Introduction
Structuralist economists challenged orthodox liberal economic theory’s predictions that the price mechanism, specialization, and trade, provided sufficient impetus to stimulate investment and sustained growth across the globe. As these scholars emphasized, despite Latin American economies’ lengthy specialization based on commodity exports, they failed to grow and industrialize as predicted by orthodox theory. This body of work, which highlighted politico-economic structural distortions as impediments to investment, was central to the case for state-led industrialization based on the theory of infant industry (Prebisch 1963). Initially, import substitution (IS) programs fostered high rates of growth and were considered remarkably successful, but within a few decades, most faltered. Implementation choices characterizing most programs led to mounting economic distortions and in many cases culminated in stagnation and debt crisis (Huber 2002).
With much water now under the bridge in development studies, it is important to underscore that this outcome was not inevitable or inherent in an IS program. The conclusion that inward orientation is inherently fatally flawed is a-historical. A few early and late developers employed import substitution in a fashion that avoided crisis and laid the foundation for sustained growth,1 and a strong case has been made that implementation choices were a central cause of the postwar IS programs’ flaws.2 The widespread preference was for a highly distortionary set of policy tools that opted for maximizing domestic production at any cost and that eschewed infant maturation. Policymakers opted for high and highly variable protection, continued subsidization of increasingly capital-intensive production, and provided no program for sunsetting incentives. Despite repeated calls from economists of all stripes to move toward lower and more uniform protection, to tie incentives to more efficient use of capital, and to adhere to the original infant industry prescriptions by gradually reducing government support, most policymakers studiously ignored these recommendations. These choices strike at the heart of the possibility of sustained growth and raise the question: why did policymakers so resolutely avoid these near-universally recommended policy changes within the rubric of ISI?3
In order to explain variation in the choice for an extreme versus a milder IS policy, I turn to a new literature on politico-economic equilibria developed by North, Wallis & Weingast (2009, hereafter NWW), which highlights heretofore largely overlooked political structural constraints shaping politicians’ incentives. By endogenizing the role of organized violence to a theory of social order, this work delineates the central role of “domesticated” versus “natural” violence in shaping the strategic environment in which politicians structure access to economic and political opportunity. This work elaborates two equilibria (Open Access Order [OAO] and Limited Access Order [LAO]) and characterizes the distinct role of organized violence within each, and how this variation constrains politicians’ choice for political and economic institutions. Most political-economy analysis procedes based on the implicit assumption that all states provide centralized neutral organized violence as characterizes OAOs. NWW’s framework illuminates how variation in the form and role of organized violence critically alters the context in which politicians make policy and other decisions. In LAOs, the absence of state monopolization and specialization of organized violence (“domesticated” violence) creates a strategic context that compels politicians to limit and directly distribute access to economic and political institutions in order to incentivize dominant players to support the regime and produce rather than fight. In this competitive context, politicians’ ability to directly allocate production rights and associated rents to powerful players becomes an indispensable political tool. When implementing import substitution programs in the context of a limited access order, politicians are compelled to carve up the economy and distribute opportunities to produce directly to powerful elites, turning IS programs into vehicles for creating production rights that can be exchanged directly in return for regime support.
As I will show, these competitive imperatives neatly explain each of the extreme IS policies, and their associated failure to reform. Extreme implementation choices allowed politicians to transform the original infant industry prescriptions into a program that multiplied the points of access to economic rights and associated rents that the state could directly distribute. Reforms that were possible within the rubric of ISI, and which appeared unassailable from the point of view of sustaining growth, were in fact precluded by the structural constraints of a limited access order. Reducing distortions and promoting infant maturation would have required dismantling these excludable production rights in favor of market mechanisms. When violence is controlled by binding powerful groups directly to the social order through exclusive access, dismantling exclusive production rights in favor of market-driven production opportunities is political suicide, and tantamount to dismantling the politico-economic glue upon which regime support is built. This argument shows how the economic inefficiencies characteristic of extreme ISI, paradoxically, were highly politically efficient in the context of a limited access order.
The paper proceeds as follows. First, I delineate the “extreme” nature of the IS policies in three key countries, Brazil, Chile, and India. Next, I introduce NWW’s two equilibria, and develop how variation in the politico-economic equilibrium leads to variation in how politicians structure access to economic opportunity. The next section shows how direct distribution of economic rights characteristic of LAOs can provide a detailed explanation for each of the extreme policies and the failure to reform. A final section discusses the argument’s consistency with the original structuralist emphasis on underlying politico-economic conditions as impediments to growth.
The Economically Inefficient Choices Within ISI
In the postwar period, the vast majority of developing countries adopted some form of import substitution industrialization, and nearly the same percentage opted for a highly distortionary version and stayed the course to the point of crisis and stagnation. In almost all developing countries, ISI was implemented in what I will call an “extreme” version of an inward-oriented strategy. These choices included: high and highly variable levels of effective protection rather than lower and more uniform protection; a continual march toward substitution of more capital-intensive goods rather than maintaining emphasis on labor-intensive production and productivity improvements over time; and providing no incentives for infant industry maturation. These “extreme” policies entailed a pronounced trade-off that was well-articulated by an ECLA economist:
While it is true that import substitution necessarily brings about a rise in prices, and that protectionism conduces, by definition, to the inefficient allocation of resources, those higher costs and this inefficiency, as well as the bottlenecks and distortions in the structure of prices and production, might on the other hand be reduced to reasonable levels and temporary status by virtue of a far-sighted and properly programmed substitution policy, and protection on rational lines. Moreover, industrialization would not then militate against the export trade, but on the contrary, would stimulate it, since the development of efficient industries would be promoted. (Macario 1964, 83; emphasis added)
Policymakers across the postwar developing world paid little heed to the many admonitions echoing this one to minimize distortions and avoid permanent infancy. As the ECLA quote succinctly notes, the “milder” version of the inward-oriented strategy could have, at a minimum, staved off the most damaging results of the typical IS program. Why did politicians nearly ubiquitously choose to implement import substitution industrialization in a fashion that undermined the policy’s stated goals of ascension in the international division of labor and sustained growth? The next section will review the key distortions of the program in Brazil, Chile, and India.
High and Variable Protection
In protecting nascent firms, all three countries used a set of overlapping tools that resulted in very high effective levels of protection with very significant inter- and intra-sectoral variation. As Table 1 demonstrates, Brazil and Chile’s average levels of protection for manufacturers were over 100%, while in India they were over 300%. A comparison to Taiwan, which employed a milder ISI, highlights the extreme protection in the other three countries, ranging from more than three to roughly ten times Taiwanese levels.
Country | Year | Average Level of Effective Protection for all Manufacturers (percentage) |
---|---|---|
Brazil | 1957-1963 | 118 |
India | 1961 | 313 |
Chile | 1961 | 133 |
Taiwan | 1965 | 33 |
Source: for Brazil, see Bergsman (1970); for India, see Bhagwati & Srinvasan (1975); for Chile, see Behrman (1976); for Taiwan see Little, Scitovsky & Scott (1970).
Levels of protection were not only very high; the inter-sectoral variation directly contradicted the prescriptions of an infant industry policy. As shown in Table 2, in all three countries, the sectors that were most fully substituted after several decades of ISI, and in which these countries had a comparative advantage, paradoxically received the highest levels of protection. Similarly, the sectors in which these countries were least competitive received the lowest levels of protection. As Bergsman (1970) aptly put it in his study of Brazil, the “Daddy” sectors, which were the most established and the least in need of special treatment received the greatest protection. Why would policymakers opt for this pattern of protection that runs directly counter to infant industry prescriptions? In the case of Brazil, Bergsman states: “One conclusion [...] is that removing all instances of very high protection could have cost Brazil very little of its industrialization and import substitution, and also could have produced great benefits by forcing older firms to improve their efficiency[...]” (1970,173). He estimated that, for the Brazilian economy, the cost of excessively high protection was approximately 8-10% of GNP over the eight years of his study. In her classic study, Krueger (1974) estimated that Indian quantitative restrictions reduced national income by 7.3 percent in 1964 alone.
Brazil 1955-64 | India 1968-9 | Chile 1961 | |
Consumer Durables and Non- Durables | 190 | 90.8 (food) 165.1 (other consumables) | 2884 (food) 672 (textiles) 386 (clothes) |
Intermediates | 50 | 142.5 (agro-based) 106.5 (other) | 198 (basic metals) 227 (mineral prod.) |
Capital Goods | 15 | 77.9 | 85 (non elec. mach.) 111 (elec. mach.) |
Source: for Brazil, see Bergsman (1970, 48-52), for Chile, see Behrman (1976,138-9), and for India, see Bhagwati & Srinivasan (1975, 91).
Inter-firm variation in protection was another feature of the policy that can be characterized as extreme in the sense of generating unnecessary distortions and social welfare losses, even within the rubric of import substitution. In a multi-country study conducted by the National Bureau of Economic Research (NBER), Bhagwati (1978) emphasized that there were not only very high differentials in domestic resource costs (DRC), or efficiency of production across sectors, but that DRCs varied quite significantly within sectors as well. As the author states: “It is noteworthy that in these studies the inter-firm variations in DRCs were almost as large as the inter-industrial variations in DRCs so that there is no easy way out of the conclusion that the pattern of investment allocations was less than optimal” (Bhagwati 1978, 91).
Some observers argued that using high levels of protection to eliminate foreign competition need not be as damaging to overall social welfare as it might first appear, as long as there is robust domestic competition. The high intra-firm variation in DRCs discussed above, however, suggests that such domestic competition was anemic at best. Robust domestic competition should reduce variation in productivity within a sector, yet this is not what was observed. Moreover, an analysis of prices across ISI-implementing countries led to a similar conclusion: researchers found that most prices tended to rise to the tariff equivalent, indicating that there was little internal competition (Little, Scitovsky & Scott 1970, 41). In short, in virtually every aspect of protection -its effective level, the inter- and intra-sectoral variation, or the role of foreign versus domestic competition- politicians opted for extreme choices that were not inherent to an ISI program. Policymakers prioritized protection per se over managing its use to reduce distortions and gradually foster competition and productivity gains, as infant industry prescriptions would dictate.
Capital Intensity
The continual march toward substitution of increasingly capital-intensive goods was also a near ubiquitous choice in postwar IS programs, and Brazil, India and Chile were no exception. Policies provided generous subsidies to almost any new domestic producer and would often require existing producers to incorporate newly substituted (high cost and often low quality) products as inputs. In addition to protection on final products, subsidies granted to firms moving into more capital-intensive production included exchange subsidies and tariff reductions on inputs, and loans at negative real interest rates, all of which reduced the costs of investment.5 The requirement to continually incorporate domestic inputs, which over time were increasingly poor substitutes for what could be imported, added another layer of inefficiency in the productive structure. In other words, this policy mix virtually ensured the proliferation of increasingly inefficient firms. One of the most unfortunate results of these policies was limited job creation. As Table 3 and the discussion below shows, manufacturing growth far outpaced employment growth in each of the three countries.
India 1950-60 | Chile 1950-60 | |
---|---|---|
Manufacturing Growth | 6.8 | 5.4 |
Employment | 3.3 | 1.7 |
Source: Baer & Herve (1966, 89-90)
These same trends occurred in Brazil, where value added in manufacturing rose from 22 percent of GDP in 1949 to 30 percent in 1959, but employment remained constant at 9 percent of the total (Bergsman 1970, 159-160). Moreover, in Brazil, industrial wages were roughly twice as high as the opportunity cost of labor. Bergsman estimates that changes in these two policies alone, aimed at reducing the cost of labor relative to capital, would have increased workers employed in manufacturing by 130 percent (1970, 162).
It is important to underscore once again that reducing the bias toward capital-intensive production is not antithetical to the industrialization goal. It could have forced industrialists to consider real opportunity costs more fully in their investment decisions, and therefore would likely have increased employment as firms specialized in manufacturing while using abundant labor. Moreover, reducing the bias toward capital intensity would have spread the benefits of industrialization more widely by spurring externalities such as higher skills. A capital-intensive strategy for promoting industrialization, however, created distortions that led directly to the exhaustion of growth. Perhaps more than any other aspect of the policy, this emphasis was clearly self-defeating. Why would politicians make this choice out of the menu of options available under the rubric of ISI?
Permanent Infants
There is now considerable consensus that the key flaw of the postwar ISI programs was the failure to adjust policies over time to promote infant maturation (Huber 2002). No matter how inefficient the original policies were, a clearly defined program of incentive reductions over time could have delivered results similar to those of a less distortionary initial policy. According to infant industry arguments, this was the essential second phase of the policy. Yet, politicians provided subsidies and protection but failed to tie those incentives to any kind of performance criteria or set a timetable for phasing them out.
More specifically, in choosing which firms would receive incentives, officials in all three countries employed some type of historical shares criterion and eschewed any kind of performance or efficiency metric.6 Officials administering the scheme often argued that this approach had the advantage of simplifying the process and provided continued support to the most “trustworthy” firms.7 Officials in charge of implementation opted to trade efficiency in the use of domestic resources in exchange for stability in the identity of the recipients.
As the extreme contours of the ISI policy choices emerged over the course of the program, economists of all stripes advocated modifications within the rubric in order to reduce inefficiencies, raise productivity, and sustain growth. The calls for milder and more carefully managed protection, less capital-intensive investment, and a planned phase-out of incentives to foster infant maturation came not just from liberal economists, but also from the intellectual parents of inward oriented-development: economists at The Economic Commission for Latin America, or ECLA. These admonitions reflected the fact that the problem with ISI as it was being implemented was not the focus on domestic markets, nor the size of those markets, but the failure to foster any kind of competition. Yet, nearly all policymakers in developing countries ignored these recommendations on how to bring about such domestic competition and sustained growth. Politicians single-mindedly pursued policies that maximized the number of products produced domestically and eschewed options within the ISI rubric that could have increased employment and spread the benefits of rising income. They paid little heed to the overall social welfare effects of the policies they had chosen, or their inevitable denouement. Why did the vast majority of developing country policymakers adopt such distortionary versions of the policy, and seemingly marched inexorably to crisis, rather than pursuing reform within ISI? The next section will present a model that provides a clear link between these apparently self-defeating economic policies and political success.
Access Orders, Organized Violence and the Strategic Environment
As with the original structuralists, the new institutional economics literature refocused attention on the underlying political distortions and away from the earlier emphasis on policy-induced economic distortions. The literature on credible commitments highlighted the absence of political institutions that undergird anonymous economic transactions -such as neutral justice and contract enforcement- as key constraints on investment and development (North & Weingast 1989; Haber 2002). The most recent work in this vein has moved beyond simply highlighting the absence of neutral justice in developing countries and has instead moved to endogenize the role of neutral violence monopolized by the state against non-neutral organized violence, into a theory of social order. NWW’s (2009) limited access order (LAO) and open access order (OAO) specify how “domesticated” (neutral, monopolized) violence and “natural” (non-neutral, non-monopolized) violence constrain how politicians structure access to economic and political institutions. As I will discuss below, in LAOs, where the state does not monopolize and neutralize violence, politicians limit violence through the construction of a hierarchical social order which distributes access to political and economic organizations based on violence potential. In order to create stability and foster production in LAOs, politicians essentially “pay off” powerful elites by directly distributing access to production rights and associated rents. In LAOs, in the absence of domesticated violence, development policy becomes a powerful tool for distributing exclusive access to the economy, creating production rights that can be directly traded with powerful groups in return for regime support.
NWW begin with the premise that the control and organization of violence is central to the role of any government. They state that “the possibility that some individuals will be violent poses a central problem for any group” and that “no society solves the problem of violence by eliminating violence; at best, it can be contained and managed” (2009, 13). All governments attempt to control violence (or the conflict between individuals or groups for resources) within society through the creation of state and society institutions and organizations. What differs dramatically across the two social orders is how they structure access to economic and political opportunity in order to address the problem of domesticated versus natural violence, respectively.
The link between the type of access created by a given set of institutions and the control of violence is perhaps most clear in the case of the LAO. Limited Access Orders are built around institutions that directly allocate privileged control of social functions based on violence potential. This organizational tenet encourages those with the ability to disrupt the social order to support the regime and invest and produce rather than fight. These elites, known as the dominant coalition, use their exclusive control of political, economic, religious, military, and educational institutions to structure patron-client support networks. These institutionalized patron-client structures are the mechanisms that determine access to economic and other types of opportunity down the social hierarchy. By controlling and directly distributing access to necessary social functions, limited access order organizations generate large rents for those who control them. Any introduction of competition for these opportunities would threaten the direct and privileged access and associated rents that tie these elites and their networks to the regime.8
In contrast to LAOs, violence in OAOs is monopolized and neutralized by the state, and politicians structure political and economic institutions to create open access. OAOs are built around institutions that create competitive political and economic markets with free entry. Economic and political winners are selected by the great mass of consumers or constituents who choose the goods and services they find most valuable. These competitive institutions and structures align individual and social rates of return, meaning that individuals and groups that produce and exchange goods and services that raise overall social welfare the most, receive the highest rewards. The benefits are diffuse and impersonal, and they flow from unfettered competition and innovation in economics and politics. In other words, politicians in OAOs “domesticate” violence and use it to support a system that picks winners through impersonal and pareto-optimizing mechanisms such as competitive markets.9
In LAOs, in contrast, those who can most effectively wield violence directly determine who has access to political and economic institutions and opportunities, and distribute rewards commensurate with each individual or group’s role in maintaining the violence wielder’s dominance. In a nutshell, in LAOs, politicians directly control and distribute access to economic and political institutions to powerful elites in order to incentivize them to produce rather than fight over available resources. As NWW note, “Systematic rent creation through limited access in a natural state […] is the essential means of controlling violence” and maintaining social order. “Rent-creation, limits on competition, and access to organizations are central to the nature of the state, its institutions and society’s performance.” (2009, 17, emphasis added). Both the powerful elites receiving direct access in exchange for regime support, as well as those integrated into their patron-client networks are trading their support directly for access to whatever benefits they receive. They cannot opt to oppose this system of institutionalized direct exchange and the outcomes it produces, without giving up whatever benefits they are receiving in exchange.
Since all systems produce some concentrated benefits, it is important to underscore how their distribution differs across OAOs versus LAOs. In OAOs, politicians distribute rents through indirect mechanisms, and for this reason their ability to generate excessive rents is far more restricted than in LAOs. In OAOs, territorial distribution of pork and sectoral distribution of rents do not directly confer access to territorial goods or production rights on individual constituents or firms. Rather, constituents and firms receive these benefits by virtue of membership in a given political constituency or economic sector, and their receipt of these benefits has no relationship to who they supported politically. When rent-provision becomes noticeably detrimental to overall welfare, political entrepreneurs can challenge the policies, and constituents can oppose rent-provision without fear of loss of access, as is the case in LAOs. As NWW point out, much of the scholarship on rent-seeking in OAOs inappropriately conflates these relatively small and indirectly distributed rents with directly allocated production rights and access to other benefits characteristic of LAOs. In NWW’s words, much scholarship “ignores the ability of political institutions to reflect the interests of unorganized groups [in OAOs]. When too many rents threaten an open access order, the opposition party has Schumpeterian incentives to make this a central issue, publicizing this problem and advocating its eradication” (2009, 141). The open access political order, and the indirect nature of the links between politicians and constituents means that competition for political support constrains the level of rents politicians can distribute. In LAOs, the limited access and the direct exchange of political support for that access (and associated rents or other benefits) hamstrings constituents’ ability to sanction existing policies by opting for a pareto-optimizing political entrepreneur. Since opting for such a reformer requires relinquishing one’s access and benefits, the institutionalized patron-client networks characteristic of LAOs eviscerate the Schumpeterian countervailing forces characteristic of OAOs described by NWW above. Thus, this difference in access and direct exchange versus indirect exchange alters the degree to which politicians can impose social welfare losses on the public with impunity.
The Political Efficiency of Extreme ISI
Most generally, NWW’s argument implies that in LAOs, all social functions, including production, are subordinate to the need to control violence, the overriding political distortion in these systems. The political structure consisting of a dominant coalition and supporting patron-client networks leads to distinct political implications from a given policy change in comparison to an OAO system. When production rights are distributed directly to tie powerful groups to the regime, a reform that introduces market forces to weed out economically inefficient firms produces pure political loss. Such a move sacrifices political support from those economically inefficient firms without any countervailing political gain. The expected winners -new efficient producers- do not emerge in a context in which production rights are directly determined by the state. Similarly, the expected political gain from constituents based on the diffuse benefits of increased job creation and rising standards of living fails to materialize when patron-client networks aggregate constituent support through the direct exchange of jobs and other benefits. As long as most producers and constituents are unwilling to risk their direct access in order to reward reform, the diffuse benefits that such reform would produce fall upon tied hands, so to speak.
More specifically, understanding the imperatives of competing within LAOs highlights crucial aspects of how politicians implemented and shepherded ISI policy. The creation of production rights that could be distributed to powerful players was essentially an unmitigated political gain as long as it did not impinge on the existing distribution of access and rents.10 As I will discuss below, this was precisely the difference between each of the extreme versus mild choices -the milder policy options would have required dismantling the direct distribution of production rights in favor of market mechanisms, and thus would have severed the ties that bound new manufacturing groups to the LAO coalition. In the detailed examination of extreme IS policy below, I will discuss infant maturation first, as this policy epitomizes the tension between government or market-determined firm success and is the crux of the question of sustained growth. I will then turn to how an understanding of the underlying LAO equilibrium can explain other puzzling details of policy choices related to protection and capital intensity.
Limited Access Orders and Infant Maturation
An underlying politico-economic equilibrium that drove politicians to directly distribute access to the economy in order to preserve social order and their power within it explains their failure to foster infant maturation. The heart of the prescription for development through infant industry promotion is the requirement for the government to eventually relinquish support and control over industries, and allow competitive markets to drive efficiency gains or ruin. Such a prescription, however, is anathema in a system where politicians maintain order and power through directly controlling and distributing economic opportunity. Any move to relinquish the carefully constructed and directly distributed production rights upon which these social orders were based would have directly threatened these producers and their rents.11
In addition to explaining what policymakers failed to do, the argument also provides a positive explanation for the choices made regarding the distribution of incentives. When questioned by researchers, policymakers argued that the use of “historical shares”12 was efficient because this allowed politicians to identify and work with the most “trustworthy” firms. Absent an understanding of LAO dynamics, one might question the definition of trustworthy, and wonder why efficiency might not fit the bill, or be at least as important as some other notion of trustworthiness. With an understanding of LAO imperatives, however, the logic of this distribution criteria becomes crystal clear -existing producers were conferred exclusive production rights in return for their political support, and the criteria were constructed to maintain that relationship. In other words, the criteria demonstrate that implementing officials opted to trade efficiency in the use of domestic resources for stability in the identity of the recipients precisely because these policies were designed to deliver rents to specific firms in return for political support. As NWW articulate: “The natural state (LAO) cannot support creative destruction because the [open] creation of new economic organizations directly threatens existing economic organizations and their patterns of rents” which tie them to the existing social order (2009, 116). The logic of how to implement an infant industry policy is not arcane. As the quote from ECLA above indicated, many were baffled by politicians so stubbornly ignoring the clear economic imperatives to the point of driving their economies into crisis, but under the rationale of this argument, this choice no longer appears irrational, but in fact politically efficacious.
Limited Access Orders and Protection
If politicians are distributing production rights and associated rents directly to firms, then policy must adjudicate the generation and distribution of rents among each of these groups. Herein lies an explanation for each of the three highly unusual and/or distortionary aspects of protectionist policy. First, the very high average levels of effective protection were a result of the need to maintain rents for firms in a given sector, even while they were required to absorb the high costs of a newly substituted good used as an input. The policy that maintained substitution into inputs and capital goods, and also required older firms to incorporate these inputs, clearly required ever-increasing levels of effective protection in order to maintain their rents. When we consider that the firms in which countries had a comparative advantage were the first to be substituted, we also have an explanation for the inter-sectoral pattern that protected “Daddy” industries first and foremost. Rather than graduate early firms with the greatest initial comparative advantage by reducing support and introducing market mechanisms, and then shift promotion policies to the next level, as a milder ISI with an eye on social welfare might have done, early beneficiaries retained their subsidies and protection even as new firms, with less comparative advantage, were being promoted. In order to remain viable, while absorbing ever-more expensive domestic inputs, the early firms had to have their protection continually increased. Thus, this particularly puzzling direct contravention of infant industry prescriptions comes into focus based on the need to manage and adjudicate the multiplying production rights and rents. The high inter- and intra-sectoral variation in DRCs is also of a piece. If policymakers are conferring production rights directly at the level of the firm, then subsidies must also be adjusted on a firm-by-firm basis in order to adjudicate inter- and intra- sectoral rents. When production rights are distributed directly, the expected intra-sectoral competition that would normally equalize DRCs across a sector is eliminated.
Finally, the same logic also explains another common aspect of most programs that has puzzled analysts: the lack of domestic competition that might have lowered prices, and the choice to allow prices to rise the tariff equivalent. Control of economic access through firm-level distribution by definition obviated any kind of competition among domestic firms. And the price increase to the tariff equivalent follows directly from the direct distribution of production rights. This was both understood and intentional, since distributing production rights and strictly limiting competition allowed political leaders to maximize the rents created and distributed to their supporters. The high inter- and especially the intra-sectoral variability once again reflects politicians’ need to distribute access and strictly limit competition, even within a given economic sector. This is one of the key distinguishing features of distributive politics often seen in OAOs, characterized by the distribution of rents through regulatory rules that apply to a sector as a whole, and that does not preclude new entrants, and the distributive politics characteristic of LAOs. In the former, the maintenance of free entry leads firms to converge on similar levels of productivity, while in the latter, control and distribution of production rights conferred at the level of the firm explains why firms in the same sector do not converge on similar levels of productivity.13 In sum, with an understanding of the constraints imposed by a limited access order, these aspects of the policy emerge as intentional features driven by the LAO strategic context, rather than mistakes, oversights or excesses.
Limited Access Orders and Capital intensity
The decision to continuously substitute imports of increasingly capital-intensive goods was one of the policy trade-offs that appeared most perverse, as discussed above. Once we consider the incentives within LAOs, however, we can see why this is a winning political choice. In LAOs, politicians prosper by tying powerful groups to the existing order through the creation and direct distribution of production rights and associated rents. At the same time, the possible countervailing pressures from constituents paying the costs of these high rents are siphoned off through patron-client networks, and do not redound to politicians’ detriment. Thus, political imperatives drove politicians to continue fostering new domestic production for as long as macro conditions allowed. When the benefits to creating new production rights are unmitigated by the diffuse costs of those rents, then the decision to proceed into ever more inefficient capital-intensive production can be understood. Each increment of new substitution meant that new production rights could be distributed to build political support, despite the associated mounting distortions, because the diffuse costs and possible opposition they might engender were mitigated through patron-client networks.
In concluding this section, the argument may be further clarified by comparing the LAO constraints with a rough sketch of how a reform toward a milder ISI could redound to politicians’ benefit in an OAO. In an OAO, all political benefits provided, whether narrow or broad, are indirect. Political support is not exchanged directly for a given benefit or policy, and thus politicians can substitute policies providing one kind of indirect benefit for distinct policies providing another kind of indirect benefit. Thus, politicians can shift the policy mix and still provide an overall package that a given producer may support. This provides an explanation for how some governments were in fact able to reform IS policies over time to sustain growth. When producers prosper not because they receive exclusive production rights, then, even if they receive a rent or a subsidy, they can be persuaded to relinquish that rent in return for policies that would assist them in widening their markets. Moreover, they do not have to relinquish direct access to production rights in order to throw their support behind such a reform. They only have to adjust their firms to the new terms of competition. Similarly, in an OAO, workers and the general public do not have to relinquish access to specific benefits provided by their patron-client network in order to shift their support to a politician championing reforms that raise employment or standards of living. For this reason, in an OAO, the diffuse benefits created by such policies will redound to politicians’ benefit. These diffuse benefits do not fall on tied hands, as they do in LAOs, but rather fall on free hands that can pull the lever in favor of such reforms without risk of losing access. In short, if policies promoting domestic industry are not built on direct distribution of production rights and patron-client networks, then politicians can reap political benefits from altering the policy mix to reduce protection and subsidies in order to foster competition and the diffuse benefits of higher growth.
In sum, LAO constraints and their implications for politicians’ ability to reap political gains from reform illuminates how political distortions drove policies that were ultimately disastrous economically. The argument not only explained broad patterns of failing to sunset incentives and the continual substitution of ever more capital-intensive goods, but also some of the more puzzling specific details of the incentives, including protection structured directly counter to infant industry prescriptions, inter-sectoral and intra-sectoral differences in productive efficiency, the historical shares criteria, and prices rising to the tariff equivalent.
Conclusion: Structuralism, Access Orders and ISI
Structuralist economists highlighted politico-economic constraints in developing countries that invalidated underlying assumptions of orthodox models and impeded predicted investment and development. As these scholars argued, the problem of late development was not one of simply “getting the prices right”. Among other factors, structuralists emphasized institutional distortions limiting savings and hindering productive investment of available capital. The argument presented here is consistent with the broad emphasis of original structuralist thinking on institutional distortions, and further specifies the political structural constraints on low savings and unproductive investment. In LAOs, the key underlying political distortion is the absence of domesticated violence and neutral rule of law. In the absence of specialized and neutral state-monopolized violence, politicians are driven to limit access and distribute exclusive production rights to powerful elites and thus incentivize them to support the regime and produce rather than fight. This basic constitutive fact of limited access orders precludes and is directly antithetical to the competition-productivity gains-savings-investment cycle characteristic of an open access order with free entry and competitive markets. When such an underlying pervasive political distortion compels politicians to organize production to directly distribute rents to powerful players, the sources of savings and the mechanisms of investment assumed by orthodox models fail to materialize.
In order to bind powerful elites to the social order, politicians implemented import substitution programs in a fashion that allowed them to multiply and diversify points of access to production rights that they could in turn distribute directly to powerful elites. Reform toward a milder ISI would have necessitated dismantling this direct exchange of production rights for political support and allowing market mechanisms to determine firm viability. Relinquishing state distribution of production rights in exchange for open competition aimed at driving infant maturation would have destroyed the basis upon which politicians bind powerful groups to the regime in a limited access order. With an understanding of the strategic context of a limited access order, it becomes clear how the economic losses generated by extreme ISI, paradoxically, drove political gain.