4 - Distributive Threats, Elite Cohesion and the Emergence of the ‘Developmental State’

Summary

Chapter 4 focuses on the political dynamics that shaped the EPRDF’s development strategy from 1991 to the death of Prime Minister Meles Zenawi in 2012. Despite military dominance, the EPRDF was vulnerable due to its lack of a political base outside Tigray. The new government sought to consolidate control of the ethnically diverse peasantry through a dual strategy: establishing federalism that provided for ethnic self-determination; and a broad-based development strategy to secure compliance through mass distribution. From early on, however, the government recognised that the main form of distribution – land access – would be undermined by population growth, necessitating industrialisation and mass employment creation to maintain mass acquiescence. During the early 2000s, the EPRDF leadership experienced a series of crises that resulted in increased elite cohesion and a shared sense of the threat facing the ruling elite. The result was Ethiopia’s ‘developmental state’ model, comprising infrastructure development, industrialisation and agricultural commercialisation to deliver the economic transformation required to meet the mass distributive challenge facing the regime.

Chapter 2 argued that the main political drivers of state-led development entail the formation of a state with a high degree of autonomy and capacity, and elite threat perceptions that shape how political leaders use the state for developmental ends. The EPRDF inherited from the Derg a state with a high degree of state autonomy and considerably enhanced capacity and territorial reach compared to its Imperial predecessor. This chapter focuses on how elite political dynamics shaped the government’s development strategy from EPRDF’s arrival in office in 1991 to the death of the Prime Minister, Meles Zenawi, in 2012. Despite being militarily dominant on taking office, the EPRDF remained politically vulnerable due to its lack of a political base outside Tigray. This perception of vulnerability amongst the political elite resulted in the new government employing an approach derived from their successes in Tigray, seeking to consolidate control of the ethnically diverse Ethiopian peasantry through a dual strategy: establishing a federal system that, in principle, provided for ethnic self-determination; and pursuing a broad-based development strategy that aimed to secure compliance through mass distribution. Importantly, this development strategy was motivated in part by the challenge of population growth. From early on, the government recognised that land access would, in the medium term, be undermined by rapid population growth, necessitating industrialisation and mass employment creation to maintain mass acquiescence.

Progress with each prong of this strategy was initially limited, however. The federal system provided for the use of local languages in administration and the recruitment of regional officials from the titular ethnic group of the region, an important break from the past. Yet the reality was that centralised, top-down control within the ruling party and the requirements of a centrally defined development strategy restricted regional autonomy. Moreover, growing divisions within the ruling elite, resource shortages and limited state capacity undermined implementation of the development strategy. During the late 1990s and early 2000s, the EPRDF leadership experienced a series of crises, which the leaders themselves referred to as ‘Armageddons’, that threatened their hold on power. These crises resulted in increased elite cohesion through the centralisation of power around the Prime Minister, Meles Zenawi, and a shared sense of the threat facing the ruling elite. The result was what became known as Ethiopia’s ‘developmental state’ model, with a renewed focus on infrastructure development, industrialisation and agricultural commercialisation, as means of delivering the economic growth required to meet the mass distributive challenge facing the regime. The government also pursued a massive expansion of party-state infrastructural power with a view to delivering both rapid development and political control.

The analysis begins by reflecting on the political challenge facing the EPRDF on taking national power before examining how the transition from a rural insurgency to a national government led to growing fragmentation within the ruling party and a series of political crises in the early 2000s. The chapter then examines how these crises resulted in the formation of a coherent political elite with a common vision for development and a commitment to expanding the capacity of the state to realise its ambitious plans.

The EPRDF in Power: Consolidating Dominance

The EPRDF entered Addis Ababa and seized national power in May 1991. The Front was politically and militarily dominant thanks to its victory in the civil war. However, the EPRDF was also extremely politically vulnerable since it had spent almost the entire civil war mobilising the Tigrayan peasantry along ethno-nationalist lines, had no established political base outside Tigray and had only expanded the EPRDF coalition beyond the TPLF in the final years of the conflict. The challenge facing the new government therefore was, on the one hand, to forge a cohesive ruling coalition amongst the EPRDF parties and, on the other, to secure a mass base amongst an ethnically diverse population.

The EPRDF took a number of immediate actions to consolidate its position. The former Ethiopian army was demobilised and the TPLF army became the Ethiopian National Defence Force (Berhe 2017). In contrast, the civil service was retained, with bureaucrats requested to return to work and continue their duties. For an insurgency that had secured political power through force, the EPRDF’s initial approach to national politics was surprisingly open and inclusive. Rather than seeking to govern alone, the EPRDF organised a National Conference in July 1991 inviting a range of predominately ethnic-based parties – from long established movements like the Oromo Liberation Front (OLF) to more ad hoc creations with little established support – to participate in a debate about the country’s future (Vaughan 1994, Lyons 1996). The National Conference established a Council of Representatives and a Transitional Government, with representation of a range of parties, albeit that the EPRDF had a majority of seats and cabinet posts. The Transitional Charter acknowledged ethnic groups’ right to self-determination including secession, laying the groundwork for Eritrean secession following a 1993 referendum and the creation of a federal system in Ethiopia under the 1994 Constitution.

The inclusive spirit did not last long, however. The most established parties, other than the EPRDF, were the OLF and the Ogaden National Liberation Front (ONLF), which was formed from a splinter group of supporters of the Somali invasion in 1977–1978 (Woldemariam 2018). Though both initially participated in the Transitional Government, they departed acrimoniously due to the EPRDF’s creation of competitor ethnic parties in their home regions and what they considered to be unfair restrictions on their local election campaigns. Both were suppressed militarily. Most of the OLF’s armed forces were wiped out and some 19,000 Oromo were arrested in 1992 (Tronvoll 2009a, p. 454). While the OLF was severely limited as a military force from that point, it retained significant popular support in the region. The military attacked the ONLF in 1994, leading to the arrest or death of most of its leadership, reducing the organisation to small scale operations on the Somalia border (Woldemariam 2018). The Transitional Government ended in 1995 with the election of a new national government, and the elections duly confirmed the de facto distribution of power, with EPRDF parties and affiliates winning almost all seats and forming the new government with EPRDF chairman, Meles Zenawi, as Prime Minister. Most opposition parties boycotted the polls, a situation that was repeated in 2000.

While the EPRDF was quickly able to assert its dominance militarily and in national politics, it nonetheless faced a major challenge in terms of its lack of a popular base in the country. For almost the entirety of the civil war, the TPLF mobilised the peasantry in rural Tigray. To a lesser degree, its junior partner in the EPRDF coalition, then named the EPDM, had carried out similar mobilisation efforts in Wello. A very different situation applied in the rest of the country, where newly created EPRDF-linked parties had not spent time mobilising the ethnic groups they sought to represent and lacked local legitimacy, being viewed, not unreasonably, as top-down creations of the TPLF.

In seeking to build a popular base in the country as a whole, the coalition applied a very similar class analysis to that used previously by the TPLF (see Chapter 3). As laid out in a 1993 party document, the coalition viewed the peasantry – accounting for some 85 per cent of the population – as the Front’s key constituency, alongside the far less numerous urban proletariat,¹

Just as the center of our economic development program is the rural part of the country, so, too, the focus of our political work is the peasantry

(EPRDF 1993).

Indeed, the overwhelming political importance of the peasantry was reiterated by Meles Zenawi in a 1995 address to the House of Peoples’ Representatives,

Let the peasants never be disaffected. Once they are disaffected, it will be the end of the world. But whatever happens, with the support of the peasantry we may stagger, but we would surely make it

(Meles, cited in Markakis 2011, pp. 248–249.)

In crafting a political and economic strategy to secure the control of the peasantry, the EPRDF applied a variation of the same approach that the TPLF had used with great success in Tigray: ethno-nationalism; broad-based distribution; and, where necessary, coercion. The EPRDF built on its origins in the Ethiopian student movement and the debates about the national question, with the party convinced that peasants could be best mobilised in their own language and by members of their own ethnic group (Vaughan 2003, Zeleke 2019). The result was the reconfiguration of the Ethiopian state based on ‘the wholesale takeover of Stalin’s theory of the national question as an approach to the problem of Ethiopia’s ethnic diversity’ (Clapham 2006, p. 148).² Not only was the EPRDF coalition constructed along ethnic lines, but the Ethiopian state was reconstructed as a federal system by the Transitional Charter and subsequent 1994 Constitution, with the country divided into ethno-linguistic regional states (see Figure 4.1). The government delineated ethnic regions based primarily on the dominant language in an area, providing for self-administration of these regional states and the right to self-determination, with the new constitution guaranteeing the right to secession. Each of the newly created regions was ruled by an EPRDF party or affiliate.

Figure 4.1 Ethiopia’s federal boundaries

Source: author/Manchester Cartographic Unit.

The federal system constituted a major practical and symbolic shift from the nation-state building project of past regimes. For the first time, local languages, rather than Amharic, became the working language of the state and were expanded as means of primary education in each of the ethnic regions.³ Moreover, ethnically defined regional states, rather than the federal government, became the central reference point of day-to-day state–society relations. Federalism signalled a sharp shift from previous regimes regarding the official narrative of Ethiopian history. Haile Selassie had sought legitimacy in the Solomonic myth that linked the Emperor to Menelik I, 3,000 years previously, while the Derg had maintained an emphasis on the historic integrity of Ethiopia. In contrast, the EPRDF took the view – common within the student movement – that Ethiopian history was primarily a story of Amhara expansionism and exploitation over the past century (Tronvoll 2009b). Indeed, Meles Zenawi explicitly questioned key Ethiopian national symbols when he was quoted as arguing,

The Tigreans had Axum, but what could that mean to the Gurague? The Agew had Lalibela, but what could that mean to the Oromo? The Gonderes had castles, but what could that mean to the Wolaita?

(Lyons 2019, p. 52)

As argued in Chapter 3, Ethiopian history is characterised by a high degree of fluidity in territorial control, language, and the degree of ethnic self-identification. Yet, the ethnic federal system applied a rigid, primordialist conception of ethnicity that drew heavily on Stalinist theory and used language to identify ethnic territories and individuals’ singular ethnic identity (Vaughan 2003, Kymlicka 2006).⁴ This simplistic approach was inevitably a poor fit for many parts of the country where history was considerably more complex than the nineteenth century Shewan-Amhara colonisation and oppression of distinct nations, nationalities and peoples. Prominent flashpoints have included: predominately Amharic-speaking urban centres in southern Ethiopia (see Chapter 6); significant Oromo populations in Wello in Amhara region; ethnically diverse Raya and Welkait which were included in Tigray, but in past periods had been administered by other provinces; the numerous ethnic groups with very small populations which were abruptly grouped into the Southern Nations, Nationalities and Peoples Region (SNNPR) for administrative convenience (Vaughan 2003); and contestation over questions of indigeneity in and political boundaries between most regions, but particularly those where mobile livelihoods are common: Afar, Benishangul-Gumuz, Gambella, Oromiya and Somali (Young 1999, Feyissa 2005, Kefale 2013).

As with the TPLF’s peasant mobilisation in Tigray, the EPRDF’s attempts to consolidate a base amongst the peasantry built both on the symbolic appeal of ethno-nationalism and material distribution. Mass distribution in Tigray entailed land reform and modest attempts to expand agricultural extension, health, education and emergency assistance within the constraints of the conflict and the imperative of self-reliance, all channelled through party structures. Similar principles underpinned the EPRDF’s distributive approach nationally, which, as discussed at length in Chapter 5, rested on a strategy of Agricultural Development-Led Industrialisation (ADLI). The EPRDF inherited a country in 1991 in which low productivity smallholder agriculture accounted for 77 per cent of employment and 59 per cent of GDP, and in which GDP per capita was lower at just $187 than it had been in 1981. As a result, 71 per cent of the population fell below the World Bank’s $1.90 poverty line in 1995.⁵ The ‘fiscal situation inherited from the Derg was disastrous’ as a result of the limited tax base and massive military expenditure during the civil war (Mascagni 2016, p. 30). Moreover, as Marxist-Leninist rebels, EPRDF had little expectation of securing large-scale support from western donors. ADLI therefore sought to combine mass distribution aimed at securing the control of the peasantry, with productive investment, promoting ‘self-reliance, meaning dependence on national resources and independent national development’ (MoPED 1993, p. 16). The main thrust of the strategy was to expand access to improved agricultural inputs such as fertiliser, seed and irrigation for peasant farmers who already depended on the state for their access to land. Improved agricultural productivity, in turn, was expected to produce a surplus that could be used to support industrialisation, thereby creating dynamic linkages between agriculture and industry. Expansion of infrastructure, education and health aimed to support growth in both sectors. As such, the strategy was explicitly focused on economic growth and structural transformation, but also ‘equity, including regional equity’ (MoPED 1993, p. 16). ADLI therefore moved beyond the land question of the 1970s in an attempt to address the challenge of the agrarian question in Ethiopia.

ADLI’s political importance derived not only from the EPRDF’s need to consolidate a peasant base outside Tigray in the short term, but also recognition of what the government described as ‘the specter of accelerating population growth’ (MoPED 1993, p. 14) that represented the ‘single overriding challenge of Ethiopia in the next three decades’ (TGE 1994, p. 8). Following the land redistribution of the 1970s, state power rested to a considerable degree on the enmeshment of the peasantry through state land ownership and the distribution of usufruct rights. By the time the EPRDF took office, however, Ethiopia’s population was growing at more than 3 per cent per year, with rapid growth in both rural and urban areas (Hailemariam 2019).⁶ Consequently, the National Population Policy of 1993 viewed population growth as a major threat to stability, leading to rural landlessness and food insecurity, and migration to urban centres where employment was limited (Office of the Prime Minister 1993). As such, the Population Policy laid out plans to ease the problem by reducing fertility through increased use of contraceptives, raising the minimum age of marriage and education campaigns promoting small family sizes, as well as through broader initiatives to increase women’s education and labour force participation. Ultimately, however, the government recognised that upkeep of the system of coercive distribution and political control in the context of population growth required the distribution of new resources beyond the agricultural sector, in a context in which state revenues were extremely limited. ADLI was therefore explicitly conceived as a strategy for agricultural intensification to support a growing rural population in the short-term, while enabling a shift to an industrial economy over a period of 20 years, creating mass employment to absorb surplus labour. The political imperative of ADLI’s success was succinctly summarised by Newai Gebre-Ab, one of the Prime Minister’s leading economic advisors,

The race between demography and growth in agriculture has to be won in favour of growth.⁷

Elite Fragmentation and Stalled Development

Despite this attempt to frame national development as a key means of addressing mass distributive pressures, the EPRDF’s transition from a Maoist insurgency mobilising the Tigrayan peasantry to a national government promoting structural transformation brought to the fore a number of tensions that gradually undermined the cohesiveness of the ruling coalition. These tensions concerned the relationships between the political leadership and foreign donors; between the party and the state bureaucracy; and within the EPRDF coalition itself.

The first challenge was that the EPRDF came to power in 1991 at a time when socialism was widely discredited and at the height of the Washington Consensus and its focus on market liberalism. The new government desperately needed external support to finance reconstruction and its plans for economic development. The EPRDF strategy was to shift its discourse, at least as presented externally, from the Marxist language of the past to that of markets and democracy in an attempt to placate western donors. However, this raised important, unresolved questions about the new government’s true intentions and the appropriate role for state and market in its development strategy. The EPRDF government signed up to modest structural adjustment programmes with the World Bank and IMF in 1993 and 1996 that reduced tariffs and taxes, removed price controls and privatised some state enterprises (Borchgrevink 2008, p. 202). However, implementation lagged and the government refused to go further, insisting on a managed exchange rate and refusing any discussion of privatisation of key sectors such as finance, land, energy and telecommunications. The government’s resolve was tested in the late 1990s when IMF loans were conditioned on reduced government spending and further privatisation of the financial sector (Stiglitz 2003, Manyazewal 2019). However, the government, and Prime Minister Meles in particular, refused to back down, despite the IMF cutting off lending, sending a signal to donors that it was willing to make sacrifices to defend its policy autonomy.⁸ The result was that the EPRDF government received relatively little aid in its early years, while the confrontation raised questions within the party regarding the role of the state and market going forward.

The second challenge concerned the relationship between the new government and the bureaucracy. Retaining Derg era officials had facilitated the maintenance of order and service delivery. Nonetheless, many bureaucrats were sceptical or outright resistant to the new government (Vaughan and Tronvoll 2003, De Waal 2015). The EPRDF’s response was to strengthen party control over the bureaucracy, putting party officials through rapid training in the Civil Service College and placing them into key positions, while also extending the gim gema evaluation system used by the TPLF to the entire civil service (Tadesse and Young 2003, Vaughan 2011, Weis 2015, Berhe 2020). The creation of regional governments from scratch presented a distinct challenge, particularly regarding the recruitment of sufficiently qualified and capable officials proficient in local languages. A common practice was for TPLF officials to be allocated to regional governments to support this regional state and party building. Inevitably, however, the formulation of a national development strategy by the federal government, the EPRDF’s concerns about maintaining political control and the weakness of the new regional administrations meant that there was a continual tension between the formal devolution of powers under federalism and the de facto distribution of power between the federal government, political party and regional administrations. In practice, decision-making remained highly centralised (Fiseha 2006, Aalen 2011).

The third source of tension concerned growing divisions within the EPRDF itself. A central challenge concerned the incorporation into the EPRDF of ethnic-based parties that did not share the TPLF’s revolutionary history and which lacked a popular base in the regions they claimed to represent. The EPRDF consisted of four parties, namely the TPLF, ANDM, OPDO and the SEPDM, along with a range of affiliated parties created to administer the peripheral regions. While subordinate to the TPLF within the EPRDF, ANDM nonetheless built on the EPDM, an established organisation prior to joining the EPRDF coalition. In contrast, the TPLF created the OPDO and SEPDM in 1989 and 1992, respectively. In establishing these parties, the TPLF avoided reaching out to social actors with independent legitimacy in their communities since most had ties to either the Imperial government, the Derg or the OLF. Instead, EPRDF parties recruited first among re-educated prisoners of war and then those with basic education, such as teachers who would come to form a new political class which depended on the party, not popular legitimacy, for its power (Vaughan 2003, 2011). Relations between the EPRDF and affiliated parties formed to govern the peripheral regions of Afar, Benishangul-Gumuz, Gambella and Somali have proven even more problematic, with leaders recruited primarily to keep the peace locally and execute federal orders, but with little concern for popular representation (Young 1999, Feyissa 2005, Hagmann 2005). The inevitable result was both the ‘ideological watering down’ of the TPLF, as well as growing frustrations among other ethnic groups at TPLF dominance and the illegitimacy of their assigned regional leaders (Tadesse and Young 2003, p. 395).

These strains were also reflected within the TPLF itself, with the unity and collective leadership forged during the insurgency in Tigray gradually fragmenting during the 1990s (Milkias 2003, Tadesse and Young 2003). The main fracture point was between the federal government in Addis Ababa and the TPLF party structure in Tigray. Meles Zenawi became chairman of the TPLF and EPRDF in 1989 and took the leading role in the federal government after 1991, first as President of the Transitional Government and then, from 1995, as Prime Minister. Meanwhile key officials in the TPLF leadership in Tigray began to form an alternate centre of power, particular after Meles’ decision to post critics such as Siye Abraha – the former Minister of Defence – out of the way to Tigray (Africa Confidential 2001). Moreover, many senior TPLF officials were tasked with the separate job of building the regional states in Oromiya, SNNPR and elsewhere (Vaughan 2011). The result was an emerging division between two factions, one involving Meles and his allies and another in Tigray involving Siye and others.

On coming to power, therefore, the new government identified mass distributive pressures as a major threat to its power and a combination of rapid, broad-based development and a federal system as means of securing mass acquiescence. Nonetheless, the government’s ability to act on this strategy was constrained not only by the challenge of post-conflict reconstruction and the shortage of resources, but also elite factional and ideological divisions. Addisu Legesse, ANDM chairman and deputy Prime Minister from 2001, highlighted the difficulties at this time,

During the struggle we had a socialist or communist ideology. But after the fall of the USSR this was not possible anymore, it didn’t have a place. But it did contribute a lot in terms of being pro-poor. The westerners had doubts about whether we were communists. So they pushed privatisation, land reform, power. There were so many challenges. There was confusion within the party at that time regarding which way to go. We needed to clarify this internally and externally … we didn’t have the same ideas on development, about how to go forward.⁹

Elite Threat Perceptions: Armageddons and Renewal

The EPRDF leadership faced a series of crises from the late 1990s that came to be known internally as the ‘Armageddons’.¹⁰ These crises resulted in a major split in the ruling coalition, an increase in elite cohesion and renewed commitment to a slightly revised development strategy as an imperative for the EPRDF’s political survival.

The first crisis was the Eritrean war (1998–2000). Despite collaboration between the TPLF and EPLF in the final stages of the civil war, interactions between the parties had always been problematic. Relations deteriorated significantly in the 1990s, with disputes over economic policy between the newly separated states playing an important role. In Ethiopia’s efforts to protect the domestic economy and promote industrialisation, it limited Eritrean imports, investment in key economic sectors such as banking, electricity and air transport, as well as rejecting Eritrea’s proposal that its new currency would trade at parity with that of Ethiopia (Negash and Tronvoll 2000). Tensions came to a head over disputed territory along the border, which had never been adequately demarcated since the Italian occupation of Eritrea. The Eritrean Army invaded and made initial gains in 1998, but was subsequently repelled by an Ethiopian counter-offensive in 1999, which re-captured lost territory and advanced into Eritrea, forcing the Eritreans to accept a previously rejected peace treaty. The war served to exacerbate the emerging division within the TPLF, with Prime Minister Meles marginalised during the conflict by the establishment of a Central Command in Tigray, which was dominated by the leadership in Tigray (Tadesse and Young 2003, Vaughan 2011). In negotiating the end of hostilities, Meles, sensitive to external pressures, took a more conciliatory approach, while the dissidents were more aggressive, seeing the opportunity to demolish the Eritrean military.

Following the war, the TPLF held a meeting in early 2001 to evaluate the party’s performance over ten years in office and this meeting became the battleground for a far-reaching debate over the future of the party that pitted the two factions against each other. The debate was framed in Marxist terms. On one side, Meles submitted a 700-page document that drew on Karl Marx’s The 18th Brumaire of Louis Bonaparte (1869), accusing his opponents of Bonapartism – essentially a regime which relies on political support from the peasantry and as a result of a lack of opposition can use its power to enrich itself through corruption (Tadesse and Young 2003, Simon 2011). In contrast, for his opponents, Meles’ camp was ‘selling out to capitalism’, abandoning the party’s anti-imperialist stance in its subordinate relations with the USA and IFIs, and guilty of being far too lenient on Eritrea (Africa Confidential 2001). After a lengthy debate, Meles’ faction prevailed by the narrowest of margins through a combination of procedural manoeuvre and power grab, with the dissidents expelled from the party.

Subsequently, Meles launched a process of renewal (tehadso) intended to purge his opponents first from the TPLF and then other EPRDF parties. The ANDM leadership, including Addisu Legesse and Bereket Simon, came out strongly in favour of Meles and, as a result of their loyalty, was spared (Tadesse and Young 2003). In contrast, key figures in the other parties such as Kuma Demeksa – the OPDO leader since its founding in 1989 – and Negasso Gidada – then President of Ethiopia, as well as a key OPDO leader – and Abate Kisho – President of SNNPR – supported the TPLF dissidents and were expelled. Kuma and Negasso had, in the years leading up the split, become increasingly frustrated at TPLF’s dominance and the lack of Oromo autonomy (Milkias 2003). As such, Kuma and Negasso were labelled as ‘narrow nationalists’ and their removal from office and the party quelled for the time being the push for greater regional autonomy. Moreover, their dismissal enabled the promotion of a new leadership loyal to Meles, including: Abadula Gemeda, a trusted military leader since the struggle, first as Minister of Defence and then Oromiya regional president, and Girma Birru in the OPDO; and Kasu Ilala and Hailemariam Dessalegn – Meles’ eventual successor as Prime Minister – in SEPDM. This new leadership’s political fortunes depended heavily on Meles, limiting potential for further splits and heightening the sense of shared threat in the face of the Armageddons that followed. The result was the erosion of the TPLF’s collective leadership and the centralisation of power around Meles and a handful of advisors, with ‘Meles assuming a position of unchallenged supremacy’ over party and state (Tadesse and Young 2003, p. 397, Vaughan 2011). While these crises served to strengthen elite cohesion, they did so through the consolidation of power by one individual, rather than the institutionalisation of party.

In the wake of this rupture, the EPRDF leadership faced another series of ‘Armageddons’ seen as existential threats to their authority and indeed the integrity of the country. In April 2001, protests at Addis Ababa University demanding an independent students’ union and civilian policing on campus turned into riots across the city that were forcibly put down with the loss of thirty-nine lives (BBC News 2001, Gebremariam 2017). For Bereket Simon, a close confident of Meles, this was interpreted as a ‘wake up call’ regarding the dangers of urban unemployment (Simon 2011). This was followed by protests that were violently suppressed in western Oromiya (March and May 2002) and in Hawassa in the first half of 2002 (May 2002). In the midst of these protests, in 2002/03, the country experienced the most serious food crisis since the 1984/85 famine precipitated by a severe El Niño. The crisis left some 14 million people requiring emergency assistance and forced a re-evaluation of the emergency food system and, indeed, the very viability of ADLI’s focus on raising the productivity of all smallholder farmers, many of whom were struggling for subsistence (see Chapter 5).

Finally, in 2005, following the most open electoral campaign in Ethiopian history, the EPRDF was surprised by the coherence of the opposition coalitions and the support they received. With the EPRDF party structure, from the elite to grass roots, undermined by the 2001 split and the process of renewal, the elections took place at a moment of relative party weakness (Tronvoll 2009a), undermining its ability to deploy low visibility and low intensity forms of coercion to control the electoral process. Moreover, the Eritrean War had resulted in resurgent Ethiopian nationalism, which translated into increased support for the pan-Ethiopian opposition Coalition for Unity and Democracy (CUD) (Clapham 2009, Bach 2013, Lyons 2019). A series of opposition events drew large crowds, culminating in a rally attended by a million people in Addis Ababa shortly before the election. Official results gave the opposition nearly one-third of the seats in the new parliament, and almost all seats in Addis Ababa and several other urban centres. Moreover, large-scale opposition demonstrations were organised in Addis Ababa and elsewhere in May and June 2005 to protest against apparent electoral manipulation by the government and to claim overall victory for the opposition.

For the EPRDF leadership these crises came one after another, leading to the impression that they were indeed facing an existential threat. In the words of then deputy Prime Minister Addisu Legesse,

The renewal was critical, that was an Armageddon, then the drought was another Armageddon, the 2005 elections was another.¹¹

The response to these crises was twofold. First, a cohesive and united EPRDF leadership took decisive action to suppress electoral protests forcibly in May and June 2005, resulting in the deaths of more than 200 protestors and arrest of tens of thousands, including many of the opposition leadership (BBC News 2006).¹² Subsequently, the opposition was dismantled, with the government employing a range of strategies to fragment the parties, harass their leaders and limit their ability to campaign, with most key figures ultimately forced into exile (Aalen and Tronvoll 2009). Moreover, legislation in 2009 on Anti-Terrorism and Non-Governmental Organisations was widely interpreted as means of suppressing political and civil society organisation. The result was that within a year of the elections the opposition had been effectively neutralised and ceased to pose any electoral threat. Regional elections in 2008 and federal elections in 2010 and 2015 were a non-event, used primarily to project EPRDF dominance (Aalen and Tronvoll 2009, Tronvoll 2011, Arriola and Lyons 2016).¹³

The use of coercive powers to suppress the protests and prevent future political mobilisation grabbed the headlines and this has been the main focus of much academic research. Yet, the Armageddons also spurred the now cohesive elite to reflect on the sources of the political upheaval. The EPRDF interpreted these crises through essentially the same lens that it had used during the TPLF struggle and the early years of the EPRDF, with the food crisis and mass unrest seen as distributive crises – a symptom of the slow pace of economic development and rapid population growth – while rapid agricultural development and structural transformation were the solutions, providing the resources required to secure mass acquiescence. Nonetheless, the crises cast doubt on the viability of the existing development strategy, with the food crisis in 2002/03 demonstrating the limits of ADLI’s primary focus on smallholder agricultural productivity, while urban protests in 2001 and 2005 highlighted the dangers of treating urban development, industrialisation and employment creation merely as a long-term outcome of an agriculture-focused development strategy. The crises also underscored the danger of regional autonomy fostering potential rival centres of political power and the potential power of nationalism as a means of popular mobilisation. The result was a gradual change in the orientation of the development strategy and a massive expansion of the party-state to deliver on these plans, while further restricting regional autonomy in the interests of a state-led development strategy.

As Gebresenbet (2014) has argued, after 2001 the government framed rapid and broad-based development as an issue of national security, with the failure to realise these goals presenting an existential threat both to the EPRDF and to the country as a whole. Bereket Simon – a senior ANDM leader, Minister of Information and close confidant of Meles – wrote that,

the dangers these issues brought to the ruling party were described by Meles as being an impending avalanche and Armageddon in the making. Not only was growth needed but rapid growth. Otherwise, the avalanche will swallow the system resulting into chaos. Therefore this became a question of survival for the EPRDF.

(Simon 2011)

Moreover, repeated statements by senior officials, in policy documents and internal party documents make it clear that the government’s concern was not just any form of economic growth, but ‘broad-based growth in which the majority of the population is able to benefit’ (MoFED 2003, p. 9). Since, development ‘that leaves the bulk of the population as mere spectators of a well-endowed few would not rescue the country from engulfing crisis’ (EPRDF 2010, p. 3). Indeed, Meles and many other government statements tied the mass distributive threat faced by the EPRDF to the fate of the country as a whole, ‘I am convinced that we will cease to exist as a nation unless we grow fast and share our growth’ (De Waal 2013, p. 154).¹⁴

Despite the EPRDF’s intensified commitment to rapid, broad-based development and mass distribution as a solution to its political problems, the government’s fiscal position in the early 2000s remained extremely weak. The EPRDF had made some efforts to increase tax mobilisation during the 1990s. However, this was constrained by the small tax base in an economy dominated by peasant agriculture, and the Eritrean war had drained state coffers. While the government undertook a series of tax reforms throughout the 2000s aimed at increasing compliance and broadening the tax base, the only way to significantly expand tax revenues would be through economic development (Mascagni 2016). Moreover, foreign aid receipts remained modest, reaching a low point during the Eritrean war, as donors pushed for an end to a conflict regarded as pointless and avoidable (see Figure 4.2).¹⁵ Aid receipts only slightly increased following the war and were again reined in as donors protested the government’s handling of the elections in 2005. Indeed, the donor response to the election fallout was to divert general budget support previously allocated to the federal government to wereda administrations through the Protection of Basic Services programme, supporting service delivery while limiting federal government discretion over spending. Moreover, the general inclination of Ethiopia’s main donors continued to be the promotion of a market-based approach to development, albeit in the context of the post-Washington Consensus focus on poverty reduction and social services (Amsden 2010, Mkandawire 2010). State-led development and industrial policy were not in fashion and there was widespread pessimism about the prospects of an East Asian-style developmental state in Africa (Mkandawire 2001). Certainly Ethiopia, widely regarded as a ‘basket case’, did not seem a likely candidate for successful state-led development. Consequently, the development strategy laid out in the early 2000s ‘was very much criticised and ridiculed’ by donors.¹⁶ The experience served to reinforce the government’s longstanding view that while donor support might provide supplementary revenues, it was unreliable and the government’s strategy should focus first and foremost on self-reliance.

Figure 4.2 Net overseas development assistance to Ethiopia

Source: author, based on World Development Indicators, World Bank.

As Figure 4.2 shows, OECD donors did significantly increase support to Ethiopia from the late 2000s. However, this was primarily in response to what they considered to be an emerging success story and a government capable of delivering on its promises, rather than aid being a driver of developmental performance (Feyissa 2011a, Fantini and Puddu 2016).¹⁷ The result was that although total aid increased, aid as a percentage of GDP was in decline. Moreover, the emergence of Chinese financial assistance provided an important alternative source of finance, particularly for infrastructure development projects, including roads, bridges and dams, that were key pillars of the government’s plans but which western donors were reluctant to fund (Cheru and Oqubay 2019). Ultimately, domestic politics and EPRDF ideology, rather than donor influence, were the central factors that shaped the government’s development strategy, with the government aggressively defending the policy autonomy carved out during the 1990s. Moreover, the government explicitly sought to manage donors strategically (MoI 2002a), seeking donor support for sectors with shared priorities, including primary education and basic health, capacity building and social protection, while reserving government resources for priorities that donor were unwilling to fund such as infrastructure, agriculture and industry (Furtado and Smith 2009). The result has been that while donors were quite influential at points within particular sectors, they had relatively little insight into and little influence over the big issues shaping the government’s development strategy.

By the early to mid-2000s, therefore, Ethiopian politics came to approximate the conditions identified in Chapter 2 as being suited to state-led development. The state autonomy and capacity forged through the revolution of the 1970s was now accompanied by a coherent political elite that viewed rapid and broad-based development to be the only feasible solution to the existential threat posed by growing mass distributive pressures and severe resource constraints.

The Emergence of the ‘Developmental State’

While the TPLF split was primarily a power struggle between contending factions, many of the dissidents expelled from the party were also among the more committed Marxists. Consequently, the removal of the left wing of the party ultimately contributed to an ideological shift and consensus on a new approach. This ideological consensus leant heavily on the party’s longstanding commitments to self-reliance and the importance of state intervention, but also represented a notable change. In particular, the EPRDF embraced capitalist development, rather than socialism, opening up the party to the national bourgeoisie and pursuing integration into the global economy, changes that would surely not have been accepted by the expelled dissidents on the party’s left wing (Tadesse and Young 2003, p. 392, EPRDF 2006, 2010, Simon 2011). In Bereket Simon’s words,

Marxist Leninist ideologies were replaced by a capitalist outlook. While capitalism may not be the perfect system, the leadership agreed that there is no better alternative. It is the only system that was tested and succeeded. It is the only system that can take Ethiopia out of its centuries old backwardness. The leadership reached a consensus on this.

(Simon 2011)

It is important to acknowledge here that while structural factors shaping the ruling elite were hugely important, this is not to deny the importance of agency and ideas. As much of the existing literature has argued (De Waal 2015, Clapham 2018), Meles Zenawi, his political position now solidified, was enormously influential in outlining the future direction of the national development strategy, apparently drafting most policy and party documents himself, as well as taking the lead in educating his colleagues. Indeed, according to one senior official in the Ministry of Capacity Building,

At a certain point Meles decided that he was the one to have the ideas, from his books and his reading and even his close colleagues became implementers. He was training senior colleagues by video link from the palace – Addisu [Legesse], his deputy, and Bereket [Simon].¹⁸

Meles drew heavily on his reading of the experiences of South Korea and Taiwan in his efforts to build an Ethiopian ‘developmental state’, framed in contrast to a rather caricatured ‘neo-liberalism’ (EPRDF 2006, Zenawi 2006a, 2006b, 2012). This shift in development model built seamlessly on the EPRDF’s longstanding commitments to state-led development and the importance of self-reliance dating back to the TPLF insurgency, with all available resources to be mobilised by the state in the pursuit of development. This shift to a capitalist model of development also entailed a significant change in the class analysis underpinning the Front’s political and development strategy. The private sector had previously been viewed negatively. However, drawing on the developmental state literature on state–business relations, Meles’ statements and other government documents began to highlight the important role for the private sector. From this perspective, the central function of a ‘developmental state’ was to direct the private sector towards productive activity and away from ‘rent seeking’,

One of the distinguishing characteristics of a developmental state is its relations with the private sector. While the developmental state has to establish a very close relationship with the private sector engaged in productive activities, it has to curtail the activities of those engaged in rent-seeking activities and, hence, has to have adequate autonomy from both sections of the private sector, to provide directive support to the first and restrict and ultimately overcome the second.

(Zenawi 2006a)

The result, following the renewal, was greater openness to private investment from both foreign and domestic sources that had previously been held in ‘suspicion’ (MoI 2002b, p. 2, EPRDF 2006, 2010). While the EPRDF continued to see itself as ‘fundamentally an organization of the peasantry’, which also enjoys the support of the ‘broad masses in urban areas’ (EPRDF 2010, p. 6), the national bourgeoisie was classified as a key constituency and a group that would have to be won over given their ‘special (vital) role in our development endeavor’ (EPRDF 2010, pp. 6–7). As noted above, the EPRDF had long distinguished between developmental and ‘rent seeking’ private sector actors (TGE 1994, p. 2). However, from the early 2000s the distinction between productive investors and ‘parasitic’ rent seekers became central to the government’s ‘developmental state’ discourse (EPRDF 2010, p. 7). Moreover, the government began to label almost any form of corruption or political opposition as ‘rent seeking’ and ‘anti-development’, implying that political opposition reflected not just a different viewpoint, but a threat to national security (Vaughan 2011, Gebresenbet 2014).

Ethiopia’s ‘developmental state’ therefore identified a central role for ‘constructive government’ in the development process, centralising rent allocation to enable the state to prioritise ‘developmental’ private enterprises pursuing ‘productive’ activities and limit access for parasitic rent seekers (EPRDF 2006). Moreover, where the private sector was unable or unwilling to engage in sectors identified by the state as being of developmental importance, the state would need to intervene directly (EPRDF 2010, p. 5). The state’s ability to influence and discipline the private sector depended on state control of two key resources that were likewise essential to state-led development in Korea and Taiwan: finance and land (Woo 1991, Amsden 1992, Castells 1992). Both finance and land were nationalised under the Derg, and the EPRDF had resisted donor pressure in the 1990s to relinquish state control. In the 2000s, donor pressure for privatisation periodically continued, yet government resistance was increasingly justified with reference to the ‘developmental state’. Newai Gebre-Ab, one of Meles’ main economic advisors throughout his time as Prime Minister, was clear that,

on finance our views were very different from the World Bank and IMF. We thought the state should have a strong influence on finance. That was prior to the developmental state. It was the same with land, the thinking was before. But yes, when thinking about the developmental state, those two factors were instrumental.¹⁹

The EPRDF also responded to the resurgent nationalism evident during the Eritrean War and 2005 elections by framing its renewed developmental project within a form of Ethiopian nationalism. Meles cast the ‘developmental state’ as a means of realising an ‘Ethiopian Renaissance’, restoring the glories of past Ethiopian civilisations (Bach 2013, Orlowska 2013). In his speech on the occasion of the Ethiopian Millennium in 2007 he argued that,²⁰

We do not have written records telling us how Ethiopians celebrated the end of the first millennium and the banging of the second one a thousand years ago. We can therefore not be certain as to what their festivities might have celebrated with a well deserved pride in their achievements. They after all could rightfully have claimed to be the makers of one of the most advanced nations of the day … A thousand years from now, when Ethiopians gather to welcome the fourth millennium, they shall say that the eve of the third millennium was the beginning of the end dark ages in Ethiopia. They shall say that the eve of the third millennium was the beginning of Ethiopian renaissance.²¹

Having previously argued that Ethiopia had only one hundred years of history, dating from Menelik’s conquests, and that Aksum had no relevance to people from southern Ethiopia, this marks a fairly radical change of perspective. Meanwhile, the symbolic embrace of Ethiopian nationalism and further centralisation of power by the federal government as part of the ‘developmental state’ raised additional concerns among many federalists about the limits to ethno-regional autonomy.

These ideological changes had a gradual impact on the details of the government’s development strategy. Shortly following the TPLF split, the federal government released a series of sectoral strategies outlining the government’s priorities. Beginning with the Rural Development Strategy in November 2001 (translated into English in 2003), these sectoral strategies expanded on ADLI and its focus on raising the productivity of peasant agriculture as a means of stimulating industrial development and structural transformation (MoFED 2003). Subsequent strategies focused on building state capacity to deliver these ambitious development plans, the first industrial strategy (MoI 2002b) and the Foreign Policy and National Security Strategy, which framed economic development as the main driver of national security (MoI 2002a).²² Indeed, one former minister, deeply involved in the planning process, emphasised that rather than a major change in direction, the initial impact of these crises was,

Urgency. It shows that we are still not making progress. So we had better organise a support system to raise productivity. We need to place a weight on the implementation speed and the design of institutions.²³

However, over subsequent national development strategies, beginning with the Plan for Accelerated and Sustained Development to End Poverty (PASDEP, 2005–2009) (MoFED 2005) and continuing with the Growth and Transformation Plans (2010–2015 and 2016–2020) (MoFED 2010, NPC 2016) the ‘developmental state’ model came to the fore. The result was an important evolution of the government’s developmental and distributive strategy well beyond anything envisaged in ADLI. The changes related to agriculture, urban development, industrial policy and social protection are analysed in detail in the following chapters. In summary, the government came to focus to a much greater extent on urban development with a major push for infrastructure expansion and industrialisation fuelled by foreign investment rather than an agrarian surplus. This did not mean the neglect of the agricultural sector, but rather a new focus on commercialisation of high potential smallholder farmers and the promotion of capital-intensive agricultural investments. Vitally, this change in approach entailed a shift in the government’s distributive strategy, which to this point had focused on guaranteeing mass access to land and agricultural inputs. From the mid-2000s, the government came to accept the limits of land access as a means of mass distribution in the light of population growth, displacement for agricultural investments, infrastructure and urban expansion. Instead, industrialisation was to secure mass acquiescence through large-scale employment creation. These changes ultimately also translated into recognition of the need for some limited form of social protection for those unable to secure a livelihood through access to land or employment.

Building the Infrastructural Power of the Party-State

The elite cohesion and sense of shared threat in the wake of the Armageddons not only led to consensus on the new development and political strategy, but also a focus on building the capacity of the party-state to deliver on these plans. This meant strengthening capacity in federal ministries and agencies but also the administrative capacities of regional and local state structures. Moreover, this expansion of infrastructural power went well beyond formal state structures, with an expansion of the party-state as a means of reaching into communities and households, thereby channelling the symbolic, coercive and, in particular, distributive powers of the state towards securing EPRDF control.

With the evolution of the EPRDF’s guiding ideology and development strategy, there was a concerted attempt to instil the paradigm of the ‘developmental state’ throughout the party-state. The EPRDF remained a vanguard party within a framework of revolutionary democracy and democratic centralism (EPRDF 2010, p. 36). As such, the expectation was that key decisions made by the leadership – in this case the turn to developmentalism – would be accepted without question by the rank and file. Indeed, Meles explicitly sought to make developmentalism a ‘hegemonic’ ideology within the party, state and society (Zenawi 2012, p. 167). To this end, pamphlets were distributed within the party, outlining the successes and limitations of various examples of historical development, the successes of East Asian state-led development and the applicability of these models to Ethiopia. Elsje Fourie cites one senior EPRDF MP who commented on one such party document, Development, Democracy and Revolutionary Democracy (EPRDF 2006), claiming, ‘We believe in this, as Christians believe in the Bible’ (Fourie 2012, p. 139). As such, this approach to capacity building in Ethiopia differs markedly from the standard picture of East Asian state-led development that emphasises a Weberian bureaucracy with high levels of technical expertise based on meritocracy and a strong sense of purpose, which is politically empowered to drive decision-making based on technical criteria (e.g. Johnson 1982, Evans 1995, You 2015). In Ethiopia, in contrast, it was the political elite and the party that provided ideological direction and policy detail, with the bureaucracy focused on implementing politically defined strategies (see Weis 2015, Gebresenbet and Kamski 2019, Lavers et al. 2021).

The civil service was also reformed with a view to improving implementation and alignment with government plans. After 2001, what became known as ‘superministries’ were formed focusing on the key priorities of the development strategy and headed by senior party officials close to Meles, namely: Rural Development (Addisu Legesse, deputy Prime Minister), Infrastructure Development (Kasu Ilala), Federal Affairs (Abay Tsehaye) and Capacity Building (Tefera Walwa). Furthermore, as the federal government institutionalised the five-year national development plans, ministries were required to produce their own five-year sectoral plans, setting targets aligned with national strategy. The introduction of a Result Oriented Performance Appraisal System (World Bank 2017, p. 18) meant that annual and five-year targets were cascaded down through the tiers of state administration providing the basis for performance assessment of each official and department. The key means of evaluating performance continued to be the system of gim gema adopted by the TPLF during the struggle and extended to government since.²⁴ In 2006, the government also adopted Business Process Re-engineering, a new public management tool intended to re-design bureaucratic processes to focus efforts on service delivery and performance (World Bank 2017). These reforms were seen by a former senior minister as key steps that expanded the state’s capacity to deliver,

In 2003/04 we also started the civil service reform to promote efficiency, effectiveness and service delivery. Every ministry should have a strategic plan. The public services and servants now, compared to the past, are duty bound and service oriented. We established the Ministry of Capacity Building up to the wereda level and also federal and region. It completely re-oriented the work attitude—we are talking in a relative context—for the first time modern management concepts were introduced.²⁵

Similarly, there was considerable investment in increased capacity within regional governments, with regional bureaux established to mirror federal ministries. Despite the important symbolic shift that came with the introduction of federalism, regional administrations had little autonomy and primarily acted as implementing agencies, transmitting federal instructions to lower levels of the state and coordinating their activities. This focus on building capacity also focused heavily on lower levels of the state. A central priority in the reforms launched after 2001 was the wereda (district) decentralisation process that established representatives of line ministries at the wereda level and financed wereda activity through central block grants that provide a degree of budget responsibility (Vaughan 2011). Indeed, the donors’ decision in 2005 to redirect foreign aid from the federal government to wereda administrations through the PBS programme actually served to support this important government agenda of building local state capacity.

Under the EPRDF, the kebele – or sub-districts that built on the Derg-era peasant associations – also greatly expanded in capacity. Initially consisting of unpaid, elected leaders, these unpaid posts have expanded considerably with the creation of various committees to oversee issues such as land administration and peace and security, a local militia, as well as the expansion of the kebele council to comprise 200–300 elected members (Emmenegger et al. 2011, Lefort 2012a). Moreover, paid positions for a kebele manager, 3–4 development agents (agricultural extension officers), two health extension workers and schoolteachers were created in every kebele from the early 2000s. The result was to increase the number of development agents nationally from 2,500 in 1995 to more than 47,500 in 2008 (Spielman et al. 2011, p. 26, Berhanu and Poulton 2014, p. s198), while by 2009/10, the government had trained 30,000 health extension workers and constructed 14,000 health posts nationwide (Bilal et al. 2011, p. 11).

The result was a tripling of the number of public employees between 2004/05 and 2013/14, following a more steady increase since the 1990s (see Figure 4.3). Yet, this huge expansion of the civil service did not lead to a major increase in the wage bill thanks to efforts to suppress wage growth, as well as mass recruitment of relatively junior officials at wereda and kebele level who were at the low end of the salary scale. Regular salary increments for civil servants were stopped in 2003 in favour of periodic ad hoc increases that have fallen well below inflation (World Bank 2017, p. 9). Indeed, low salaries for all officials, from junior civil servants to ministers, was a point of pride within government, maximising the resources that could be directed to key projects of the ‘developmental state’. Addisu Legesse recalled a visit to Vietnam, where for the first time he found a country that paid its ministers less than Ethiopia,

I went to Vietnam and found that our government ministers are better paid by $20. So I came back and told Meles, ‘there is somewhere that pays ministers less than we do!’²⁶

While this enormous state expansion improved service delivery, it has also enhanced the reach and control of the EPRDF, with almost all employees also party members with party political responsibilities. This expansion of the party-state reached beyond the kebele level, with the government establishing sub-kebele structures at the village level (kushet in Tigray, known by other names in other regions) and attempting to mobilise every household in the country through the creation of development teams (comprising thirty households) and 1-to-5 networks (five households) for men, women and youth. This development team structure – also described as a development army – resembles a military hierarchy and forms part of the ‘securitisation of development’, whereby poverty and backwardness are themselves framed as a key security threat, justifying the mobilisation of the entire population (Gebresenbet 2014).²⁷ Moreover, these developmental structures are fused with the party, with leaders of development teams and 1-to-5s invariably also party members.

Figure 4.3 Civil service employment

Source: author, based on World Bank (2017, pp. 5–6). No data available for wages, 2005/06 to 2009/10.

In principle, then, the EPRDF’s party-state building project established a direct chain of command through 6–7 tiers of administrative hierarchy from the federal government through the region, zone, wereda, kebele, kushet, development team and 1-to-5 network to every individual in the country.²⁸ As discussed in Chapter 9, the effectiveness of these structures varied in practice, working most effectively in Tigray, to a moderate degree in Oromiya and little more than an administrative fiction in Afar (Lavers et al. 2020), for example. Nonetheless, where functional, these structures provided the party-state with a vital infrastructure with which to promote the EPRDF’s developmental ideology, distribute the material benefits of the ‘developmental state’ in ways that tied individuals to the regime, and monitor and control the population. As one Oromo farmer evocatively described the system,

The government has its hands and the garee [development team in Afaan Oromo] are like the fingertips…. The garee is very near to you. The kebele was not.

(HRW 2005, p. 26).

The development team structure advanced farthest in rural communities where individuals and households were most dependent on the resources they distributed. Nonetheless, the EPRDF aimed to extend this mobilisation strategy to the entire population, which was divided up into three ‘wings’ – the party, state and public (World Bank 2017). As a Ministry of Capacity Building official noted, ‘civil servants, universities, students – all were formed into 1-to-5s’.²⁹

In addition to the development teams, the ‘public wing’ was mobilised through mass associations and cooperatives that engage in regular consultations with party and state representatives (World Bank 2017). As noted in Chapter 3, the TPLF had formed mass associations during the struggle, but the EPRDF made the decision in 1998 to separate these organisations from the party. Following the shock of the 2005 elections, this decision was reversed, with the 2006 party congress announcing the creation of new women’s and youth leagues (Berhe 2020). Furthermore, major efforts were made to promote mass membership in the party in response, with party membership rising from 760,000 in 2005 to more than four million in 2008, and more than five million by 2010 (Aalen and Tronvoll 2009, Vaughan 2011). As well as party branches at kebele level, party cells were formed in local communities in parallel with the development teams. In practice, many development team leaders – as party members – also became cell leaders, with the result that the two parallel structures are often used interchangeably, with both involved in party political as well as developmental activities.

The overall result of these efforts was a massive expansion of the party-state, both in terms of the technical skills required to implement the national development strategy and, in particular, the territorial reach of the fused party-state to reach into communities and channel the symbolic, distributive and coercive powers of the party-state towards the maintenance of political order.

Conclusion

The series of political crises or ‘Armageddons’ in the early 2000s were a key turning point for the EPRDF and Ethiopia. The political upheaval of this period resulted in a high degree of elite cohesion within the ruling party, a sense of collective threat from mass unrest and a shared vision for achieving rapid economic development as a means of transforming the country and meeting the distributive needs of a growing population. The result was that the government was finally able to build on the legacy of state autonomy and capacity inherited from the Derg, further strengthening party-state capacity and utilising these structures to pursue a project of state-led development. The headline results of this political shift are undeniably impressive, as shown in Figure 1.1, with an unprecedented period of economic growth from 2004 onwards, a massive expansion of infrastructure and the reduction of poverty in every region of the country.

The ‘developmental state’ model built heavily on the class analysis and ideological orientation of the party going back to the insurgency against the Derg, with a focus on self-reliance, securing acquiescence of the masses and framing developmental progress as a political imperative. Nonetheless, the embrace of capitalism and attempts to fast-track development had vitally important implications for the state’s distributive powers that were key to the EPRDF’s strategy of political control. The following chapters explore these developments and their distributive implications across sectors, starting with access to land and the agrarian question (Chapter 5), before moving to industrial policy and access to jobs (Chapter 6), urban expansion and land expropriation in the urban periphery (Chapter 7), access to social protection (Chapter 8) and the ways in which the expanded infrastructure of the party-state was used to channel the symbolic, distributive and coercive powers of the state towards the objective of political control (Chapter 9).

An important point to reiterate with respect to the arguments developed in Chapter 10, however, is that while the crises of the early 2000s resulted in a high degree of elite cohesion, this was achieved primarily through the concentration of power in one individual, Prime Minister Meles Zenawi, rather than the institutionalisation of the party as a means of limiting elite divisions and binding together a cohesive elite. While this centralisation of power temporarily addressed the elite divisions that had undermined the government in the 1990s and enabled the rapid development that followed, it left the EPRDF highly exposed to the challenge of succession that is common to all authoritarian regimes. The party was forced to confront this challenge in 2012 when Meles Zenawi passed away. The EPRDF initially managed a smooth transition with deputy Prime Minister, and Meles’ chosen successor, Hailemariam Dessalegn, taking over. However, despite the appearance of continuity and stability, Meles’ death re-opened longstanding faultlines between and within the EPRDF parties. This erosion of elite cohesion would prove a key factor in the EPRDF’s subsequent collapse, as discussed in Chapter 10. First, however, I turn to the ‘developmental state’ and its distributive implications.