo and Parra-Lancourt, 2009;Ram, 2004;UNCTAD, 2002UNCTAD, , 2016. The dependency research programme allows us to explore these phenomena through a global historical approach, pointing to how relations of production have evolved over time, taking different forms, but for the large part maintaining peripheral characteristics. Indonesia has among the largest labour share of GVCs in the world, after India and China (Suwandi, 2019). Trends in FDI and employment in the industrial sector have been increasing in Indonesia over the past decades, and manufacturing value added has increased as a share of GDP from approximately 9 to 24 per cent from 1960 to 2012 (ibid.). As the GVC literature on Indonesia has been mostly on Indonesia's opening up to foreign companies and firm-level analysis, it has remained silent on the role of the state in mediating development strategies, national processes that affect GVC participation, power asymmetries within chains, and global circumstances that shape upgrading possibilities (Neilson, 2014).
A global historical approach is crucial for understanding Indonesia's transformation, as both its colonial legacy as well as its developmentalist period post-independence have had lasting structural impacts on the economy. The formation of Indonesia's extractive regime began in the colonial period, and strikingly the five leading exports that accounted for 68 per cent of exports in 1900 still accounted for 65 per cent of exports in 1990 (Gellert, 2010). After independence from The Netherlands in 1945, Indonesia experimented with economic nationalism and anti-imperialism and the domestic capitalist class and state-owned capital were crucial factors in laying the foundations for Indonesian industrialization (Robison, 1986). In the 1950s and 1960s, developmentalism was considered Indonesia's collective project (Evans, 1995), espoused by the state elites, where by Sukarno pursued efforts to create indigenous national economic capacity, including The Relevance of the Dependency Research Programme 23 by nationalizing key industries (Robison, 1993). After Suharto took over in the late 1960s, he promoted large-scale exports and resource-based industrialization to build the legitimacy of his regime (Neilson, 2014).
If one wants to understand the successful upgrading from logging to plywood that Indonesia underwent, which represents a significant and impressive structural transformation and upgrading, one certainly must look beyond firm-level developments. For example, Gellert (2003) points to the historical timing of its entrance into global markets (when sources were dwindling), the fact that Indonesia had the second largest resource base in the world, state support, and collaboration with private firms. However, this successful upgrading in the logging sector took place alongside the persistence of low productivity sectors, including raw materials and petty production, and a small domestic market, reminiscent of the dualism dependency theorists observed in the periphery that constrains sustainable industrialization (Boeke, 1953;Singer, 1953;Sunkel, 1969).
While Indonesia faced several important constraints to industrialization, such as technological dependence, high inequality and a small domestic market, during its structural transformation in the developmentalist era it saw some of the common financial constraints relaxed. Both oil revenues and geopolitics were important in this regard. The oil revenues made it possible for the government to finance a decade of ISI policies, although when push came to shove during the balance-of-payments and fiscal crises of the 1980s, Indonesia was forced into structural adjustment reforms and shifted towards more of a manufacturing-led export-oriented industrialization along with raw material exports (Neilson, 2014). The geopolitical and ideological importance of Indonesia to the West during the Cold War was also key, which ensured that Suharto's regime received substantial foreign aid from Western countries and was the leading recipient of Japanese aid and loans (Gellert, 2010).
However, important limits to industrialization persisted, despite some constraints being relaxed. An example is that the diffusion of industrial activities was conducted within hierarchical structures of corporate control (Hart-Landsberg and Burkett, 1998;Suwandi, 2019). Because of this, development within Indonesia was not determined according to the industrial needs of the economy, but rather in line with the interests of foreign capital -in particular Japanese MNCs (Gellert, 2010). Illustratively, the manufacturing exports' explosion from 1980 to 1992 was not so much due to an internal transformation of the Indonesian economy, but rather in large part due to Japan's use of export platforms in Indonesia. This export production continued to depend on Japanese technology, and export earnings and profitability eventually began to decline by the mid-1990s due to overproduction and competition from China and elsewhere (Glassman, 2003). It is therefore important to challenge the 'multi' part of multinational, given that MNCs generally remain national in their governance structure, as the centres of management and advanced 24 Ingrid Harvold Kvangraven technological research of these corporations are still concentrated in the centre (Screpanti, 2014).
Because of this, the manufacturing sector of Indonesia is characterized by limited technological capability and Indonesia remains a net importer of advanced technologies (Jacob, 2005;Wie, 2005). As recognized by Amin (1976), Furtado (1956) and many others in the dependency tradition, the development of such technologies is crucial to generate and sustain industrialization and growth. Because of Indonesia and other peripheral economies' lack of relative technological capabilities, the transformations of global production networks are perhaps more in line with Arrighi's (1990: 24) observation that the spread of industrialization 'appears not as development of the semiperiphery but as peripheralization of industrial activities'.
Meanwhile, in contrast to a lot of developing countries which are still exporting unsophisticated, highly standard products that offer little potential for upgrading through diversification (Felipe et al., 2012), China has made significant advances in terms of upgrading. However, even in the case of China, although it has massively expanded its manufacturing exports based on its integration into global production networks, this expansion has involved a strong dependence on FDI, rapid denationalization of the export-oriented manufacturing sector and relatively low levels of domestic innovation incorporated into exports (Ferrarini and Scaramozzino, 2015;Fischer, 2015). Indeed, Fischer (2015) makes the argument that even the development of China can be fruitfully understood through a lens of peripherality.
This brings us to the importance of a perspective that seeks to explain the polarizing tendencies of capitalism to shed light on why upgrading may not automatically occur in manufacturing sectors of the periphery. There are several perspectives that could be pursued in this regard, for example theories of imperialism (which Suwandi, 2019 applies to Indonesia), monopoly capitalism (Durand and Milberg, 2019), real competition in international trade (Shaikh, 2016), or other dependency-related theories (e.g. drawing on Latin American structuralists or neo-Marxists). In addition, the dependency research programme requires attention to what were previously key tenets of the GVC literature, such as the institutional context of the chains, the role of the state, industrial upgrading as a development strategy, analysis of ownership and how power is exerted in the chain, and a concern for world-historical processes that shape opportunities for inclusion and exclusion in global development (see also Neilson, 2014). As the case of Indonesia demonstrates, development strategies are contingent on a range of factors, such as world economic conditions, national history, resource endowment and institutions. Recognizing these issues within a dependency research programme leads to a strong argument against a one-size-fits-all approach to GVCs.

CONCLUSION: REFLECTIONS ON THE PROGRAMME'S RELEVANCE AND DEMISE
This article has demonstrated that defining dependency theory as a research programme provides an alternative way of categorizing dependency scholarship that captures the breadth of the scholarship as well as its strengths. This research programme -characterized by 1) theorization on the persistence of uneven development; with a focus on 2) the specific constraints peripheral countries face; and 3) structures of production; with 4) a global historical approach to these issues -provides a foundation from which to fruitfully explore important questions related to development and global inequality. While each of these elements may be found in other theories of development, it is the combination of the four characteristics that makes the dependency approach especially powerful and relevant for understanding development problems.
Despite the marginalization of dependency theory since the 1980s, this article demonstrates that the research programme holds important contemporary relevance for a renewal of development studies. In fact, analysis within the research programme allows us to unpack and explain the persistence of asymmetries and constraints in the global economy that adversely affect the development prospects and opportunities of peripheral economies. What is more, this article also demonstrates that important empirical patterns associated with dependency theory, such as technological lagging, declining terms of trade, and the uneven nature of capitalist development, have remained relevant despite the changing global economy. Furthermore, the hierarchy of forms of production and innovation that characterize the new division of labour can be fruitfully assessed through a dependency research programme, as demonstrated by the case of Indonesia.
The fact that dependency theorists do not self-identify as part of a larger research programme, but rather focus on debates within the programme and defending or refining their own position, has led to a gap in the literature with no one taking interest in defining or defending the programme as a holistic and systematic approach to development. Rather than focusing on what the underlying commonalities are, the debates have been focused on identifying differences. This has likely made dependency theory as a family of theories easier to marginalize and stereotype.
As this article shows, the main critiques of dependency theory -tautology, economic reductionism and the absence of agency -do not necessarily apply to the research programme, but are largely based on either a misreading of dependency theory or on specific theories within the tradition, and not the tradition as a whole. Therefore, they do not justify the marginalization of a dependency research programme. That said, there are things to draw on and learn from the critics, for example, regarding the role of race and gender in shaping structures of production historically, and how this affects constraints to development in the periphery. Indeed, given that epistemological and empirical critiques of dependency theory are largely misguided and misdirected, it is more likely that dependency theory was marginalized through a political process. This is in line with both Lakatos ' (1978) and Kuhn's (1962) observations that science does not necessarily move forward based on an objective measure of what the best scientific programme or paradigm is. As Lakatos noted, new programmes do not necessarily explain the same questions better, but rather explain different things from the ones previously considered. As we see with mainstream development economics today, the questions are largely different from the dependency research programme, centred around poverty reduction, basic needs and constraints at the individual or firm level -rather than structural transformation, the dynamics of capitalism, and constraints to industrialization and technological upgrading. As Chang (2011) puts it, the field of development economics has turned into Hamlet without the Prince of Denmark, if we compare it to previous development paradigms where production structures were at the core of development debates, and poverty, basic needs, and so forth were considered symptoms of structural underdevelopment.
Analysis within the dependency research programme is particularly urgent in a time when research and policy prescriptions regarding development have become increasingly segregated. For example, you will rarely have the same scholars conducting research on how capitalism has formed in a particular context, on constraints to fiscal policy in the periphery, and on technological development and dependence. The research programme on dependency can help shed light on how these elements can be intricately connected to the extent that studying only one aspect may give us inadequate answers. A comprehensive approach to understanding constraints to development necessarily goes beyond analyses of specific sectors and is therefore more likely to result in relevant knowledge production and policy recommendations.
Given that between-country inequality is significant for explaining the high global inequality levels, a dependency research programme opens the door for investigating how the changing organization of global production structures may still reproduce such inequalities. Although manufacturing production has shifted to the periphery, unequal relations of exchange continue to prevail with the dominance of a few major corporations from a small number of countries (Norfield, 2017;Suwandi, 2019). While several critics have recently pointed to the weaknesses of the mainstream GVC framework as a signal that there is a need for a new approach to global production (Bair and Werner, 2011;Selwyn, 2015), this article argues that a dependency research programme is an ideal approach to understanding the possibilities and limits that GVCs offer for development processes.
Finally, the fact that elements of dependency theory have shown up in a variety of guises, at different times and in different regions, can be seen as a testament to the relevance of concepts related to dependency theory beyond The Relevance of the Dependency Research Programme 27 discussions of Marxism or Latin America. Other emerging areas related to development can also be fruitfully addressed through this programme, such as the effects of the increase in global financial integration or uneven ecological patterns. Regarding financial integration, a subordinate financialization perspective highlights how the subordinate locations that firms in developing countries occupy in global production networks also mediates their relations to financial markets (Bonizzi et al., 2019). Regarding ecological patterns, an ecologized history of capitalism offers a promising avenue for understanding the uneven drivers and effects of climate change through an ecological unequal exchange perspective (Hornborg, 2014;Nordlund, 2014) or a 'Capitalocene' framework (Bonneuil and Fressoz, 2017;Moore, 2016). Approaching these new developments within a dependency research programme would open up interesting new directions within development studies and development economics.