What Can Subnational Regions in Developing Countries Teach Us About the Middle-Income Trap?

Scholars often portray Mexico as caught in a “middle-income trap” of slow growth, with its economy still too unproductive to rival advanced countries through specialization but also too high wage to compete with poorer countries on cost (e.g., Hanson 2010; Gallagher 2011; Levy 2018). Insofar as escaping the “trap” requires upgrading into new industries with higher value added, productivity, and technological intensity (Doner and Schneider 2016), Mexico’s economic situation can also be viewed as an example of the broader problem of structural transformation in developing countries.

One benefit of looking at such transformation through the lens of a “trap” is that it connects the problem to an extensive debate on the practical challenges faced by a wide range of developing countries seeking greater, and more broadly shared, prosperity. Two major approaches emerge from the literature. First, a “voluntarist” view holds that escape from the trap is achieved by leaders who select policies that prioritize industrial support and investment in public goods and infrastructure (Hausmann and Rodrik 2003; Paus 2012). And second, a “structuralist” view contends that voluntarist policies, while generally desirable, will fail to produce their intended effects (or even be politically feasible) without representative “growth coalitions” to provide the necessary consensus, impetus, and coordination to make these policies a reality (Foxley and Sossdorf 2011; Doner and Schneider 2016).

Between these views lies a proverbial “gap in the literature,” while the voluntarist view is criticized for lacking a framework specifying how policies are successfully implemented in diverse and contentious settings, the structuralist school offers mostly a list of complex challenges (in particular, the various “cleavages” between political organizations, social groups, and institutions) but little in terms of concrete alternatives. Moreover, both perspectives’ default framing of the middle-income trap as a national problem obscures significant subnational variation. This prevents a full evaluation of the available escape pathways.

To address these theoretical and empirical limitations, we focus on Querétaro, a state in Central Mexico that by most measures is a positive outlier for development. Over the past 40 years, Querétaro, with a population of just over two million (INEGI 2021), has outperformed the national economy (> 3.8% annually in Querétaro since 1981 vs. 2.3% nationally) and posted higher average growth rates than 29 of Mexico’s 32 states. It simultaneously diversified into more capital- and skill-intensive industries, creating strong demand for high-skilled workers. This trajectory poses two questions: how has this state accomplished sustained high performance in economic development? And, by extension, what applicable lessons for elsewhere can be gleaned from Querétaro’s development trajectory?

To begin to address these questions, this paper examines two key industries that have successively spurred Queretaro’s growth and transformation since the late 1970s: automotive and aerospace. The former long held a leadership position in the state’s economy, whereas the latter has recently come to the fore by combining new employment generation with relatively high levels of technology and skill intensity. A review of their progress over time reveals that while both have cumulatively delivered real developmental gains to Querétaro, each industry also exhibits a complex mix of successes and limitations in its developmental progress. A comparative analysis of the emergence of these strengths and weaknesses over time offers some potential generalizable common explanatory factors.

The advancements and blockages that have manifested in both industries appear closely related to the kinds of “partial” coalitions — groups of actors from the state, private firms, and civil society organizations such as labor unions — which support each industry. Both industries saw periods of “upgrading” toward more sophisticated skills and technology when the coalitions supporting them were expanding by adding participants from a broader variety of social sectors and backgrounds (referred to here as “cognitive diversity”) while also avoiding gross imbalances in the distribution of power among coalition members. Furthermore, both industries stagnated when their partial coalitions became more cognitively homogenous and/or one or more coalition members enjoyed overwhelming power, marginalizing the rest.

These observations suggest a more incremental approach to the middle-income trap — one that focuses on how ongoing changes in political coalitions affect policy models and upgrading. Such a perspective adds to the voluntarist school by offering a more concrete method by which to analyze how local environments affect the generation, selection, and implementation of policy options. It also contributes to the structuralist school by decomposing the problem of cleavage bridging and institution building into more discrete pieces, wherein pathways to construct more competent, capable coalitions may be discerned.

To establish these contributions, this article proceeds by first reviewing the literature on the middle-income trap and related issues of structural transformation in developing economies. Next, it introduces the case selection and methods employed. The fourth section describes the trajectories of each industry in Querétaro and considers the explanatory role of changes in their partial coalitions. The conclusion offers an interpretation of the key variables that emerge from the analysis. It also assesses how they might contribute to further research on the middle-income trap, structural economic transformation, and industrial policy.

The Middle-Income Trap and Partial Growth Coalitions

The notion of a “middle-income trap” rests on a pattern observed among developing countries since World War II; while many saw phases of rapid growth from the lowest to more moderate levels of national income, the movement from medium to high incomes and productivity has been much slower and rarer (Kharas and Kohli 2011; Eichengreen et al. 2011; Aiyar et al. 2013). Although the exact empirical pattern is somewhat contested (Pritchett and Summers 2014), the underlying fact remains that bringing structural change to economies in the Global South — especially change that brings with it wide-scale economic mobility toward what is normally associated with the notion of the “middle class”Footnote 1 — encompasses an interlocking set of puzzles whose solutions are poorly understood and/or difficult to implement (see, e.g., Whittaker et al. 2010; Rodrik 2016; Pritchett et al. 2017).

Research on the causes of, and responses to, the middle-income trap falls into two main camps: voluntarism and structuralism. The “voluntarist” approach suggests that countries are caught in this trap because of “policy misdiagnosis,” which leads them to misallocate labor and talent by insufficiently promoting local innovation, diversification, and industrialization (Agénor et al. 2012; Felipe et al. 2012; Paus 2017). From this perspective, the solution is for states to adopt the “right” policies, prioritizing industrial support and investments in public goods, social development, and infrastructure (Hausmann and Rodrik 2003, Cimoli et al. 2009, Foxley and Sossdorf 2011, Paus 2012, Rodrik and Sabel 2019).

For its part, the “structuralist” alternative holds that while voluntarist policies may indeed offer cogent responses to the middle-income trap’s economic challenges, the real problem lies deeper, at the level of political and social conflict. Even if the right policy solution can pass through the gauntlet of political disagreement, implementation will still likely fail to produce the intended effects absent broadly representative “growth” or “upgrading coalitions” that include the state, business, and labor (Foxley and Sossdorf 2011; Doner and Schneider 2016). Without coherent, encompassing coalitions, this approach claims, a collective inability to strike intertemporal bargains between groups, transfer information effectively, negotiate, monitor performance, and/or tolerate time lags in the returns for the intended beneficiaries will undermine productivity-enhancing policies (Doner and Schneider 2016:618).

The structuralist approach wraps an economic problem with a political blanket. It suggests that prior advancements from low- to middle-income status necessarily diversify the economy and local institutions. These processes, in turn, generate structural “cleavages” (e.g., between domestic and foreign firms, formal and informal workers, high- and low-income earners) that produce conflicting interests and deep political fragmentation (Doner and Schneider 2016). And those cleavages prevent growth coalition formation and effective social coordination, at least at the scale needed to solve the problems associated with the middle-income trap. The challenge from a structuralist perspective, then, is to overcome those various society-level cleavages.

While this structuralist approach correctly identifies the voluntarist perspective’s omission of the political underpinnings required for effective policymaking, it raises important questions of its own. For one, it offers little in the way of solutions to the problem it defines, other than noting that the process of coalition building feeds into institution building as a matter of “incremental interactions [among] a small number of large, mutually vulnerable actors” (ibid. p.637). More generally, the approach implies a high bar for escaping the middle-income trap: the resolution of deep political fragmentation at the national level by integrating divergent factions of capital and labor institutionally to collaborate on common projects of economic transformation.

More room to maneuver within this apparent world of structural constraint might be found by considering the formation and evolution over time of “partial” growth coalitions at the subnational level.Footnote 2 By definition, these partial coalitions overcome only some of the structuralist perspective’s cleavages by bringing together a subset of relevant actors (e.g., formal workers, domestic and foreign firms, state bureaucrats, educational institutions). Such “cleavage bridging” may occur episodically so as to produce cumulative progress toward the structuralist approach’s high ideals. Observing partial coalitions therefore allows for more diverse means which may be sufficient to deal with the structuralist approach’s concerns — i.e., “second-best” forms of monitoring, negotiation, and information transfer that nevertheless are enough to produce sought-after gains (Rodrik 2008; Chang 2010). It also extends the insights from research on the political economy of industrial policy and economic transformation at the subnational level. By emphasizing the non-reducibility of subnational economic units to national trends, this research enriches our understanding of the mechanisms and contingencies through which different groups and factions come together to resolve development challenges, as well as how they mediate external economic shocks and influences (Locke 1995; Snyder 2001; Montero 2002). We build on such research by training our lens specifically on the problems of the middle-income trap.

The empirical analysis of Querétaro’s automotive and aerospace industries reveals that two conditions significantly impact the extent to which partial coalitions make headway against the middle-income trap: the cognitive diversity among coalition members and the balance of power between them. The concept of cognitive diversity draws primarily from research on democratic deliberation and is defined as the breadth of approaches to problem-solving (e.g., problem interpretation, problem-solving tools, heuristics, and predictive models) represented within a given decision-making group (Landemore 2017). This literature finds that cognitive diversity confers a problem-solving advantage by both bringing more relevant information into the problem-solving process and using it to develop solutions that are more appropriate for all of the actors and contexts involved. That is, more diverse groups integrate a more complete set of relevant information, better identify and resolve the weaknesses of a given idea or position, and produce solutions to problems that outsiders assess to be better and more innovative (Page 2006, Landemore 2017, Farrell and Shalizi 2015).

By increasing their cognitive diversity, coalitions integrate novel perspectives, interpretations, heuristics, and models regarding the wide range of obstacles and opportunities across institutions, social groups, and geographic locations. Coalitions that fail to continually broaden their coverage of this “problem space” (Farrell and Shalizi 2015) will be less capable of identifying and resolving the key bottlenecks that are the locus of voluntarist policy intervention. Here, we assess a partial coalition’s cognitive diversity by primarily examining the differences among coalition members in terms of the institutional sectors represented (e.g., state, private sector, labor unions, universities), as well as the geographic and class differences among them (e.g., foreign vs. domestic capital, working class vs. elite political base) (see Table 1). The key issue is the direction of change within a coalition. Therefore, if a coalition adds members from a new institutional sector, or even representatives of the same institutional sector but from different geographic and/or socioeconomic backgrounds, these can all increase cognitive diversity.

Table 1 Key independent variables: concepts, dimensions, operationalization, and types of variation

If a partial coalition’s cognitive diversity constitutes its pool of knowledge and collective skills, its power balance has a strong effect on their deployment. This balance of power concept suggests that to effectively increase its coverage of the problem space, a coalition must expand by inviting actors who, while increasing its cognitive diversity, do not overwhelm the preexisting power base and therefore the voice and contributions of existing coalition members. Under “grossly imbalanced” compositions of power, one or a few actors within a coalition (e.g., a single multinational corporation) dominate the priority-setting and discussion processes, undermining even “second-best” versions of information transfer, negotiation, and coordination needed to successfully execute productivity-enhancing policies. On the other hand, coalitional expansions that preserve or increase overall balance allow a more mutual discussion and the development of coordinating capacities across the coalition’s membership.

This notion of balanced power also undergirds the capabilities approach to development and related egalitarian theories of democracy for similar reasons: as Sen observes, the reason functioning (as opposed to just formal) democracies prevent famines is because the balancing of power among groups holds politicians accountable, preventing them from allowing large segments of the population to suffer without assistance or attention (Sen 2009:343–44). In the empirical literature on development, such equalization of power (with or without formal democracy) has been cited as necessary for various notable policy outcomes, including sustained growth while reducing inequality in Malaysia (Kuhonta 2011), expanded access to education in a number of developing countries (Kosack 2012), and Brazilian institutions’ greater capacity relative to Mexico’s to enact economic policy in favor of rural and poor populations (Maxfield 1991). In all of these cases, it is the relatively greater equalization of power across groups, not perfectly balanced power, which produces policies that are better suited to foster structural transformation.

Much of the literature on coalitional power focuses on observing some aspect of “unity,” “coherence,” or “density” (Doner and Schneider 2000; Domhoff 2012; Huber et al. 2019). To adapt these ideas to the context of “partial coalitions” in Querétaro, we focus on whether industry-level coalitions become grossly imbalanced in their power among members. This is observed by observing whether they recruit actors with overwhelming resources compared to the rest of the coalition. Such resources include, but may not be limited to, economic capital (e.g., revenues, ability to access credit) and constituency (e.g., size of workforce, union membership, and/or voter base) (see Table 1 above).

Overall, we expect that industry-level coalitions which grow while maintaining a balance of power, and selecting for cognitive diversity, will make greater strides in their capacity to elicit increased growth and industrial upgrading. By contrast, a significant decrease in either cognitive diversity or power balance will likely observably reduce a coalition’s collective problem-solving capacity. The case studies from Queretaro’s automotive and aerospace industries support these claims.

In measuring changes in the explanatory variables, our case-based approach currently allows only for categorical observations (e.g., “increased coalitional diversity” or “shift to power imbalance”). That is, the framework is configurational in that it focuses on cases holistically as comprising combinations of a small set of variables. It is also non-probabilistic in that it refrains from inferring continuous or average coefficient-type effects of a given independent variable across settings (Ragin 1987; Mahoney and Goertz 2006).

This data and approach carry a few implications. For one, a discrete shift in any dimension of any key variable will be largely analytically equivalent. Therefore, we do not have a basis for discerning a difference in the impact from, say, a given institutional sector joining a coalition (e.g., representation of academic institutions) versus increasing the socioeconomic or geographic diversity of an already-represented sector (e.g., adding local or nonlocal representation to participating businesses). The framework solely distinguishes between discrete changes toward and away from greater diversity and/or a gross power imbalance in a coalition. This also forecloses claims about the relative magnitude of the contribution from one group versus another. For example, we remain agnostic regarding the developmental benefits of adding representation from organized labor versus academia to a coalition. That said, if the addition of one group broadens diversity on more dimensions than another (e.g., sectoral and socioeconomic vs. only sectoral), our cases are suggestive that this might make a difference in their respective cognitive contributions.

Given these considerations, generalizability from this framework arises not in terms of overarching single-variable effects but, rather, from an understanding of how key factors combine to produce causal pathways. At the same time, the specific dimensions of each of these two factors may vary across contexts. This means that while this framework, with its emphasis on cognitive diversity and power balance, can potentially apply broadly across different settings, scales and industries, the relevant dimensions and operationalization of those two variables (e.g., extent and types of geographic, socioeconomic, or other kinds of diversity) may vary based on the institutions and parties involved.

Querétaro’s Economic Record and a Methodological Approach to Understanding It

Despite its relatively small size, Querétaro figures prominently as one of the Mexican states that has most progressed toward an escape from the middle-income trap (Table 2). Over the past 40 years (1980-present), it grew at an average yearly rate of 3.8%, far surpassing Mexico’s national average growth of 2.3% (authors with data from INEGI 2021). Only two states, Aguascalientes (3.9%) and Quintana Roo (3.9%), grew at a higher rate throughout the period. As a result, Querétaro is now the sixth richest state in Mexico, only falling behind such high-income states as Mexico City and Nuevo Leon (SEDESU 2020, authors with data from INEGI 2021). This has been accomplished seemingly without disproportionate favor from the federal government — public social spending per capita, which is heavily reliant on federal revenue in Mexico, shows Querétaro at or near the bottom of Mexican states from 2000 to 2018 (Medrano 2021). In 2020, Querétaro’s total federal expenditures per capita were in the bottom quartile of Mexican states (CIEP 2021, with authors’ elaboration).

Table 2 Average yearly growth rates, by decade (1980–2019)

Querétaro also underwent structural transformation. To be fair, it entered the 1980s with a growing manufacturing sector that, benefitting from federal, nationwide support policies, specialized in food processing, textiles, other consumer goods, and, increasingly, heavy industry (Keren 1987; Kahn 2019). Yet, during the 1980s, as the debt crisis ensnared other comparable states, Querétaro continued diversifying and upgrading rather than losing its local firm base to the economic upheaval. It expanded its metal-mechanic and auto part industries during the 1980s and 1990s, specializing in auto parts production. It later added the aerospace industry in the 2000s, as advanced manufacturing activities grew to comprise close to half of the state’s total manufacturing output (authors with data from INEGI 2021). Today, much of its manufacturing output originates in the production of transportation equipment (25% of manufacturing GDP) and machinery/electronic equipment (13.3% of manufacturing GDP) (SEDESU 2020).

In achieving this structural transformation, Querétaro has, for the most part, avoided the type of low-cost “maquiladora” model that dominated Mexico’s national model of economic growth during the post-liberalization era. Its high productivity is reflected in its attainment of the third highest level of GDP per member of the economically active population, behind only Mexico City and the historically wealthy state of Nuevo León (IMCO 2022). Beyond Querétaro’s relatively skilled workforce, which has rapidly grown and received substantial training since the 1980s, what stands out is a dynamic and prolific public/private research and development (R&D) infrastructure that works closely with different manufacturing industries in the state. A recent report by Querétaro’s Science and Technology Council (CONCYTEQ) counted 52 R&D centers. Twenty-two of these were affiliated with academic institutions, and twenty were stand-alone organizations. Jointly, they solicited the third highest number of patents of any Mexican state in 2014 (Kahn 2019). Indeed, local patenting activity has increased rapidly — from an average of 1.77 patents per million inhabitants from 2001–2009 to 4.64 from 2010–2015 (compared to national averages of 1.62 and 2.27 during these same time periods) (OECD 2022) — and has provided indispensable support for Querétaro’s industrial upgrading.

Data Collection and Analysis

To begin to explain Querétaro’s performance, and better grasp the factors favoring (or hindering) the escape from the middle-income trap, this research employs parallel historical analyses of the state’s automotive and aerospace industries. Each of these industries has, at different times since the 1970s, served as the state’s high-profile economic engine by incorporating some of its most complex, capital-intensive production activities. Both are still relevant for the state’s development (see Table 3). Through their production activities, the two industries have offered employment opportunities for skilled labor while also fomenting and building close connections to R&D centers to foster innovation. At the same time, both have displayed notable limitations that prevented even greater success. A systematic assessment that addresses the commonalities across their emergence and evolution over time may therefore shed light on some of the main factors impinging upon the broader development of the state and its performance vis-a-vis the middle-income trap.

Table 3 Automotive and aerospace industries in Queretaro: basic indicators, 2019

To obtain the empirical evidence, we visited Querétaro in the summer of 2019. We conducted 13 interviews with various actors from the state government, academia, and industry. We also collected organizational documents from public agencies, business associations, and research centers, culled through local newspaper reports, and retrieved government documents from local archives. Additionally, we benefitted from material included in prior industry- and state-level studies, as well as federal and state statistical and documentary data.

The primary means of analysis is a process-tracing approach (George and Bennett 2005; Collier 2011). We use multiple within-case observations at different points in time, assessing periods in each industry that display distinctive policy approaches and upgrading patterns. The purpose is to use these multiple within-case observations to reveal consistent relationships between independent and dependent variables. As the following sections detail, the case observations suggest a central role for specific traits of industry-level partial coalitions in accounting for development trajectories.

Tracing and Interpreting Industrial Trajectories in Querétaro

The historical evidence on Querétaro’s economic performance since the 1980s reveals two periods of rapid structural change (i.e., 1980s, 1990s) and one of stalled upgrading and inertia (i.e., 2000s-present) in the state’s automotive industry. It also sheds light on one period of notable structural change (i.e., 2000s) and one of inertia (i.e., 2010s-present) in the aerospace industry. Research further suggests that industry performance during those periods was related to their partial coalitions’ patterns of cognitive diversity and balance of power (see Table 4). The following sections describe these relationships in detail.

Table 4 Overview of the industry cases

The Automotive Industry

A cursory review of current information suggests that Querétaro’s automotive industry is unexceptional when compared to its dynamic Bajío regional neighbors.Footnote 3 Lacking a large assembler, it ranks among the region’s smaller automotive industries in terms of output, number of firms, average firm size, and employment. The industry’s value added is close to the mean for Mexico and the Bajío, while its labor conditions, reflected in its expenditures per worker, fall about 10% and 2% below the regional and national averages, respectively (authors with data from INEGI 2021). Even in its specialization area of auto parts production, the state’s total output now trails Guanajuato’s (in the Bajío), as well as at least five other states’ nationally (i.e., Coahuila, Nuevo León, Puebla, Chihuahua, Estado de México).

Yet, this lackluster assessment conceals the industry’s significant cumulative contributions to the state’s economic development since 1980, as it evolved through three distinctive phases. During the 1980s, auto parts production served as Querétaro’s cutting-edge industry and engine of growth. Querétaro stood out as a first mover in the automotive segment, especially auto parts, among its Bajio neighbors. By 1988, the industry included at least 260 firms, employed more than 5000 relatively high-skilled workers, and accounted for over a third of the Bajío’s total automotive production (authors with data from INEGI 2021). Moreover, that growth in the 1980s was neither propelled by the type of “maquiladora” model that prevailed in the northern border, nor disrupted by the widespread labor strife and economic disarray that engulfed auto production in Central Mexico. Instead, Querétaro followed an alternative path, offering expenditures per worker that exceeded the regional and national averages by 16% and 6%, respectively, by 1988 (authors with data from INEGI 2021). In return, employers benefited from labor peace and an attractive business environment, and some of the larger, mostly foreign firms took the opportunity to boost their productivity by introducing total quality management (TQM) practices at the shopfloor level.

During the 1990s, the industry remained an important source of Querétaro’s growth, but two key decisions affected its structure. First, the industry diverged from neighboring states (e.g., Guanajuato, Aguascalientes) by rejecting the aggressive pursuit of a large-scale OEM assembler and instead prioritizing auto parts production by small- and medium-sized firms. It captured a niche as both a major supplier for final assembly sites throughout the Bajío region, and an auto parts exporter (Covarrubias 2014; Banda et al. 2016). Second, the total quality management (TQM) model, with its focus on “greater productivity, technology adoption, and skills upgrading…” (Kahn 2019, 102) diffused broadly across the industry.Footnote 4 Local quality control technicians were sent to train in Japan, while the industry expanded its industrial park infrastructure and spawned several public and private R&D centers (Kahn 2019).

Those decisions shaped this second phase of the industry’s development. For one, in contrast to the “enclave” auto industry model that came to prevail in the booming industries of neighboring states (e.g., Guanajuato, Aguascalientes) (Martínez et al. 2017, Mejia 2017, Martínez and Carrillo 2017), Queretaro’s industry became a prominent hub in the state’s (and larger Bajío’s) economic network. To this day, the industry maintains thick connections to a range of economic activities in the state, including R&D, aerospace, plastics, logistics, metal mechanic, higher education, and software. Furthermore, the industry emerged as a “seedbed” for new, dynamic Queretano economic activities, such as an R&D network and aerospace. Its suppliers, workers, and institutions served as the foundation for these ascendant industries.

During the 2000s, the automotive industry in Queretaro entered its third and most recent historical phase — one more characterized by structural inertia than transformation. In contrast to most other Bajío states (with the exception of Jalisco), the automotive industry lost its position as the state’s economic “crown jewel.” That distinction now belongs to the aerospace industry. Nor did it remain Querétaro’s most important platform for the foundation of other industries. Rather, it has become a major part of the state’s economic “brick and mortar”; while it undoubtedly attracts substantial FDI and still fosters R&D, the industry also distinguishes itself for the overwhelming presence of medium and small firms with lower productivity,Footnote 5 the employment of close to a tenth of the state’s workforce,Footnote 6 and a slower paced growth.

How to account for the industry’s evolving role as an early engine of growth (1980s), a seedbed and economic hub (1990s), and, eventually, a more conventional part of the state’s economic brick and mortar (2000s–2010s)? We suggest that the changes in the industry coalition’s cognitive diversity and power balance at three crucial formative junctures clarify much of its trajectory over the past four decades (see Fig. 1). The first of these junctures unfolded in the early 1980s. Like some other Mexican states, Querétaro entered the decade with a small metal mechanic and auto industry supported by federal policies of import substitution and stabilizing development.Footnote 7 And, like some other Mexican states, the industry in Querétaro faced an existential crisis in the early 1980s as demand plummeted, government support collapsed, and labor mobilized. Yet, unlike most of those other states,Footnote 8 Querétaro emerged from this crisis with a rapidly growing auto industry that would support the state’s development and enhance worker conditions, during the 1980s and 1990s (Fig. 1).

Fig. 1
figure 1

Automotive industry: case junctures, coalitional dynamics, and development impacts

A major feature setting Querétaro’s auto industry apart during this period was the coalition of industry decision-makers that responded to the crisis. In contrast to states such as Puebla (Middlebrook 1995), Chihuahua, Estado de Mexico (Carrillo 1995), or Jalisco (author interviews), this “partial coalition” effectively bridged the cleavages between labor and capital and among domestic and foreign investors, harnessing institutional, geographic, and socio-economic sources of variation. It brought together representatives from the Mexican Workers’ Confederation (CTM), a range of domestic and foreign firms (e.g., Tremec, Singer, Engranes Conicos/Dana, Kostal), and the Querétaro state government under prolabor PRI (Institutional Revolutionary Party, which governed Mexico as largely a one-party state for most of the twentieth century) Governor Rafael Camacho Guzmán (1979–1985). The varying perspectives and priorities of these actors, particularly labor, offered a greater degree of cognitive diversity than in the past, when discussions were limited to government and business. The absence of any single dominant member of the coalition ensured a balance of power. That led to industry-level policies that considered the priorities of capital (e.g., labor peace), labor, as represented by the CTM (e.g., improved working conditions, employment for Queretanos), and government (e.g., avoiding an economic collapse).

The formation of this diverse and balanced partial coalition was a direct result of two distinct conditions: the labor strife that affected Querétaro toward the late 1970s and early 1980s and the election of the labor-affiliated PRI governor Camacho in 1979. As Kahn (2019) explains, the labor strife was largely driven by new “radical” unions unaffiliated to the CTM and largely led by non-Queretanos. Their organization of strikes in some of the auto industry’s leading firms, such as TREMEC, increasingly alarmed not only business in Querétaro but also the CTM, which saw its monopoly over worker representation slip. That shared concern, reinforced by a parochial disdain for the non-Queretanos leading the dissident unions, initially offered business and the CTM a common cause.

Yet, it was the initiative of Governor Camacho, elected in 1979, that translated that common concern into the formation of a partial coalition. An outsider to Querétaro’s ruling patricians, Camacho, a former union leader in the state, enjoyed a close relationship with the CTM’s national leader, Fidel Velásquez. In fact, Velásquez, along with other Mexico City allies, had lobbied to win Camacho the PRI nomination for the governorship, much to the dismay of Querétaro’s political and business leaders, who openly opposed him.Footnote 9 Once in office, Camacho brought business and labor together to respond to the emerging challenges through tripartite actions. The combination of Camacho’s political position and experience as a union leader — a rarity for a PRI governorFootnote 10 — provided an additional check on business power, ensuring a balance that allowed for relatively equal influence. His background as an outsider to Querétaro’s ruling class further insulated him from the pressures of Querétaro’s high society and added cognitive diversity to the coalition.

The tripartite response to the existential crisis devised by this partial coalition of labor, business, and government involved voluntarist-style institution building to support the industry’s growth. It built on the coalition members’ priorities of avoiding economic collapse (government), maintaining labor peace and surviving the crisis (business), and avoiding mass layoffs, particularly of Queretanos, and defeating the “radical” unions (labor). This response first focused on the labor challenge. With support from the governor, the more management-friendly CTM reasserted control over dissident “radical” and newly formed unions and brought the strikes to an end. The response then turned to the economic shock of the debt crisis and the ensuing retreat from Mexico’s stabilizing development strategy. The CTM accepted wage restraints and promised continued labor peace. In return, domestic and foreign-owned business avoided mass layoffs, with Queretanos being dismissed last. They also offered improved conditions once the economic problems abated. Some firms went further, introducing changes in production processes and technology at the shopfloor level by adopting Japanese-inspired total quality management practices (Kahn 2019).Footnote 11

While the industry, buoyed by its labor peace and incipient production process changes, grew and provided “shared gains” during the early- to mid-1980s, its partial coalition came to a second juncture in the late 1980s and early 1990s. By then, some of Querétaro’s neighboring Bajío states, most notably Aguascalientes, Jalisco, and Guanajuato, were attracting investment from large foreign assemblers.Footnote 12 However, Querétaro shunned this option. Interviewees cited several reasons for this reluctance, including the necessary fiscal sacrifice, opportunity cost (better to diversify into other industries), high land and labor demands, fear of overreliance on one or a few dominant foreign firms, and the attractive possibility of serving as a supplier base for multiple assemblers from different states.

Rather than pursuing an assembler-centered strategy, the industry’s coalition during the terms of the more technocratic and academically connected Governor Palacios Alcocer (1985–1991) and the labor-friendly Governor Burgos (1991–1997) favored an industry-wide shift to total quality management and improvements in productivity of auto part production through R&D and infrastructure development (Kahn 2019). This shift partly responded to the expansion of the coalition’s cognitive diversity, particularly along the dimensions of institutional sector and socio-economic background. Governor Palacios Alcocer’s cabinet incorporated individuals linked to the state’s main public university, the Universidad Autónoma de Queretaro (UAQ), to work alongside the labor-business-state government coalition members (ibid.). These new “academic” members, who continued to participate in the coalition under Governor Palacios Alcocer’s successor, the labor-friendly Governor Burgos, brought an emphasis on local research and training. Yet, the shifting automotive industry strategy of the 1990s also built on the coalition’s continued power balance, which allowed the different cognitive perspectives of business, government, labor, and academia to influence policy. Had large-scale assemblers arrived in the state in the early to mid-1990s, their economic and political clout likely would have likely upset that balance (something interviewees for this research seemed to intuit), foreclosing the implementation of this approach.

During this phase, the goal was to continue improving labor conditions and industry competitiveness by transforming production processes across the industry’s firms, fostering local research and innovation, raising worker skills, and attracting investment and buyers. The industry built close collaborations with new public and private R&D centers. Prominent among these centers were the Mexican Institute of Transportation (IMT) (1988), a public agency under the Mexican Secretary of Communications and Transportation (Instituto Mexicano del Tranporte), the Center for Electrochemical Research and Technological Development (CIDETEQ) (1991), and the National Metrology Center (CENAM) (1992). Furthermore, two business groups, Tremec and Grupo Carso’s Condumex, founded stand-alone research centers in the early 1990s. Finally, the industry strengthened its ties to older centers in the state, including the CONACYT-affiliated Advanced Technology Center (CIATEQ) and Engineering and Industrial Development Center (CIDESI).

The industry also benefited from the aggressive modernization of the state’s infrastructure and a series of initiatives to raise the workforce’s skill level. The state’s industrial park area increased by a factor of four, with the construction of state-of-the-art facilities such as Bernardo Quintana or Queretaro’s Industrial Park. Alongside this infrastructure development, a variety of worker training programs emerged, including those offered by the Monterrey Technological Institute Queretaro Campus (ITESM-CQ) — founded in the 1970s — or the more recently created Universidad Tecnológica de Querétaro (UTEQ) and Universidad Tecnológica de San Juan del Rio (UTSJR).

This strategy of the 1990s positioned the industry as a local R&D hub and a seedbed for new economic activities. Yet, the industry reached a third juncture in the late 1990s and early 2000s, as labor, represented by the CTM, lost its prominent political position, reducing the partial coalition’s cognitive diversity — both along socio-economic and sectoral dimensions — as only the state government, business, and academia remained active. Labor’s withdrawal can be largely attributed to the CTM’s weakening as an actor in national Mexican politics throughout the 1990s. In Querétaro, it led to a transition from the CTM’s unifying control to a weaker decentralized structure of local unaffiliated unions. Though labor continued to participate in the auto industry’s partial coalition for a few more years, its “presence was merely formal… with very little power…” (Interview, June 2019). The election of four business-friendly PRI and, especially, PAN (National Action Party, a center-right party with strong business backing) governors after 1997 further eroded that power, as labor lost the backing that governors Camacho and Burgos, especially, had provided.

With labor’s decline came the loss of its distinctive perspective. Its concerns with working conditions, steady and well-remunerated work for Queretanos, or the unionization of new production facilities, faded. Labor “peace” survived not as a result of an agreement between the CTM, business, and government but because firms successfully undermined and resisted unions. Conditions of work worsened in the industry, with Querétaro’s average expenditures per worker, once far higher than the Bajío and national average, falling below both.

The industry’s institutions in the new millennium also reflect the changing composition of the partial coalition. For instance, the “triple helix” model, which currently frames much of the industry’s activities, only incorporates business, government, and academia. As Governor Domínguez (2015–2021) explained, this model involves a “government that facilitates and creates conditions for development of the state, while business and academia work together to address the needs of local industry” (cited in Márquez 2021). Much the same applies to the “Automotive Cluster,” which institutionalized the industry’s partial coalition in 2013. Headquartered at ITESM-CQ, its members include tier 1, tier 2, and tier 3 auto suppliers, R&D centers and universities such as CIDESI and the Polytechnic University of Querétaro (UPQ), and Querétaro government agencies such as the Sustainable Development Secretary (SEDESU). Gone are any references to organized labor.

As the coalition’s cognitive diversity has narrowed, its approach to industry development has lost some of its distinctiveness. A business-led focus on FDI attraction, cost reduction, and labor peace, reminiscent of other Mexican auto sectors, currently prevails.Footnote 13 At the same time, two conditions have prevented a full transformation into the more extractive, less integrated, enclave approach observed in neighboring states. First is the established presence of R&D and academic centers, which ensures the continued emphasis on specialization, quality, skills, and productivity. Second is the absence of a large, dominant foreign assembler. Querétaro’s auto industry retains a diverse base of firms in terms of size, origin of capital, and degree of specialization, with no individual or oligopolistic set of foreign firms exerting overwhelming political and economic power.

This coalitional profile goes some way toward explaining the industry’s current economic “brick and mortar” phase. It accounts for the recent creation of new industry-specific R&D centers, such as Continental’s “Research and Development Center” (2018) and Harman’s “Advanced Manufacturing Engineering Center” (2019), as well as the support for employer-defined skill development programs, such as ITESM-CQ’s “Center for the Automotive Industry Development” (CEDIAM) (2013) or the joint Brose-UPQ “Center for Automotive Innovation and Development” (2018). It similarly explains efforts to expand the industry’s thick connections to a range of other industries in the state, as well as terminal industries in the Bajío. And it clarifies the policy emphasis on attracting a large number of firms, which employ a significant share of Querétaro’s industrial (and overall) workforce.

The Aerospace Industry

As of 2019, the aerospace industry in Querétaro comprised approximately 80 firms and 9500 direct employees (SEDESU 2020). Widely regarded as an important source of growth and development in the state, it is described as “creat[ing] new productive and technological capacities for the region, [including] in the realm of institutional relations and public–private collaboration” (Villavicencio et al. 2013). The industry is a major focus of the state’s economic development strategy, having been included, alongside the automotive industry, among SEDESU’s four top-priority industries (SEDESU 2020:157). It is prized as a beachhead from which to attract and develop more skilled labor and specialized niches than what commonly serve as the main employment bases for middle-income countries. Overall, the aerospace industry carries great economic, political, and symbolic power in the state.

The fact that Mexico might fully assemble airplanes typifies a long-standing fantasy of rising economic power and specialization for the country. Advocates of Querétaro’s aerospace industry point to several factors as evidence that this achievement is well underway; besides the thousands of aforementioned jobs, the production work involves significant, sophisticated aircraft components such as fuselages, wings, and landing gear. There are also close to a dozen research and development centers, ranging in staff size from very small to over 1000 researchers, which devote some or all of their work to this sector. Collectively, this suggests that the aerospace sector offers much more for Mexico’s development than the classic maquiladora model.

Indeed, these and other features distinguish Querétaro’s rapidly rising aerospace cluster from those found elsewhere in the country. Large clusters in states like Sonora and Chihuahua, while older, are more reliant on basic assembly. Furthermore, Querétaro has invested far more than any other state in Mexico in developing workforce skills specific to the aerospace industry. Its National Aeronautics University (UNAQ), the only industry-specific university in the country, is a product of significant state-level investment and trains workers at levels from skilled assembly to postgraduate (master’s) degrees in engineering (Aragón 2014).

Local research and development institutions are also connected to the aerospace industry to a degree not seen elsewhere in the country, with the state’s industry-related R&D centers far outnumbering those found in other states, and actively involved in the industry’s cluster association (Casalet 2013; ProMéxico 2017). Perhaps most famously, GE-IQ (GE Infrastructure, formerly CIAT) is a joint venture between the Center for Advanced Technology (CIATEQ), a national science lab, and General Electric (GE). Focusing on turbine design, it has been described as GE’s “most important aviation engineering center outside of the United States.” It is the largest standalone research and development institute in the state, employing approximately 1100 investigators (SEDESU 2020:153–55). There is also the Laboratory for Aeronautical Testing and Technologies (LabTA), a joint project of the state’s three national science labs which provides laboratory services and product design and prototyping. Another prominent example is the Center for the Development of the Aeronautics Industry (CEDIA). Based out of the aforementioned ITESM-CQ, it focuses on supply chain consulting and management training.

For these reasons — higher-skilled workforce, more sophisticated functions performed across a diverse array of aerospace product subsystems (e.g., fuselage, navigation, landing), significant integration of a large base of local R&D capacity into the industry, and a high degree of state-level coordination with the private sector — the Mexican federal government has recognized Querétaro as one of two “strategic poles for innovation,” alongside Baja California Norte’s more established cluster (ProMéxico 2009, 2017). Yet, the industry also faces some limitations.Footnote 14 First, the production work in Querétaro’s aerospace firms may require higher skill levels than basic assembly but often lacks pathways for what would broadly be considered high skill or high-value added for this industry. Although some of the Mexican employees of these production firms will serve as engineers or production managers, far more are high school graduates with basic in-house training (Casalet et al. 2011; Salinas-García, 2012).

Second, the role of R&D in Querétaro’s aerospace sector is important but needs to be examined beyond its mere existence. Many of the local research institutions provide services to firms that require only iterative improvements to existing products and production systems — a more heat-tolerant chemical treatment, perhaps, or a more automated, ergonomic workstation to increase production efficiency and safety. Most fail to register patents, and for those which do, patenting is still a new frontier (Interviews, Jun 8, 2019, and July 11, 2019). Furthermore, it is unclear whether these R&D activities represent the best use of local engineering and research resources to advance Queretaro’s position within a global “knowledge economy.” Most of the benefits appear to accrue to foreign multinationals. For instance, the activities of GE-IQ, perhaps one of the most advanced local institutions, are at the dictate of and redound almost to the sole benefit of GE. Its engineers work under confidentiality contracts and licenses to transfer their intellectual property directly to GE, which registers their patents in the USA. Designs originating in GE-IQ are manufactured in the USA (Casalet 2013:118).

Lastly, both industry advocates and critics agree that linkages to local firms in Querétaro have been meager, their general absence a hindrance to the industry’s development contributions. As one business consulting report puts it, “Top-heavy would best describe Querétaro’s aerospace sector today. It seems that the state has invited too many people to a party with too few drinks to go around” (Global Business Reports 2016: 127).Footnote 15 This is despite repeated attempts over the past 15 years to facilitate local firms’ entry into the supply chain.Footnote 16 As a result, nearly all of the components used in local production facilities are imported, and the participation of local firms at the industry level remains sparse.

How might we account for this industry’s mixed development impacts, combining high dynamism as the state’s leading economic engine with labor, R&D, and local supplier limitations? Similar to the automotive industry, its history can be divided into two major phases: a first phase of structural transformation in the late 1990s and into the 2000s, followed by a second phase of structural inertia since the 2010s (see Fig. 2). An examination of the changes in the characteristics of the industry’s coalition over this period reveals key insights to account for the evolution of these two phases and their consequences for the industry’s structure. The early period of developmental efflorescence unfolded as a result of the actions of a balanced and increasingly diverse coalition under the successive administrations of Ignacio Loyola Vera (PAN, 1997–2003) and Francisco Garrido Patrón (PAN, 2003–2009). The second phase, spanning from the later Garrido Patrón years through the Calzada Rovirosa (2009–2015) and Domínguez Servién (2015–2021) administrations, in turn, has been marked by a high degree of policy continuity that has allowed a growing imbalance of coalitional power in favor of foreign multinationals.

Fig. 2
figure 2

Aerospace industry: case junctures, coalitional dynamics, and development impacts

The proximal cause for the industry’s fast growth in Querétaro, a large international airport, was initiated under the governorship of Ignacio Loyola Vera. While this airport provided elements that would later be crucial to Querétaro’s ability to distinguish itself in pursuit of foreign investment, that was not the original intention. According to Loyola himself, the airport was a constituency service project in response to local voters’ demands for better transportation and mobility, rather than a tool for industrial diversification or upgrading. The Loyola administration’s industrial policy instead focused on maintaining largely the same industrial mix that it inherited — automotive with other forms of traditional manufacturing, such as consumer goods — while expanding its base geographically into parts of the state outside the capital where labor was cheaper. Yet, this policy was largely a failure.Footnote 17

The initial notion of the airport as a matter of public works, and not industrial policy, was in keeping with the Loyola governorship’s coalition, which mostly drew from traditional Querétaro political elites, placing constraints on at least its geographic dimension of cognitive diversity. Although he was the first candidate from the right-leaning PAN-elected governor in Querétaro, Loyola came from a prominent local family with strong ties to the PRI. Ignacio Loyola’s background was more in business than in politics, while his brothers, Roberto and Rodolfo, both had ties to the governorship of Enrique Burgos (1991–1997, PRI). Ignacio’s administration reflected these relations, with several of his most prominent cabinet members having served under Burgos. Several more were members of the PRI and/or had worked with the Burgos administration in a less formal capacity. Crucially, when it came to the airport, Loyola’s coalition lacked broad support. Indeed, despite a 30% subsidy from the federal government, the project was largely mocked as a white elephant not only by business interests in the state but also by members of Loyola’s own party, the business class-based PAN (Kahn 2019: 207, 211).

The election of Loyola’s successor Francisco Garrido brought increased cognitive diversity — in particular, on geographic and sectoral dimensions — while at least initially maintaining a balance of power among its members. Garrido also ran under the PAN but stood in contrast to Loyola in almost every other way. Whereas Loyola came from a long-standing political family from Querétaro, Garrido was the first modern Querétaro Governor not from that state, having moved from Mexico City in the 1980s. He was more of a PAN purist and decried Loyola’s non-PAN appointments. His administration brought in more actors from the business sector, some with less political experience than in Loyola’s cabinet. In addition, instead of sharing Loyola’s leanings toward ideological moderation and strong commitment to public works, Garrido hewed much more strictly to the PAN party line and was an outspoken critic of government intervention. In fact, Garrido loudly condemned Loyola’s airport project, and for a time, it appeared that he might prevent its opening as governor (Kahn 2019: 203–4, 211).

Yet, once in power, Garrido abandoned his prior criticism and puts the airport to strategic use when Bombardier Inc., a Canadian business jet firm, announced that it would no longer consider China and India as potential sites for a new production facility, focusing only on Mexico. The Garrido administration lobbied aggressively for Bombardier to locate in Querétaro, placing the airport at the center of their proposal. They offered land on the airport site free of charge along with a proposed institution for customized training — initially an extension of the Technological University of Querétaro (UTEQ), but eventually the standalone Aeronautical University of Querétaro (UNAQ) — whose facilities and instructors would be paid for by the state. The proposal vaulted Querétaro ahead of the other Mexican statesFootnote 18 vying for the high-profile investment, which Bombardier ultimately sets at US $200 million and 1500 jobs over several years. This deal — the airport land, the tailor-made training institution, and state subsidies — became the basis of attracting further foreign investment, including the Safran Group (from France, associated with Airbus) and Aernnova Aerospace (from Spain, associated with Bombardier). These firms became the Querétaro aerospace industry’s “anchors,” generating the lion’s share of local production employment and demand for local R&D capacity.

This first phase in Queretaro’s aerospace development illustrates how the sequencing of leadership between Loyola and Garrido brought growing coalitional diversity, and at least for a limited period of time, balance, to the formulation of industrial policy for the sector. Loyola’s signature contribution was a commitment to public investment that would distinguish what Querétaro had to offer for the aerospace industry. This policy was in line with a traditional local political elite view in Querétaro that strong public intervention — from railroads to water systems — was essential for economic and social well-being. While Garrido initially derided the airport as a boondoggle, once it became a potential asset to attract foreign investment, he embraced it enthusiastically, using it as the core of a “made to measure” deal with Bombardier (Salinas García 2016:60) and other major employers that catalyzed the industry’s growth. That such a policy success could have been achieved by integrating ideas spanning otherwise conflicting political factions illustrates how cognitive diversity does not necessarily coincide with harmonious social relations.

However, over the course of establishing roles, institutions, and arrangements among leading actors in the aerospace sector, the Garrido administration also sets in motion dynamics that would lead to the sector’s partial coalition power relations becoming increasingly imbalanced.

The sector’s dynamics were more balanced, while the government was actively building up the intensive support infrastructure of organizations and institutions that nurtured the sector into rapid emergence. Yet once the state allowed these organizations and institutions to govern themselves, they became dominated by large foreign “anchor” firms, precipitating a gross imbalance. These eventual gross imbalances constrained efforts that might have otherwise enhanced innovation and fostered structural change. They thus narrowed the coalition’s ability to generate further economic dynamism — especially in terms of the initial core goal of increasing local participation in the higher value-added segments of the aerospace sector — even as the sector’s employment and exports grew, and its mix of skills and industry-enhancing public goods were the envy of much of the rest of Mexico.

During this second phase of the industry’s development, the “made to measure” investment attraction policy gradually weakened the state vis-à-vis large private sector actors in the coalition. The Garrido administration’s assumption, shared by Mexican President (and fellow PAN party member) Vicente Fox, who lobbied for Bombardier to commit to investing in Mexico (Rangel-Padilla 2021:15), was a notion known as escalamiento or “climbing up.” This assumed that once Bombardier located operations in Mexico, a natural process would be set in motion in which Mexican workers and firms would take on increasingly sophisticated, higher value-added activities. This general model has been preserved beyond Garrido’s administration by maintaining in stable positions the high-level personnel who structured the initial Bombardier deal.Footnote 19 But it has also allowed the large foreign firms, with overwhelming economic capital, to become the most influential private-sector representatives in industry-level collaborative efforts, whether in terms of associations to promote the industry (e.g., Aeroclúster Querétaro, founded in 2012) or other organizations and initiatives to enhance the industry’s growth and capabilities (e.g., the Querétaro Network for Aerospace Research and Innovation, or RIIAQ, and the Aerospace Laboratories for Technical Services and Technology, or LabTA, both founded in 2010).

The subsequent plateauing of the sector’s contributions to local specialized skills and high-value-added economic niches could have been anticipated — as previous research shows, these predominant private sector partners for the state favored a cost-based, rather than an upgrading, competitive strategy (Casalet 2013). If the state wanted to opt for the latter, it stands to reason that it should have brought actors for whom those issues were more urgent more to the center of its coalition. The fact that the state built its industry associations, research initiatives, and other cross-sectoral partnerships around these cost-focused foreign firms entailed a tacit acceptance of their competitive strategy, something the state most likely optimistically assumed could be balanced out by building up a local supplier base in parallel.

Yet, despite the state’s repeated efforts to encourage local supplier development, the entry of local firms into the aerospace supply chain has been disappointing. Without prominent representation of these groups’ interests, the mission of the aerospace policy coalition became one of servicing foreign firms interested primarily in Querétaro’s capacity to combine low costs with efficient logistics. This observation speaks to recent arguments highlighting the positive potential for MNC contributions. For example, Kahn (2019) suggests that MNCs add more to local development if they are embedded in a more sectorally diversified economy, whereas Rangel-Padilla (2021) finds that similarly positive results are more likely when business actors at the industry level are more coherently organized. The Querétaro aerospace industry presents a puzzle insofar as it fulfills both of these criteria — after all, Querétaro is the former’s prime example of a sectorally diversified local economy, while its strong aerospace cluster organization would seem to fulfill the latter’s criteria for coherence. The puzzle may be resolved by drawing historical comparisons to both the state’s aerospace industry in the 2000s and its automotive industry since 1980. The differences in coalitional power balance between MNCs and other actors (e.g., local business, labor, the state) in the 2010s aerospace industry vis-à-vis the automotive industry and the 2000s aerospace industry are strongly suggestive of the need for other sources of power and voice in the private sector beyond those of foreign firms. These contrasting examples within Querétaro bring to the fore the potential downsides of grossly imbalanced coalitions and highlight that multinational firms in particular can easily cause such imbalance. This does not contradict arguments that a diversified industrial profile and coherent business organizations matter, but it does expand the types of contingency involved in how local embedding factors mediate the influence of foreign capital.

Discussion: Coalitions and the Wealth of Regions

If one reads about Querétaro in the international business press today, it is largely as a successful Mexican anomaly: a state with a strong local industrial base and professional class, high growth in industries involving higher technology than the rest of the country, high standards of living, and a thriving higher education and research sector. This broad-stroke characterization, while largely correct, motivates closer examination of the kind undertaken here. Our review, in turn, reveals a more nuanced view of Querétaro’s developmental strengths and weaknesses, which we study through the lens of two of the state’s leading industries. It also uncovers two factors central to both the successes and failures: the cognitive diversity represented within the industry’s partial growth coalition and the balance of power among its members.

In the automotive industry, labor emerged as a central actor in industrial policy and business-state relations during the last two decades of the twentieth century, buoyed by the support of two governors — Camacho (1979–1985) and Burgos (1991–1997) — with strong ties to the Mexican Workers’ Confederation (CTM). Its participation brought a different problem-solving perspective to address firm performance and economic downturns, resulting in “high-road” investments in worker skills and productivity, a robust cadre of local firms capable of weathering economic crises and the preservation of job tenure and worker experience through uncertain times. In the late 1980s and 1990s, the more university- and technocracy-oriented Palacios Alcocer (1985–1991) administration further broadened the scope of the coalition and integrated the labor-supported high-road model with local investments in universities, technical training, and state-supported research and development, all with a focus on industrial applications. The additions not only made the Queretano auto parts sector more robust and specialized. They also preserved a role for local capital, maintained a wary eye toward potentially unbalancing foreign final assemblers, and invested in local public goods and institutions which supported the growth of many subsequent industries, including aerospace. However, when this coalition narrowed — in particular, when labor’s role decreased by the late 1990s — the auto parts industry began to ossify, with high-road investments suffering and little in the way of novel responses to the breakneck growth of the industry in neighboring states.

Meanwhile, the aerospace industry began with a cross-pollination of coalitional views: two distinct, sometimes conflicting factions under the same party (PAN) each brought key contributions, namely large public investments in infrastructure and a “made-to-measure” industrial policy to woo major foreign investment. This sequencing accounts for the emergence of a sector that has both pushed Querétaro toward international attention, as well as generated somewhat of an anomaly within the state — an industry driven by foreign anchor firms with essentially no domestic counterparts. In turn, the increasingly imbalanced power after this initial phase appears important for understanding several features that have been hallmarks of a structural inertia that seems to have set in: anemic development of a local supplier base and a bracketing of the contributions of local labor at both ends of the skills spectrum within the prerogatives of foreign firms, i.e., reducing costs and increasing efficiency within existing product lines. Although Querétaro has seen different administrations from different parties since the emergence of its aerospace sector, these administrations all have seemingly viewed the initial “made to measure” anchor firm-oriented model, despite its power imbalance, as a valuable commodity to be preserved by maintaining continuity among the state representatives who oversee the key sectoral policies and institutions.

These findings hold significant implications for scholarship on the middle-income trap and, more broadly, structural transformation in developing economies. First, they offer a more concrete basis upon which to fill gaps in the voluntarist view. They suggest specific, observable political dynamics that are relevant to understanding how local coalitions select and implement economic development policies. Future research could, in this sense, take specific industrial policies of interest (e.g., investments in public goods for industry clusters, support for local firm entry into supply chains) and compare their different manifestations across local coalitions based on cognitive diversity and power balance.

Second, these findings allow for a more iterative approach than what the structuralist view currently offers. From the structuralist perspective, the middle-income trap is something escaped by countries, and the political fragmentation which constitutes the key obstacle to progress tends to be viewed at the same national scale. But this does not specify much beyond an all-or-nothing proposition: either a country has very broad, highly organized, cross-class, inter-institutional coalitions capable of making widespread, highly impactful investments in infrastructure, education, and more or its extant cleavages will perpetuate the middle-income trap. By taking subnational units of analysis with their own observable economic trajectories, the approach outlined here significantly increases the number of cases, and the degree of empirical variation, which can be drawn upon to induce potential solutions. Furthermore, by positing that incremental changes in a coalition’s diversity and power balance can produce corresponding improvements in its ability to intervene effectively in the local developmental “problem space,” this approach offers the possibility of partial movements in the right direction.

To advance the type of case-based agenda proposed, qualitative or mixed methods such as “typological” theory building and qualitative comparative analysis (QCA) may be relevant. The former would facilitate enlisting a diverse array of cases into understanding the key causal pathways precipitating different or the same (equifinal) outcomes (George and Bennett 2005). The latter would facilitate more “middle-n” comparisons that bring statistical procedures to bear on arrays of contextual conditions, preserving the case-based logic of understanding observable relations between sets of co-occurring conditions, as opposed to a more variable-driven logic of inferring average causal effects of isolated factors (Ragin 1987, 2008). These expanded approaches could clarify how the composition of a local regime, as well as the participation and composition of key actors in the private sector and civil society, affect a local coalition’s ability to detect and resolve the coordination challenges presented by the middle-income trap.

At the same time, the approach proposed here also brings some limitations. First, the coalitions that emerge at the subnational level can at times be rather small. Thus, if coalitional composition is a key determinant of problem-solving capacity, then the subtraction (or addition) of perhaps just one organization, institution, or even individual may undermine this crucial variable (or suddenly render it constructive). In that sense, the “partial coalitions” discussed here may be fragile. That said, this fragility may be common, regardless of scale. National-level coalitions may prove just as fragile, perhaps even more so at times, due to the necessary overall scope and strong internal organization of key players involved. Thus, while fragility may imperil the predictive power or robustness of any one coalitional design, it may nevertheless be more an artifact of empirical realism than of conceptual weakness.

A second limitation of the approach comes as the flip side of the coin of one of its strengths: the focus on incremental changes. Insofar as “partial coalitions” can help regions move along a path toward structural transformation, each progressive move may establish new constraints or liabilities. Development researchers as far back as Albert Hirschman (1958) have recognized this double-edged aspect of developmental progress. Therefore, any “partial coalition” approach that claims any meaningful contribution to the overall resolution of the middle-income trap must extend beyond the preceding analysis. It must explain how coalitions form sequences that allow them to overcome emerging obstacles and continue to move in the right direction. That raises several questions. For example, how do existing coalitions decide who is the next interest group or institution that they will try to recruit? What can be done when a coalition becomes overly constrained in its view of the problem map, or imbalanced to the point where crucial information from among its members is not being properly integrated? Such questions require more longitudinal cases to illustrate the dynamics of such sequence building.

Finally, a “partial coalitions” approach as elucidated thus far does not hold obvious implications for how to achieve ultimate solutions to the “middle-income trap” at the national level. The national scale involves new actors, conflicts, institutions, and challenges without obvious analogues in Querétaro. In some ways, this is a trade-off inherent in subnational analysis: we gain more cases and add variation but also limit some aspects of external generalizability.

At the same time, there are some mitigating factors relevant for future research. First, just because states or provinces do not have all of the equivalent institutions of a national government does not mean that insights from progress in the former cannot meaningfully speak to the latter. It may well be that certain configurations of groups, or certain initiatives to grow a coalition, are quite relevant at the national level. In Mexico, the model of a small political clique, or camarilla, has been prevalent in the structuring of state and national ruling regimes through the nineteenth and twentieth centuries, arguably even to the present day (Camp 2002). This type of political grouping is well-known in Mexico but not specific to it and could be said to operate on similar principles to a “power elite” model (Mills 1956; Domhoff 2018). And while the exact formulation and dynamics of small political cliques will certainly vary across time and place, such structures can nevertheless be expected to be fairly common. Thus, the caveat of scale is important, but should not be taken to suggest a total non-transposability of insights. Second, learning or coordination among subnational units could ultimately influence national-level political dynamics. One could imagine a situation in which other states adapted Querétaro’s successes and subsequently worked together to seek national-level coalitional and policy choices reflective of their collective learning. Such “scaling up” is not without precedent — for example, Jalisco’s industrial policy for software became the template for a national-level policy (PROSOFT) once its successes became more widely known (Palacios Lara 2008:38). In that sense, future research on “partial coalitions” should consider the potentially positive (and negative) aspects of interlocal dialogue and diffusion.