The prospects of cross-class alliances in former bureaucratic development societies: comparing Taiwan and Burkina Faso

Abstract This article examines cross-class alliances in former ‘bureaucratic development societies’. We look at middle-class mobilisational efforts aimed at the lower classes in Taiwan and Burkina Faso. In particular, we analyse their capacity for challenging the socio-political dominance of the remnants of post-colonial state classes and hypothesise that the transformative potential of such class alliances depends on the development of the productive forces. Specifically, we argue that the formation of competitive middle-class-led political parties with a mass base independent of the clientelist networks of post-colonial state classes hinges on an early empowerment of labour. We test the validity of our argument using a ‘diverse selection’ comparison of Taiwan and Burkina Faso.


Introduction
The aim of this article is to shed light on the political prospects of middle-class-led cross-class alliances in the Global South. Class coalitions and their transformative potential have been studied by different authors. For instance, in their seminal monograph Capitalist Development and Democracy, Dietrich Rueschemeyer, Evelyne Huber Stephens and John D. Stephens (1992) identify alliances between organised labour and middle-class representatives, ie urban professionals in the public and private sector, as key enablers of democratisation. While, according to Rueschemeyer et al., in Europe the push for political liberalisation tended to be spearheaded by the working class, in Latin America the middle classes initially acted as the leading force against authoritarianism.
In a recently published book, Lisa Mueller (2018) investigates contemporary political developments in sub-Saharan Africa through a comparable class-focussed lens. Challenging the notion of ''classless Africa" (Mueller 2018, 8), she finds that urban coalitions between middle-class activists and lower-class citizens are a characteristic feature of a 'third wave of protests' that unfolded in the region in the early 2010s.
Building on the work of Mueller and Rueschemeyer et al., we probe middle-class efforts to mobilise the lower classes. Specifically, we study divergent constellations of middle-class-led class coalitions in former 'bureaucratic development societies' . The concept of the bureaucratic development society has been developed by Elsenhans (1996;see also 2021, 125-131). Rooted in Keynesian macro-economics, it revolves around the hypothesis that in the post-colonial societies of the Global South, the emergence of dynamic profit-based economies has historically been blocked by structural deficits. These deficits, Elsenhans argues, paved the way for the rise and expansion of bureaucratic state classes appropriating the economic surplus via the state apparatus.
Drawing on Elsenhans' theoretical and conceptual propositions, we probe the potential for a middle-class-led class coalition to turn into a viable political alternative to the remnants of state classes and their organic clientele. We do so with the help of a comparative case study of Taiwan and Burkina Faso, two former bureaucratic development societies. Our selection of these two cases follows the logic of a diverse selection design (see Seawright and Gerring 2008, 300). In line with this logic, we focus on highly contrasting paths of the economic transformation of bureaucratic development societies and explore their impact on coalitional class politics.
While Taiwan's state class managed to bring about a comprehensive development of the productive forces, its Burkinabé counterpart did not even come close to inducing the emergence of a fully fledged capitalist market economy. Instead, Burkina Faso's bureaucratic development society transitioned towards neoliberal structural adjustment. Echoing Rueschemeyer, Huber Stephens and Stephens' propositions on the political implications of capitalist development, we contend that this divergence crucially shaped the political prospects of middle-class-led alliances emerging in the two countries at different points in time.
In particular, we argue that the capacity of such alliances for challenging the socio-political dominance of the state class depends on the make-up of the subaltern classes. In Taiwan, the lower classes have been dominated by a strong and assertive working class emerging prior to the onset of multi-party politics. Under such circumstances, class coalitions are able to assert themselves as independent and competitive political forces seeking to take power, pushing through what we designate as 'political liberation' . By contrast, where -as in Burkina Faso -an empowerment of labour has not taken place, middle-class-led alliances tend to form after the emergence of multi-party politics. State class patrons and their clientelist networks continue to dominate the liberalised political sphere. Under these conditions, middle-class activists may successfully mobilise lower classes to push for political reform. However, given the weakness of labour, the resilience of clientelism and the bad reputation of conventional party politics, these activists are inclined to refrain from trying to transform cross-class alliances into organised and viable alternatives to state-class dominated parties.
The article is divided into four parts. In the next section, we review the work of Mueller (2018) and Rueschemeyer, Huber Stephens, and Stephens (1992), and present our argument. In the third section, we then proceed to the comparison of Taiwan and Burkina Faso. We make use of primary sources and the comparatively rich secondary literature on the political and economic history of Taiwan and Burkina Faso to identify the mechanisms underlying class coalitions' dynamics. The purpose of the fourth section is to discuss the different trajectories of these two former bureaucratic development societies and to reflect on two further cases similar to Taiwan and Burkina Faso.

Class coalitions and political change
Over the past four decades, cross-class alliances have been identified as key drivers of political change by various authors (see Dix 1984;Goodwin and Skocpol 1989;Goldstone 2011;Beissinger 2013;Durac 2015). One of the most seminal theoretical frameworks revolving around such alliances has been developed by Rueschemeyer, Huber Stephens, and Stephens (1992). Focussed on historical developments in Europe, Latin America, and the Caribbean, they study the relationship between capitalist development and democracy. At the heart of their macro-historical theory lies the assumption that 'it is power relations that most importantly determine whether democracy can emerge, stabilize, and then maintain itself even in the face of adverse conditions' (Rueschemeyer, Huber Stephens, and Stephens 1992, 5). In the light of this premise, they present the balance of power among different classes and class alliances as 'a factor of overwhelming importance' (Rueschemeyer, Huber Stephens, and Stephens 1992, 5).
Emphasising that class interests are socially constructed, rather than pre-given, Rueschemeyer et al. investigate divergent configurations of inter-class relations involving labour movements, middle classes, the industrial bourgeoisie and landed elites. Their main argument is that capitalist development is conducive to democracy. Specifically, they posit that capitalist development 'weakens the landed upper class and strengthens the working class as well as other subordinate classes' (Rueschemeyer, Huber Stephens, and Stephens 1992, 271). However, while they identify labour as the most reliable pro-democratic force, they also detect that 'in all regions […] pressure from the organized working class alone was insufficient to bring about the introduction of democracy; the working class needed allies' (Rueschemeyer, Huber Stephens, and Stephens 1992, 270).
Indeed, regarding Latin America, Rueschemeyer, Huber Stephens, and Stephens (1992, 155-226) find that cross-class pro-democratic pressure was largely spearheaded by members of the middle class, including state employees, private-sector professionals and intellectuals. This regional specific outcome, they argue, can be attributed to the deficiencies of dependent import-substituting industrialisation limiting working-class power. Furthermore, pointing to the causal relevance of divergent industrial structures, they also find that in Latin America, '[w]ithin the industrial sector, the predominance of small enterprises with paternalistic labor relations hampered labor organization' (Rueschemeyer, Huber Stephens, and Stephens 1992, 180).
Inspired inter alia by Rueschemeyer et al., Lisa Mueller (2018) provides a pioneering analysis of the class dynamics underlying political change in contemporary sub-Saharan Africa. Exposing the notion of ''classless Africa" as a myth (Mueller 2018, 8), she stresses that the region's post-colonial leaders have historically headed corrupt ruling classes using the state apparatus 'to accumulate wealth and oppress the poor' (Mueller 2018, 65). She further contends that on the basis of their clientelist state-centric economic policies, these state classes blocked the emergence of an autonomous middle class, ie a 'stratum of Africans who meet their basic material needs with income from sources outside the state' (Mueller 2018, 9, emphasis in original).
According to Mueller, such a middle class eventually emerged in the late twentieth century following the economic crises of the 1970s and the 1980s and the neoliberal structural adjustment policies subsequently enacted. On the one hand, Mueller (2018, 67) states that 'economic liberalization was initially painful for a middle class built on patronage and urban bias' . On the other hand, she finds that as private capital accumulation was becoming possible, 'in the long term, the middle class flourished' (Mueller 2018, 67).
The intellectuals, business owners and salaried professionals constituting this middle class, Mueller argues, have acted as driving forces behind political opposition movements. Most crucially, according to Mueller, their mobilisational efforts have decisively contributed to the emergence of the 'third wave of protests in sub-Saharan Africa' (Mueller 2018, 67) beginning in the early 2010s. 1 Mueller identifies middle-class representatives as the main organisers and spokespersons of protests. yet she also finds that the rank-and-file of protest movements in Africa is primarily made up of 'an economically marginalized lower class' rooted in 'the uneven distribution of rising incomes' (Mueller 2018, 54). Thus, according to Mueller, middle-class-led cross-class coalitions are a key feature of contemporary African politics.
While Mueller embraces the concept of the 'state class' , she does not refer to the state class theory of Elsenhans (1996). In line with other authors, such as Sklar (1979), Diamond (1987) or Boone (1998), Elsenhans discusses the genesis and evolution of bureaucratic post-colonial ruling classes appropriating economic surplus via the state apparatus. However, in contrast to these authors, Elsenhans' framework possesses a crucial macro-economic component. Central to this component is the distinction between two types of economic surplus: profit and rent. While, according to Elsenhans, profit is the outcome of apolitical market competition, rents are appropriated via political means.
Building on the classical Keynesian assumption that the total volume of profits in a given period depends on the size of investment spending (and not vice versa), Elsenhans proposes a model wherein the development of the productive forces follows from rising mass incomes. Guided by this assumption, Elsenhans contends that in the post-colonial societies of the Global South, the predominance of rents has been rooted in a structural deficit: low agricultural productivity implies huge labour surpluses, which inhibit real wage increases, as there is a large reserve army of labour willing to work for subsistence wages. As a result, demand for machine-produced consumer goods is low so that private entrepreneurs are kept from appropriating the existing economic surplus as profit via investment spending. Under these conditions, Elsenhans (1996) argues, economic surplus has to be primarily appropriated politically, ie in the form of rents. Against this background, Elsenhans diagnoses the postcolonial genesis of bureaucratic development societies dominated by state classes. He states that '[t]he state class includes all those employed in the state sector (administrative set-ups and state-owned enterprises) who, as compared to the average worker, enjoy higher incomes, greater opportunities for participation and greater prestige' (Elsenhans 1996, 177). More precisely, the state class comprises those parts of the bureaucracy that oversee the appropriation and/or distribution of rents or are involved in decision-making on the executive level in the bureaucracy, enterprises owned by the state and organisations tied to the state (Elsenhans 1996, 177). Further personnel employed by the state are considered the organic clientele of the state class. It is labelled 'organic' since it depends on the growth of the state class and 'clientele' as it enjoys privileges -job security and high incomes -of which the large part of the population is deprived (Elsenhans 1996, 177f ). Trade unions that aggregate interests of the personnel employed by the state class in such bureaucratic development societies can be considered the organic clientele.
In Elsenhans' Keynesian framework, state classes face the challenge of re-orienting bureaucratic development societies towards capitalism by implementing specific economic policies. In particular, he attaches great importance to agrarian reforms, coupling productivity growth with an egalitarian income distribution and rising mass incomes. At the same time, he highlights the self-privileging tendencies of state classes resulting from their clientelist grip over the masses and their use of economic surplus free from market pressures. Thus, Elsenhans generally attributes state classes an ambivalent role when it comes to overcoming economic underdevelopment.
Drawing on Elsenhans' theoretical and conceptual propositions, we define two opposing poles of a spectrum of former bureaucratic development societies. At one end of the spectrum are societies where a state class successfully brought about a comprehensive transition to capitalism. At the other end are countries where the state class' economic interventionismas well as the subsequent neoliberal structural adjustment -did not even come close to achieving this transition. On the basis of this conceptualisation, we now investigate the potential of middle-class mobilisation of the lower classes to challenge the socio-political dominance of the remnants of the state class and its organic clientele. This implies going beyond the study of democratisation as political openings may 'merely' lead to competitive electoral politics between political parties dominated by state class patrons and their clientelist networks. Thus, our dependent variable differs from those of both Mueller (protest) and Rueschemeyer et al. (democracy).
We hypothesise that in former bureaucratic development societies where state classes bring about a comprehensive development of the productive forces, the empowerment of labour provides a mass base for the formation of middle-class-led competitive political parties. Such parties possess a mass base independent of the clientelist networks of the state class. Under these conditions, cross-class coalitions operate as competitive challengers to the socio-political dominance of the remnants of the state class.
By contrast, we contend that in former bureaucratic development societies where the transition to capitalism has drastically failed, the sphere of multi-party politics tends to be dominated by state class patrons and their clientelist networks. As shown by Mueller (2018), in these societies middle-class efforts to mobilise the lower classes for political change normally only arise in the aftermath of the onset of multi-party politics. Building on this finding, we argue that while middle-class protest organisers may successfully push for political reforms, cross-class alliances tend to fray quickly. The weakness of labour, the resilience of clientelism and the bad reputation of conventional party politics incentivise middle-class leaders to refrain from trying to transform alliances into organised and viable alternatives to state-class dominated parties.

The divergent paths of cross-class coalitions in Taiwan and Burkina Faso
In line with our focus on the opposite poles of a developmental spectrum of bureaucratic development societies, we compare Taiwan and Burkina Faso according to a diverse selection logic. Taiwan's state class managed to bring about a comprehensive development of the productive forces. It pursued a strategy of import substitution combined with an export-led growth strategy via small-and medium-sized enterprises (SME) to facilitate technology learning. Rising mass incomes with an egalitarian and productive agricultural sector allowed Taiwan not only to subsidise industrialisation but also to reach selfsufficiency of foodstuffs and the undervaluation of its currency to make products more competitive. In Burkina Faso, in order to transform the economic structure, factions of the Burkinabé state class sought to implement similar policies that aimed to increase mass incomes but failed for various reasons. As a result, Burkina Faso transitioned towards neoliberal structural adjustment.
Echoing Rueschemeyer et al. 's propositions on the political implications of capitalist development, we contend that the contrasting developmental trajectories of Taiwan and Burkina Faso crucially shaped the political prospects of middle-class-led alliances that emerged in the two countries at different points in time.

Rapid economic development in Taiwan
Taiwan's state class introduced economic policies that changed the structure of the economy fundamentally in a short time span, making labour scarce within only two decades.
The Kuomintang (KMT) had lost the Chinese civil war and took refuge in nearby Taiwan. They established authoritarian rule in 1949, banning labour strikes and the formation of political parties. Furthermore, independent media outlets were forbidden (Chu 1996, 496). However, these measures were complemented by policies to create legitimacy such as an egalitarian land reform (Gold 1986, 8, 123f;Wong 2003, 242). The land reform served to cut off the landed elite from its power base and assuage peasants (Chu 1996, 496;Wong 2003, 243). By nationalising large-scale industries, the KMT also gained control of the industrial economy located in urban Taiwan (Chu 1996, 496;Ranis 1995, 511). Against the backdrop of communist advances in Southeast Asia, Taiwan received considerable economic support from the US, just as its neighbour South Korea did. However, Taiwan was much less dependent on this outside intervention by the US than nearby South Korea (yi 1988, 131-36).
Although Taiwan had relatively good preconditions for the development of its productive forces, the country still had a labour surplus in the 1950s (Ranis 1995, 511). Nevertheless, excess labourers were soon to be absorbed by the market due to the burgeoning agricultural and industrial sectors. Taiwan significantly increased its agricultural productivity after the land reform process. It witnessed remarkable growth rates from the 1950s onwards, and also in non-agricultural output once it had shifted from an import-substitution policy to an export-oriented strategy. Overall, economic transformation on the island was institutionally assisted by the Taiwanese state, relying on various sector-specific support services and policies (Brautigam 1994). One prominent example of this transformation from the primary sector to industrial manufacturing took place in the food processing sector, which represented the fastest growing sector in Taiwan in the early 1960s outside agriculture (Ranis 1995, 511f, 518).
Three features of Taiwan's economic transformation merit more attention here: first her egalitarian income structure due to the land redistribution and the rising mass incomes, second the reliance on SMEs and third the burgeoning export sector. First, incomes were very evenly distributed due to the land redistribution. The Gini coefficient for rural incomes declined from approximately 0.50 in 1950 to 0.31 in 1970 (Ranis 1995, 515), meaning that Taiwan's farmers saw their real incomes more than double in roughly the same time frame (Ho 1979, 91). Higher incomes meant more purchasing power for consumer goods. The rural industrial sector could supply these goods, especially those that were income elastic, such as furniture (Ho 1979, 94;Ranis 1995, 515).
Second, the remarkable aspect about Taiwan's economic transformation was its highly decentralised industrialisation. SMEs mostly located in the countryside were at the forefront of industrial development. The heavy reliance on SMEs facilitated rapid technological learning, as imported technology had to be adapted to the smaller local market and its specific needs (yi 1988). Taiwan's reliance on SMEs in industrialisation also guaranteed the rural population the ability to earn a livelihood in the countryside. From 1956 to 1966, 46% of the new workforce in manufacturing was hired in rural SMEs as rural-based industries were highly labour intensive (Ho 1979, 83, 86). The overwhelmingly rural distribution of the SMEs ensured that labour did not have to migrate to the urban centres in pursuit of a better livelihood (Ho 1979, 78, 93). The reliance on small-and medium-sized businesses impacted the method of workforce organisation, a matter to which we will turn in the sub-section hereafter. According to Minns and Tierney (2003, 109), in 1971, almost 70% of all enterprises had fewer than 20 employees. In the 1980s, 90% of the companies had a workforce of fewer than 30 employees. Furthermore, in the early 1990s, 98% of the total number of firms were SMEsabout 700,000 firms -employing 70% of the workforce (Minns and Tierney 2003, 109).
Third, the shift from an import-substitution policy towards an export-oriented strategy based on SMEs and the economic boom based on this export strategy saw a 'new commercial sector [emerge that was] relatively independent of the party-state' (yang 2007, 509f). From 1952 to 1980, the non-agricultural output always saw double-digit percentage growth, with an all-time high of 18.5% from 1960 to 1970 (Ranis 1995, 512). The export-oriented strategy, which was accompanied by a devaluation policy, radically transformed the economy. In 1952, Taiwan's exports consisted of almost 92% of agricultural products, but in 1988, almost 95% of the exports were industrial goods (Ranis 1995, 522). The transformation was reflected in the occupational transition of the workforce. Employment outside agriculture rose from 29% in 1956 to 67% in 1980 (Ranis 1995, 514). The above-mentioned dynamic firms were mainly controlled by native Taiwanese (yang 2007, 510). The relevance of this factor will become more evident when we turn to the convergence of middle-class and working-class interests.
Due to the export-oriented strategy that focussed on rising mass incomes, first through the land reform and later through its decentralised industrialisation, Taiwan experienced labour scarcity only two decades after it had embarked on a process of industrialisation (Ranis 1995, 512-522). The economic transformation had changed the social structure of Taiwan extensively. These changes later allowed for democratisation. As Wong (2003, 244) highlighted, 'equitable growth under authoritarianism meant that economic inclusion in Taiwan was realized before the extension of political citizenship' . Economic inclusion, in turn, facilitated a stable and sustainable class coalition between scarce labour and the middle classes, as we illustrate in the next section.

Labour empowerment and the emergence of a sustainable class coalition in Taiwan
Only a decade after Taiwan experienced labour scarcity in the 1970s, civil society's increased self-confidence led to the founding of several organisations that established close ties with oppositional forces. These organisations were led by the middle class and brought human rights, feminist, environmental and labour issues to the fore. As Ho (2010, 6) observed, '[m]edical doctors, journalists, college professors and lawyers were instrumental in establishing pioneer social movement organizations' . More importantly, however, was the emergence of oppositional candidates first in local and then later in national politics. In addition to these movements and new oppositional forces, government-critical media outlets increasingly demanded political reforms from the KMT government. However, the KMT government sought to undermine these movements in the 1980s, and repressed or/and co-opted some of its leaders into the state class (Ho 2010, 6-8;Jacobs 1981, 37). Nevertheless, the oppositional movement asserted itself against the state class, since the cross-class alliance between the working class and the middle class was durable.
Oppositional politicians already participated in national and local elections in the late 1960s. The organisation of oppositional forces, however, only began to gain momentum in the 1970s, which correlated with the achievement of labour scarcity. The main oppositional challengers, the so-called Tangwai (people outside the KMT), had won several seats in the 1977 local county and township executive elections (Lee 1980, 49-52). Already, support for the Tangwai mainly came from the popular classes (yang 2007, 520f, 524f ).
Early on, the KMT had opted for a corporatist labour regime, which tied the working class under a KMT-controlled labour union to the state class (Buchanan and Nicholls 2003, 224;see also Minns and Tierney 2003, 111). In 1986, almost one-fourth of the workforce was unionised in these labour unions (Chu 1998, 191). Discontent with this arrangement was evident, however, as 'a 1980 university survey of union members found that less than 10 percent expressed satisfaction with union cadres and nearly 90 percent believed that unions were not capable of representing their opinions' (yang 2007, 524). Workers' waning support for this regime was reflected not least by increased strike activity. Using data on strike activity in Taiwan from 1984 to 1998, Buchanan and Nicholls (2003, 227) showed that there was a consistent upward trend of stoppages, increasing from 907 strikes in 1984 to 4138 stoppages in 1998. The strikes mobilised only 9000 workers in 1984, but this number increased to 104,000 employees in 1998 (although the average number of workers participating in these strikes in the years between 1985 and 1997 was usually much lower than during 1998) (Buchanan and Nicholls 2003, 227). However, it is important to note that stoppages in Taiwan were of a briefer nature, included fewer workers and hence meant fewer lost work hours than in South Korea. Also unlike in South Korea, they were also more locally bound and limited to certain firms (Buchanan and Nicholls 2003, 233).
The mostly urban-based social movements recruited heavily from the middle class. 'While the numbers pale in comparison with the great 'people power' movements of the Philippines and South Korea, their very emergence in normally placid Taiwan shocked the regime to the core' (yang 2007, 522). Together with the increased self-confidence of labour, the increasingly self-confident politicians fed into the foundation of organised oppositional parties (Chu 1996, 496). Despite the prohibition of oppositional political parties, the first formalised oppositional party, the Democratic Progressive Party (DPP), was founded in 1986. The martial law introduced in 1949 was abandoned in 1987, a year after the foundation of the DPP (Ho 2010, 8;Chu 1996, 496).
The relationship forged between the popular classes and the middle classes was reinforced during the abolition of the martial law. 'Especially in the heady days just before and after the lifting of martial law, workers mobbed the offices of DPP officials and flocked to legal seminars organized by opposition labour groups such as the Taiwan Labor Legal Assistance League' (yang 2007, 525). The labour movement established a close connection with the political party DPP, lobbying for political transformation through these party organs (Jung 2011, 398). Minns and Tierney (2003, 116) illustrated this clearly when citing an activist who reflected on the relationship between labour organisations and the DPP: 'They have tied their support organisations to the electoral ambitions of the Democratic Progressive Party' . Although the labour movement tried to form its own interest association with the foundation of a Labour Party in 1987, this political entity broke in two and was no longer a significant actor on the political stage (Buchanan and Nicholls 2003, 229). Labour's interests were thus closely bound to the DPP.
Following student protests and subsequent reforms of the electoral system, the DPP eventually entered parliament in 1992 and even won the presidency in 2000 2 (Ho 2010, 9-13;Arrigo 1993, 35;yu 2005, 111). The class-coalitional project that led to the formation of the DPP achieved this 'political liberation' approximately half a century after the KMT had devised and implemented policies seeking legitimacy but ultimately giving rise to the workers' strong bargaining power.
It is noteworthy, as yang (2007,525) writes, that the DPP's programme on socioeconomic matters 'was virtually indistinguishable' from the state class. The DPP had shifted its campaign from a leftist agenda to promoting an explicitly Taiwanese ethnic identity calling for the independence of Taiwan. yang (2007, 528f ) showed that class issues were in fact expressed in ethnic terms, as especially the working class was an ardent supporter of the Taiwanese national identity.
The presence of a nationalist discourse and the increasing absence of a distinct leftist discourse to mobilise the electorate becomes comprehensible once we reconsider the peculiarity of Taiwan's path to industrialisation. Due to the unique industrialisation strategy via SMEs, a considerable part of the country's workforce was employed by mostly small, family-run firms. The company structures and the dynamic growth of these companies meant that becoming a company manager or owner of one of these firms seemed feasible (Chu 1996, 496f ). Furthermore, until 1986 trade unions remained under the close control of the KMT government. When labour strikes rose, the KMT intervened immediately to resolve labour issues (Chu 1998, 191-195). Chu (1998, 195-197) highlighted that workers experienced the state as pro-business and hostile to their demands. Therefore, he writes, they drew a connection between their subjugation as workers and their political exclusion through the authoritarian regime. The mobilisation along nationalist lines and the DDP elite's awareness of the ability of such a discourse to mobilise a large part of the electorate, which yang (2007, 533) but also Minns and Tierney (2003) pointed out, now becomes intelligible. The two groups that benefitted immensely from the KMT's outward-looking production strategy were the workers and capitalists in the export sectors, who were mainly native Taiwanese.

Burkina Faso: economic stagnation
Burkina Faso's labour surplus has remained significant since independence mainly due to the fact that attempts to transform the economic structure through redistributive and mass income-oriented economic policies failed. In the late 1950s, 96% of the population was employed in the agricultural sector. The primary sector contributed close to two-thirds of the gross domestic product (GDP), while the industrial sector's share of GDP was only 3.5% (Ebong 1967, 9;Samo 1967). The predominance of the agricultural structure and low productivity levels persist to date (Speirs 1991, 94f). In 2018, manufacturing still only contributed about 5% of the country's GDP (World Bank 2019). Labour statistics from 2001 show the overwhelming importance of the informal sector, which employs 74.3% of the total workforce (Institut national de la statistique et de la démographie 2018, 64). The primary export commodities -primarily cotton and gold -have remained largely unchanged since independence (Samo 1967, 50;Lange 2018).
With the ascension of Thomas Sankara in the 1980s, Burkina Faso saw its first attempt to transform the country's economic structure. The Sankarist government wanted to reach self-sufficiency in the production of staple foods and standardised industrial goods through a redistributive policy called Réorganisation agraire et foncière. The aim was not only to increase production but also to include the peasantry in state-society relations by increasing agricultural output and nationalising and reorganising land allocation while subordinating middle-and upper-class consumption desires by reducing imports (Reza 2016, 97-101;Otayek, Sawadogo, and Guingané 1996, 10). Besides import controls and the attempt to harmonise demand, the core of Sankarist policies aimed to increase farmers' incomes through price increases and better commercialisation of agricultural products (Tallet 1996, 121). Agricultural support services such as sinking wells and planting trees to stop desertification were further important cornerstones of this policy (Reza 2016, 98). However, these policies failed not only due to the inefficiencies of para-statal marketing boards but also because of political resistance from powerful groups. Efforts to restructure the primary sector and attempts to create value chains in the various industries in the 1980s were mainly in vain.
Sankara had introduced wage freezes for civil servants and tight budgetary controls to finance agricultural development (Speirs 1991, 104), to narrow income gaps and orient the incomes towards mass consumption. Trade unions disapproved of these policies as wage freezes and cracking down on corrupt officials affected a large share of its members, the formal workforce. Sankara, in turn, brought the dissenting labour unions under his control. Nevertheless, resistance to redistributive policies also came from such powerful elites as the chefs de terre. The establishment of the Comités de défense de la révolution aimed at restructuring the system of land allocations and empowering farmers through redistributive land policies. The committees ultimately failed to sideline the influential chefs de terre, who have played a vital role in controlling access to land (Gausset 2008, 60;Tallet 1996, 123;Kuba 2004, 66f ).
The growing resistance of the elites and discontent with the leftist policies fed a coup d'état and the assassination of Thomas Sankara in 1987. Blaise Compaoré, a long-time ally of Sankara who is believed to have orchestrated the overthrow, changed the government's course immediately. His government re-integrated the alienated clientele of the state class as it raised wages, abandoned import controls and cut prices (Speirs 1991, 101-105). Further, Compaoré rehabilitated the chiefs as they provided vital support to mobilise the rural population (Reza 2016, 100-102).
As early as 1991, Compaoré reached a loan agreement with the International Monetary Fund (IMF) and committed himself to the liberalisation policies of the IMF. The country witnessed cuts to its welfare budget, while previously nationalised state companies were again privatised. Burkina Faso then became one of the top target countries for development aid in Africa and was one of the first countries to become part of the Heavily Indebted Poor Countries Initiative. Western donors lauded Burkina Faso for its privatisation and deregulation policies. Despite the impressive GDP growth rates in the first two decades after the liberalisation policies were implemented, the liberalised Burkinabé economy is still dominated by Compaoré and his inner circle (Reza 2016, 102f ).
Burkina Faso experienced periodic food shortages (Speirs 1991, 89) and staple foods still constitute a sizeable part of the country's imports. While under the prevailing circumstances self-sufficiency of staple foods could theoretically be achieved despite Burkina Faso's seemingly unfavourable conditions (Herrera and Ilboudo 2012, 88), the goal of autonomy in food production was no longer a priority of Compaoré's, in contrast to Sankara's governments. The high but volatile export-earnings of cotton and gold -in which the Compaoré family holds a significant stake -currently preclude a transformation of the agricultural sector in favour of the production of staple foods (Herrera and Ilboudo 2012, 87-90;Speirs 1991, 97;Reza 2016, 103).
The lack of structural transformation and the persistently high labour surplus had severe implications for the constitution of a class coalition. The weak bargaining power of labour in Burkina Faso only allowed the formation of a fragile middle-class-led alliance, which quickly frayed rather than transforming into a viable political alternative.

Failed empowerment of labour and limited impact of middle-class mobilisation
Protests and mass mobilisation are a constitutive element of Burkina Faso's post-colonial political history (Harsch 2017;Engels 2019). The most important actors driving these uprisings in the post-colonial era were labour unions representing a small and privileged formalised workforce eager to be co-opted by the state class.
As Burkina Faso's economy went into stagnation in 1966 shortly after independence, the government reacted with austerity measures and cut civil servants' wages. Labour unions mobilised against this policy and called for the military to oust the then president (Reuke 1979, 19-21). The organised workforce represented in these labour unions would once again become pivotal in exerting pressure on the regime, leading to another coup d'état, when Sankara's efforts to industrialise the country and to mobilise the peasantry were met with resistance by the trade unions. The opposition to these policies was triggered by the fact that they lowered the living standards of the unions' members. More importantly, the Comités de défense de la révolution, which aimed to absorb trade unions and traditional institutions (Loada 1999, 138), even threatened to render the unions irrelevant. The coup d'état against Sankara in 1987, which was likely led by Compaoré, his former deputy and minister of state at the presidency, marked the end of the attempts to restructure the class composition of Burkinabé society.
Demonstrators also took to the streets in 1998, 2006, and 2008, protesting against corruption, repression and persistently high poverty. Trade unions representing the tiny and privileged formalised workforce always played a pivotal role in these protests (Engels 2019(Engels , 2013Loada 2019). When president Compaoré tried to lift the constitutional presidential term limits in 2014 to extend his 27-year rule beyond 2015, protests erupted again. However, this time a cross-class coalition between the middle class and the lower classes, ie the urban poor, lobbied for change. In 2014, the movement mobilised large crowds against Compaoré's plans to lift the constitutional term limits and extend his rule beyond 2015.
Apart from trade unions and political parties that had already mobilised against Compaoré in the previous years and hence also played a crucial role, a central actor in this class coalition of 2014 was Le Balai Citoyen. According to Frère and Englebert (2015, 303) this movement 'made the fall of the regime possible' . Founded by two popular musicians Smockey and Sams K'le Jah, Le Balai Citoyen, especially popular with the young population, interwove political and economic issues. This enabled a broad coalition of forces interested primarily in the removal of Compaoré (Harsch 2017, 193, 198;Touré 2017;Gorovei 2016;Citoyen 2014). Not only its leadership but also its organisational structure was overwhelmingly urban-based (Commeillas 2015). Also, other organisations such as the Collectif anti-référendum and the Mouvement 21 avril mobilised against Compaoré's plan to prolong his stay in power (Carayol 2018).
The foundation of Le Balai Citoyen has to be situated in the changed political context of Burkina Faso, as Frère and Englebert (2015, 301) highlight, when they write: The birth of Balai Citoyen was an important development as it provided a hitherto largely alienated and disconnected youth with a means of political mobilization at the same time as it illustrated the relative impotence of regular political parties to genuinely represent the aspirations of the youth and, in general, to aggregate the demands of citizens beyond the small cliques of their members.
In mid-2000, Compaoré had already planned for his brother François Compaoré to succeed him as president after his last constitutional term officially ended in 2015. However, Compaoré's party, Congrès pour la democratié et le progrès (CDP), opposed his plan. Compaoré subsequently founded the Fédération Associative pour la Paix et le Progrès avec Blaise Compaoré (FEDAP-BC) in order to support François Compaoré as presidential candidate for the 2015 elections. After FEDAP-BC figures took over the CDP to push through Compaoré's brother as his successor, key politicians from Compaoré's inner circle -such as Roch Marc Kaboréwere pushed aside and subsequently defected. In 2014, they founded the party Mouvement du people pour le progress (MPP) (Frère and Englebert 2015, 300). Kaboré had served as prime minister in the 1990s and later as president of the parliament. In 2010, the CDP had already suffered a high-level defection by Zépherin Diabré, former minister for commerce, industry and mines under Compaoré in the 1990s. After his defection, Diabré founded the party Union pour le progrès et le changement (UPC) (Jeune Afrique 2020). Following the 2012 elections for the legislature and municipalities, the UPC took over the role of the main opposition party from the Union pour la Renaissance/Parti Sankariste (Harsch 2017, 194f ).
In the period directly before the ousting of Compaoré, these newly founded opposition parties also mobilised against the long-standing president. Recently founded opposition parties such as the UPC and the MPP, as well as trade unions, which had initially opposed the early mobilisational efforts of Balai Citoyen, joined the movement against Compaoré which then culminated in the latter's ouster in October 2014 (Carayol 2018;Frère and Englebert 2015, 296).
In 2015, after a short period with a transitional government, 3 the two former key figures from Compaoré's inner circle, Roch Marc Kaboré and Zéphirin Diabré, ran against each other in the presidential elections, which Kaboré won (Roger 2015). Le Balai Citoyen's co-founder Smockey had explicitly decided against transforming the momentum of the movement into a political party: 'We are a political movement, but we don't want to come to power or access any political office. We intend to represent a civic strength that can pressure the authorities to get them to work towards the people's interests' (Saddier 2014).
Later, Smockey explained that since Burkina Faso already had several parties, he preferred to remain independent so as to be able to control the government. However, he considered supporting a party standing for real change a possibility (Gänsler 2019;Bationo 2019). In fact, Le Balai Citoyen had called for the youth to cast their ballot in the 2015 elections and overcome their disillusionment with party politics (Harsch 2017, 223). Voter registration in the 2015 elections increased by 70% compared to registered voters in the 2010 elections (Harsch 2017, 223).
The ousting of Compaoré did not lead to a replacement of the state class with autonomous interest bodies of the lower classes. Even the return of the former long-time ruler is currently under discussion. After his re-election in 2020, Kaboré co-opted the former main opposition leader Zépherin Diabré 4 into his government and appointed him as minister for reconciliation (Roger and Duhem 2021). Amid the heavily deteriorating security situation in Burkina Faso, Diabré is reaching out to the exiled Blaise Compaoré to discuss his return to the country as the ruling elite envisages him negotiating with insurgent groups. 5

Discussion
Shortly after independence, state classes in Taiwan and Burkina Faso implemented economic programmes that aimed at catching up with industrialised nations and in the case of Taiwan eventually succeeded in orienting the economy towards a profit-based system.
Taiwan encountered labour scarcity only about two decades after it had embarked on its industrialisation strategy. The country's land reform and the various agricultural services raised mass incomes and subsidised the industrial sector. The chosen path of industrialisation via SMEs to better serve the demands of the local markets facilitated technological learning through adaptation. Furthermore, self-sufficiency in foodstuffs allowed the devaluation of the country's currency to increase the competitiveness of exports. The industrialisation strategy impacted on the class composition of the country. Workers already had a strong bargaining position in the 1980s. The middle class and its various movements also became more self-confident in the 1980s, staging protests and strikes actively criticising the KMT government, but more importantly, they channelled these forces into founding the opposition party DPP while martial law still prohibited any free political association. The state class initially tried to repress and co-opt the movement, but could not impede these movements to achieve 'political liberation' . The middle-class-led coalition first enforced the liberalisation of the polity and then replaced the remnants of the state class in 2000 by winning the presidential elections.
While Burkina Faso's government under Thomas Sankara also tried to implement policies to raise mass incomes of the rural sector, the country has had a considerable labour surplus since independence. As the organic clientele of the state class, Burkina's labour unions represented only the privileged and narrow interests of a small fraction of a formal workforce in the public and private sectors. Attempts to restructure and empower the lower classes were actively opposed by this labour aristocracy. The failed attempts at restructuring the Burkinabé economy made the state class dependent on donors and forced it to liberalise its polity and economy in the 1990s. While economic liberalisation gave rise to a middle class independent of the state class, the political liberalisation did not lead to competitive party politics with political forces independent from the state-class patrons. Given the tarnished image of party politics, the leading middle-class activists refrained from trying to transform cross-class alliances into organised and viable alternatives to state-class-dominated parties.
We selected Taiwan and Burkina Faso as two diverse cases of former bureaucratic development societies. We regard these cases as representative of a larger sample of the two distinct trajectories of bureaucratic development societies. Here, we briefly reflect on two other countries on the spectrum of the divergent trajectories of former bureaucratic development societies.
In other newly industrialised countries, such as South Korea, the state class pursued an equally successful strategy of industrialisation as Taiwan's ruling class. However, Korea's industrialisation process via conglomerates and not via SMEs meant that the workforce was more concentrated geographically and in fewer firms and had fewer prospects of becoming entrepreneurs themselves. Although the middle class initially mobilised to push for political reforms, 'the main driving force for democracy was transferred from the student movement and the Chaeya groups to the labor movement' (Jung 2011, 397, emphasis in original). The labour movement that was 'known for its militancy' (Jung 2011, 392) organised a hitherto unknown number of stoppages over several months in 1987, the year regarded as the breakthrough on the way to 'political liberation' from Korea's state class. The Korean case points to the fact that the impetus to replace the state class can also come from the labour movement. The interests of the middle class and the lower class can also converge more consistently at a later stage to vote the state class out of office via its own political force (as was the case in Taiwan).
Senegal's socio-economic and political conditions starkly resemble those of Burkina Faso and also gave rise to a middle-class-led class alliance disenchanted with the stalled economic and political developments in the country and disillusioned with party politics (Touré 2017, 64, 67). The class alliance mobilised against President Abdoulaye Wade's (2000)(2001)(2002)(2003)(2004)(2005)(2006)(2007)(2008)(2009)(2010)(2011)(2012) efforts to remove the constitutional presidential term limits in 2011. Demarest (2016, 70) showed that the class coalition itself was heavily dependent on financial contributions from former state-class patrons. These patrons broke away from the ruling party and then founded their own political parties as a vehicle to compete in presidential elections. The class alliance initially threatened to abstain from the elections if Wade's candidacy was not withdrawn -the Constitutional Court later confirmed Wade's candidacy for a third term as constitutional. However, the coalition soon faltered when former state-class patronsamong them Macky Sall, who then was elected president in 2012 -withdrew its support for the alliance and seized the opportunity to run in the presidential elections (Demarest 2016, 72). Notwithstanding the differences between Burkina Faso's and Senegal's political context, the dynamics of the cross-class alliance in the two countries are similar. The momentum of the cross-class coalition could not be transformed into a viable political alternative to the remnants of the state class, which allowed former state-class patrons to further dominate politics in Senegal.

Conclusion
Where labour is empowered prior to the onset of multi-party politics, as was the case in Taiwan, it can enter into a durable middle-class-led coalition capable of effectively challenging the remnants of bureaucratic state classes. It only took Taiwan two decades to reach labour scarcity and for a self-confident civil society to emerge, which in turn achieved a 'political liberation' from the state class. This fact underlines the relevance of Elsenhans' emphasis on the importance of egalitarian income policies coupling productivity increases with rising mass incomes.
In the absence of the successful implementation of such policies, in Burkina Faso multiparty politics has historically been dominated by former key members of the state class. Middle-class representatives have eventually resorted to mobilising the lower classes to push for political change. However, faced with workers' weak bargaining power and a political process characterised by the resilience of clientelism and patronage politics, they have been reluctant to work towards the establishment of a viable political alternative.
We have also shed light on the dynamics of South Korea's (successful development of productive forces and empowerment of labour) and Senegal's (very limited economic transformation and significant labour surplus) cross-class alliances as two further cases on the spectrum of bureaucratic development societies. South Korea's cross-class coalition equally pushed through a 'political liberation' of the former state class, although here the central impetus came from the lower class, and it was only at a later stage that middle-class interests converged with the working-class demands. The cross-class coalition in Senegal which emerged in the early 2010s did not transform into a viable political entity and lead to a replacement of the remnants of the state class. Former state class patrons further dominate the country's political landscape.
Indeed, as observed by Mueller (2018, 190), politically assertive middle-class activists in the former bureaucratic development societies of contemporary sub-Saharan Africa often 'are disillusioned with party politics as a way of effecting political change' . This disenchantment manifests a fundamental dilemma because, as inadvertently demonstrated by Taiwan's and South Korea's state class, any political project aiming to sustainably empower the masses ultimately needs to be based on state power.

Disclosure statement
No potential conflict of interest was reported by the authors.

Notes on contributors
Julian Friesinger is a researcher at the Institute of International and Intercultural Studies (InIIS) and a PhD Candidate in political science at the University of Bremen. He holds a master's degree in peace and conflict studies from the Philipps-University Marburg, Germany. His research interests include democratisation and social movements, political economy, and peace and conflict in sub-Saharan Africa.
Jannis Saalfeld is a researcher and PhD Candidate at the Institute for Development and Peace (INEF). He holds a master's degree in international relations and development policy from the University of Duisburg-Essen, Germany. His research interests include social movements, politicised religion and political change in sub-Saharan Africa.