In 1985, Julius K. Nyerere, one of the South's most articulate and influential critics of Northern policies resigned as president of Tanzania. Soon after, his successor – Ali Hassan Mwinyi – would reach agreement with the IMF and World Bank. Tanzania, for many, the experiment in African socialism and the high-profile heretic of the 1980s had fallen and recanted, implementing a structural adjustment programme. Tanzania was in many ways exceptional. A former World Bank favourite, it became one of the few cases in sub-Sahara Africa where the IMF and World Bank (the IFIs, International Financial Institutions) brought their full coercive power to bear in a protracted struggle. In understanding this apparent contradiction, this paper discusses the role played by economic discourses in shaping the positions taken by two key actors in the policy process: the Government of Tanzania and the World Bank.
The capitulation of the Tanzanian government in 1985 to the demands of the International Monetary Fund (IMF) and World Bank (hereafter the Fund and the Bank), stands as a landmark in relations between the Bank, Fund and sub-Saharan African (SSA) states. After six years of bitter struggle, Nyerere's Tanzania, leader of the Non-Aligned Movement, and standard bearer of African socialism, had given into the demands of international capitalism (McHenry, 1994). Thereafter, in the era of adjustment, there would be few SSA states willing to resist the apparently overwhelming coercive power wielded by the Bank and Fund over impoverished aid-dependent states (see e.g. Hayter and Watson, 1985; Onimode, 1988).
This paper does not seek to retread the well-worn ground debating the costs and benefits of structural adjustment (see e.g. Stewart, 1995; Cornia et al. 1987; Demery and Addison, 1987). Nor does it seek to explore the ‘interests’ the policies of the Bank and Fund served (see e.g. Woods, 2000; Escobar, 1995; Payer, 1991, 1983). Instead, the paper focuses upon why the Tanzanian government resisted neo-liberal policies promoted by the Bank and Fund for six years before acceding to them in 1985.
The majority of the literature examining policy change in the period 1978-85 treats the adoption of Bank/Fund-sponsored neo-liberal policies in the 1985 budget as an example of an alien set of ideas (neo-liberalism) imposed upon an unwilling government left with no choice but to capitulate (see e.g. Green, 1995; Svendsen, 1995). The literature on this period also highlights the relative autonomy of the state vis-à-vis other classes (see e.g. Kiondo, 1989; Coulson, 1982; Shivji, 1976) and the struggle within the state between pro- and anti-reform factions (see e.g. McHenry, 1994; Lofchie, 1994; Campbell and Stein, 1992). Taken together, this body of literature highlights the ‘coercive power’ of donors, most notably the Bank and Fund over the government, identifying them as the key players who ‘made’ policy within Tanzania. This paper complements this account. It accepts that as an aid-dependent country, the ability of the Bank and Fund to mobilise and withhold aid was a decisive factor: by 1985 the government was essentially bankrupt and had little or no choice but to accede. Nevertheless, this does not explain why Nyerere, in near suicidal fashion, drove the country right to the very brink of collapse, before turning back.
The paper contends that in order to more fully understand policy-making in Tanzania, we need to pay more attention to the institutional realms within which policy was made. In understanding how two key institutions – the Bank and the Government of Tanzania – moved from broad agreement and support in the 1970s, through bitter recrimination to uneasy agreement in the 1980s, the paper will focus upon the evolution of discourses within each institution. The paper will argue that the discourses within each institution came to circumscribe the realm of ‘rational’ policy, and therefore their ability to compromise. By drawing upon the work of Peter Hall (1993), the paper will examine how discursive change within the Government that could have allowed agreement to be reached in 1979, was blocked. The paper will argue that it would take another six years before another political window of opportunity would open the door to the discursive change that enabled agreement to be reached.
The paper is divided into four sections. The first outlines the theoretical framework used, the second outlines the existing literature on the crisis. The third section discusses how an understanding of discursive change can enrich and complement this literature, and concludes with a brief summary.
Policy paradigms, experts & discursive change
Peter Hall's (1993) work on the transformation of economic policies within the British government in the 1970s and 1980s uses the concepts of ‘social learning’ (developed by Theda Skocpol, and Paul Sacks) and ‘policy paradigms’ (developed by Thomas Kuhn) to help explain how and why the three different ‘orders’ of policy (‘instruments’, ‘setting[s]’ and ‘the overarching goals that guide policy’) he identifies change (Ibid. pp. 278-279). Hall's work suggests that because ‘policy paradigms’ come ‘to circumscribe the realm of the politically feasible, practical and sensible’, until the paradigm changes, only first and second order changes are possible. Hall argues that because third order change is a radical break with past policy, it has far reaching consequences and is more likely to be overtly politicised. Therefore, third order change depends not only on the outcome of conflict between experts pushing for and against third order change, but also upon their ability to mobilise wider support by having their view accepted as ‘authoritative’ (Ibid.).
Hall's analysis of economic policy change in the UK in the period 1970-1989, indicates that the radical change in policy after 1979 was catalysed by a ‘policy network’. Hall contends that this network's ability to change policy depended upon political change: it was Margaret Thatcher's political success that enabled them to implement the (paradigmatic) change they pushed for. Hall concludes that the prevailing discourses that make up the dominant paradigm legitimates some policies, privileging them, while delegitimating others. Thus, actors within the policy process, be they organised interests, politicians or policy experts do not simply ‘exert power’, ‘they acquire power in part by trying to influence the major discourses of their day’ (Ibid. p. 290).
Hall's work complements and can itself be enriched by much of the literature on policy-making processes. For example, Hall (1993) identifies the role played by ‘policy networks’, but his treatment of them remains rather underdeveloped. As this paper will show, of the different types of ‘policy network’ (see e.g. Stone, 1996), Haas′ (1993) work on ‘epistemic communities’ is the most relevant to Tanzania. Hass (1992) defines an ‘epistemic community’ as the characteristics of a network that has:
a shared set of causal principles (analytical and normative) and beliefs, a consensual knowledge base, and a common policy enterprise (common interests) that distinguishes epistemic communities from other groups (Haas, 1992:18).
Policy change within the south
The work of both Hall and Haas explores policy changes within a liberal democratic society with a well-developed bureaucracy. In contrast, under Nyerere, Tanzania became a one-party state and its bureaucracy remained weak (Bigsten, 1999). Power was centralised, and the bureaucracy did not emerge as a powerful independent actor. Instead, the air-conditioned Weberian bureaucracy would became ‘corrupted’ by the politics of the ‘veranda’ (Terray, 1986). The bureaucracy – a prime source of political patronage – would become subsumed within, and controlled by, the broader ruling class. Dubbed the ‘bureaucratic bourgeoisie’, it would come to dominate both political and economic life in post-colonial Tanzania, suppressing the emergent private bourgeoisie (Hyden, 1994; Kiondo, 1989; Coulson, 1982; Shivji, 1976).
The bureaucracy in Tanzania may therefore have been less institutionalised and political conflict less open than in the UK, nevertheless, this paper will demonstrate that Hall's, and by extension Haas’ work, is still applicable. This paper's case rests upon the central proposition that economic policy in Tanzania was ‘rational’. That is to say, economic policy was not arbitrary, but informed by discourses (Hyden and Karlstrom, 1992). The absence of multiparty politics and a well-developed bureaucracy did not render Tanzania's politicians any less idealistic or self-interested than their British counterparts. Political actors supported economic policies that they expected would both develop Tanzania (McHenry, 1994) and secure their political position (Kiondo, 1989). Crucially, their expectations of the outcomes of policy were shaped by discourses (Cf. Escobar, 1995; Ferguson, 1990), and therefore the major shift in economic policy in the mid-1980s was dependent upon the discursive change examined by this paper. This discursive change led politicians to reassess the policies that best served both their own interests and those of their country.
The dominant narrative of the Tanzanian crisis & its resolution
As the following section outlines, there is broad consensus in the literature that the seeds for Tanzania's economic crisis in the late 1970s were sown by the policies of the government. By the end of the decade, internal policy failures were compounded by external shocks, leaving Tanzania in desperate need for aid and therefore, donor support.
The roots of the economic crisis, 1970-79
In the post-colonial era Nyerere's bold vision of ‘self-reliance’ somewhat ironically attracted aid from donors keen to help a recipient who was ‘helping themselves’. This was readily accepted as Tanzania sought to transform both its economy and society (Samoff, 1992). Within the Bank, the appointment of Robert McNamara as president in 1968 brought a significant change in Bank thinking. Placing ‘poverty’ at the forefront of the Bank's activities, in 1973 McNamara launched the Bank's first major poverty-focused programme of lending: ‘Integrated Rural Development’. This was followed by a degree of North-South consensus around the ‘Basic Needs’ agenda (Kapur et al. 1997).
With the Bank keen to burnish its ‘apolitical’ independent credentials during the cold war, Tanzania's moderate socialism, far from ruling out aid, proved a positive boon (Lancaster, 1997; Rugumamu, 1997). Decisively, McNamara's vision of promoting growth and equity was matched by Nyerere's vision. The confluence of interests between the Bank and Tanzania was cemented by the personal bond between their leaders (Svendsen, 1995). As a World Bank Country Program Note in 1973 put it:
From a donor's point of view Tanzania comes close to being a model developing country in the sense that the Government is seriously committed to development in a climate of political and social stability … Development policies and priorities are generally well though out and well conceived … We are inclined to give Tanzania an excellent performance rating (Duncan, 1997:394).
In the mid-1970s, the initial shocks caused by the first OPEC oil crisis (1973-74) were eased with Fund credits1 and were later offset by the coffee and cotton price booms (1975). However, the financial windfall brought by the booms served to disguise the extent of the underlying structural problems, most notably the stagnation of the agricultural sector (Lofchie, 1994; Hyden and Karlstrom, 1993). In 1977, the East African Community broke up disrupting trade and by 1978 the coffee boom had ended. In the wake of a brief but costly war with Uganda, and the onset of the second OPEC oil crisis (1979), foreign exchange reserves that had peaked at US$281.8 million in 1977 (equivalent to 4.5 months of imports) following the coffee and cotton booms slumped. By 1980 they stood at just $20.3 million (equivalent to 0.25 months of imports) (World Bank, 1995) and Tanzania was facing a worldwide recession, rising interest rates and the onset of the debt crisis (Hyden and Karlstrom, 1993).
In response to the mounting economic crisis, Tanzania like much of SSA, sought donor aid to ease the economic problems. A number of bilateral donors, most notably the Scandinavians, were sympathetic. However, as the 1970s drew to a close, the political and economic climate in the North was changing. Many Northern governments proved less disposed to their ‘obligation’ to the South; its problems increasingly blamed upon internal policy failures. The Bank and Fund were enrolled as the North's prime instruments for promoting, and where needed, imposing reforms (Kapur et al. 1997). By 1979, both the Bank and Fund had become increasingly critical of Tanzania's agricultural policies, domestic pricing, BIS and the exchange rate. Their criticism and advice flew in the face of Nyerere's vision and policies, and their previously healthy relations with the government began to deteriorate (Green, 1995; Svendsen, 1995; Barkan, 1994).
The breakdown in relations between the fund & government, 1979-80
As the economic crisis deepened, Tanzania turned to the Fund for financial assistance to complement bilateral aid as it had in the mid-1970s. An interim programme was agreed in 1979 requiring devaluation and credit controls. Attempts to reach agreement on a more comprehensive programme faltered though, breaking down in the autumn 1979 in the face of Fund conditionalities, including a 15 per cent devaluation (Stein, 1992). Nyerere, strong both domestically and with many bilateral donors held fast, blocking reforms (Bigsten and Danielson, 1999; Biermann, 1988).
There have been a number of reasons put forward for the breakdown in negotiations and consequent rejection of the Fund proposals. The literature discussing the period has tended to emphasise two largely complementary issues that motivated the government's rejection: the approach of the Fund and the content of its policies. The Fund presented its package on a ‘take it or leave it’ basis, bypassing protocol to go straight to Nyerere with a ‘final offer’. As Reginald Herbold Green, then a key government advisor put it:
it was not so much what was in the IMF package [in 1979] that brought about the sharp breakdown of negotiations as the manner in which it was presented and the tactics used in the negotiations (Green, 1983:357-358 cited in Kiondo, 1989:151).
The literature may emphasise differing motivations for the government's rejection of the Fund's proposals, but is in agreement that the atmosphere within government became increasingly poisonous in late 1979. Those supportive of reforms were denigrated for ‘allying themselves with the IMF’ (Kiondo, 1989). The Fund's mission was ordered to leave the country and the pro-reform Minister for Finance, Edwin Mtei, was replaced by Amir Jamal, in a move designed to suppress pro-reform elements within the CCM Party (Bigsten et al. 1999; Yeager, 1989). The hardening of anti-reform sentiment within government was followed by attempts by Nyerere to rally Southern solidarity, with a conference on ‘The International Monetary System and the New International Economic Order’ in Arusha (Kiondo, 1989). Despite Nyerere's stubborn public stance, the continuing economic crisis forced Tanzania back to the negotiating table. Nyerere's position as a leading voice in the Non-Aligned Movement may have entrenched his inability to compromise with the Fund. Nevertheless, Nyerere's high profile position and the conference helped ‘internationalise’ the debate, and appears to have contributed to the relatively ‘soft’ terms Tanzania was subsequently offered, leading to a three-year standby agreement with the Fund (Loxley, 1989). Despite the softer terms, Tanzania was unable to comply2 and the programme was soon abandoned (Green, 1983 cited in Stein, 1992).
The entrenchment of the crisis, 1980-84
In the wake of the breakdown of the Fund agreement, Tanzania turned to the Bank in light of its support during the 1970s. As noted, the Bank had already by this stage become increasingly critical of the policies it had previously supported. Although the Bank was more conciliatory than the Fund, in the prevailing international political and economic climate, the Bank was reluctant to release aid without a commitment to reforms on the part of the government. In what would prove a decisive move, it sided with the Fund, effectively becoming its ‘junior partner’. The Fund would focus upon ‘stabilisation’ while the Bank focused upon ‘structural adjustment’ (Kapur et al. 1997). The previously bilateral relations between donors and government were increasingly ‘multilateralised’ as the Fund, with the support of the Bank, took a leading role in co-ordinating aid. A number of bilateral donors (e.g. the UK, USA and the Federal Republic of Germany) made their aid conditional on agreement with the Fund (Kiondo, 1989). In effect, donors delegated responsibility for negotiating adjustment policies to the Bank and the Fund, considerably amplifying their coercive power. Nyerere castigated the shift as ‘de facto recolonisation by fairly crude financial blackmail’ (Green, 1995:102). As Nyerere put it:
And when we reject IMF conditions we hear the threatening whisper: ‘Without accepting our conditions, you will not get our money and you will not get no other money’ (Nyerere, 1980:8 cited in Kiondo, 1989).
The government's initial attempts to respond to the economic crisis – the ‘National Economics Survival Plans’ (NESP I, 1981 and NESP II, 1982) – contained some new measures and were effectively de facto applications to the Fund, but they remained ‘more exhortations than policy change documents’ (Bigsten et al. 2001:317). Crucially, they did not challenge any of the fundamentals; problems were to be dealt with by adjusting state intervention. Structural reforms such as liberalisation and deregulation that would have transformed the Tanzanian economy were ruled out (Bigsten et al. 1999; Kiondo, 1989). This position only reinforced the view held by the Bank and Fund that Tanzania did not understand the depth of the crisis and the changes required (Svendsen, 1995).
In the increasingly poisonous atmosphere, the Fund submitted a memorandum for a possible programme in 1982 (Stein, 1992). Attempts to agree on a programme stalled over the Fund's demands for a 50 to 60 per cent devaluation, significant cuts in the budget deficit, removal or reduction of consumer and producer subsidies, positive real interest rates, higher agricultural prices and import liberalisation. The proposed devaluation emerged as the major point of contention. It appears that Nyerere did not fully understand the economic implications of devaluation (Helleiner, 2000). With his Minister for Planning, Kighoma Malima, supported by key expatriate economists such Ajit Singh, forcefully arguing that devaluation would be inflationary (Singh, 1986), Nyerere blocked a major devaluation, and hence blocked agreement with the Fund. Meanwhile, as in much of SSA, as the economic crisis deepened, public support for the government's position softened, with many turning to the ‘exit’ option (Bigsten et al. 1999). Although this forestalled direct confrontation with the government, it sapped the government's strength by further exacerbating its revenue problems (Bigsten et al. 1999).
With the country in a ‘desperate situation’, former Bank President Robert McNamara, supported by a number of bilaterals, helped put together the ‘Tanzania Advisory Group’ (TAG). The TAG was funded by the World Bank but composed of both expatriate and Tanzanian economists hired by the Government of Tanzania. It was intended to help broker rapprochement between the Bank, Fund and government, and lay the groundwork for a Structural Adjustment Programme (SAP) (Bigsten et al. 1999; Svendsen, 1995). Many of the TAG's authors were sympathetic to the Nyerere's vision, and the first draft was in many ways a ‘Socialist Structural Adjustment Programme’ (Loxley, 1989), yet the TAG would ultimately fail. Decisively, the (unpublished3) report, which in many respects was quite different from the recommendations of the Bank and Fund, proposed a modest devaluation but this was once again rejected by the government (Loxley, 1982 cited in Kiondo, 1989).
The government used the TAG's report as a basis for its own structural adjustment programme (1982): increased prices for producer goods, cut-backs on government expenditure and curtailed monetary expansion. Nevertheless, under the influence of Kighoma Malima, a noted anti-reformer, the government diluted many of the TAG's proposals. As a consequence, the SAP did not adequately address issues like the liberalisation of agriculture – areas that remained key concerns of the Fund and Bank. Despite further conciliatory gestures on the part of the Government, most notably a series of small exchange rate devaluations,4 the government's proposals were rejected as inadequate, a decision apparently motivated by the desire ‘to teach Tanzania a lesson’5 (Loxley, 1989:11; see also Campbell and Stein, 1992; Kiondo, 1992, 1992; Kiondo, 1989).
As the 1980s dragged on, the economic and political situation worsened (Lofchie, 1994; Yeager, 1989). The limit of the SAPs reforms were further undermined by the failure to secure external finance to support them, and already weak per capita GDP growth of 1.1 per cent (1965-1980) slipped to -2.4 per cent (1981-1985)6 (Barkan, 1994:22). It is tantalising to speculate what might have happened had the Bank not sided with the Fund and supported a SAP that had attracted significant foreign aid. In the event though, the Bank stood with the Fund. By 1983, although the Fund remained the ultimate ‘gatekeeper’, the Bank had expanded its traditional role helping co-ordinate donors, to one where it actively mobilised other donors to grant – or in cases like Tanzania – withhold aid (Kapur et al. 1997). By 1984, the Scandinavian donors who had been sympathetic to Tanzania's resistance, made it clear that additional aid would be conditioned on Fund approval, leaving the government bereft of aid and facing a united donor front (Bigsten et al. 2001; Kiondo, 1992). In 1985 Nyerere resigned. His inexperienced successor – Ali Hassan Mwinyi – was left facing a deepening economic and political crisis. With the Tanzanian government facing bankruptcy and with no prospect of aid without reform, Mwinyi was forced to capitulate in the face of the coercive power wielded by the Bank and Fund (see e.g. Bigsten et al. 1999; Svendsen, 1995; Stein, 1992).
Re-interpreting the crisis: Coercive & discursive power
In the context of relations between the Bank, Fund and African states in the early 1980s, the acrimonious debates in Tanzania and the posturing of both sides were not of themselves unusual. Indeed, there were number of other countries (including Zaire, Zambia, Zimbabwe, Ethiopia and Sudan) threatening to reject Bank/Fund sponsored reforms in the early 1980s (Lancaster, 1997). What marks out countries like Tanzania, and to a degree, Ghana, is the length of the dispute. In most of SSA, the Bank and to a lesser degree the Fund, were in sufficient internal need of a deal (given pressures to disburse) to compromise. Meanwhile, African elites slowly adjusted to adjustment, using the aid to sustain their regimes, and the Bank and Fund as convenient new scapegoats. In countries like Kenya, compromise was fairly swiftly secured (even if not always honoured7 (Chabal and Daloz, 1999; Mosley et al. 1995; Ndulu and Mwega, 1994). In contrast, Tanzania and Ghana dug their heels in. The Bank and Fund brought their full coercive power to bear against a recalcitrant government by issuing an effective ultimatum: no aid without reforms. Having exhausted all other options, their governments apparently reluctantly acceded to reforms to sustain themselves. The following section will discuss the discursive roots of these conflicts by examining the evolution of economic discourses and their relationship to economic policy in Tanzania.
The limits of coercive power
Tanzania and Ghana can be used to support the literature arguing that conditionality (coercive power) doesn't work (see e.g. Dollar and Burnside, 2001). Because the Bank and Fund took a hard line which they had not taken elsewhere, they stand as an exception to the general principle (Mosley et al. 1995). Nevertheless, this coercive model of ‘policy transfer’ leaves a number of questions unanswered. In particular, we might ask, why did their Governments resist so fiercely? Conversely, why did both Ghana and Tanzania, standard bearers of African socialism, execute such a dramatic volte-face to become celebrated Bank ‘success stories’, while their more ‘capitalist’ brethren like Kenya acquiesced more rapidly, but implemented reforms with far less vigour (see e.g. Devarajan et al. 2001)? In the case of Ghana, it has been persuasively argued that the economic crisis created political space. This space enabled an indigenous pro-reform movement aided and abetted by the Bank, to champion a pro-reform agenda that transformed economic discourses and hence policy (see e.g. Devarajan et al. 2001; Mohan, 1994). As this paper will demonstrate, Tanzania followed a broadly similar path.
The Bank's and Fund's hard line toward Tanzania appears to have been motivated by two key factors: first, the need to ‘make an example of Tanzania’ given both its very public anti-neo-liberal stance and a sense of pique at the way the Fund was ejected in 1980 (Loxley, 1982, cited in Kiondo, 1989); and second, their incomprehension that the Tanzanian government couldn't ‘understand’ the problems they faced (Svendsen, 1995). On the government's side, Nyerere and the party, unlike many of his contemporaries, was not facing a political and economic crisis, nor were they bereft of ideas. Bolstered by a discourse that ruled out the reforms promoted by the Bank and Fund, the self-confident regime was able to resist. The remainder of this paper focuses upon one of the key reasons underpinning the incomprehension and intransigence of both sides: the opposed and non-communicating discursive positions the two sides occupied.
Discursive change in Tanzania, 1963-80
After independence in 1963, the economic modernisation discourse practised by the British Colonial Service and promoted by advisors from among others the Bank, continued within government. ‘Modernisation’ would be the hallmark of Tanzania's first development plan (1964-69) (Yeager, 1989; Coulson 1982). Notwithstanding a policy of ‘Africanisation’ of the public sector, the private sector remained important, and the economy remained fairly open although industrial development was promoted by an import substitution policy. The policies brought macroeconomic stability, low inflation and growth, but progress was made from a low base and Nyerere became increasingly concerned about increasing inequality (Bigsten, 1999).
Throughout the 1960s, the foundations of Nyerere's political thinking which were laid out in ‘Ujamma: The Basis of African Socialism’ (1962) were consolidated. By 1967 he published perhaps his most famous work, ‘The Arusha Declaration’ (1967). Nyerere's idealisation and vision of a more egalitarian communal traditional African society (ujamma 8) as the solution to Tanzania's problems remained constant (see e.g. Nyerere, 1962, 1968). However, by 1967 his earlier belief that ‘the basic difference between a socialist society and a capitalist society does not lie in their methods of producing wealth, but in the way in which that wealth is distributed’ (Nyerere, 1962:1) had hardened into the belief ‘that in order to ensure economic justice the state must have effective control over the principal means of production’ (Nyerere, 1968:231-232).
Nyerere had come to believe that market determined prices were simply an indication of profit making and hence a capitalist outlook and barrier to the realisation of ujamma (Erixon, 2003; Hyden and Karlstom, 1996).
Nyerere's ‘Arusha Declaration of 1967’ heralded a radical change of direction for the government. The hitherto fairly cautious approach of economic modernisation was cast aside in favour of the grand vision of political and economic transformation in order to realise ujamma. State control was increased in an attempt to hasten industrialisation and reduce external dependence. By the 1970s the government controlled the ‘commanding heights’ of the economy, and state pricing had largely replaced the market system. Foreign exchange controls were introduced, the ambitious villagisation programme was launched, national marketing boards were established and the BIS launched (Yeager, 1989; Coulson, 1982). By the third Five Year Plan (1976-81), ‘autarky and comprehensive state control over the domestic economy’ had become the government's policy (Hyden and Karlstrom, 1993:1396).
Nyerere's political power, derived in part from his charisma, helped secure and was itself secured by this discursive vision for the transformation of Tanzania (see e.g. Jackson and Rosenberg, 1982). Its diffusion through the Party and bureaucracy was speeded by political indoctrination. By the end of the 1970s, Nyerere's ‘strong antimarket philosophy’ had colonised both the Party and government (Hyden and Karlstrom, 1993).
Discursive change within the bank, 1970-80
The Bank under President Robert McNamara supported a broadly mixed economy in line with the post-war Keynesian settlement (Ruggie, 1995). Although broadly sceptical of large-scale government intervention, in practice, when faced with the enthusiasm for national economic planning, the Bank acquiesced and in some cases even advocated comprehensive state-led planning and industrialisation. McNamara's championing of the ambitious Integrated Rural Development lending of the 1970s, saw the Bank's and Tanzania's discourse converge (Kapur et al. 1997).
Discursive convergence would break down at the end of the 1970s as the limitations of McNamara's ambitious rural development programme became increasingly clear and the ‘Basic Needs’ agenda stalled. In 1979 the second OPEC oil crisis hit. The economic shock would soon gather political momentum with the ascension of monetarists in both the White House and Downing Street (Reagan and Thatcher). The consequent ‘sea change’ in political and economic discourses among the Bank's masters would effectively kill off McNamara's ‘Basic Needs’ agenda (Kapur et al. 1997). The Bank's conversion to the new political and discursive order epitomised in the ‘Berg Report’ (World Bank, 1981), pilloried the interventionist rural development pioneered by McNamara, and heralded the neo-liberal onslaught. It advocated rolling back of the state and ‘structural adjustment’ at the expense of the interventionist rural development previously promoted by the Bank and implemented by Tanzania (Lancaster, 1997). As the World Bank's Official History put it:
if there was one country that epitomized the Bank's disillusionment with the past, the break from it, and a preview of the future, it was Tanzania (Kapur et al. 1997:712).
Discursive divergence & convergence, 1979-80
By 1979, both the Bank and Fund and the government were operating within competing discursive frames of references. Each discourse offered mutually incompatible constructions of the problems facing Tanzania (endogenous vs. exogenous causes) and ‘appropriate’ policies to rectify the problems (stabilisation and adjustment vs. more aid, trade reform and state intervention). Nonetheless, there is a strong case for arguing that in 1979 the depth of the discursive divide need not have condemned both sides to the bitter dispute that followed. In 1979, the University of Dar es Salaam as a whole was weak and the intellectual capacity within key government ministries remained limited (Bigsten, 1999). However, a nucleus of pro-reformers remained while a generation of Tanzanian economists had departed to complete their Ph.D.'s in North America. The ideas they returned with had ‘transformed’ internal policy discourses within the faculty of economics to one sympathetic to the Bank and Fund's neo-liberalism; the change was marked. Previously considered to the left of Nyerere, by the late 1970s, key members of the faculty were urging reforms and rapprochement with the Fund (Lipumba, 1988 cited in Campbell and Stein, 1992).
Within the government and the Party there had always been people who were less ‘ideological’ than Nyerere. In part, this reflected those cases where the Arusha discourse had not been internalised (e.g. because it conflicted with other discourses) and in part because support from some quarters was motivated by the rewards that support brought, rather than belief per se (McHenry, 1994; Kiondo, 1989). Whatever their motivation, in 1979 the government was not as yet united against the reforms proposed by the Bank and Fund. In this context, it was by no means certain that the negotiations were doomed to fail. The government's decision to approach the Fund appeared to portend the ‘beginning of technocratic rule’. At the very least, it opened up political space for debate and discussion (Biermann, 1988). Agreement was reached in 1979. If it had not broken down, it might well have sustained the political space needed for a discursive transformation of the bureaucracy away from the ‘Arusha discourse’ toward a position closer to the Bank's and Fund's ‘neoliberalism’. In the event, the agreement broke down, precipitating a bitter split, closing the window of opportunity and forestalling the emergent technocracy's rise (Bigsten and Danielson, 1999).
With Nyerere publicly castigating the Fund and later the Bank, the political space for the development of economic alternatives was closed. The dominant Arusha discourse delegitimated the radical third order called for by the Bank and Fund as well as pro-reformers within government. The government's initial attempts to respond to the economic crisis, the NESP I (1981) and NESP II (1982), inevitably only contained ‘first’ and ‘second order change’ (cf. Hall, 1993). The government firmly believed that the problems facing the economy were primarily exogenous, and that the Bank's and Fund's proposals, which entailed cutbacks in state control and social spending, would not help resuscitate the economy (Bigsten et al. 1999; Kiondo, 1989); problems were to be dealt with by adjusting state intervention (the ‘instrument’ and its ‘settings’). As noted, this position only reinforced the view held by the Bank and Fund that Tanzania did not understand the depth of the crisis and the changes required. In a sense they did not. Operating within a different discourse to the Bank and Fund, they ‘saw’ quite different problems and identified quite different solutions. For Nyerere, the Bank's and Fund's policies were at best ‘irrational’, and at worst, part of a neo-imperialist plot to oppress the South.
Throughout the early 1980s the discursive positions of both sides became entrenched, curtailing their ability to compromise. Within the government, Nyerere's political dominance was exacerbated by the weak intellectual capacity within key ministries where a more pragmatic (more pro-Bank/Fund) position might otherwise have been strengthened. In crude terms, the capacity to develop independent analyses and policy options outside the parameters of the Arusha discourse was simply absent (Hyden and Karlstrom, 1993). The problems were compounded by Nyerere's own weaknesses: his political astuteness did not extend to the economic sphere (Helleiner, 2000), contributing to his ‘excessive reliance’ on a small number of expatriate advisers, including Reginald Herbold Green and Ajit Singh, who were firmly against any devaluation. With Nyerere's support, they were able to resist those within the Party and government calling for devaluation (Hyden and Karlstrom, 1993). This issue, upon which the Government, Bank and Fund were all unyielding, helped block ‘meaningful debate’ or negotiation, and the discursive change that could have broken the deadlock (Bigsten et al. 2001; Gordon, 1994; Kiondo, 1989).
In contrast to the government's position, the Bank's and Fund's monolithic neoliberal discourse confidently constructed the problems as primarily endogenous, indicating the need for structural adjustment and stabilisation (Campbell and Stein, 1992). In the prevailing political climate of the 1980s, this cause was aggressively championed by neo-liberal economists eager to bring down statist development polices, and reluctant to give an inch, lest this be interpreted as a sign of weakness or uncertainty (Kanbur, 1999). The resulting confrontation with the Tanzanian government was predictably heated and the discursive conflict became ‘almost theological’, with little scope for compromises that would not undermine the integrity of either discourse or the political capital each side had invested in it (Hyden and Karlstrom, 1993; Loxley, 1982 cited in Kiondo, 1989). Without any change in their respective discursive positions, each side became imprisoned by them and unable to compromise. The only way the deadlock between the government and the Bank/Fund could be broken was by policy makers adopting ‘irrational’ polices that ran counter to their discourse (e.g. in the face of coercion), or through a further discursive transformation that would change what would count as ‘rational’ policy.
Reform from within, 1982–85
The network of pro-reformers may have failed in 1979 as the political window swiftly closed, but by 1982 a number of them were working on the Bank-sponsored Tanzania Advisory Group (TAG). This helped consolidate links between them and gave its members an early taste of policy work (interview;9 Bigsten et al. 1999; Svendsen, 1995). Within both government and the Party, those of a more ‘pragmatic’ outlook (to use McHenry's, 1994 phrase) helped establish an informal network, or ‘epistemic community’10(cf. Haas, 1992). United by the neo-liberal discourse and an advocacy of economic reform, the network brought together economists from the University's economics department and Economics Research Bureau (ERB) together with private consultants and staff from key ministries such as agriculture, transport, industry and trade (Bigsten et al. 2001; Svendsen, 1995; Lofchie, 1994).
The pro-reform epistemic community went on to organise a series of seminars, targeting policy makers, politicians and a few business leaders, that would prove ‘very influential’ (interview;11 Bigsten et al. 1999). This process helped consolidate the still nascent, neo-liberal counter discourse within Tanzania (Green, 1995; Lofchie, 1994). At the time, their attempt to build support for their position set off a vicious struggle at the University of Dar es Salaam (described by those who took part in it as a time of ‘war’, interviews12) and parts of the ruling party. The reform movement was heavily criticised. The prevailing mood was summed up in the belief that the reformers were part of ‘a systematic manoeuvre by the imperialist West, dominated by the US, to extend their ideological, political and economic influence over Tanzania’ (Bagachwa, 1992:40).
Given this divide, most analyses of resistance to reform in Tanzania (including the Bank's13) identified two broad competing groups within the Party and bureaucracy: ‘pro’ and ‘anti’ reformers (e.g. Kiondo, 1989); ‘reds versus experts’ (Lofchie, 1994:165); ‘marketeers’ versus ‘planners’ (Campbell and Stein, 1992:16); ‘directive’ versus ‘liberal’ factions (Stein, 1992:60); or ‘ideological socialists’ versus ‘pragmatic socialists’ (McHenry, 1994:217). Despite the differences in terminology, most of the analyses agreed that the ‘anti-reform’ faction included Nyerere, major sections of government, including the National Executive Committee, large parts of the Party (CCM), most government ministries (including Labour, Commerce and Industry), and the employees in parastatals corporations and state industries. This anti-reform grouping clung to the ‘Arusha discourse’ that defined the problems facing Tanzania as predominantly exogenous. This discourse indicated that the costs of adjustment were greater than continued resistance, favouring economic autarky (Green, 1995; Lofchie, 1994; Kiondo, 1989). On the pro-reform side, following the Fund's rejection, Cleopa Msuya, formerly the Prime Minister, was appointed Minister of Finance, in 1982. Supported by his permanent secretary, Rutihinda, and the Bank of Tanzania, he pushed for a rapprochement with the Fund, and for reforms. Further support emerged within the Ministries of Planning and Economic Affairs and Agriculture and from the economists at the university (Bigsten et al. 2001; Lofchie, 1994; Hyden and Karlstrom, 1993).
The cabinet remained dominated by ‘hard line left-wingers opposed to the reforms’ leaving Cleopa Msuya largely isolated. Malima brought in Ajit Singh from Cambridge to write a rebuttal of Msuya's proposals, while Malima called Msuya a ‘traitor’ in a memo to Nyerere (Bigsten et al. 1999). Malima went so far as to argue the cause of the problems facing the parastatals was the private sector, suggesting the private sector should therefore be eliminated to allow the parastatals to recuperate. Public opposition grew particularly in the coffee growing areas of Arusha, Moshi and West Lake. Generally though, in the wake of the repression or co-option of civil society in the 1970s, there was little organisational base for resistance, and most Tanzanians simply withdrew into subsistence production and the informal sector, rendering state-directed planning increasingly ineffective (Bigsten et al. 2001; Hyden, 1994; Yeager, 1989).
In the early 1980s, a confident political leadership fortified by the certainty offered by the ‘Arusha discourse’ was able to reject serious reforms without discussion. Yet as time passed, it became increasingly clear that maintaining the existing exchange rate was untenable. The debate slowly began to shift ground, with conflict emerging amongst those seeking to salvage what could be left of Tanzanian socialism and those like Edwin Mtei who advocated its dismantlement. Nevertheless, even those like Cleopa Msuya also seeking to salvage Tanzania socialism, recognised the need for reform, and faced tough opposition within the Party. Although major change was not forthcoming, by this stage incremental (first and second order) change was possible – both advancing and retarding reform depending on who could get the President's ear (Bigsten, 1999).
Alongside the partial reforms and against a worsening economic backdrop, the government continued its exhortations to ‘tighten belts’ with slogans like hali halisis (life is hard) and hali mbaya ya umchumi (bad economic conditions). These, however, were undermined by the increasingly visible inequality dividing the government from the increasingly impoverished populace. As both the economic and now political situation deteriorated, Prime Minster Sokoine initiated harsh security measures against alleged ‘smugglers’ and ‘profiteers’ believed to be stockpiling goods (Bigsten et al. 1999; Shivji, 1992). These campaigns received widespread public support until it became clear that the drive would not actually extend to those within the ruling class. In the face of mounting popular pressure, the government backed down (Campbell, 1992).
The mounting political crisis would create the space for reforms, but did not of itself dictate them. The ever increasing economic hardship and the shortage of cash to fund the patronage that underpinned Party legitimacy left politicians increasingly desperate for a solution. However, it was the apparently ‘perverse’ results of Sokoine's measures – they served mainly to increase shortages, a material result in direct contradiction to the ‘Arusha discourse’ – that helped stimulate a more fundamental rethink within government (Bigsten et al. 1999; Svendsen, 1995). In a situation analogous to Hall's ‘third order change’, the economic and consequent political crisis created the political space in which new ideas could be put forward by pro-reform technocrats (Mans, 1994). A limited import liberalisation was introduced and in 1984 Sokoine widened the liberalisation starting with transport in a bid to aid distribution. The positive response in both economic and popular terms helped consolidate a process of social learning, further strengthening the neoliberal discourse and the technocrat's position (Bigsten et al. 1999; Biermann, 1988).
By 1983/84, with the necessity of devaluation increasingly accepted and the limited economic liberalisation bearing fruit, the neo-liberal discourse gained strength (Green, 1995; Lofchie, 1994). By this stage it was also clear that without a ‘proper’ Fund-sponsored SAP, much needed donor aid would not be forthcoming. Increasingly politicians reassessed where both their country's and their political interests lay. In crude terms, the reforms appeared to offer the prospect of reviving the economy, releasing much needed aid that would help sustain the political system of patronage, and provide the basis for a new settlement between the ‘bureaucratic’ and ‘private’ bourgeoisie (Rugumamu, 1997; Kiondo, 1989). With political support growing, an inter-ministerial technical group co-ordinated by the Ministry of Finance was formed, bringing together key government members with the emergent ‘epistemic community’. Tanzania initiated ‘discussions’ rather than ‘formal negotiations’ with the Fund, first in Washington and later in Dar es Salaam (Svendsen, 1995; Kipokola, 1993). By this stage both agreed on the need for reform, but there were important disagreements about the timing and extent of reforms, with a perception that the Fund favoured ‘shock treatment’ and painful reforms before funding would be forthcoming. In contrast, the Tanzanians feared that this could stimulate still higher inflation and cause political instability (particularly if subsidies for basic foodstuffs were removed) (Ibid.).
As the state teetered at the edge of financial collapse, Nyerere conceded defeat (Svendsen, 1995; Lofchie, 1994). In a bid to secure essential external financing, the 1984 budget, crafted without external direction, albeit with an eye to acceptability, contained most of the reforms demanded by the Fund and advocated by pro-reformers (Green, 1995; Kiondo, 1989). The budget attempted to reduce public expenditure by 30 to 35 per cent, remove subsidies and introduce user fees for the health sector. In line with the ‘National Agricultural Policy’ (URT, 1983), there was also a substantial devaluation (35 per cent), agricultural prices were increased by 30 per cent, and agricultural expenditure doubled (Ponte, 2002; Stein, 1992). Perhaps crucially an own fund importation scheme was introduced which acted as a further de facto devaluation14 (Biermann and Wago, 1986, cited in Kiondo, 1992).
The Fund was approached in late 1984 to evaluate the programme and, hopefully, secure financial support. Meanwhile, the Bank began to take steps to mobilise aid pending an agreement with the Fund (Kipokola, 1993; Loxley, 1989). By 1984, the Fund was attempting to be more ‘accommodating’ in a bid to regain influence. Consequently, the Fund took a less hard-line neo-liberal stand (softening its discursive stance), opening the space for compromise and its support.
In 1985 Nyerere retired as President, though he continued to hold the powerful position of chair of the Party. Opinion is divided about how decisive his retirement was. The majority view is that it opened the door to further liberalisation and devaluation (see e.g. Bigsten and Danielson, 1999; Hyden and Karlstrom, 1993). The minority position maintained that many of the reforms commenced in 1984 before Nyerere's resignation, while his commitment to the ‘Arusha discourse’ had only slowed the pace of change (see e.g. Green, 1995; Morrisey, 1995). In either case, commentators are agreed that the period 1984-86 was the turning point undermining the Arusha discourse and transforming the economy from statism to neoliberalism (Bigsten et al. 2001; Kiondo, 1989). Once Tanzania took this step, and agreement with the Fund was reached, the aid monies mobilised by the Bank flowed once again, saving the state (Mukandala, 1999; Rugumamu, 1997).
Conclusions
The experience of Tanzania and Ghana challenges the dominant view in much of the literature of apparently overwhelming ‘coercive power’ on the part of the Bank and Fund. The increasing multilateralisation of donor relations behind the Bank and Fund leadership in the 1980s may have increased their coercive power, but there were very few cases in the 1980s where they actually appeared to exercise it. Both Tanzania and Ghana are exceptions but they also indicate clear limits to that power. Where resistance was motivated by a discursive disagreement as it was in both Ghana and Tanzania, the discourse also served to stiffen resistance; agreement depended upon either discursive change or one side adopting ‘illogical’ policies. This conclusion is in line with the current literature on aid effectiveness which emphasises the importance of genuine ‘national ownership’ of reforms (see e.g. Dollar and Burnside, 2001, 2000; Collier and Dollar, 1999; Killick, 1998). What this paper offers is a discussion of how that ownership developed in Tanzania.
This paper has demonstrated that Hall's (1993) work on the link between discursive change and policy change in the UK is applicable to Tanzania. In both, policy decisions were considered ‘rational’, that is to say informed by discourses which came to circumscribe the realm of both the political and economic possiblities. Therefore, while the political processes in the UK and Tanzania may have differed, in each case policy change was championed by an ‘epistemic community’ that pioneered the emergence and diffusion of the neo-liberal discourse which flourished when their ideas were taken up by politicians.
By exploring the process of discursive change in Tanzania, this paper has also explored both the limits of the Bank's coercive power and the potential of what might be thought of as its ‘discursive’ power (Holtom, 2003). Within Tanzania, donors like the Bank supported discursive transformation. By suspending aid, they exacerbated the financial crisis that helped spark a political crisis; they incentivesed change (through the promise of aid for reform); supported pro-reformers; and dominated both economic analyses and training. Ultimately though, much of the discursive change within Tanzania was ‘authorless’ and not only beyond their direct control, but all the more powerful for being so (cf. Ferguson, 1990). Moreover, although economic discourses are important, they are not determinative. Agent's subjectivities may encompass multiple (discursive) rationalities. For example, economic policy making in Tanzania, has as elsewhere, never been solely an economic affair. Political considerations have mediated the impact of economic discourses upon policy. Nyerere's and the Party's resistance to economic reforms was bolstered by political considerations. Therefore, even if they had shared the Bank's discourse, political considerations might still have trounced economic ‘rationality’, leading to conflict with the Bank and Fund.