|
|
|
|
Hemmed In: Responses to Africa's Economic Decline
Thomas M. Callaghy and John Ravenhill, editors
New York
1993
4. Neither Phoenix nor Icarus: Negotiating Economic Reform in Ghana and Zambia, 1983-1992
M ATTHEW M ARTIN
| Mr. Kaunda has never known the first thing about economics... the IMF's liberalism-plus-austerity is the right remedy. |
| Ghana, which led the revolt against economic reason, now gets full marks for trying to restore it.... Ghana has done well because its disaster was complete. 1 |
These quotations typify the opinion of IMF-sponsored economic reform programs in Ghana and Zambia held by many in the international financial community. 2 Ghana's post-1983 program is seen as the major Sub-Saharan success story, as the country rose, like the phoenix, from the ashes of 15 years of economic and political turmoil, thanks mainly to its "political will" to implement IMF-style policies. Zambia is a "bad boy": the collapse of its Economic Recovery Program (ERP) in May 1987 is often ascribed to poor implementation or "lack of political will" by the Zambian government. It is believed, like Icarus, to have ignored advice from those who knew better-the IMF and World Bank-and strayed from the path of economic righteousness.
That view, I argue, is simplistic. Overall, Ghana has implemented far more economic policy reforms than Zambia, and with more positive results for the economy. 3 However, this outcome was not determined by "political will." It depended on the degree of domestic political support for reform; economically and administratively appropriate design, and flexible implementation of reform; and availability of foreign exchange. In turn these factors depended on how reform and external finance were negotiated between Ghanaian and Zambian governments and external actors. The negotiating procedures of the 1980s did not guarantee any of the above factors in a predictable or sustainable way, and was not in the interest of any party to the talks. However, it began to change in 1986, and such change must continue if Ghana and Zambia are to sustain "adjustment with growth" in the rest of the 1990s. The following sections will deal with each of these issues in turn.
Programs, Successes, and Failures
Ghana and Zambia had IMF programs before 1983, but in 1983 both embarked on new periods of reform. In 1983-92, Ghana had six IMF programs: three standbys, followed by an Extended Fund Facility (EFF) combined with a Structural Adjustment Facility (SAF) in 1987, an Enhanced SAF in 1988-91, and a "Fund-Monitored Program" (without a loan) from 1992. Zambia also had six programs: three standbys during 1983-87 before abandoning the IMF for its own Interim National Development Plan (INDP) in 1987-89, and since 1989 a Policy Framework Paper in 1989-90, a Fund-Monitored Program in 1990-91, and a Rights Accumulation program in 1991-92. 4
Several excellent studies of adjustment in each country 5 conclude that Ghana has implemented much more adjustment than Zambia. This is certainly true in terms of continuing with adjustment programs. None of Zambia's programs before 1987 were completed, formal talks were abandoned in 1987-88, and the renewed program had to be suspended in the second half of 1991. All of Ghana's programs were finished as planned, and during a nine-year period there was no breakdown in talks.
More detailed analysis of whether program targets and objectives were implemented during 1983-90 shows that Ghana's record was not as perfect-and Zambia's not as appalling-as the progress of talks indicates. Ghana breached many of its monetary and reserves targets, and often fell short of its inflation and real GDP growth objectives. Zambia met many of its fiscal targets, and most of its current account objectives. 6 However, in general, Ghana breached conditions by far smaller margins than Zambia. In addition, as discussed below, the Ghanaian government was much more capable of explaining the reasons for breaches, and the Fund reacted more flexibly to them partly because economic indicators and policy were moving in the right direction. These vital additional factors decided that Ghana's programs continued while Zambia's did not.
Ghana's policies have produced positive economic trends, as shown in table 4.1, below. Per capita GDP and consumption have risen substantially (as have real wages). Budget revenue has trebled as a proportion of GDP, producing a surplus in spite of higher spending. Exports and imports have grown in value and volume. In the early years of adjustment, the programs failed in several ways. Inflation and money supply growth were high. Debt and debt service rose rapidly, requiring large new aid to allow debt to be paid on schedule, and arrears of US$440m to be cleared. Export earnings had a narrow commodity base of cocoa, gold, and timber, because agricultural and manufacturing supply response was disappointing. Food self-sufficiency was elusive, and soaring timber exports caused environmental damage. Savings and investment were worryingly low. However, most of these faults have been remedied (see below). Inflation, money supply growth, debt, and debt service are falling. Nontraditional exports are growing as supply response improves. More attention is being given to food and the environment. Savings and investment have risen. Yet budget revenue, savings, investment, and imports remain dependent on large aid flows and, excluding aid grants and interest payments, the current account deficit has risen considerably. Stabilization and recovery have almost been achieved, but self-sustaining domestically financed development is a long way off.
In contrast, most of Zambia's economic indicators have moved in the wrong direction. Per capita GDP and consumption have fallen almost as fast as before 1983. Savings and investment are barely above the levels of 1983-84. Budget revenue and expenditure have both fallen by 4% of GDP, leaving the fiscal deficit unchanged. Money supply growth and especially inflation are far higher than at the start of adjustment. Exports and imports collapsed until 1987, and by 1990 exports were only 10% above the value of 1983, and imports were 25% lower. Debt and scheduled debt service grew rapidly, payments were less reliable, and arrears mounted; though scheduled debt service apparently fell back at the end of the decade, amounts due to clear arrears meant that even large new aid flows could not prevent negative net transfers. Real wages fell, social services were hit and income inequalities grew; food self sufficiency occurred only in 1986. Agriculture and manufacturing failed to grow as hoped. In mid-1992, the Zambian economy still relied on copper exports and aid for imports; neither stabilization, recovery nor self-sustaining growth were in sight.
Overall, Ghana's adjustment was much more successful than Zambia's. Three factors explain this: domestic political support; the design and implementation of programs; and availability of foreign exchange. It is impossible to separate them or apportion responsibility, because they interact; nor is it necessary in order to show that all three should be (and are usually not) considered when negotiating reform.
Domestic Political Support
7
The successes or failures of the programs cannot be ascribed to blunt factors such as the "political will" of either the government or the broad type of regime. Until 1990, the government of the Republic of Zambia (GRZ) was one-party civilian, under President Kaunda, with periodic elections; in 1991, multiparty elections resulted in a new government under President Federick Chiluba. Throughout 1983-92, Ghana was ruled by the unelected Provisional National Defence Council (PNDC), headed by President Rawlings; but by mid-1992 the PNDC had made rapid progress toward multiparty democracy, with presidential elections planned for November. More specific political characteristics of each country affected support for IMF programs over time: the motivations, cohesion, and power of each government; and the interests, beliefs, and power of extra-governmental forces.
Government Motivations, Cohesion, and Power
The ideologies of Kaunda and Rawlings could theoretically have opposed IMF policies. Kaunda was "humanist," welfarist, and statist, and the PNDC proclaimed itself Marxist revolutionary. "Humanist" concern for growing social problems (malnutrition, poverty) did reduce Kaunda's commitment to IMF-style policy, but Marxism had no influence on Ghanaian policy after 1983. The Chiluba government had no ideology opposed to adjustment.
In practice, both governments and the IMF agreed on the long-term aim of self-sustaining growth through adjustment, by rehabilitating and diversifying exports. The PNDC negotiating positions agreed with the principles of many IMF proposals, but disagreed on their speed, sequencing, and severity. The GRZ under Kaunda often opposed both the principles and the details, but under Chiluba it was closer to the position of the PNDC.
In both countries, the most powerful motivation for agreements with the IMF was a longstanding economic decline that had recently accelerated and was not projected to go away (a "trough" factor). By 1983, both countries had a decade of falling GDP, real wages at 20% of 1975 levels, and real imports 80% below 1980s peaks. Both faced economic collapse just before agreement. By mid-1982, drought in Ghana brought a food shortage (requiring imports), a power shortage (reducing manufacturing), and fires (cutting timber and cocoa exports). Foreign exchange and gasoline supplies were drying up, and a million migrant workers returned after expulsion by Nigeria. By the end of 1982, Zambia's reserves covered only seven weeks of imports, debt arrears were mounting, and food and foreign exchange were short; by mid-1985, foreign exchange and gasoline were vanishing. Similarly, the Chiluba government inherited an economic mess from Kaunda in 1991.
Both countries' governments publicly announced a need for policy change in 1983. The PNDC had taken drastic measures in early 1982, but they did not produce major economic benefits, largely because of a shortage of foreign exchange. The GRZ equally believed that only the IMF could mobilize foreign exchange. Both governments tried to find other funds, but in 1982, Eastern Bloc countries gave Ghana only token loans, and Arab and Indian central and commercial bank bridge loans to Zambia became prohibitively expensive by 1983. Most Organization for Economic Cooperation and Development (OECD) governments and commercial lenders were making new funds conditional on IMF agreement. Zambia's commercial bank oil import loans were suspended in 1985, and Nigeria and Libya refused to give Ghana cheap oil in 1982-83.
Many proponents of reforms in the GRZ used a "no alternative" argument in 1983, and prevailed for long enough to achieve agreement. However, many Zambian leaders continued to believe until 1987 that the "trough" was due to a temporary fall in world copper prices, rather than their own past economic policies, and lobbied for abandoning the IMF because it did not bring economic results. After Zambia's independent recovery program ran out of steam in 1989, the "no alternative" argument became even stronger, creating virtual unanimity behind a new Fund program by mid-1989. Ghana's leaders-and Chiluba's government in Zambia-were free to blame past governments' policies for the "trough," admit its seriousness, and dismiss alternatives; this gave them a "honeymoon period" with the electorate and reinforced their commitment to reform.
Some officials and politicians in each country (more in Ghana) had higher expectations, due partly to Fund, World Bank, and OECD government salesmanship. They believed IMF measures would produce economic recovery, and mobilize huge net flows of aid, bank loans, and foreign investment. This was also true of many in the Chiluba government, who believed that they would implement the measures more coherently than their predecessors. But such expectations could breed rapid disillusion. One previously pro-reform Zambian official said in March 1987: "The Fund doesn't care why or how we are adjusting: it just thinks our previous policies were wrong. The means have become the ends." 8
As is already apparent, neither country's governments had a united "political will"; all were to some extent unstable and divided. In Ghana, political instability delayed the initial agreement in 1983. The previous government had been too unstable to agree. The PNDC took 6 months to resolve internal divisions. Negotiations with the IMF (from April 1982) prompted IMF opponents to attempt coups in October-November 1982. These gave Rawlings the excuse to push them out of government. By the end of 1983, the PNDC had lost almost all anti-IMF members. Many backed coup attempts in 1983-85, but had no influence on economic policy. 9 Most political leaders believed there was no alternative to the IMF, and that the economic benefits outweighed the political problems. In 1986, popular protest and foreign exchange shortfalls increased internal division, but this receded with the changes to the program in 1987-88. Rawlings was able to keep the same senior economic team throughout, enhancing policy continuity, implementation experience, and public credibility.
During 1983-87, Zambia's top policy-making body, the National Council, opposed the IMF by growing majorities. Many IMF opponents were the GRZ's strongest supporters, preventing greater potential threats to political stability (e.g. ethnic tensions). They could not be removed, and remained to speak against the IMF. 10 Their dominance peaked in 1986-87, after IMF opponents replaced Finance Minister Mwananshiku, Central Bank Governor Phiri, and Presidential Economic Advisor Mulaisho in April 1986. After the relative failure of Zambia's home-grown adjustment program in 1987-88, the National Council supported a new program with the Fund, but in 1991 internal opposition grew again in advance of the elections.
In Ghana, a clear, simple policy-making structure eased agreement and implementation. A few pragmatic military leaders made the decisions, advised by a small group of united and extremely competent economic officials (who also negotiated with the IMF). Senior military leaders were also united in supporting Rawlings and most IMF policies. In Zambia, complexity exacerbated division. The National Council, party central committee, cabinet, ministries, and parastatals had a say, reflecting decision-making processes institutionalized since independence and Kaunda's need to balance all groups. Final decisions rested largely with Kaunda, advised by confidants and economic officials. But economic officials were less powerful or united in supporting the program. By 1986-87, as the program failed to produce major economic benefits, their cohesion waned, and opponents of the IMF became more vociferous. In 1989-90, structures were simplified, with the cabinet and the National Economic Monitoring and Implementation Committee (NEMIC) taking almost all decisions. With the arrival of Chiluba's government in late 1991, the new governing party (MMD) lost all formal involvement in economic decision-making: decisions became more centralized in a small group of economic officials around the president.
Division within both governments continued in the mid 1980s while they implemented the measures. There were disputes over many measures at four stages: approving measures in principle; approving timetables and implementation details in principle; announcing measures; and actual implementation. These were more common in Zambia: for example, disputes delayed the ending of consumer maize subsidies for almost two years. They also led the PNDC to reverse civil service benefit cuts in 1986. Implementation exacerbated intra-government division, especially in Ghana in 1983-84 and in Zambia in 1985-87.
Heads of government resolved or overrode division. Rawlings and Kaunda were both politically shrewd, and cared about the economic future of the country; but neither was a trained economist. The main differences were their beliefs, their freedom to exercise absolute power, and their will to be ruthless when necessary. Rawlings was pragmatic, and saw no alternative to IMF-mobilized money and measures; Kaunda was constrained by humanism and wavered in his belief in the efficacy of IMF reforms or the ability of the IMF to mobilize foreign exchange. Rawlings was able to take rapid decisions; Kaunda arbitrated among sectional groups. When necessary, Rawlings was more willing to detain trade union and student leaders, and overrule or remove anti-IMF politicians. Kaunda initially overrode GRZ opposition to the IMF, but in January-May 1987 gradually decided not to expend political capital, pushing implementation past the GRZ and popular dissent, because the program was not producing large economic gains and the IMF did not soften conditions sufficiently. He believed the political and social costs of the program outweighed its economic benefits, and ended talks. 11 In 1989, Kaunda saw that the independent New Economic Reform Program had failed, as foreign exchange shortage and inflation increased rapidly, and agreed to another formal Fund program; but in 1991 he decided to suspend it as elections approached. After the elections, Chiluba played a crucial role in rapidly uniting the new government behind a new program.
Extra-Governmental Influences
Government division often reflected extra-governmental pressure. Both societies contained similar groups that supported or opposed the IMF depending on their interests and belief in IMF measures (or the lack of an alternative). 12 No group was unanimous or favored/opposed all adjustment measures. However, in Ghana, all groups were disorganized, leaderless, and disunited, and the many changes of government in the 1970s and 1980s disrupted lobbying channels. On the other hand, in Zambia, until 1991 opponents were well-organized, well-led, and united, and had well-established channels through which to lobby the GRZ. In contrast, with the arrival of the new Chiluba government in 1991, lobbying channels had to be rebuilt.
Politicians and civil servants were the most powerful group in both countries. They resisted or delayed changes not only because they believed they would not work, but also because the innovations damaged their interests by reducing their living standards, bringing layoffs, depriving them of "rents" from existing policies, or reducing their prestige and status. This applied especially to reforms in major parastatals, with delay in reform of the Cocoa Board in Ghana, and in restructuring Zambia's copper mining, maize harvesting, and agricultural marketing policies. It also delayed budget reform in Zambia and reversed civil service benefit cuts in Ghana in 1986. Many interviewees saw resistance by such groups as a major barrier to reform in Zambia; but after the 1991 elections, a gradual replacement of most senior civil servants relaxed this constraint on adjustment. 13
The other "semi-governmental" group was the armed forces. In both countries, the higher ranks could be relied upon to support the government; they were replaced when they did not. The support of the lower ranks was also crucial in Ghana: though opposed in principle to many adjustment measures, they were convinced by consistent large real wage and benefit increases, and new equipment purchases which maintained their prestige. The lower ranks were much less well-rewarded in Zambia: planned cuts in military spending partly precipitated the coup attempt in 1990-but after the 1991 elections the new Chiluba government had sufficient mandate to ignore them.
Multinationals, white and expatriate businessmen, and large-scale farmers were consistent public supporters in both countries. They gained more access to foreign exchange for imports, and increased production and profits; but their economic gains were resented by other groups. Most local businessmen initially backed reform, but (notably in Zambia) many were disillusioned by 1986-87, because of inflation, lack of credit for working capital, rapid devaluation, and foreign competition from trade liberalization. 14 Farmers were potential supporters of measures that increased their income, but in both countries smallholders were unorganized. The PNDC initially tried to mobilize their support for the 1987-88 local government elections, but eventually fell back on large farmers and migrant workers. In Zambia the influence and organization of the minority of commercial farmers led agricultural reform to concentrate on providing foreign exchange and raising producer prices rather than structural problems. 15
Trade unions, students, academics, the urban poor, and many in the middle class generally opposed IMF agreement in both countries, becoming more vociferous when they perceived no short-term economic benefits or when measures were hastily implemented. Trade unions were largely ineffectual in Ghana, but had more power in Zambia, especially in the copper industry-the country's key foreign exchange earner. They resisted wage and staff cuts and mine closures. Their power determined government reaction (which in turn determined the effect of the opposition): this combined consultation and repression in varying degrees. After mid-1983, the GRZ negotiated with them, pushed the IMF for higher wage rises, and tried to involve them in designing reform; but over time union protest grew, often led by Chiluba who was then head of the largest union federation, culminating in major strikes in January-May 1987. The PNDC combined some concessions while ignoring, detaining, or expelling union leaders. 16 Equally, it closed the universities to quell student protest more readily than the GRZ. Only in 1986, when civil servants, the army, trade unions, and students united, did the PNDC take notice and adapt the program. In 1991, the elections in Zambia brought into government many former trade union and university leaders; as a result, strikes and protests were less well-organized, united, and led-and had less effect on government policy.
In Zambia, 53% of the population was urbanized by 1987, compared to only 32% in Ghana. 17 Urban popular protest in Zambia was the main reason for abandoning the program. In 1983-84 it led to some backtracking on reform, but the December 1986 maize riots were a profound shock to GRZ's (especially Kaunda's) commitment to rapid and dramatic adjustment; they changed the political climate in Zambia far more fundamentally than the IMF realized. They reflected the accelerating collapse of urban real incomes due to rising inflation, and the inept introduction of the maize price increase. 18 The final straws in Kaunda's decision to abandon the IMF were an escalating strike wave in January-May and Lusaka riots over gasoline price increases in April. 19 Again in mid-1990, maize price increases led to riots, but this time their impact was offset by the government's more unified stance behind the program.
In the context of divided societies, the governments' public explanation of policy was vital to successful reform. In Ghana, division was kept private by policymakers' restraint, government control of the media, and PNDC proclamation that the program was its own. The PNDC united in presenting the IMF program as pursuing the revolution and, insofar as ideology mobilized support for the government, it also helped the program in 1983-84. On the other hand, the GRZ was divided in public over whether the IMF reforms were humanist. Ministers publicly criticized reforms and the IMF throughout 1983-87, reflecting a freer press, Kaunda's reluctance to suppress dissent, and the identification of the program with the IMF. This division undermined the credibility of reform. Changing the economic team in 1986 shook donor and business confidence, and accelerated the fall of the currency in the foreign exchange auction. 20 Changes of policy direction also damaged government credibility. This was particularly true when Zambia's IMF talks broke down in May 1987, and when it later returned to the IMF-but it also applied to the PNDC's reversal of civil service benefit cuts in 1986.
Consulting interest groups did little to reduce opposition. The GRZ held a National Economic Convention in July 1984 to increase consensus behind reform and sent out leaders to explain the 1985-86 program. But many were unenthusiastic proponents, or oversold potential benefits, damaging GRZ credibility when they did not materialize. The PNDC allowed less public debate, instead launching a campaign to convince people that reform was the only viable and revolutionary route to recovery, and that it was Ghanaian-designed. This became less effective as memories of the "trough" faded and interest groups did not reap benefits of GDP growth-but after 1987 genuine benefits arrived for most groups. As one IFI official said: "Attempts to convince people inside and outside governments are counteracted if programs fail to produce short-term economic benefits. Consultation exercises only identify the government more closely with the program and increase its instability." 21
The ultimate form of consultation was election. Zambia's one-party elections delayed agreement in 1983 and influenced the decision to end the IMF link in 1987. The multiparty elections of 1991 led to the suspension of the adjustment program by the government, to allow the GRZ more room to increase budget expenditure and imports. Ghana's 1987-88 local elections led the PNDC to slow the pace of adjustment and boost social spending; but even so, economic benefits for the voters did not materialize, and the PNDC had to resort to administrative manipulation of the election results. By mid-1992 it was still difficult to judge the possible effects of the approaching multiparty presidential and parliamentary elections on adjustment implementation, but it is striking that no major party formally declared itself opposed to the PNDC's economic policies, although major political figures did speak against them. 22 Indeed, economic policies seemed to be a relatively minor issue compared with the proliferation of parties and whether Rawlings would stand for the presidency. In September 1992 he announced that he would do so. Some IFI staff saw elections as "a major barrier" to implementation, 23 though post-election honeymoon periods can give new governments more leeway. Rawlings won the November presidential election with about 60% of the vote, and the opposition boycotted the late December parliamentary elections, allowing Rawlings to maintain complete control of the government. The PNDC may have shown that growth-oriented adjustment can facilitate at least partial political liberalization, particularly when substantial economic reform comes first and the opposition is badly fragmented.
A failure to improve conditions can also be a powerful contributory factor in government instability. Yet the relationship between reform and instability is much more complex. In Ghana, prior government instability speeded agreement because the PNDC needed to produce economic results, and enhanced implementation because it allowed the PNDC to blame economic problems on predecessors, and made coup attempts unpopular. This "honeymoon period," the disorganization of opposition and the "trough" effect produced a "culture of silence" in Ghana, which allowed the government to implement adjustment with relatively little instability until 1986. In Zambia, prior stability speeded agreement by enabling Kaunda to override the National Council. However, it weakened implementation because of discontent with past economic performance, and a legacy of GRZ internal disagreements on other issues. Each period of instability (the strikes of 1985, the riots of 1987 and 1990, and the coup attempt of 1990) directly reflected the introduction of new adjustment measures, and was in turn followed by slippage in implementation. The ultimate instability of a pre-election period led to the suspension of the adjustment program in mid-1991.
There was nothing in the state or society of Ghana or Zambia that necessarily precluded reform. They shared political and social characteristics that made reform more difficult, but these were more prevalent in Zambia. In particular, Ghana had a stable team of military leaders who supported extremely competent economic officials, which was one major reason why Ghana achieved more reform. Members of governments and the population acted according to their beliefs, interests, and power. Reconciling these to support reform depended on "carrot" (perceived economic benefits from programs) and "stick" (the government's will and ability to force measures through). Ghana's programs produced more carrot (imports, budget expenditure increases, real wage rises) than Zambia's, and the GRZ was not prepared or able to use as much stick against more formidable opposition. Zambian politics could tolerate the rapid pace of adjustment demanded by the IMF only for short periods, leading to the repeated suspension of programs, the suspension of talks in 1987, and ultimately the fall of Kaunda's government. As one World Bank official involved in Ghana's 1982-84 talks has said, "While the Zambian authorities concluded that the adversity caused by the economic reforms was clearly unsustainable, the leadership in Ghana remains undeterred in undertaking reforms." 24 It remains to be seen whether the new Zambian government will be any more resilient, and whether the post-election government in Ghana will be as successful in designing and implementing reform.
Program Design and Implementation
The second factor in success was whether the program was correctly designed, appropriate to the country's economic and administrative capacity for reform, and implemented flexibly to overcome design faults (a section below explains why programs had design faults). Nobody, including the IMF, knew the exact policies needed to produce "adjustment with growth" in Ghana or Zambia, given limited data and previous research.
Yet three types of design faults could have been remedied regardless of country context:
Four other faults made programs inappropriate to the economic and administrative capacity of Ghana and Zambia: 28
These four trends increased strain on the economy, and political and administrative demands on both governments, and decreased their freedom to implement overall reform targets in ways they chose, without any extra disbursement of external finance. They also strained IMF and World Bank administrative capacity to design conditions appropriate to Ghanaian and Zambian circumstances; to avoid or reduce experimentation, incorrect sequencing, and conflicting conditions; and to monitor compliance or suggest ways of fine-tuning implementation.
Both countries had low economic capacity to adjust. This was due to imperfections in capital, factor, and product markets, including poor communications and incomplete or delayed information flows, an inadequate transport system, declining provision of education and training, low social and labor mobility, dualistic capital markets, and private and public sector monopolies and monopsonies. In addition, protracted recession and import strangulation cut availability of capital and consumer goods; massive excess demand for foreign exchange caused rapid devaluation in an auction; capacity underutilization and decay of capital equipment reduced manufacturing supply response; and low living standards and unemployment cut the tax base and propensity to save. There were also sectoral, distributional, and seasonal constraints on adjustment. These caused "lags" and shortfalls, and unexpected results from reform policies, reducing political support.
However, there were six factors that explained differences in economic capacity between the two countries:
Such major differences in economic capacity required tailoring programs to country circumstances. Yet programs often implemented lessons of one country in another regardless of different circumstances: Zambia's foreign exchange auction was an idea copied from Uganda's 1983-84 program, though the two economies were fundamentally different. However, learning lessons could be positive when they led to caution. Ghana's (and Nigeria's) auctions learned from Zambia's problems: they ensured prior monetary and fiscal stability to reduce demand for foreign exchange, and adequate prior supply of foreign exchange.
On the other hand, important differences of design between the two countries' programs also explain much of their success and failure. Ghana's greater economic and administrative capacity led the PNDC and IMF to have higher expectations of output and supply response in Ghana. Together with rising net foreign exchange inflows, this encouraged them to design a more growth-oriented program after 1984. In turn this design enhanced Ghana's capacity to adjust, with several measures that increased supply-instead of compressing demand, the main focus in Zambia.
As nobody was sure whether measures would work, they often needed fine-tuning through flexible implementation. Proliferating, tighter and deeper conditions and preconditions strained Ghanaian and Zambian administrative capacity to implement them, even with committed reformists in power. This capacity also varied with the number, training, experience, commitment, and morale of the staff involved; the complexity of the decision-making structure; and above all the speed of economic collapse during the program. One pro-reform GRZ official said: "In order to meet monetary and fiscal targets (let alone the others), negotiate external finance, and incidentally run the economy, we needed to be in five places at once and have 36 hours in a day." 34
Monetary and fiscal targets caused the greatest administrative problems for both countries. Both foreign exchange auctions were administrative nightmares, requiring complex documentation and monitoring procedures, and reports to aid donors to gain more aid. Administrative incapacity was due to poor organization, coordination, delegation, monitoring systems, and technical understanding of the wide range of new measures. To the extent that the IMF and World Bank ignored these, or implemented unsuccessful administrative reform, they were partly responsible. It was particularly acute in Zambia, where skilled or experienced professional staff were scarce; conditions like the import tax for the 1986 standby, which needed complex constitutional approval, took 14 months to implement. In contrast, Ghana achieved remarkable improvements in fiscal and monetary monitoring and aid processing by 1988; but lack of permanent institutional mechanisms to adjust agricultural, fertilizer and gasoline prices caused delay. 35
The Ghanaians proved much more capable at fine-tuning-and, by sustaining adjustment, at convincing the IMF to allow changes in mid-program-as emerges from comparison of the two foreign exchange auctions. 36 Both governments managed the auctions, but in different ways. Both initially excluded some foreign exchange transactions (debt service and oil/fertilizer imports); Ghana also excluded pharmaceutical imports and cocoa exports, and Zambia imports for the copper mines. But Ghana initially used a lower exchange rate for excluded items, which reduced the negative impact of devaluation on budget expenditure, whereas Zambia had a single rate. The Fund agreed to Ghana changing auction rules in order to limit speculative bidding and reduce demand for foreign exchange. The PNDC used "Dutch" bidding (where the successful bidder paid the rate bid), 100% prior deposit of local currency by bidders, strict and changing eligibility and documentation requirements which reduced the number of bids accepted (and excluded consumer goods imports in 1986-87).
On the other hand, the IMF opposed rule changes in Zambia, because it did not trust the GRZ to devalue in a managed auction. Though the Zambian pro-reform team asked the IMF to allow rule changes early in 1986, IMF "theologians" refused. This was a major cause of the ensuing depreciation below what the GRZ-including the pro-reform team-saw as an acceptable level, which gave opponents powerful ammunition for the removal of that team. This was then followed by hamfisted GRZ management of the auction, including large changes in the amount of foreign exchange supplied, tighter documentation requirements, Dutch bidding, and a 30% deposit-all without IMF approval. In addition, Ghana satisfied auction demand for foreign exchange at key stages by borrowing from Standard Chartered Bank, while Zambia's auction faced inadequate foreign exchange supply throughout. 37
There were also negative examples of implementation by both governments, announcing or reversing measures before details had been finalized. Sometimes this was done against Fund and World Bank advice. Due to the collapse of the currency, rising inflation and growing political opposition, the GRZ first changed the rules of the foreign exchange auction, then pumped large amounts of foreign exchange into it, and then suspended it in January 1987 with no clear idea of how to replace it. The announcement of an interim fixed exchange rate with two days of warning allowed speculators to make a killing. The PNDC reversed cuts in civil service wages and benefits rapidly in April 1986, in order to dampen popular protest, without sufficient time to assess the implications for fiscal and monetary targets. The IMF suspended the program and insisted on compensating cuts in other government expenditure. Its inflexibility almost led a government wholly committed to adjustment, with remarkable economic and administrative capacity and little political instability, to suspend talks with the IMF. Though (as discussed more below) the Fund soon softened its position, this incident showed how vital careful implementation was to successful adjustment.
However, poor implementation often followed IMF or World Bank advice. The Zambian auction was restored in March 1987 without a complete relaunch of the reform program to boost its credibility, or huge inflows of foreign exchange to pay off arrears. The rate was bound to collapse, and did. In December 1986, Zambia's 120% price rise for the expensive grade of maize meal was announced without properly informing the private maize millers of how they would be compensated for loss on the cheaper grade. They reduced production of cheaper meal (the GRZ argues that they were acting irresponsibly, and should have requested more information), and resulting shortages were the true cause of riots in which 17 people were killed. The fault lay on both sides. The GRZ had delayed price rises for 2 years. The IMF and World Bank were exasperated at this delay, and insisted on a price increase by the exact amount immediately as a precondition for standby resumption, in order to close the budget deficit. Pushing this rapidly through a divided GRZ meant that neither the IFIs nor the GRZ had time to monitor maize availability. 38
There were design faults in both countries' programs. Yet the faults were greater in Zambia, affecting key measures (the foreign exchange auction and maize price). Given Zambia's lower economic and administrative capacity, they made a key contribution to the stop-go nature of reform. Because the Fund and the GRZ had lower expectations of output or export supply response in Zambia, they decided that aggregate demand had to be cut sharply, reducing budget expenditure, imports, and real wages. 39 When compounded by inflexible implementation and inadequate fine-tuning, it was not surprising that the program lacked political support. In contrast, Ghana's program was designed as more growth-oriented and supply-enhancing, increasing budget expenditure, imports, and real wages. Greater economic and administrative capacity, and flexible implementation, produced economic and political benefits which sustained the program.
Foreign Exchange Availability
40
The amount of foreign exchange mobilized by reform was a third crucial factor in the relative success of the programs. Most importantly (as already discussed), it influenced the design and implementation of the program-indeed it often prevented governments from meeting specific program targets. In addition, the extra amount of aid and debt relief pledged helped to reinforce government commitment to reform.
Zambia's inflows of new foreign exchange fluctuated wildly, but exceeded pre-1983 levels only in 1986 and 1991. The IMF and multilateral lenders provided an annual average of 70% of the total (though World Bank lending rose more slowly than in Ghana). Bilateral donors did not believe until 1985 that reform would be sustained, and then lost faith early in 1986; they provided much lower gross flows than in 1980-81, and commercial flows collapsed. 41 Many in the GRZ and World Bank argued that the international community was not responding sufficiently to Zambia's reform efforts. 42
However, Ghana's gross inflows rose sharply, doubling between 1982 and 1989. In 1983-86 the IMF (average US$200mpa) and other multilateral lenders (US$200mpa) led the way, financing 70-80% of imports in 1983-84. Bilateral donors increased commitments only by limited amounts, through antagonism to the PNDC and uncertainty that the program would continue. By 1987-88, encouraged by large new World Bank adjustment loans and the PNDC's ability to sustain adjustment, they were disbursing US$250mpa. Gross inflows rose fast in 1984-5, providing substantial demonstration of the international community's support. A slower rise in 1986 coincided with adjustment problems, but thereafter gross flows rose much faster.
The more important figure for each country was net inflows, after debt service, debt relief and arrears accumulation or reduction. This was the key difference between Ghana and Zambia. Ghana's debt rose from US$1.65bn in 1983 to US$3.50bn in 1990, by a higher percentage but a lower actual amount than Zambia. Its debt service-export earnings ratio rose from 32% in 1983 to 49% in 1990, peaking above 55% in 1987-89. Because most of its debt to governments had been rescheduled in the 1970s, and it owed negligible amounts to commercial creditors, only about US$250m of oil import payment arrears to Nigeria and Standard Chartered Bank were rescheduled (reducing the debt service ratio by only 6-8%pa). Service to the IMF was 40% of the total in 1987-88, because of the short repayment periods of IMF loans. Until 1986, debt service rose even faster than gross inflows, leading to a fall in net inflows from US$131m in 1983 to US$54m in 1986. The falls in net inflows were in 1985 and 1986, which was also the most difficult time for adjustment. 43
Between 1983 and 1990, Zambia's debt rose from US$3.80bn to US$7.22bn, and its scheduled debt service-export earnings ratio from 60% to 66%; however, it peaked in 1986-87 at 100% or higher. 44 Most service due in 1983-86 on medium and long-term debt to governments and commercial creditors was rescheduled or fell into arrears, reducing the actual debt service ratio by 20-30%. But the low and stagnant gross inflows meant that net financial inflows were positive only in 1983, 1985 and 1990-91, and were highly negative in 1986, 1988 and 1989, because of rising debt service and (in 1986) net outflows to the IMF of US$141m. In April 1987, the GRZ faced a 1987 scheduled debt service ratio of 190% including arrears, and a projected net financial outflow of SDR25m. 45 This gave supporters of adjustment little basis for arguing in 1986-87 that it mobilized additional external finance.
In turn net inflows largely determined the volume of imports financed by available foreign exchange, which was a key element in growth and government commitment to adjustment; this also depended on export earnings and terms of trade. Ghana's export earnings rose 89% between 1984 and 1989, because of successful reform, local currency price incentives and foreign investment, while Zambia's fell 29% between 1983 and 1986. In the crucial early reform period, Ghana's terms of trade improved by 47%, while Zambia's fell. Higher net financial inflows in Ghana widened the disparity in import volumes: Zambia's fell by 51% in 1983-86; Ghana's rose by 46% in the same period. 46 Slower import growth in 1985-86, caused by lower net financial flows and export growth rates, coincided with adjustment problems, as the PNDC questioned whether enough growth was being allowed. Similar questions in the GRZ contributed to the abandonment of talks with the IMF in 1987.
Expectations were also important: both countries had large shortfalls of foreign exchange compared to program targets. By 1986, slow disbursement of aid by bilateral donors had created a backlog of US$950m for Ghana. Zambia had even larger shortfalls: gross flows were US$302m less than projected in 1986 alone. In both countries, this reflected donor and technical problems in processing funds; but in Zambia the problem was compounded by suspension of IMF and donor disbursements while adjustment programs were renegotiated. 47
Programs also consistently underestimated scheduled debt service for Zambia (in five of seven program years). This reflected underestimates of short-term debt and arrears, and overestimates of the amount of relief creditors would provide (especially the amount the Paris Club of creditor governments gave Zambia in 1984 48 ). On the other hand, better debt records meant that overestimates of debt service were the norm for Ghana after 1984.
Projections of other balance of payments variables were also wildly optimistic. 49 Prices and production fell short for copper and cobalt in each program year for Zambia, and prices for cocoa in 1986 for Ghana. Zambian copper earnings shortfalls, due largely to price falls, were the largest single cause of collapse in all three programs. They were compounded by SDR-US$ and UK£ -US$ exchange rate changes: as the dollar fell, copper earnings were worth less. 50 Import volumes were lower than projected in every year of both countries' programs during 1983-87, but to a lesser degree for Ghana after 1984.
As with macroeconomic policy, Ghana was more successful at fine-tuning external finance availability, by filling temporary shortfalls with short-term bridging loans. Its loans paid for oil imports in 1983 and 1986, supported the foreign exchange auction, and prevented arrears to the IMF. They were a vital source of stability in IMF programs, covering delays in foreign exchange receipts and enabling more long-term planning of imports and debt service. In addition, in the context of growing foreign exchange and import availability, Ghana was able to repay them all on time, usually out of IMF funds, and new loans remained available in 1987-88. On the other hand, Zambia had to juggle larger and more persistent shortfalls of imports and foreign exchange day-to-day, with increasing desperation. It borrowed more than US$400m of bridging loans in 1983-86, but because shortfalls were much larger than in Ghana, these went immediately to pay growing IMF or World Bank arrears. By 1986, Zambia had insufficient foreign exchange to repay the bridging loans. It took out new loans to repay them, but in late 1986 and early 1987, banks refused to lend.
Overall, optimistic balance of payments projections and external finance shortfalls had crucial effects on the different abilities of each country to implement adjustment targets, and on government and donor commitment to the adjustment program. Higher gross financial inflows and export earnings allowed Ghana to pay most debt service on schedule, and to reduce its arrears. They helped it to implement IMF fiscal and current account targets, thereby increasing donor faith in the program and beginning to restore commercial lender confidence. It also allowed imports, investment, and budget expenditure to grow, with knock-on positive effects on GDP and per capita consumption. This dampened political opposition and reinforced government commitment to adjustment, but even so, persistent shortfalls of aid were one reason for the acrimonious IMF-Ghana negotiations in 1986.
In contrast, Zambia's gross financial inflows stagnated and exports fell sharply. In spite of rescheduling, debt service arrears grew, donor commitment wavered, and commercial lender confidence collapsed. Imports halved, and except in 1986, investment and budget expenditure stagnated. This "undermined many of the production and export benefits of courageous economic reforms," as one World Bank staff member has said. Political opposition grew and GRZ commitment to reform wavered. Persistent shortfalls undermined GRZ belief in the international community's ability to mobilize enough finance to support "adjustment with growth," and gave program opponents powerful arguments to use against reform supporters. By April 1987, most in the GRZ (using data agreed by the IMF) believed Zambia would have higher import and GDP growth if they abandoned the IMF program and limited debt service to 10% of export earnings. 51
The Faults of the Negotiation Process
52
Because of the way reform and external finance were negotiated between the PNDC and GRZ and external actors, both countries' programs had political problems: economically or administratively inappropriate measures, which were inflexibly implemented; and foreign exchange shortfalls. 53 There were four stages in IMF talks--preparation, negotiation, approval and renegotiation-which overlapped in practice.
Preparation: neither PNDC nor GRZ had time to design a complete program independently. The PNDC at least prepared its own Economic Recovery Programs, but with IMF and especially World Bank advice. The GRZ prepared only positions on negotiating issues, though it drafted a program in February 1987. Both spent half their preparation time compiling fiscal, monetary, and balance of payments data requested by the IMF. This left them little time to construct detailed arguments about the political, administrative, and economic feasibility of measures. The GRZ was also delayed by internal negotiations and staff changes; but PNDC preparation improved substantially with staff continuity and experience, helping it to gain marginal concessions by 1985-86.
The IMF prepared a draft program or "letter of intent," containing precise measures and targets, based on IMF staff analysis and calculations of the "distortion" of the economy. These were often based on faulty data provided by the GRZ/PNDC or estimated by Fund staff: notably on inflation, budget expenditure/revenue, and external financing. 54 Some draft letters of intent were "mere intellectual exercises," abstracted from Ghanaian or Zambian circumstances; 55 others took some account of World Bank or PNDC/GRZ views, depending on the flexibility of individual IMF staff. Nevertheless, they tended to define the negotiating agenda, especially in Zambia's talks.
Both sides also concentrated on targets, dates and wording of measures rather than implementation methods. Given the need for rapid agreement to prevent further economic collapse, neither IMF nor PNDC/GRZ had enough administrative capacity to prepare for negotiations by determining domestic political support for measures or their economic or administrative appropriateness; or by preparing accurate data on likely foreign exchange available.
Negotiation: comparing documents on IMF initial negotiating positions with final agreements shows that until 1986 the initial position of the IMF determined 90-95% of Zambia's programs, but only 80-85% in Ghana's.
The IMF position was decided by negotiations within the IMF, between missions, resident representatives, departments and senior management, as well as the strength and views of individual staff and departments, and the backing they received from top management; the Fund's "institutional memory" of Ghanaian and Zambian compliance with past programs; and the views of major creditor governments on the IMF Executive Board.
IMF staff working on both countries varied considerably in their views and flexibility in negotiations, but those on Ghana were in general more flexible (in IMF parlance were more "pragmatists" than "theologians"). 56 Resident representatives and African department staff were usually more flexible than technical departments. Actions of individual staff had considerable influence on talks, largely because the IMF had overwhelming power. Theologians caused GRZ (and occasionally PNDC) resentment by "take it or leave it" presentation of draft letters of intent; this delayed new standbys and weakened the position of pro-agreement GRZ officials. They also rejected changes to Zambia's foreign exchange auction during February-June 1986, pushing the GRZ to take unilateral action. In intra-IMF disputes, senior IMF management tended to back theologians, especially on Zambia; this reflected the IMF's institutional view of government "political will" to implement programs and the views of Board members.
On the whole, Fund staff and management had negative memories of both countries' implementation before 1983, but while Zambia's image stayed negative (due largely to continued implementation problems, but partly to some IMF staff views), Ghana's fulfillment of successive programs improved its image. Meanwhile, powerful IMF Board members became more hostile to Zambia, not only because of implementation problems, but also because Kaunda abandoned mediation in Southern Africa and vociferously supported sanctions against South Africa. The U.S. and U.K. did not intervene to soften conditions on Zambia's behalf in 1986, as they had done in 1983-84. In contrast, by 1986 Canada and the U.K. backed concessions for Ghana, because of Ghana's adjustment record and Canada's and the U.K.'s trading and investment interests, and U.S. and FRG hostility to the PNDC had waned. Due largely to its good performance, Ghana became a "test case" for IMF-style reform in Sub-Saharan Africa by 1985-86; Zambia held that status only briefly in 1986.
Such views of individual countries were largely overruled by a hardening of Board attitude to Sub-Saharan Africa as a whole in 1982-86 because of program breakdowns and growing arrears to the IMF. This produced many of the problems described in earlier sections. Net foreign exchange inflows fell as IMF loan amounts were reduced compared to the period before 1981, and implementation capacity was strained as conditions or preconditions tightened, proliferated, and deepened. The Board attitude also supported "theologians" in intra-IMF disputes. Even Ghana gained few major concessions by 1986, producing "adjustment fatigue." 57
Meanwhile, as discussed previously, domestic political factors explained the different negotiating abilities of the two governments, and thereby their success in adapting the IMF's initial positions. PNDC negotiating ability improved considerably in 1983-86, due to less division, a simpler policy-making structure and staff continuity. Its credibility with the IMF also improved, thanks to sustained adjustment and the persuasiveness, experience, politico-economic skills, preparedness, and IMF knowledge of its officials. Several GRZ negotiators also had these qualities but, because of greater division and extra-governmental pressure, complex policy-making structures, and staff changes, they were less well-prepared for negotiations. Their credibility with the Fund was also undermined by Zambia's poor record of implementing programs.
Yet negotiators from both countries faced acute strain and overwork from talks within government and with the IMF and other creditors, and day-to-day economic management. They were also short of support staff, as well as financial and technological resources. In addition, even when debtor negotiating capacity improved, superior IMF power generally made domestic political support for prospective measures an insignificant consideration in IMF positions. The IMF and most OECD governments took a simplistic view, relying on a fragile coalition of support from Kaunda and a few economic officials; and, fortunately, a less fragile coalition of Rawlings, the army, and economic officials. 58 Politicians were often pigeonholed as "opponents," to be talked into seeing the light, bypassed, or replaced; or "proponents" who were given no major concessions to help them convince opponents or keep their jobs. 59
With the IMF Board's hardening attitude to Sub-Saharan Africa, and past breakdowns seen as due to insufficiently "deep" conditions, many IMF staff saw economic and administrative feasibility as a matter of a "comprehensive" program (i.e. as many "deep" conditions as possible). This imperative took precedence over designing conditions appropriate to the Ghanaian or Zambian economy, especially because data were so unreliable. Data were often disputed, with the IMF overruling GRZ and PNDC calculations of budget expenditure or monetary growth as "optimistic," or balance of payments figures as "pessimistic." 60
World Bank influence on IMF programs was limited. In both countries the Bank, especially its resident mission, played a key role in convincing governments to initiate reform. On some issues, Bank and Fund staff agreed; where they disagreed, Bank staff were often divided. The Fund and Bank agreed to an informal division of labor: the Bank left monetary and exchange rate policy largely to the Fund, and concentrated on sectoral reforms. However, they had severe disputes over both countries, notably on fiscal policy and import levels. The Bank also suggested key compromises, including a foreign exchange auction for Zambia in 1984 (instead of devaluation), a dual exchange rate for Ghana in 1983, and the more flexible form of (and changes in) the Ghanaian auction in 1985-86. However, in general the Fund was not seeking compromise, and Fund views prevailed in Fund-Bank disputes. 61
The dominance of IMF and OECD government views over the outcome of talks reduced GRZ and PNDC commitment to reform. GRZ officials (whether broadly pro- or anti-agreement) felt they had virtually no say, and the IMF made significant concessions only in March 1987. Most PNDC officials felt somewhat exasperated at IMF inflexibility, especially in 1985-86. This had a negative effect on implementation, especially if opponents of measures were unconvinced. When it made negotiations acrimonious (as for Zambia in 1984-85 and Ghana in 1986), it delayed agreement and reduced attention to political, economic, or administrative feasibility. All sides knew this made programs more vulnerable to political problems and misdesigned or misimplemented measures, but the IMF saw no other option and decided to "hope for the best." 62
Approval and Calculation of External Finance Needs: negotiating procedure also left less time to ensure adequate foreign exchange, to arrange repayment of arrears to the IMF, and for the PNDC and GRZ to fulfil any "preconditions," all necessary before formal approval of the program by the Board. Above all, IMF staff often believed there was little additional last-minute foreign exchange to support the program. Thus, instead of calculating a "financing gap" (tailoring external financing to a model of how much was needed to provide a target level of growth), they tended to calculate an "adjustment gap," tailoring balance of payments projections to an estimate of the amount of external finance available.
As a result, export prices and production had to be overprojected-one Fund staff member said of Zambia: "we had to close the gap, and copper projections were the largest variable, so we made them look as good as possible." They restrained projected rises of imports and reserves, which caused acrimonious disputes with the Bank and PNDC/GRZ, who generally wanted higher imports to boost growth and higher reserves to guard against foreign exchange shortfalls. In addition, projections of external finance were often wildly inaccurate, due partly to faulty data on debt and new finance, and partly to optimism needed to close the gap. They took aid disbursement pledges at face value, in spite of a history of disbursement shortfalls throughout Africa. They overestimated debt relief that would be (or had been) provided, by not allowing for downpayments on arrears and short-term debt; this made Zambia's 1984 program underfinanced before it started. Finally, balance of payments projections were juggled at the last minute, especially for Zambia in early 1987, without changing the adjustment program targets to ensure overall coherence. 63
The "adjustment gap" calculation also allowed creditors to pretend that programs were sustainable without any major concessions on debt relief or new funds. 64 Finally, import cuts and optimistic export projections were sometimes forced past PNDC and GRZ opposition; this reduced their commitment to programs. Many authors use the term "external shocks" to imply that foreign exchange shortfalls are external to the program and cannot be predicted or guarded against. This description of the procedure used to calculate external finance needs shows that this is not true: the procedure made programs vulnerable to foreign exchange shortfalls. 65
Implementation and Renegotiation IMF reaction to noncompliance with program conditions was inconsistent. There were no formal rules on when the IMF could waive conditions or had to negotiate a new program. There was only a limited correlation between the number of conditions missed by the GRZ or PNDC (or the amount by which they were missed) and suspension of disbursements, waiver of conditions, or cancellation of programs.
The Fund took note of political problems, administrative or economic capacity to implement conditions, misdesign of conditions, and foreign exchange shortfalls. However, it usually downplayed administrative or economic incapacity, added new conditions to reform administrative procedure or institutions, or insisted on more technical assistance. It saw foreign exchange shortfalls as "external shocks," which could be only partly compensated by more external finance and therefore demanded more adjustment. It denied that the programs were misdesigned, and blamed most problems on the "low political will" of senior politicians (ignoring complex political factors) or misimplementation by the PNDC or GRZ. It therefore continued to insist on the same conditions (including reintroducing Zambia's foreign exchange auction), or added conditions to make programs "comprehensive."
Yet the Fund did not assess which factors were responsible for noncompliance with programs, or vary its response accordingly. Zambia's 1983-84 standby was canceled because of debt service arrears and an excessive budget deficit/GDP ratio. These were due respectively to a foreign exchange shortfall and a budget revenue shortfall, both caused by copper price and production shortfalls; the budget deficit also caused the third breach of central bank credit to the GRZ. In 1986, almost all conditions were breached, but again partly as a result of foreign exchange shortfalls. Fiscal and monetary breaches were due to exceptional extra expenditure through devaluation, higher debt interest payments than expected, and revenue shortfalls. IMF documents for December 1986 showed that excluding unforeseen foreign exchange shortfalls, the GRZ easily met its budget deficit and current account targets-but recommended further budget and import cuts. Other issues causing cancellation (further devaluation in 1983-84 and 1984-85, and disagreements over future budget expenditure in 1984-85) were due to political and technical disagreements.
The IMF did not distinguish causes partly because the key factor in decisions was the subjective view of powerful governments on the Board, and of IMF staff and management. Zambia's 1983 and 1984 standbys were canceled partly because staff and Board saw Zambia as having a poor implementation record, and wanted to have new standbys to make conditions "deeper" and "tighter." Ghana's 1985-86 standby ended not because Ghana had breached conditions, but because the Board wanted softer conditions.
GRZ-IMF negotiations in 1986-87 are an excellent illustration of inadequate negotiation procedure during programs. In February-April 1986, GRZ officials and some Fund and Bank staff argued strongly for more flexible implementation of the foreign exchange auction and maize price increases, partly to keep the pro-reform economic team in place. They were overruled by IMF "theologians" and the Board. The GRZ, facing rising inflation and devaluation greater than the IMF had projected in March, replaced the pro-reform team, introduced auction changes, and delayed the maize price increase without IMF approval. The IMF then suspended disbursements and insisted on tighter conditions for resuming the standby. By December, the GRZ agreed, but one of the tighter conditions was the maize price increase. When riots resulted, the talks collapsed. By March-April 1987, several Fund and Bank staff knew the GRZ was edging away from a program; they felt that IMF concessions in April came "too little and too late," due to failures of communication between IMF and GRZ, and internal disputes within both. 66
The method of treating noncompliance made programs more vulnerable to the causes of noncompliance. It undermined debtor commitment to reform by its subjectivity, inflexibility, and simplistic analysis of political problems. It increased foreign exchange shortfalls by suspending disbursements and further "juggling" adjustment gap data. It cut administrative capacity to implement by distracting all sides in new talks. Finally, the greater pressure for rapid agreement increased program misdesign and misimplementation.
Faults in the procedure of IMF negotiations exacerbated political opposition to economic reform, and encouraged foreign exchange shortfalls and economically or administratively inappropriate (or otherwise misdesigned or misimplemented) programs. They did not create these problems: similar difficulties arose in implementing programs independent of the IMF (Ghana in 1981-82; Zambia in 1987-88). However, to the extent that the procedure did little to overcome them, the problems were symptoms of its faults.
Negotiating procedure was therefore not in the interest of any party to the talks. By 1986, it was damaging the credibility of IMF-sponsored reform even in "good" countries like Ghana: it was amazing that the Ghanaians achieved as much adjustment as they did. Even so, after the IMF and World Bank suspended their loans in 1986, Ghana's Finance Secretary Botchwey (a reform supporter) declared the economy "dead in the water," and adjustment "in danger." 67 By 1987, by exacerbating political problems, economic and administrative incapacity, and foreign exchange shortfalls, it had led Zambia to abandon IMF talks.
Reforms During 1986-92
68
Many other developing countries, particularly in Sub-Saharan Africa, were experiencing similar problems. From 1986, the IMF, World Bank, and OECD governments acknowledged that the negotiating process did not fulfill their interests, and began to reform it. Having persisted with adjustment, Ghana could take advantage of these reforms. Zambia's talks collapsed just as reforms were being introduced; until it agreed to an IMF program in September 1989, it did not benefit. 69
Economic Reform
Between 1986 and the present, there has been growing flexibility in the design and implementation of adjustment programs. Ghana's program was one of the pace-setters for this flexibility, which came at a key time. In 1986, PNDC commitment to adjustment was waning and extragovernmental opposition growing. In protracted negotiations, Ghana (backed by the World Bank) argued for a more growth-oriented program, and achieved it. The growth, investment, savings, and import objectives were increased, higher budget expenditure and a larger deficit were allowed, and the Fund relaxed conditions on gasoline price increases and domestic credit cuts.
In addition, the program gradually concentrated more on social and environmental effects. The World Bank joined UNICEF in 1987 in promoting a Program of Action to Mitigate the Social Costs of Adjustment (PAMSCAD), a program of social sector projects to be funded separately by aid donors. Though it was small (US$80m) and took a long time to get started, it was the first such program in Africa and a major step to combatting the social costs of adjustment. In particular, it reduced the potential negative political results of reducing civil service employment. Ghana has also benefited from studies on how to incorporate antipoverty measures in adjustment programs, under the World Bank-United Nations Development Program-African Development Bank Social Dimensions of Adjustment (SDA) program, and from World Bank studies on how to increase its food security. From 1990, the government banned exports of certain types of timber, to reduce environmental damage. 70
This flexible design and implementation persisted until 1992. Design conditions became gradually less tight (in the sense of additional adjustment required), and preconditions diminished as trust grew between the PNDC and IMF. Thanks in part to larger flows of external finance, programs became increasingly growth-oriented. As an example on implementation, when the PNDC ended the auction and unified the auction rate with that of private foreign exchange bureaus in 1990, it convinced the IMF to allow continued surrender of 60% of foreign exchange to the Bank of Ghana, and other measures to limit speculation.
Greater flexibility came too late to save Zambia's program in 1987. The Policy Framework Paper (PFP) signed in December 1986 contained conditions that were tighter and more detailed than before, and it was abandoned after the maize riots. Concessions on conditions in April 1987, notably a larger budget deficit and slower rises in maize consumer prices, prompted partly by World Bank and OECD government mediation, were insufficient to enable a new PFP.
However, since negotiations reopened in 1988, program design and implementation have been slightly more flexible. In preparing the PFP in 1989, Zambia and the IMF compromised on maize prices, devaluation, and budget expenditure, with help from independent advisers and the World Bank and pressure from OECD governments. In 1990-91, a Social Action Program was designed and added to the program. There was also some leeway in implementation, allowing import liberalization and maize meal price rises to be delayed during 1990-91, but this was very limited and the Fund and GRZ formally suspended the program before the elections of 1991. In addition, the IMF has insisted on continued rapid adjustment, with tight and proliferating conditions and preconditions. The program has remained broadly deflationary and marginally growth-oriented; falls in per capita GDP, per capita consumption, import volume, and budget expenditure as a percentage of GDP were projected in 1991 and 1992 programs.
The changes in programs reflected general trends in IMF procedure: more effort to tailor conditions to national capacity, more thorough preparation and data compilation, and more support from senior management for Africa department staff and resident representatives. The greater flexibility for Ghana resulted from IMF appreciation of Ghana's past adjustment, its wish to maintain the Ghanaian program, pressure from OECD governments and the World Bank, the collapse of other Sub-Saharan programs (which was ruining the Fund's reputation in Africa in the mid-1980s); more effective negotiation by the Ghanaians; and the greater leeway provided by higher external finance inflows. Less flexibility for Zambia partly reflects the need to counteract what the Fund perceives as the "backward steps" taken during 1987-89; the historically based IMF mistrust of the Kaunda government until its fall in 1991; the Chiluba government's anxiety to hasten adjustment; and lower net external finance. The changes therefore produced a program more appropriate to Ghana's political, administrative, and economic capacity, which avoided major implementation problems, but have not had the same effect on the Zambian program.
Calculating and Fulfilling External Finance Needs
There have also been changes in the calculation and provision of external finance needs. Zambia's program in 1990 introduced a "contingency mechanism" to reduce the risk of balance of payments or external finance shortfalls compared to program projections. Copper export earnings were projected in the program at much lower levels than actually expected: any excess earnings were to offset aid disbursement shortfalls, with the remainder used to boost reserves or repay the IMF faster. In 1991 the mechanism was refined and expanded to allow for possible higher oil prices. This mechanism reflected a genuine learning process within the Fund, based on internal study and debate, because many earlier programs (especially in Zambia) had been derailed by such shortfalls. Throughout 1990-91, it kept import levels approximately as projected, and donor and GRZ staff see it as a major reason why a number of programs have not broken down.
Less formal methods have been used for Ghana. Since 1987, the IMF has agreed to less optimistic balance of payments projections to guard against probable cocoa price falls. Unfortunately, extremely low cocoa prices in 1987-88 led to the suspension of SAF and EFF disbursements because they directly caused breaches of budget and external sector conditions. They also led to shortfalls in import volume in 1987-89. Since then they have been offset informally by excess aid disbursements.
The more growth-oriented design of the Ghanaian adjustment program, and the artificial pessimism of the Zambian balance of payments projections under the contingency mechanism, both demanded greater efforts by donors and creditors to provide external finance.
The IMF seemed to be making these efforts. In November 1987, it granted Ghana a huge package of loans on softer terms: SDR143m through the Structural Adjustment Facility (SAF); and SDR245m through the Extended Fund Facility (EFF)-the first EFF loan to Sub-Saharan Africa for four years. In November 1988, Ghana gained an Enhanced Structural Adjustment Facility (ESAF) loan of SDR368m, but the amount of new money was less than it seemed, because SDR250m undisbursed from the SAF and EFF were included. These loans did not prevent net repayments to the IMF throughout 1987-92, and will not prevent net repayments in 1993-97. However, they reduced net flows to the IMF (and with other measures increased net inflows to Ghana) enough to reinforce Ghana's commitment.
Zambia was about to be granted a SAF loan in December 1986, but since then arrears to the IMF have prevented it from borrowing. It failed to mobilize enough additional aid to clear them at a donor "support group" in 1989. In May 1990, prompted largely by the Zambian case, the IMF Board introduced the "rights approach," which allowed the GRZ to accumulate rights to IMF loans, without ever receiving the funds. The "rights" were credited to the IMF, to offset arrears. Once all arrears were cleared, Zambia could draw new IMF loans. In addition, it had to pay all current IMF debt service. This combined burden of arrears clearance, current service to the Fund, and no new IMF loans implied underfunded (and therefore less growth-oriented) adjustment programs, and contributed directly to the suspension of the program in 1991. In June 1992, Zambia started the second year of its rights program, postponing the date for resuming normal relations with the Fund-and continuing underfunded adjustment-until June 1994.
The World Bank has dramatically increased its IDA commitments and disbursements to both countries (though Zambia had first to use an expensive Bank of England bridging loan to clear arrears to the World Bank). In addition, the "fifth window" initiative has given both countries relief from interest payments due on IBRD loans they received when they were middle-income countries. This has reduced Ghana's debt service by a total of US$20m in 1988-92; and Zambia's by US$70m in 1991-92.
Bilateral and multilateral aid to Ghana continued to rise sharply, as donors tried to compensate for cocoa price falls and allow more import growth: pledges at Consultative Group meetings were of US$800m a year for 1987 and 1988, US$900m for 1989-90, and more than US$1bn for 1991-92. Donors also responded for Zambia, making pledges exceeding US$700m in 1990 and US$900m in 1991. These pledges demonstrated donors' commitment, and reinforced PNDC and GRZ determination to persist with reform.
Both countries have also gained from the World Bank's Special Program of Assistance (SPA). This has mobilized additional bilateral aid, enabled the Bank to monitor bilateral aid disbursements and foreign exchange availability more closely, concentrated aid on cofinancing economically viable projects and reform programs, and reduced demands on recipient and donor staff by streamlining conditions attached to the aid. However, disbursement has remained behind schedule, largely because of delays in bilateral aid. At the end of 1991, Ghana's "pipeline" of undisbursed aid exceeded US$2.5bn. Zambia had disbursement shortfalls estimated at more than US$160m in 1990 alone. The delay is slightly greater in Zambia, reflecting both donor uncertainty that adjustment will be sustained, and GRZ problems coping with processing the complex conditions attached to aid. Aid shortfalls have been largely responsible for foreign exchange shortfalls in Ghana since 1988, and have offset most of Zambia's excess copper earnings under the contingency mechanism in 1990-91.
In addition, gross flows continue to be almost outweighed by debt service in Zambia, while net flows have risen sharply in Ghana. This reflects the end of the debt service "hump" in Ghana, and accelerated efforts to cancel Ghana's debt service by Canada, Denmark, FRG, France, the U.K. and the U.S.. Almost all aid debt was canceled, amounting to US$440m or 15% of total debt, and reducing debt service payments by US$55m a year in 1990-91 (7-8% of export earnings). Canada, Denmark, the Netherlands, and Sweden canceled loans owed by Zambia in 1986, but others (the U.K. and U.S.) held back until 1990-91 because it had no IMF program. In 1990, Zambia received the highly concessional Toronto terms from the Paris Club. But this reduced annual service by only US$14m compared to previous terms, so creditors had to go further, by deferring 30% of the interest due on the agreement, in order to provide the external finance needed to support the program. In 1992, Zambia received the "Enhanced Toronto" terms, but these saved only US$17m, and interest on the 1990 agreement had to be deferred again. 71
Commercial creditors took no initiatives. Ghana's commercial debt arrears were cleared by 1990. Though Zambia applied to use World Bank money to buy back its commercial debt, there had been little progress on this by mid-1992. The only contribution to relief by commercial creditors was to reduce debt through debt-equity swaps, which cost them nothing as they had already set money aside against bad debts.
The contrast between Ghana and Zambia is stark: in Ghana, rises in new aid are leading to larger net flows, allowing larger current account deficits and import volumes. In Zambia, aid increases are being mostly offset by rising debt service, particularly for arrears to the IMF and World Bank. Clearing these arrears forced Zambia to cut reserves temporarily, and donors to disburse aid faster, in 1991. These added to the cost of the bridging loan to clear World Bank arrears, and underfinancing caused by the IMF's "rights approach." They have prevented a substantial rise in net inflows or import volumes, which would be needed to make Zambia's program genuinely growth-oriented.
Conclusion: The Need for Further Reform
72
While initiatives since 1986 have reduced some faults in negotiating procedure, they have not gone far enough. The lesson of economic reform in Ghana and Zambia in 1983-92 is that further reforms are required in the procedure for negotiating economic policies and external finance.
Economic Policy Reform
Recent negotiations, notably those of Ghana, have taken more notice of political, administrative, and economic constraints on reform, but have not systematically assessed them. Successful and sustained reform in both countries will require basing programs on three aims: maximizing domestic political support; increasing economic capacity to adjust; and increasing administrative capacity to adjust.
The first step is to improve preparation for talks. To make data as impartial and accurate as possible, joint preparation by teams of IMF, Bank, PNDC, or GRZ and independent experts should precede formal talks. It is vitally important that both IMF and PNDC/GRZ staff write their own independent draft programs, concentrating on details and implementation methods, but leaving room for concessions by suggesting broad ranges instead of precise targets and by offering several ways to achieve them. Severe disputes on programs or individual policies in either country might be resolved by using such teams or independent advisers to report on past incapacity or program misdesign and misimplementation, and suggest broad outlines of a program or compromises on individual measures.
As to the outcome of talks, tailoring policies to Ghanaian and Zambian circumstances requires more stress on certain largely neglected policy areas. To some degree, as judged by the failures of programs described previously, these are similar in both countries: export diversification; reducing budgetary and import dependence on external finance; reducing debt stock and debt service payments; food production, social sector, income distribution, and environmental issues; and improving supply response, especially in manufacturing. More stress on these wider, longer-term and structural issues implies World Bank influence and representation equal to that of the Fund in SAF and ESAF talks, because it has more expertise and experience on these issues-and the involvement of more independent expertise from within and outside Africa.
The two economies are at different stages: while Ghana needs continued recovery and long-term self-sustaining growth, Zambia also faces enormous short-term problems of stabilization and recovery. Short- and long-term, stabilization and self-sustaining growth cannot be separated because similar political, economic, and administrative incapacity affects policies designed to achieve them. Yet for countries like Ghana, which have undertaken initial periods of stabilization, the World Bank could assume program leadership, to demonstrate the focus on longer-term issues.
The political coalition behind reform in both countries remains narrowly based, and especially fragile in Zambia. To maximize political support, governments must have a larger part in program design. To the extent that they prepare their own draft programs and negotiations are marked by genuine compromise, policies will be genuinely "domestically designed," and intragovernmental discontent reduced. During preparation, negotiation and implementation, PNDC, GRZ, and IMF/World Bank representatives should consult constantly with all parts of government and extragovernmental interest groups (especially civil service, parastatals, trade unions, and business) to take account of their views and explain proposed measures. Such consultation should concentrate on measures that have caused past political problems (devaluation, maize and gasoline price rises, civil service retrenchment in Zambia; civil service wage and benefit changes and gasoline price rises in Ghana). It would aim to gain intra- and extra-governmental support for measures that were more gradual, in order to minimize the risks of negative short-term effects. One major lesson of the fate of adjustment in Ghana and Zambia is that short-term economic gains help to improve prospects for adjustment: that economic reform is more likely to persist if it is genuinely "growth-oriented."
Even with such changes in procedure, programs would not be foolproof. Thus negotiations must leave more room for flexible implementation, and reactions to problems must be more flexible. Continuing detailed analysis of causes (by both IMF and PNDC/GRZ, or by advisers) and consultation within Ghana and Zambia will help, as will implementing measures only after GRZ/PNDC officials have been fully trained; and, where possible, without encouraging speculative activity by extragovernmental interest groups involved in or affected by implementation.
This new negotiating procedure will require internal changes on both sides, to simplify and maximize communication within and between them. Within the IMF, to encourage sensitivity to national constraints, more staff should be assigned to the African Department on each country. They should regularly review experience on political, administrative, environmental, and social issues, assessing the causes of implementation problems more precisely to ensure that the IMF's "institutional memory" of the countries is accurate. Resident representatives in Accra and Lusaka should have immediate access to (and support from) top IMF management and, together with staff of the Bank and donors in Accra and Lusaka, be responsible for warning of political or administrative problems.
Changes in debtor government negotiating procedure are also needed, particularly in Zambia, which requires continuity of experienced staff and more trained support staff, to improve administrative capacity to design, negotiate, and implement policies; a simpler policy-making structure, with a lead agency or stable small team to conduct negotiations and supervise implementation; and a mechanism for reporting to a wider coordinating group of interested agencies, to maximize political support. Ghanaian negotiating procedure needs far fewer changes, though there is room for continued training and expansion of support staff and more consultation of interested government agencies.
External Finance
Recent initiatives will increase net flows to Ghana and Zambia. However, the problem is much more fundamental: after a decade of adjustment, the international community still has no idea how much external finance any country needs to support economic reform. In order to maximize the effectiveness of future external finance for Ghana and Zambia (and minimize the additional amounts needed), the faults of gap calculations and methods used to fill them must be remedied.
The volatility of export earnings and external finance flows argues that the top priority should be to apply a version of the contingency mechanism used in Zambia to the Ghanaian program. The amount of debt reduction and new money should be based on worst case projections of the balance of payments. This should spring from much more detailed study of all components of the balance of payments, especially of how to prevent growing import dependence, which was a major failure of earlier programs; 73 and on comprehensive analysis of the relationship between external finance and development, to ensure that "gap" calculations are tailored to each country's "financing needs" (rather than the procedures and practices of creditors and donors).
Filling this genuine gap will imply comprehensive debt relief and new finance from all creditors, including OPEC governments and commercial creditors. The best way to achieve this will be by greater synchronization and coordination of decisions on debt relief and new flows by different groups of creditors (and within creditor governments). Ghana's decisions already center on Consultative Groups, because it has so little rescheduling. This saves considerable staff time on all sides, leaving more for design and implementation of adjustment. Zambia faces separate Paris Club, London Club, and other rescheduling negotiations, all of which could be combined in one meeting, maintaining confidentiality and separate decision-making by separate caucuses for different creditor groups. Joint creditor-debtor preparation of such a meeting and monitoring of its implementation would reduce delay and duplication for all sides.
For Zambia, the most crucial gap-filling issue remains reducing debt stock and service-but it is also important for Ghana. Creditor governments are the main potential source of reduction. For Zambia, they can cancel 100% of aid debt (which would reduce annual debt service by US$10-12m compared to current rescheduling terms). They can also move further than the Enhanced Toronto terms on nonconcessional debt, toward Trinidad terms (66% cancellation) or even Pronk terms (100% cancellation). Trinidad terms would save Zambia US$15m annually, and Pronk terms another US$35m. They could also reschedule debt service due during a three-year period, to match the rights accumulation program period. This would avoid expensive and time-consuming annual negotiation and enhance long-term planning.
Ghana is intent on avoiding rescheduling, in order to maintain its access to new export credit loans. Even so, cancellation of all aid debt could save it US$30-40m a year (because it is not currently rescheduling), and canceling 100% of nonconcessional debt contracted before 1986 would save US$50m a year.
Both countries would also benefit if creditor governments refinanced debt service with additional grants; accepted payment in local currency; or converted their debt into equity, development or environmental projects. These measures would not necessarily imply cancellation, and could therefore be undertaken without damaging Ghana's creditworthiness. Several bilateral donors, however, are suggesting that providing such large amounts of aid (especially if it means diverting aid from other nations) may be unsustainable in the longer-term.
Ghana's commercial creditors could also undertake debt-equity swaps, but this is unlikely given Ghana's good repayment record. Zambia needs to finalize as soon as possible its use of World Bank money to reduce commercial debt or debt service by bonds or buybacks.
Both nations have large debt service burdens to multilateral institutions at least until 1997: US$250-300m pa (not including arrears) for Zambia, and US$150-200m pa for Ghana. These institutions should provide net inflows, on softer terms, to refinance current (and reduce future) debt service. For the IMF, this would imply faster and larger disbursement of ESAF loans, with repayment periods lengthened to match IDA; for the World Bank, it will require a larger replenishment for IDA IX in real terms than IDA VIII. In addition, both the Fund and the Bank could use more of their resources to subsidize repayments of principal and interest on past expensive loans to low-income Africa, and the Fund could soften the rights approach to allow the debtor country to receive a proportion of the rights as new disbursements during the rights accumulation period. These measures need not imply cost to governments: using the undisbursed portion of SAF and ESAF resources, a small part of the IMF gold reserves, the investment income from those reserves, a small issue of SDRs, and 100% of IDA repayments due in 1990-95, would provide ample funds.
Like most other African countries, Ghana and Zambia found themselves "hemmed in" in the mid-1980s, by collapsing economies and flows of external finance. At the present time, Ghana is much less "hemmed in," because of its economic policy reforms, although these remain dependent on large volumes of aid and imports. Zambia is if anything more "hemmed in," because of a combination of intermittent reform and foreign exchange shortage. The past experience of both countries is that political liberalization has tended to undermine economic reform, by increasing the demands on government for expansionary policies. At the moment, both countries are well-placed to sever this connection. Having survived both the presidential and parliamentary elections, the Rawlings government is likely to continue to pursue vigorous economic reform, and Zambia's democratically elected government is still in a post-election honeymoon period when reform may be easier to introduce. Yet this is no reason for complacency: it will be impossible to sustain this combination of economic and political liberalization without major changes in the way reform programs and external finance are negotiated.
The lesson of the Ghanaian and Zambian experience is not that a country can escape by just maintaining its "political will" to implement incontrovertible "economic reason." It is that economic reform programs succeed insofar as they increase domestic political support; administrative and economic capacity; careful design and implementation; and foreign exchange availability. If further changes in procedure ensure that programs take full account of these factors, they will make programs much less likely to break down in all African countries. They will reinforce the political liberalization and self-sustaining development Africa needs, by showing that the international financial community realizes the complexity of negotiating economic reform in Sub-Saharan Africa and that "A once and for all judgement of the 'readiness for reform' of a government and its leaders, or a nation, is unfair to the millions of people living in poverty and seeking progress now." 74
Note 1: The Economist, May 9, 1987, p. 15 and August 20, 1988. Back.
Note 2: For an extension of this argument to all of Sub-Saharan Africa, see my book The Crumbling Facade of Africa's Debt Negotiations: No Winners (MacMillan Press, 1991). I am grateful to the Economic and Social Research Committee for funding the research. I would also like to thank the many people who helped me, particularly James Mayall, Tony Killick, Gerald Helleiner, Percy Mistry, Susan Strange, and all those who allowed me to interview them and to look at their papers or those of their institutions. In what follows, any details not sourced are from confidential documents. Back.
Note 3: My aim is to compare the two sets of IMF programs to draw lessons for future IMF negotiations. Thus this chapter does not discuss Zambia's 1987-88 Interim National Development Program, because it did not involve the IMF. Back.
Note 4: Both countries also negotiated reform programs with the World Bank. This chapter largely omits them (to avoid excessive length and complexity), but discusses the Bank's role in IMF talks. Back.
Note 5: This section is based primarily on confidential IMF and World Bank documents, and interviews with all sides in the talks. See, in addition, the the following literature: For Ghana: Sheetal K. Chand and Reinold van Til, "Ghana: Towards Successful Stabilisation and Recovery," Finance and Development 25, 1 (March 1988): 32-35; Reginald Herbold Green, "Ghana: Progress, Problematics and Limitations of the Success Story," IDS Bulletin 19, 1 (January 1988): 7-15; Reginald Herbold Green, Stabilization and Adjustment Policies and Programs: Country Study 1-Ghana (Helsinki: World Institute for Development Economics Research, 1987); Tony Hodges, "Ghana's Strategy for Adjustment with Growth," Africa Recovery 2, 3 (August 1988): 16-20, 27; John Loxley, Ghana: Economic Crisis and the Long Road to Recovery (Ottawa: North-South Institute, 1988), and Matthew Martin, "Negotiating Adjustment and External Finance: Ghana and the International Community, 1982-89," in Donald Rothchild, ed., Ghana: The Political Economy of Recovery (Baltimore: Johns Hopkins University Press, 1991).
For Zambia: Christopher Colclough, "Zambian Adjustment Strategy-With and Without the IMF," IDS Bulletin 19, 1 (January 1988): 51-60; Christopher Colclough, "The Labor Market and Economic Stabilisation in Zambia" (Washington: World Bank PPR Working Paper No. 272, November 1989); Ravi Gulhati, Impasse in Zambia: the Economics and Politics of Reform (Washington: World Bank Economic Development Institute Analytical Case Study No.2, July 1989); Tony Hodges, "Zambia's Autonomous Adjustment," Africa Recovery 2, 4 (December 1988): 6-13; Igor Karmiloff, Industrialisation in sub-Saharan Africa: Country Case Study-Africa, ODI Working Paper 26 (London: Overseas Development Institute, 1988); Jonathan Kydd, "Coffee After Copper ? Structural Adjustment, Liberalisation and Agriculture in Zambia," Journal of Modern African Studies 26, 2 (June 1988): 227-251; Martin Sakala and Manenga Ndulo, "The International Monetary Fund and the Zambian Economy," in K. J. Havnevik, ed., The IMF and the World Bank in Africa (Uppsala: Scandinavian Institute for African Studies, 1987); Hans-Otto Sano, "The IMF and Zambia: the Contradictions of Exchange Rate Auctioning and De-Subsidisation of Agriculture," African Affairs 87, 349 (October 1988): 563-577; Jurgen Wulf, "Zambia Under the IMF Regime," African Affairs 87, 349 (October 1988): 579-594; and Roger Young, Zambia: Adjusting to Poverty (Ottawa: North-South Institute, 1988).
It also draws on Thomas M. Callaghy, "Lost Between State and Market: The Politics of Economic Adjustment in Ghana, Zambia and Nigeria," Joan M. Nelson, ed., Economic Crisis and Policy Choice (Princeton: Princeton University Press, 1990), pp. 257-319; the Country Economic Review and Country Economic Profile (1983-June 1992) of the Economist Intelligence Unit, London for both countries, and many periodicals. Back.
Note 6: For a comprehensive analysis of compliance with these targets, and the effects of external finance shortfalls on adjustment, see Matthew Martin and Percy Mistry, eds., How Much Aid Does Africa Need? (forthcoming). Back.
Note 7: In addition to the sources cited in note 5, political developments in Ghana are discussed by Kwame A. Ninsin, "Ghanaian Politics after 1981: Revolution or Evolution?" Canadian Journal of African Studies 21, 1 (1987): 17-37; Donald I. Ray, Ghana: Politics, Economics and Society (Boulder: Lynne Rienner, 1986); Donald Rothchild, and E. Gyimah-Boadi, "Ghana's Decline and Development Strategies," in John Ravenhill, ed., Africa in Economic Crisis (New York: Columbia University Press, 1986). There is no comparable recent source on Zambia's politics since 1983. Back.
Note 8: Interview, Lusaka, February 1987. Back.
Note 9: Rothchild and Gyimah-Boadi, "Ghana's Decline," pp. 270, 273-274 and confidential interviews. Back.
Note 10: In January 1983 only 11 of 28 members of the Central Committee voted for an IMF program, and in June-September 1985, only 5. South, February 1987, p. 20. More precise figures from interviews in Lusaka, March 1987. Back.
Note 11: International financial institution (IFI) documents and interviews in Lusaka and Washington show that in December-January 1986 he was shaken by the maize riots, the death of his eldest son and regional problems, but by February 1987 he was back in control, deciding to end talks unless the IMF made concessions by April-May. Back.
Note 12: This description is simplified for the sake of brevity: for an excellent analysis of the complex effects on income patterns of different social groups in Ghana, see Green, Stabilization, pp. 12-21. Back.
Note 13: On Ghana, see Green, "Ghana" on the Cocoa Board and Rothchild and Gyimah-Boadi, "Ghana's Decline" on the civil service; on Zambia, Gulhati, Impasse, pp. 37-46. Back.
Note 14: Interviews with 14 prominent business leaders, Lusaka, February-March 1987. Back.
Note 15: Callaghy, "Lost"; Gulhati, Impasse, pp. 43-46. Back.
Note 16: On Ghana, see Nicholas Van Hear, "Labour and Structural Adjustment in Nigeria and Ghana" (paper to ASAUK conference, September 1988, mimeo). Back.
Note 17: World Bank, World Development Report 1989, p. 224, Table 31. Back.
Note 18: The precise political cause of the riots is disputed-Sano, IMF, p. 572, note 11, says World Bank staff blamed politicians opposed to reform for stirring them up. However, they were able to stir up opposition because of the collapse in urban real incomes, especially in the Copperbelt. See Colclough, "Zambian," p. 59. See also below for how misimplementation contributed to the riots. Back.
Note 19: Interviews, London and Washington, May-July 1987. Back.
Note 20: See especially Kydd, "Coffee"; and Murray Sanderson, "Why Zambia's Auction Failed" (paper to Economics Association of Zambia Conference on the Zambian Auction System, Lusaka, June 1987, mimeo). The replacement of the expatriate governor of the Central Bank did not have the same negative effect in 1992 as it did in 1986, even though he had been regarded as a key supporter of reform, because he was replaced by Dominique Mulaisho, another supporter (and Kaunda's former economic adviser). Back.
Note 21: Interview, Washington, July 1987. Back.
Note 22: See for example the endorsement of PNDC policies and promise of continuity by one presidential candidate, Jonathan Frimpong-Ansah, in West Africa, May 25-31, 1992, p. 876. Back.
Note 23: A view expressed by two IMF staff interviewed. Back.
Note 24: Young, Zambia, p. x. See also Colclough, "Zambian" and Gulhati, Impasse, for the view that Zambian politics could not tolerate the speed of adjustment. The view that there is nothing in states or societies which necessarily prevents reform is supported by Henry S. Gersovitz and Mark Bienen, "Economic Stabilization, Conditionality and Political Stability," International Organization 39, 4 (Autumn 1985): 729-754; J. D. Fearon, "International Financial Institutions and Economic Policy Reform," Journal of Modern African Studies 26, 1 (March 1988): 113-137; Ravi Gulhati, The Political Economy of Reform in Sub-Saharan Africa: Report of the Workshops on the Political Economy of Structural Adjustment and the Sustainability of Reform (Economic Development Institute Policy Seminar Report No.8, World Bank, Washington, D.C., 1988); Stephan Haggard, "The Politics of Adjustment: Lessons from the IMF's Extended Fund Facility," International Organization 39, 3 (Summer 1985): 505-534; and Joan M. Nelson, The Politics of Economic Adjustment in Developing Nations (Princeton: Princeton University Press, 1989). Back.
Note 25: See Africa Analysis, August 8, 1986, p. 5. Back.
Note 26: See Kydd, "Coffee", pp. 242-243; and Gulhati, Impasse. It also went against the conventional wisdom on sequencing-see Sebastian Edwards, "Sequencing Economic Liberalization in Developing Countries," Finance and Development (March 1987): 26-29; Michael Mussa, "Macroeconomic Policy and Trade Liberalization: Some Guidelines," World Bank Research Observer 2, 1 (January 1987): 61-77; Martin Wolf, "Timing and Sequencing of Trade Liberalisation in Developing Countries," Asian Development Review 4, 2 (1986): 1-24. Back.
Note 27: See Africa Analysis, on Zambia; Green, "Ghana," p. 8 on Ghana; and Gulhati, Political, p. 24 and Impasse, pp. 35 and 44-45. Other sources are confidential IMF and World Bank documents. Back.
Note 28: It is also possible to argue that certain measures were inappropriate to the political context of Ghana and Zambia: such measures have been discussed under political support. Back.
Note 29: Confidential IMF documents, including the PFP document for December 9, 1986, supported by interviews, and by UNECA Secretary-General Adedeji (BBC World Service, The World Today, June 24, 1988) and UNICEF's Richard Jolly, IDS Bulletin, January 1988, p. 75. Back.
Note 30: For more details of labor shortages, see Colclough, The Labour Market, pp. 32-33. Back.
Note 31: The role of weather is acknowledged by Chand and van Til, Ghana, p. 34, as well as Green, Stabilization, and Loxley, Ghana. Back.
Note 32: See Chand and van Til, "Ghana," p. 34, and Green, "Ghana." Back.
Note 33: On Ghana, see Green, Stabilization and Loxley, Ghana; on Zambia, Colclough, "Zambian." Back.
Note 34: Interview, Lusaka, March 1987. See also Kydd, "Coffee," p. 243 and Young, Zambia, p. ix, for the view that Zambia's 1985-86 program was too comprehensive. Back.
Note 35: Sources for administrative problems are IMF documents. On auctions, see also Charles Harvey, "Non-Marginal Price Changes: Conditions for the Success of Floating Exchange Rate Systems in Sub-Saharan Africa," IDS Bulletin 19, 1 (January 1988): 67-74; and Martin Sakala and Manenga Ndulo, "The Zambian Foreign Exchange Auction" (Lusaka: University of Zambia/Bank of Zambia, 1987, mimeo). Tony Killick, ed., The Quest for Economic Stabilisation: the IMF and the Third World (London: Overseas Development Institute and Gower Press, 1984), pp. 193-194, 211 and note 38 show administrative problems of fiscal and monetary measures. Justin Zulu and Saleh Nsouli, Adjustment Programs in Africa (IMF Occasional Paper No. 34, Washington, April 1985), pp. 14-16, show impracticable fiscal and debt management reforms. Back.
Note 36: This description of the auctions is based on Loxley, Ghana, pp. 12-13; Young, Zambia, pp. 30-33; Sakala, Zambian; 1986-89 issues of Africa Economic Digest and West Africa; and confidential IMF documents. Back.
Note 37: For more examples of fine-tuning, see Martin, Negotiating Adjustment. Back.
Note 38: See sources in note 26, and John Clark, Debt and Poverty: A Case Study of Zambia (Oxford: Oxfam, 1987); and Joan M. Nelson, "Poverty, Equity and the Politics of Adjustment" (paper to the International Studies Association annual conference, London, March 1989, mimeo). This account is confirmed by confidential IMF documents, including the IMF PFP document for December 9, 1986. Back.
Note 39: See also Gulhati, Impasse, p. 49. Back.
Note 40: Much of this section is based on Matthew Martin and Mistry Percy, eds., How Much Aid Does Africa Need? (forthcoming, 1993). Back.
Note 41: See also Young, Zambia, p. 2. These figures are from confidential 1987-88 IMF and World Bank documents. Back.
Note 42: See also Gulhati, Impasse, p. 49; and Edward K. Jaycox, et al., "The Nature of the Debt Problem in Eastern and Southern Africa," in Carol Lancaster, and John Williamson, eds, African Debt and Financing (Washington, D.C., Institute for International Economics, 1986). Back.
Note 43: Details on bridges and refinancings are from confidential interviews and documents. Joe Abbey has correctly shown that the net inflows to Ghana were modest compared to export earnings or GDP: the intention here is rather to point out that there were net inflows, while in Zambia there were not; see Joseph Abbey, "On Promoting Successful Adjustment: Some Lessons from Ghana," The 1989 Pere Jacobsson Lecture, International Monetary Fund, 1989. Back.
Note 44: Figures for debt service are author's calculations, based on confidential World Bank and IMF documents. Back.
Note 45: Figures calculated from balance of payments data given in confidential IMF 1987-88 documents. Back.
Note 46: Ghanaian figures from Loxley, Ghana, p. 24 and Africa Economic Digest, January 30, 1989; Zambian from Colclough, Zambian, and confidential 1988 IMF documents. Back.
Note 47: These figures and explanations are from confidential 1987-89 IMF documents. See also Gulhati, Impasse, p. 49, for discussion of incorrect projections. Back.
Note 48: See Jaycox, "Nature," pp. 60-62. Back.
Note 49: For a comprehensive analysis of balance of payments projections, and of the relationship between external finance and adjustment in Africa, see Martin and Mistry, How Much Aid. Back.
Note 50: Abdalla of the IMF admitted IMF "serious miscalculation" in June 1987 (Times of Zambia, June 19, 1987). Green, Stabilization, p. 53, comments on cocoa projections. Back.
Note 51: Quotation from Gulhati, Impasse, p. 50. World Bank Vice President for Africa, Edward Jaycox, expressed similar views in Africa Report, November-December 1987. Callaghy, "Lost"; Chand and Van Til, "Ghana"; Green, Stabilization; Loxley, Ghana; and Young, Zambia, also see net financial flows as a key factor in program success. Back.
Note 52: It is impossible to do full justice to these faults in a brief article: see Martin, No Winners (chapter 2 on the IMF, chapters 3-6 on debt rescheduling negotiations). Back.
Note 53: Because details of IMF and debt rescheduling negotiations are confidential, sources for this section are confidential documents and interviews unless otherwise stated. Back.
Note 54: This faulty data caused frequent revisions of data in later IMF documents, and severe disputes over data during negotiation. On external financing, see below. Back.
Note 55: Former Bank of Zambia Governor Phiri, in Gerald K. Helleiner, ed, The IMF and Africa (IMF, 1986), p. 95, confirmed by documents and interviews for Ghana and Zambia. Back.
Note 56: Terms used in interviews with three IMF staff. Callaghy, "Lost," says some Ghanaian officials gained the impression that some IMF staff were acting tough to try to get promotion: this was also suggested in my interviews with Ghanaian and Zambian officials, but cannot be verified. Other reasons suggested were unquestioning belief in IMF conditions, and overwork. Back.
Note 57: This view is confirmed by Callaghy, "Lost"; Green, Stabilization; and Young, Zambia. Back.
Note 58: This view was expressed in interviews with both sides, and is also expressed by Callaghy, "Lost." Back.
Note 59: On simplistic IMF views, the sources are confidential IMF documents and interviews. See also Africa Confidential, December 12, 1984, p. 3 on Ghana; on the failure to help "proponents," see below on implementation. Back.
Note 60: Sources are confidential IMF documents and interviews. Back.
Note 61: This paragraph is based on Fund and Bank documents and interviews. Back.
Note 62: Quotation from an interview with an IMF official, confirmed by interviews with all sides. Back.
Note 63: Evidence for this paragraph is confidential IMF and World Bank documents, and interviews with all sides. Back.
Note 64: For this attitude, see Martin, No Winners, chapters 3-6. Back.
Note 65: For much more detail, see Martin and Mistry, How Much Aid, chapters 4 and 5. Back.
Note 66: This account is based on interviews with IMF and World Bank staff, and is confirmed by documents which show failures of communication, especially in March-April 1986 and February-April 1987. Back.
Note 67: Green, Stabilization, p. 52, confirmed in interviews with Ghanaian and IFI officials. Back.
Note 68: These reforms and the reasons behind them are discussed briefly in Tony Killick and Matthew Martin, African Debt: the Search for Solutions (UN Africa Recovery Program Briefing Paper No.1, New York, June 1989), reprinted in Africa Recovery, August 1989; and at greater length in Martin, No Winners, chapter 7. Back.
Note 69: Therefore it is ironic that many of those arguing for reform in 1986-87 within the Fund, Bank and creditor governments used Zambia as their example. Back.
Note 70: On the design faults, Andrew Norton, "Ghana Social Profile" (London: Overseas Development Administration, 1988, mimeo); on the slow implementation, African Economic Digest, January 30, 1989. Back.
Note 71: Toronto terms offered creditors three options: cancel 33% of debt service and reschedule the rest with 8-year grace and 14-year maturity; extend grace to 14 years and maturity to 25 years; or cut interest rates by 3.5 percentage points or 50% and reschedule over 8 and 14 years. Enhanced Toronto terms retain the second option, but enhance the first and third options to the equivalent of 50% cancellation or 50% reduction of service payments due in net present value terms. Figures on the effects of Zambia's agreements are from IMF and World Bank calculations. Back.
Note 72: For more details and explanation of these reforms, see Martin, No Winners, chapter 8; and Martin and Mistry, How Much, chapter 10. Back.
Note 73: How to reduce import and aid dependence by changing import liberalization and aid provision is the subject of a project being conducted by Percy Mistry, myself, and government officials from six African countries. Back.
Email this citation