The Impact of Multi-Party Politics

                                   in Sub-Saharan Africa



                                                               Nicolas van de Walle

                                                            Michigan State University





                                                                September 25, 2000




















Prepared for delivery at the Norwegian Association for Development Research Annual Conference, “The State Under Pressure”, 5-6 October, 2000, Bergen, Norway.

                              The Impact of Multi-Party Politics in Sub-Saharan Africa[1]


                                                              Nicolas van de Walle



The spread of the "third wave" of democratization to Africa in the early 1990s represented the  most significant political change in the continent since the independence period three decades before.  Through out the continent, significant political liberalization resulted in the emergence of a free press, opposition parties, independent unions and a multitude of civic organizations autonomous from the state.  In 29 out of 47 states in the region, the first multi-party elections in over a generation were convened between 1990 and 1994.[1]  In a smaller set of countries, elections were fully free and fair and resulted in the defeat and exit from power of the erstwhile authoritarian head of state.  By the end of the decade, only a small minority of states were not officially multi-party electoral democracies, even if the practice of democratic politics was often far from exemplary.


Has the new, more open, political climate undermined economic reform in Africa?  Has there been a negative (or positive) correlation between economic and political reform in the 1990s?  What has been the impact of democratization on the patterns that have characterized African politics in the past?  It is important to understand the impact of democratization on African economies because it can provide real insights into the dynamics of change in Africa’s political economy.


At the beginning of the 1990s,  the dominant point of view among observers of the African scene appeared one of optimism about the region's politics, but pessimism about its economic prospects.  Rather curiously, today that assessment seems reversed. During the period 1994-1997, as documented above, Africa witnessed its best economic results in over a decade.  The IFIs have suddenly become cautiously optimistic about the region, relieved to be able to argue that their policy package is finally bearing fruit.  The usually cautious Economist went so far as to declare that "Sub-Saharan Africa is in better shape than it has been for a generation", pointing out that its overall growth rate was 4.4 percent in 1996, "faster than for two decades".[2]


On the other hand, the ethnic violence that put an end to the democratic regime in Burundi, factional street fighting in the Congo's capital in 1993 and then again through out 1997, and highly flawed second multi-party elections in Mali, Zambia and Madagascar have led many observers to belittle the degree of political liberalization actually achieved in the past and the prospects for democratic consolidation.[3]  The countries which are increasingly viewed as stable and evolving in the right direction are regimes like Museveni's in Uganda, or Rawlings in Ghana, neither of which can really qualify as consolidating democracies.


So how can the relationship between democratization and economic policy making be characterized?[4]  When the wave of democratization reached Africa in the early 1990s, most academic observers were pessimistic about the ability of African countries to combine successfully the "dual transitions" of economic and political reform.  The most common view was that "all good things" did not go together and that one reform process would undermine the other.[5]  This pessimism appeared to be based, first, on the common assumption that democratization would increase societal pressures on governmental decision making.  In this view, Africanists were echoing the view that democratization in other regions of the world resulted in a sharp increase in political participation.  As Haggard and Kaufman summed up the view, "new democratic governments face exceptionally strong distributive pressures, both from groups reentering the political arena after long periods of repression and from established interests demanding reassurance"[6]  Observers were particularly worried about the impact of more open government on policy reform in Africa, where both processes were viewed as highly vulnerable to reversal.


Second, and relatedly, by redistributing power from the executive to the more participatory legislative branch of government, democratization was believed to weaken the autonomy of the executive to design and implement policy, which also undermined the prospects for macro-economic stabilization.[7]  In a trenchant essay, for example, Jeffries argued that democratic politics would further weaken the already tottering African state and "exacerbate... problems of corruption, wastefulness and short sighted economic policy formulation".[8]  Jeffries concluded his analysis by criticizing as naive donors such as the World Bank which believed that democratization would improve the prospects for structural adjustment programs.  Instead, he concluded, donors should first promote "the development of more efficient and capable government bureaucracies" (p. 34).


These assumptions suggested that the wave of democratization that hit Africa in the early 1990s represented a sharp historical discontinuity in the mode of governance through out the continent and its relationship with the outside world.  Today, with the benefit of a couple years of hindsight, it is possible to make a first tentative assessment of the economic consequences of democratization in Africa, and examine these assumptions more carefully.  In fact, as our data will reveal, there appears to be little significant difference between the democracies and the non-democracies in Sub Saharan Africa (SSA) in terms of their economic performance over the last couple years, in a context of modest economic improvement for the region as a whole.  The more interesting issue is what this tells us about economic decision making in Africa, and the evolving nature of continental political economy.  The democratization wave that swept the region constituted perhaps the most fundamental political change since independence and, therefore, it held out the possibility that different patterns of economic decision-making would emerge. How would the corrupt and largely incompetent modes of governance described in these pages be affected by the move to multi-party electoral politics?   Examining the impact of democratization thus allows us to problematize the issue of change within the contemporary African political scene and distinguish the more permanent structures within African politics from the merely epiphenomenal.  


The analysis starts in the next section with a review of the salient dimensions of political reform in Africa in the early 1990s.   I review some of the underlying reasons for the wave of democratization that occurred in the region, and then argue the extent of political reform was significant in most of the countries in the region, even if it usually fell well short of liberal democracy.  A third section attempts a rigorous analysis of the impact of democratization by making both longitudinal and cross-sectional comparisons of economic performance.  The essay then turns its attention to explaining why democratization appears to have so little effect on economic performance.  I show that political reform in the 1990s did little to upset the major institutions of African politics, at least in the short term. 



                                          Africa’s Democratization Wave, 1989-1995

This essay’s first task is to characterize with some precision the degree of actual political change in Africa in the early 1990s.  I start by briefly examining the underlying causes of democratization and then assess the political change that occurred.



The Causes of Democratization

Democratization is always a complex process involving the interaction of agency and structural factors, domestic and international ones and economic and non-economic ones.  The wave of democratization that struck Africa in the early 1990s was no exception.  Rather than attempt a comprehensive analysis,[9] I wish here to link the events of the early 1990s to the neopatrimonial politics that have characterized African politics during the course of the last twenty years.


The term neopatrimonialism captures the thesis that most African states are hybrid regimes, in which patrimonial practices coexist with modern bureaucracy.  Outwardly the state has all the trappings of a Weberian rational-legal system, with a clear distinction between the public and the private realm, with written laws and a constitutional order.  However, this official order is often subverted by a patrimonial logic, in which office holders almost systematically appropriate public resources for their own uses and political authority is largely based on clientelist practices, including patronage, various forms of rent-seeking and prebendalism.  These regimes are highly presidential, in the sense that power is centralized around a single individual, with ultimate control over most clientelist networks.  The president personally exerts discretionary power over a big share of the state’s resources.  This is true, not only in the smallest, most backward states of the region, but also in some of the bigger, allegedly more institutionalized states, like Côte d’Ivoire or Nigeria.[10]

Neopatrimonialism undermines economic policy reform in contemporary Africa in at least two ways.  First, because clientelism is based on the extensive use of state resources for political purposes, clientelist regimes almost inevitably produce highly interventionist economic policies.  Economic liberalization  is viewed as anathema to regimes which rely on the politically mediated distribution of access to state resources.  Clientelism is viewed as critical in countries with little sense of nationhood and a tendency towards multiple ethnic and regional divisions, and in which few regimes can count on either a successful economy or electoral mandates for their legitimacy  As Sandbrook has asked,[11] “what will hold these societies together when the rulers have little in the way of patronage to distribute?”

Secondly, neopatrimonialism results in a systematic fiscal crisis.  As Callaghy has argued about the Zaire of Mobutu, “finances is the Achilles heel” of these regimes.[12]   At the outset of the African crisis, the World Bank and the IMF typically accused African states of being “too big”, but that is not really the case: relative to the economies they govern, African states are roughly comparable in size to the states in other middle and low income regions.[13] True, patronage needs swelled the ranks of the civil service at a remarkable speed following independence, but salaries were also allowed to decline precipitously in real terms, so that the cost of the civil service was not unusually high.  In Tanzania, for example, salaries in the civil service were allowed to fall by an incredible 90 % in real terms during the first two decades of independence, even as the size of the civil service more than tripled over the same period.[14]  Instead, the real cause of the endemic fiscal crisis that has plagued most African states following independence has been on the revenue side. Despite extensive state intervention in the economy, cronyism and rent-seeking have siphoned off potential state revenues.  Taxes are not collected, exemptions granted, tariffs averted, licenses bribed away, parking fines pocketed.  As a result, revenues always lag behind expenditures.  It is often said that these regimes have low levels of extractive capacity, but the problem is not one of capacity so much as it is of the political logic of a system in which the authority of the state is diverted to enhance private power rather than the public domain.  In sum, the consistent problems of unmanageable fiscal and balance of payments crises since independence was an entirely logical and predictable outcome of the manner in which politics was conducted south of the Sahara. 


Based on this all too brief synopsis of African political dynamics, an equally brief stylized account of democratization can be formulated: by the late 1980s, many regimes in Africa were undergoing a legitimacy crisis brought about their dismal economic performance and worsening economic conditions.  Restive populations were increasingly willing to contest central state power, notably through the fledgling civic associations that had begun to emerge during the 1980s in response to state decline.  The first phase of the democratization wave that hit Africa in the early 1990s consisted of political protests against incumbent governments.  Often not specifically political when they began, these protests quickly escalated into demands for regime change, often as a result of clumsy government responses to the protesters.  African politics had long been punctuated by more or less spontaneous protest, but now governments found it harder to repress or accommodate the protesters.  On the one hand, with the end of the Cold War, the donors were increasing their emphasis on human rights, and were less tolerant of government repression.  On the other hand, the economic crisis and strong fiscal pressure on states made it harder to coopt or placate dissent with access to state resources.  Perhaps more important, the economic crisis was putting pressure on the state elite accommodation processes that had long served to maintain political stability.  With fewer resources at their disposal and an increasingly decrepit state apparatus that was difficult to control or command, leaders found it harder to sustain critical clientelist networks, with the result that the old political aristocracy was more likely to fractionalize.  A growing number of old style politicians were withholding their support from governments that had excluded them from power and its perquisites.  Although the political protests that spurred on the democratization movement were typically initiated by students, civil servants and an array of civic associations, it is this class of excluded politicians that typically emerged during the transition to take control of new governments following national conferences and multi-party elections.  Far from being political novices,  they were often long-standing members of the state elite, with long careers in past cabinets, but who had had a falling out with the head of state and had been consigned to the political wilderness at some point in the 1980s.[15]  Thus, in the Central African Republic, the first democratically elected president, Ange Patassé, had been prime minister under Bokassa, as had Trovoada in Sao Tome and Lissouba in Congo.  Most of the new heads of state that emerged in the 1990s had at the very least previous cabinet experience.  Even so-called outsiders often had extensive links with the authoritarian state.  For example,  in Zambia, Fred Chiluba, a long time unionist did not have cabinet experience, but much of the rest of the MMD leadership did.[16]


In sum, and this is the first salient point regarding political reform in Africa in the 1990s,  democratization can be understood to have resulted at least in part from a crisis in the post-colonial neopatrimonial order.  Democratization often proceeded the furthest in the weakest states – for example in Benin or the Central African Republic, in which this state crisis was most pronounced.  On the other hand, in many of the more developed states – in Kenya, say or Côte d’Ivoire – the ability of state leaders to centralize power and maintain control of state resources allowed state leaders to survive the initial set of street protests and adjust to new political realities, albeit often not without having conceded some political reforms to the democratic forces. 


The second point concerns the role of the donors.  Given the importance of aid to most African economies and to the functioning of the state, one might expect that donors played a key role in shaping the democratization episodes of the early 1990s.  Yet, the evidence does not support the argument that donor pressures for political reform had a significant impact on the emergence of political protests.  Protests were not correlated either with the degree of dependence on aid, the level of international debt, or with the imposition of political conditionality by the donors.[17]  Interestingly, however, the emergence of protest was positively correlated with the number of adjustment and stabilization loans signed with the IFIs in the 1980s.[18]  The more loans a country had agreed to with the Washington institutions, the more likely there was political protest at the end of the decade.  Since there is no correlation between the number of loans signed and the amount of policy reform actually implemented, this variable can not be viewed as a proxy for policy reform or the hardship endured by the population due to structural adjustment.  Bratton and van de Walle found no statistical evidence to link the rise of political protest to such indicators of economic reform as changes in the size of the fiscal deficit or of inflation.[19]  Instead, the most plausible interpretation is that a high number of loans weakened central states by disrupting central decision-making, creating the impression that key decisions were made by foreign bureaucrats and thus undermining the little legitimacy they still possessed.


As democratization moved from protest to an intense struggle over basic political rules and debate over the nature of the regime to emerge, the role of the donors changed.  It is clear that incumbents with greater access to economic resources were more likely to negotiate more effectively with surging democratic forces.  Here the willingness of donors to support incumbents often played a key if unwitting role.  In both Benin and Zambia, the IFI decision to withdraw support from incumbents in the middle of the political transition may well have sealed their fates.[20]  On the other hand, generous French support to Biya in Cameroon and to Houphouët-Boigny in Côte d’Ivoire probably played a key role in the survival of those regimes.[21]   In Cameroon, the decision by the IMF and French government to extend new adjustment credits to the country in 1991, right at a critical period of the transition, was enormously helpful to Biya, allowing him for example to keep salary arrears down to manageable levels.  Though much more critical of the regime and eventually to close its aid mission in Yaounde in 1993, even the United States had forgiven US $73.4 million of debt as late as 1990.[22]   Similarly, the decline of aid to Kenya, which can be linked to the weight of growing political conditionality from the Western Donors, occurred only after 1993, after the democratization movement had crested.  From 1988 to 1993, the period during which the regime was under the most internal threat, Kenya averaged well over a billion dollars in annual aid.[23]  I return to the role of the donors below.



The Rise of Illiberal Democracies  

The political change in the region from the late 1980s to the mid 1990s was profound and unprecedented.  Overall, in nominal terms, the Freedom House indices for political rights and civil liberties improved by an average of 1.00 and 1.28 between 1988 and 1994 and 1992 respectively.[24]  Part of this process of political reform was a widespread movement towards multi-party electoral politics.  By the second half of the 1990s, multi-party politics with more or less regular elections had become the norm in Africa. Only nine countries had held competitive, multi-party elections in the period 1985-1989.  Between 1990 and 1998, in comparison, some 70 legislative elections involving at least two parties were convened in 42 of the region’s 48 countries.  All but two of these elections resulted in legislatures with more than a single party.  In addition, there were over 60 presidential elections with more than one candidate during this time.  Only six states in the region (Eritrea, Rwanda, Sudan, Somalia, Swaziland, Uganda, Congo/Zaire) did not convene multi-party elections.  Moreover, a process of institutionalizing regular elections appears to have begun: by the end of 1998, 26 countries had convened second elections, usually on schedule, i.e. at the end of the constitutionally fixed term of officeholders elected during the first elections.  This routinization of elections contrasts also sharply with previous periods, in which competitively elected legislatures more often than not were displaced by military coups before they finished their term.


Thus, multi-party systems have emerged all over the continent.  Much less clear, however, is the extent to which these countries are progressing toward democratic consolidation, by which democratic norms and values are institutionalized and routinized by the political system.  Instead, the exercise of democratic politics is often highly imperfect in these countries.  Several prominent observers have recently suggested that many of Africa’s new democracies may be described as increasingly “illiberal”.[25]  Regular, competitive multi-party elections are held, thereby qualifying the country as an “electoral” democracy, but the day-to-day practices of the state are marked by abuses.  Political freedoms and civil rights may be formally recognized but are imperfectly observed in practice, particularly in between electoral exercises when they are more likely to be flouted.  Human rights abuses are not uncommon, even if the worst abuses are rarer than in the authoritarian past.  A nominally free press is harassed in myriad ways, and the government retains a radio monopoly.  Certain groups, notably key members of the executive branch and the military may, in effect, be above the law.  The judiciary is officially independent, but it is poorly trained, overworked and easily compromised.


The combination of chronic “illiberalism” with regular, competitive multi-party elections lends itself to considerable cynicism about the latter.  Some observers have described as the “fallacy of electoralism” the notion that these elections constitute meaningful political exercises in giving citizens meaningful choices over  the distribution of power and resources.[26]  In a thoughtful essay, Richard Joseph has gone so far as to call these systems “virtual democracies,” in which largely meaningless elections are put on strictly for the sake of international “presentability”.[27]


The illiberal and virtual democracy theses advance two propositions that can be carefully examined.  First, just how democratic are these regimes?  Table 1 demonstrates the divergent paths which Africa’s multi-party regimes have taken.  African countries which conducted at least one multi-party election in the 1990s may be distinguished according to whether Freedom House categorizes them as “Free,” “Partly Free,” and “Not Free.”[28]  Clearly, knowing a system is multi-party says very little about the degree of freedom allowed in the regime





                                                                Table 1 about here




The combination of competitive elections and a relatively poor level of political and civil rights is striking and seems to reinforce the thesis of  illiberal democracy.  By this definition, 30 of Africa’s multi-party systems are illiberal.  Indeed, the 12 multi-party systems in the “Not Free” category may be more accurately termed “pseudo democracies”[29], so egregious is the gap between democratic ideals and current practice.  Hence, while most of Africa are having multi-party elections, there is a wide range in the actual practices surrounding these elections.  Africa offers a striking contrast with the world’s older, multi-party systems in the West, which are all classified by Freedom House as “Free.” Nonetheless, Table 1 also suggests that Africa exhibits a wide range of democratic performances, and that perhaps as many as a quarter of these regimes are in the process of consolidating democratic practices.   Moreover, overall, even Africa’s pseudo democracies typically did undergo some degree of political liberalization in the early 1990s.


Second, are the gains of Africa’s democratic transitions being eroded?  Rather than undergoing the consolidation of democratic practices, according to these skeptics, Africa is heading back to the days of uncompetitive elections, banned opposition, and a shackled press.  In fact, however, the evidence tends to suggest otherwise.  Examining Freedom House’s annual indices for political rights and civil liberties from the period of initial democratization (1989-1994) to present (1998-1999) does not suggest a net worsening of the quality of governance in the region since the peak of the democratization wave.  Between 1993 and 1998, political rights actually improved in 12 countries, stayed the same in 20 and worsened in 15.  The “third wave” may be over in Africa[30] but there is no reason to believe that all the recent gains are being eroded.  Rather, the region’s countries are engaged on a variety of distinct paths, only some of which include a marked decline in the quality of democracy. 


Although much of Africa has multi-party systems, those in which the incumbents were actually ousted are far fewer.  Table 1 identifies the countries in which incumbents were actually ousted as a result of the democratic transition and its founding election, as well as the countries in which regular multi-party elections anteceded the 1990s.  This information paints a somewhat more nuanced picture of the continent’s situation, for it shows quite clearly that the majority of the “illiberal” democracies are countries in which the transition to multi-party rule proved to be seriously flawed.  Most never really had a transition, and the turn to multi-party competition amounted to little more than an erstwhile authoritarian ruler donning the garb of democracy and tolerating regular elections as a successful strategy of holding onto power.  Convening regular elections brings with it a modicum of international respectability and the resulting foreign aid, and does not threaten these leaders.  These seem to confirm the “presentability” thesis.  The countries in which the democratic transition included the ouster of the incumbent, on the other hand have on the whole performed better.  Only two - Niger and Congo - of the thirteen are in the “Not Free” category, both as a result of military coups that overturned most of the gains of the transition.  The other eleven countries which transited have seemingly been able to maintain most of the democratic gains made during the democratic transition.


The predominance of countries in the “Not Free” and “Partly Free” categories gives the impression that things have gotten worse and that the democratic transitions of the early 1990s have been betrayed.  In fact, a majority of the countries which underwent real transitions have sustained the progress made, while even the most illiberal multi-party systems are most likely freer today than they were before the democratization wave.



                                          3. The Economic Impact of Democratization

This section reviews the actual economic record of African states following democratization at the beginning of the 1990s.  As stated above, African economies enjoyed a substantial improvement in growth rates in the 1990s.  Thus, overall, Africa's economic output rose every year between 1994-1997, with average growth rates over 5 % in both 1996 and 1997 respectively, and an estimated averaged growth rate of 3.4 % for 1998.[31]  This a short time period, but given the widespread prediction that democratization would result in a worse economic performance, the obvious question to ask is whether or not political reform in the early 1990s can claim credit for this improved performance.



Comparing Economic Performance

To examine the impact of democratization on economic performance, I create the following sub-categories of countries.  First, five countries, Botswana, Gambia, Senegal, Mauritius, and Zimbabwe, were already multi-party democracies in 1989.  These states were far from perfect democracies.  The incumbent has never been defeated in Zimbabwe, for example, while Diouf’s defeat in 1999 was the first such defeat in Senegal.  Except arguably for Mauritius, all of these governments have demonstrated authoritarian tendencies on multiple occasions.  But all of these states have had significantly greater pluralism than the remaining states in the region over an extended period, with at least two decades of relatively stable multi-party electoral politics.  I thus create a category called “old democracies” for these states. 


Of the remaining 43 countries in the sample, only Sudan and Liberia did not undergo significant political reform in the early 1990s.  Of these, I exclude eleven in which past civil war and/or current political strife would have precluded normal economic policy making.  By this criteria, Liberia, Sudan, Angola, Chad, Ethiopia, Eritrea, Rwanda, Sierra Leone, Somalia, Mozambique,  and Zaire are excluded.[32]  Several of these states achieved good economic performance in the 1990s, but this is clearly far more related to the return of peace than it is to regime characteristics.  Thus, Mozambique has enjoyed the highest growth rate in the region in the 1990s, but remains considerably poorer than it was before its devastating civil war.  Following a similar logic, I exclude Burundi, a country whose democratic transition was too quickly reversed by a military coup in October 1993 and plunged into civil strife, to allow a meaningful before and after comparison.


This leaves 31 countries in the sample.  Of these, 29 countries in Sub-Saharan Africa convened founding elections between November 11, 1989 and December 31, 1994.  The first date is that of the Namibian elections and constitutes a handy starting point for this wave of democratization in Africa.  The latter date is somewhat more arbitrary, but conveniently allows us at least two points of economic data following the transition, and covers all of the free and fair elections in which an incumbent was ousted.  These were founding elections in the sense that the office of head of government was openly contested following a period during which multiparty political competition had been denied.  For many countries, this was the first election in over a decade, and for most the first multi-party election since the immediate post-independence period.  Of these 31 states, Namibia and South Africa are excluded because the presence of an economic boycott before their transition makes a "before and after" comparison misleading, while Djibouti is excluded given the absence of adequate data.  This leaves 28 states for which an analysis of the impact of democratization can be realized.


I undertake two types of comparisons.  First, the performance of the new democratic regimes is compared with those of countries which did not engage in a full transition.  Of the 29 countries, 13 could claim to have passed a relatively stringent test of democratization: a transition election had been widely viewed as "free and fair", typically including by international observers, and the loser had publicly accepted the results.[33]  Such a minimal, procedural definition of democracy helps to distinguish these states from the others in the data set, in which the degree of political liberalization had fallen short of a complete transition to democracy in some way, despite the convening of elections.  Comparing the economic performance of these two sets of countries should generate some insight into the economic impact of democratization.


As I have already suggested, It is unfortunately difficult to distinguish all the cases as clearly as one would like.  Most of the non-democratic states can legitimately claim significant political liberalization during this period.  Typically, they held multi-party elections in which an incumbent ruler more or less stage managed multi-party elections which he could not lose and has since then tolerated some degree of opposition, again so long as it is not threatening.  The more egregious cases pseudo democracies are easy to exclude from the New Democracies category.  For instance, no one claims Cameroon and Togo to be democracies, given the obvious abuses of executive power and the continuing limitations on basic civic and political rights, and despite some real liberalization during the early 1990s.  However, the Rawlings government in Ghana, say, can claim to have won two reasonably free and fair elections in 1992 and 1996.  Despite the government's occasional crackdowns and attempts at intimidation, a free press and opposition parties do exist.[34]


On the other hand, for instance, Zambia's democratic reputation has been tarnished by evidence of the government's continuing authoritarian proclivities and various abuses of its power.[35]  The point is that honest observers can differ about which states belong in which category.  Rather than rely entirely on judgments some will find questionable, the comparison between "New Democracies" and "Non Democracies" is complemented with a second kind of comparison, across time.  Regardless of the exact nature of the regime, it can be argued that all of these regimes have been significantly liberalized, with the onset of elections since 1989 to have had an impact on economic decision making.  I thus compare the economic performance of all the 29 states of SSA that are in my sample before and after 1991, to attempt to determine the economic impact of this new political climate.


Finally, I create a smaller category of “liberal democracies”, which includes the nine states which Freedom House categorized as liberal in 1998: Benin, Botswana, Cape Verde, Malawi, Mali, Mauritius, Namibia, Sao Tome and South Africa.  These include 7 states in which incumbents were ousted in elections in the 1990s and brings back into the analysis states which I otherwise excluded from the analysis (for example South Africa).  The before and after comparison makes no sense for this category, but their performance in the 1990s can be usefully compared with the other categories.



Comparing the samples:  I first look at the economic performance of African economies.  The two indicators used are real GDP growth and inflation, two standard measures of economic performance.  The economic record for 1986-1988,  the three last years before the onset of democratization on the continent is compared with the equivalent record for the first three years after the founding election of the emerging democracy.[36]  When data was not available for 1998, the data reported covers the last available year, almost invariably 1997.




                                                               Table 2 around here




The results are provided in Table 2, for the four categories of regimes.  The table provides estimates comparing the slightly longer periods 1986-1991 and 1991-1995, and 1995-1998 to lessen the impact of single year data points.  Since these are overall growth rates that do not take into account population growth, typically well above 2 percent annually, our estimates suggest that all three categories are barely growing at all in per capita terms.  Interestingly, these numbers suggest that the current official optimism about the region appears to be based largely on the growth record in the war torn countries which I have excluded from my analysis.  In countries not previously undermined by civil war, the mid 1990s does not appear to have brought on substantially better economic performance.


Otherwise, no single pattern emerges from these estimates.  The numbers suggest the Old and Non democracies have both slightly outperformed the New democracies, both before and after the transition period, and the partly overlapping category of Liberal democracies has performed slightly above the regional average.  Unfortunately, this differential is too small to be significant in analytical or statistical terms, given the high degree of variance within each category of states, and the large number of countries that have been removed from the data set before these averages were calculated.   From this small number of observations, it is not possible to say either that there were significant differences in growth performance before and after the transition period, or that the sample of New democracies performed better or worse than the Non democracies. The Liberal democracies may well have demonstrated a superior record, but more data points are needed before such a conclusion can be confirmed.


Much the same conclusion emerges from a comparison of inflation rates, also reported in Table 2, although several factors clearly unrelated to regime condition the results.  The New democracies averaged a 19.3 % annual increases in the consumer price index for the three years before transitions began (1986-88) and a 31 percent for the three years following their founding election; the set of comparable non-democracies achieved rates of 24.4 % and 18.3 % for the same two periods.  The jump in inflation rates in the new democracies is entirely due to the effect of the January 1994 devaluation of the CFA franc in the ex-French colonies, which spurred a sharp rise in inflation in 1994-95.  Given the unwillingness to devalue and give up the fixed parity with the French Franc, during the early 1990s these countries embarked on tough internal deflation programs.  As a result, they had extremely low inflation, which averaged only 5.1 % for the three years preceding their transitions, for the countries for which data exists.  In comparison, the non-CFA countries in the sample of comparable states (again excluding Congo/Zaire) averaged 26.2%.  France and the countries of the Zone finally agreed to a devaluation in January 1994 and as a result, inflation shot up in many countries of the zone, to an average of 16.3%  for the three years after transitions.  Note that this was still less than the mean level of inflation in the rest of Africa,  where it averaged 23.8%.  The sub-category of CFA country democracies go from an average of 3.7% to 19.9% during this same period.  In sum, increases within the Zone have much more to do with the 1994 devaluation and its aftermath than with political dynamics.


The particularities of the Franc Zone almost certainly have also had an impact on the growth data reported above.  Overvaluation of the currency and attempts to force depreciation and avoid devaluation progressively dampened economic dynamism across the zone after 1988.  This effort was particularly noticeable in the richer countries along West Africa's coast (ie, Côte d'Ivoire, Senegal, Cameroon, Benin), where overvaluation was the highest.  Its negative impact on growth rates cut across political regimes, as did the salutary effect of the devaluation in January 1994.  Thus, through out the zone, 1994 and 1995 witness a sharp increase of agricultural export led growth.[37]  One can be skeptical that this growth will be sustained, but the point is that its effects are confounding for the analysis.  Given problems of reliability and comparability,[38] in sum, the only safe interpretation of this data is that it provides no evidence that increases in political competition and participation in the 1990s can be associated with lower economic growth or inflation.


Finally, Table 2 also reports on the evolution of the external terms of trade during this period.  The data suggests a significantly greater average deterioration in the terms of trade for the new democracies since the late 1980s.  This places the economic performance of the New democracies in a more favorable light, even if the quality of the data again urge cautious conclusions.  For example, during the first three years following its transition, Zambia's terms of trade underwent a 40 percent decline, as copper prices tumbled.[39]


In one sense at least, these findings are surprising.  Because the actual process of regime transition is highly disruptive, one would expect worst economic performance, ceteris paribus, from the countries having undergone significant political reform, notably with incumbents losing power.[40]  First, real transitions generate considerably uncertainty, and presumably discourage investors, either the local kind who seek to protect their capital by exporting it, or the foreign kind, which defers investments until the political picture clears up.  Second, incumbent governments in the process of losing power are unlikely to be cautious stewards of the macro-economy.  Instead, they are likely to resort to desperate splurges of public expenditure to remain in power, or, once all appears lost, to asset stripping.   Both can exert a cost on the economy with which the incoming government will be saddled.  Finally, during the transition period itself, there is likely to be an inter regnum of some length during which no government is effectively in power.  Again, it is easy to imagine such a situation resulting in sub-optimal management of the economy.  


Why does the data presented not clearly identify an economic cost to the actual transition?  It may be that transition effects are swamped by the exogenous factors just identified,  such as the CFA devaluation and weather variability.   Or it may reflect the superficial nature of democratization, which was not viewed by investors as substantially changing the economic environment in the country.



                                             4. Comparing Government Performance

Economic growth and inflation are outcome variables, and are influenced by a lot of factors other than the actions of governments.  The analysis of the last section suggests that various exogenous factors have weighed heavily on the region's economies.  In addition, it may be that there is a lag of several years before policy reform yields dividends in terms of improved performance.  To best judge the economic impact of democratization on Africa in the 1990s, it is thus appropriate to also examine the actual policies pursued by governments.  One standard measure of government policy performance is the size of the fiscal deficit, as it is widely accepted that large deficits precipitated the crisis and must be reduced for sustained growth to reemerge.  Table 3 compares the evolution of the size of fiscal deficits before and after the transition period.  The data includes estimates of the deficit both with and without the contribution of grants, most of which consists of external resource flows.  The quality and cross-national comparability of this data is questionable.[41]   It does suggest that democratization has had little impact on the size of the fiscal deficit.


Interestingly, these data appear to show substantially higher average levels of deficits among the new democracies both before and after the transitions, as well as for the category of liberal democracies in the 1990s.  These differentials turn out not to be statistically significant, however, given large variations within each group of states.  Analysis reveals that the observed difference is largely due to the presence of several highly aid dependent small states in the sample.  In particular, the extraordinarily high deficits in Sao Tome, Cape Verde and Guinea Bissau, in each case averaging between a third to half of GDP during this period, and entirely sustained by foreign aid.  I discuss the political impact of aid below.  The important point to retain from this analysis is that the data emphasize the continuing importance of international public flows to sustain what have remained extremely imbalanced fiscal systems.




                                                                Table 3 about here





Table 3 also reports data on the share of government consumption in GDP, and the share of GDP devoted to investment, two other standard measures of governmental policy performance.  Since a standard objective of reform programs has been the slimming down of the state and the reorientation of expenditures towards investment, it is a good measure of the reform progress made.  Again, there is little evidence of a distinct break between the time periods around the transition.  The share of both government consumption and of investment in GDP appears systematically higher in the New Democracies, but once again, this differential is not significant and largely disappears when Sao Tome, Cape Verde and Guinea Bissau are excluded from the data set.[42]


The evidence presented so far suggests that the democratization of African politics in the early 1990s has not had the kind of dramatic effect on economic performance and government policy making predicted by many observers in the early 1990s.  If there is any trend in the data it is towards economic improvement, although it is faint so far, and does not appear solely related to regime type.  Other, largely exogenous, factors, such as the evolution in the terms of trade and the French imposed devaluation of the CFA Franc appear to have exerted a more powerful influence on economic conditions.  The remainder of this paper will interpret this continuity.



                                           5. The Political Impact of Democratization

In sum, the significant political liberalization that occurred in the 1990s has not so far had a significant impact on economic performance.  The most one can say so far is that the new democracies may be in the process of outperforming the non democracies, but that it remains much too early to reach definite conclusions about the impact of political reform.  To understand this situation, I examine the assumptions that typically undergird the view that democratization would negatively affect economic reform and performance.  I then emphasize the institutional continuities which appear to have prevailed across the changes in regime.



A Participatory Explosion? 

The prediction that democratization was going to introduce to Africa a participatory explosion, in which populist and interest group pressures would compel governments to increase their expenditures to maintain popularity has proven to be fanciful.  There has always been something of a protest cycle in African countries, in which leaders responded to periodic protests and political crises by resorting to patronage and subsidies benefitting key constituents.[43]  But the macro-economic significance of these practices was due less to the power of interest groups than to the fragile legitimacy of leaders and the weak mobilizational capacities of their single parties.  In the 1990s, there have been several episodes of governments seeking to retain waning electoral support by temporarily opening the public purse, but these cases do not validate the political business cycle theories of western democracies, in which macro-economic stability is threatened when government respond to a mobilized electorate right before an election.[44]  The more egregious cases have been in the liberalized autocracies like Ghana (the election in 1992) or Kenya (1992), rather than in the second elections for the new democracies like Zambia (1996), Benin (1995) or Mali (1997).  It might also be noted that the early data  regarding second and third elections in Africa following democratization have been characterized by declining participation rates, relative to the earlier founding elections.[45]  Table 4 below shows that the proportion of registered voters bothering to vote has declined to below two thirds.  Given low levels of voter registration, it has not been unusual for less than a third of the adult population to bother voting in subsequent elections. 


As I have argued at greater length elsewhere,[46] African political systems have been characterized by their relatively low levels of political participation.  It was fanciful to expect democratization to change this state of affairs quickly.  Compared to the more mature democracies of the west, civil society has long been poorly organized and non-representative.  True, the political salience of associational life varies across countries in the region, and its variation will increase over time; nonetheless, non-governmental actors which could channel and mobilize participation following the political liberalization of the early 1990s were typically only recently formed, with shaky finances and small memberships.  Their clout remains often undermined by their fragmentation, notably along ethno-regional lines.[47]


A lively written press has emerged, but its circulation levels and outside surveys suggest it reaches no more than a small proportion of the urban population.  Overwhelmingly, radio remains the medium by which Africans receive their news and it typically remains safely in government hands. For instance, a survey of rural southern Cameroon in 1994 estimated that 0.2 percent of rural households read newspapers on a regular basis, while 44 percent listened to the radio.[48]  Accounts of the 1999 presidential election in Senegal emphasized the key role that newly legalized private radio stations had played in publicizing the message of the opposition and in monitoring polling stations on election day, sharply rising the cost to the government of any attempt to tamper with the results.[49]  But the privatization of the air waves remains limited to a minority of states in the region, as most governments have proven unwilling to let go of their monopoly over this strategic means of communication. 


Similarly, a small active opposition may conduct well publicized protests in the parliament or outside of the presidential palace, but it typically can not rely on the support of large segments of the population, particularly outside of the capital, and it does not mobilize large crowds with any regularity.  For the most part, in sum, the increase in participation compared to the ancien régime has been no more than modest.


 The claim is sometimes made that the democratic forces that took power in the early 1990s represented a new social coalition.  The emergence of such a coalition, representing different economic interests, might have had programmatic policy implications.  However, the language of social coalitions is in fact highly misleading for Africa, because interest groups and professional associations are poorly organized and comparatively weak there, and because the structure of national economies has tended to lessen the salience of class identities and cleavages.[50]  In some countries, nonetheless, it is true that private business had actively supported the democratization movement, and appeared ready to influence the emergent democratic governments.  They had come to be fed up with autocratic governments whose brand of desultory management, misguided policies and corruption created a climate in which it was difficult for business to remain profitable.  Business funding was critical to the success of the democracy movement in many countries.[51]  As a result, businessmen exerted considerable influence in the first cabinets following democratization in countries like Madagascar, Benin and Zambia. 


This led directly to progress on institutional reform issues of importance to private business, in particular privatization, notably of public utilities, whose inefficiencies had long hampered business competitiveness.  This influence must not be exaggerated: business rarely spoke with one voice, particularly in areas like trade reform on which it was divided.  Rent-seeking motivations did not disappear with democratization, which sometimes resulted in little more than the eviction of one set of crony businessmen in favor of another.  Governments could moreover placate business with favors, while seeking policy inspiration form other parts of the alliance that had brought them to power.


Interestingly, the move to multi-party politics did have the effect of changing the process of elite accommodation, within the state elite and this may, in time, have important political consequences.  During the 1970s and 1980s, leaders sought to build broad elite coalitions, involving all of the country’s ethno-regional groups.  Political alliances were broad, even if they were shallow, notably in social terms.  Indeed, the inclusiveness of these coalitions was one of their legitimating properties.[52]  Democratization has created a new dynamic, in the context of advanced economic austerity.  Now that state leaders can seek legitimacy from the ballot-box,  they are more likely to seek to build minimally winning coalitions that provide an electoral majority, but make no claim to inclusiveness.  In the region’s more democratic countries, characterized by more or less free and fair elections, this dynamic is promoted by the overwhelmingly ethno-regional nature of political parties and voting.[53]  Whereas the old single party had usually sought nation-wide support, the new government party in countries like Zambia, Malawi or Benin may prove content with the construction of a multi-ethnic coalition that will gain somewhere between half and two thirds of the votes.  In pseudo-democracies like the ones prevailing in Cameroon, Kenya or Côte d’Ivoire, this dynamic is emerging even more clearly: the composition of the cabinet and official pronouncements suggest that Paul Biya no longer seeks Bamileke support,[54] for example, while in Côte d’Ivoire the Bedie regime’s anti-northern rhetoric was striking in its stridency before the December 1999 coup toppled him,[55] while the Moi regime has made fewer efforts to court Kikuyu voters as the 1990s progressed.[56] These are dangerous strategies: there are inherent costs in excluding such key parts of the electorate, which may in the long run undermine national stability.  At the same time, for politicians operating on a short time horizon, they are tempting solutions to the need to  maintain patronage based political machines in an environment of growing resource scarcities.



A Weakened Executive? 

The second assumption was that democratization would  result in the weakening of executive authority and this worsened the prospects for sound economic policies.  During the transition, pro-democracy forces in a number of African countries had indeed voiced the ambition to weaken the institutional power of the presidency, in order to reduce the abuses of power that had characterized the ancien régime. As a result of these transition debates, checks on executive power were instituted in a number of countries. 


In Madagascar, Congo, CAR, Benin, and Niger, the executive did not secure a clear legislative majority and the national legislature used this new power to significantly check presidential action.  The result of  weak government parties lacking a clear majority following the transition was the near paralysis of legislative politics.[57]  In none of these countries did the same party lead a legislative majority by the end of 1998,  except in the CAR, where a much weakened Patassé barely survived thanks to French support and despite repeated military mutinies.  Military interventions ended the democracy in Congo and Niger, while opposition parties won nominal control of parliaments following second elections in Madagascar and Benin, albeit with no more impressive a majority.


In the endless parliamentary squabbling that occurred in these countries following the transition, sustaining progress on economic reform proved very difficult.  In Benin, for instance, President Soglo and the Parliamentary opposition sparred for most of 1992-1994 over the extent of presidential prerogatives.  Lacking a solid majority in the legislature, Soglo had difficulties even passing his budget in 1994, when the legislature rejected his proposals for civil service salary and student stipend increases and proposed substantially higher ones that the government protested would have undermined its adjustment program.[58]  The deadlock was eventually overcome through the intervention of the Supreme Court, but not before paralyzing reform for many months and leaving the government's capacity to undertake policy reform compromised.  But was this evidence of political participation playing itself out in the more open climate of multi-party politics?  Magnusson's careful account suggests instead factional infighting within a fairly narrow political class in the midst of growing popular alienation.  He points out, moreover, that civil service and student organizations were actually weakened by the fall of the single party; the new democratic order had brought about the multiplication of weak and competing unions and student organizations and the end of corporatist arrangements between privileged single peak organizations and the government.[59]





                                                                Table 4 about here




In most countries, however, the movement to multi-party politics has been characterized by the relatively quick emergence of a dominant party system,  following the first and or second multi-party election.  Indeed, as shown in Table 4, the average share of legislative seats going to the winning party was strikingly high in first elections (63.1 %) and then even higher in second elections (69.7%).  The party that thus emerges may owe its dominance to ethnic reasons or to superior leadership, or just as often to its ability to manipulate electoral rules, and the advantages of incumbency.  It wins a strong majority of seats in parliament and then systematically uses the advantages of incumbency to strengthen its position.  As Van Cranenburgh has argued with respect to Tanzania, "the superior organizational, material and symbolic resources of the governing party" are likely to result in the emergence of a dominant party system.[60]  Just as important in the democracies in which a new majority emerged in the 1990s,  the new government party thrives and gets stronger electorally, despite the inevitable fractionalization of the democratic alliance that spearheaded the transition.  Thus, the MMD in Zambia retained its solid legislative majority following the 1996 elections, despite the progressive defection to the opposition of most of its original tenors.[61]  These dynamics largely explain why so few incumbents have lost free and fair elections following the transition, with only presidents in Benin and Madagascar suffering electoral defeats.       

Hence, far from a weakened executive, with a small number of exceptions, the continuation of executive dominance of the political system following the introduction of electoral politics remains patent.  It is indeed striking and puzzling that not a single democratizing state chose to move to a parliamentary form of government that enshrined parliamentary dominance over the executive.  It is hard to avoid the sentiment that an opportunity was lost in many of these transitions to undermine the neopatrimonial tendency towards presidentialism and introduce parliamentary control of government, which is inherently less prone to individual abuses of power.  This is particularly true in some of the smaller countries in the region, like Mali, Benin or Malawi, which seemed well suited to parliamentary government.  The fact that many of these countries did engage in constitutional engineering during their transitions, in some cases undergoing substantial changes in electoral rules, only makes the failure for a single one to adopt some form of parliamentary system more puzzling.[62]  Surely, it speaks at least in part to the continuing attraction to African politicians of the neopatrimonial model of presidentialism. 


This executive dominance is enhanced by the state apparatus. Whatever new formal mechanisms of accountability and transparency that have been adopted by the new regime are likely to be overshadowed by a state apparatus which democratization has not purged of its longstanding authoritarian tendencies.  Administrations have too long been used to functioning with few external constraints to change quickly.  Parliaments often lack the technical expertise and administrative savvy to exercise whatever new powers of control they have gained over executive decision making.  Judiciaries, suffering from legacies of inadequate resources and political interference, also need time to assert their prerogatives.  As Kwasi Prempeh notes, the African judiciary have often continued to implement a “jurisprudence of executive supremacy”, instead of a new “jurisprudence of constitutionalism”, even in countries in which constitutional reform gives it considerable independence.[63]



Institutional Continuities 

In sum, the evidence suggests that democratization has not altered longstanding political patterns in African politics. In time, democratization may profoundly change these patterns, but by itself, the onset of multi-party electoral politics left unchanged many of the defining characteristics of African politics, notably presidential dominance and low participation.  Much the same could be said about other features of these neopatrimonial rule.  Whatever the initial intentions of the new regimes, they were susceptible to the same pull and push factors which have long favored neopatrimonial practices in the region.  On the one hand, the transition did little to change the enduring weakness of vertical and horizontal accountability mechanisms facing executives.  On the other hand, clientelism, and rent seeking have continued to be attractive to poorly integrated political systems, with weak interest aggregation institutions, ethnic divisions and under-performing economies.


One highly imperfect but suggestive measure of this continuity is the size of the Cabinet of Ministers. Large ministerial cabinets have long been a feature of the African landscape, a key locus of patronage politics.  Comparing the size of cabinets over time and across regime type may be instructive regarding the persistence of anti-developmental neopatrimonial practices.  It might be argued that democratic regimes would result in bigger cabinets as they would face pressures to be inclusive.  On the other hand, it might be argued that their cabinets would be smaller, given a reaction against the neopatrimonial tendencies of their predecessors.  I collected data on cabinets at two points of time, July/August 1988 and August/September 1996.  New democracies had cabinets averaging 19.5 and 20.6 members respectively, while non democracies averaged 26.4 and 24.8 members for the two periods. The two biggest cabinets were in authoritarian states: Gabon, with 47 members at the earlier date, and Cameroon's cabinet in 1996, which included 42 members.


The stability in cabinet sizes over time is striking, as if most countries had a fixed cabinet size, through a political tradition respected by successive regimes.  Ethnic diversity in a country probably increases cabinet size, as the government seeks to preserve national unity by including elites from all salient ethno-regional community.  Interestingly, this dynamic has not increased with the return of elections.  Only in Madagascar and Malawi, both countries with serious ethno-regional divisions, have cabinets sharply increased in size over this period, from 22 and 11 to 34 and 30 respectively.  Overall, the practice of bloated cabinets, following longstanding national neopatrimonial traditions, has remained remarkably stable through the transition period.


Where there have been political discontinuities around the transition, they have favored improved policy making, at least in the initial days of the new government.  A major theme of the fight against the old ruler had typically been the need to lessen corruption, mismanagement and the politicization of decision-making.  This was in part self-serving: the democratic alliance was in part composed of men and women left out of the presidential patronage system.  At the same time, clean government was one theme that could unite the often disparate democratic coalitions.  Intellectuals, church and human rights activists could agree with the out of power politicos on the need to stamp out corruption, if nothing else.  Thus, in countries like Benin, Mali, Madagascar, or Zambia, the first set of economic appointments emphasized a-political technocrats, in part to reassure the donors and in part to lead this fight for good government.  In most cases, however, these technocrats have seen their influence wane over time.  The best example of this phenomenon may be Zambia.[64]  The first MMD cabinet in 1991 included prominent human rights activists, such as the first Minister of Justice, Roger Chongwe. It included several apolitical good government advocates such as the first Minister of Agriculture, Guy Scott, a white farmer with no previous political experience.  Dipak Patel, a south Asian businessmen who  represented the private sector community as Minister of Industry, was known to be a critic of the corrupt and inefficient parastatal sector.


By the end of 1994, all of these men had left government.  The Zambian Cabinet was soon dominated by older politicians who had once played a prominent role in the Second Republic under Kaunda, and who often had a well deserved reputation for corruption.  Once the voice of good government rhetoric of the most uncompromising variety, Chiluba put together a patronage machine to rival Kaunda’s.  By 1994, for instance, he had named 43 MMD MPs as deputy minister (albeit without cabinet status), in addition to the 23 ministers of government and the handful of MPs in official parliamentary functions.  The perks that came with this nomination were in sharp contrast to the net salary of roughly 150 dollars a month earned by regular backbenchers, and would presumably be an instrument of presidential control of parliament.


Finally, policy continuity was more likely because few of the democratic governments reached power with a mandate to implement a specific set of economic policies.  The literature has sometimes wrongly portrayed the democratic movements as motivated largely by opposition to structural adjustment.[65]  Indeed, the dynamics that propelled the transition process forward often had little to do with economic policy issues.  Transition politics was more typically concerned as much with such issues as how to reign in executive abuses of power, ensuring ethno-regional balance and, particularly towards the end of the transition process, constitutional debates.  Although many of the initial protests that had begun the process of democratization expressed the economic grievances of groups like the civil service and students and a general anger about corruption and mismanagement, economic policy choices rarely featured prominently in national conferences or in the elections that marked the end of the transition.  Instead, the democratic forces typically came to power with a vague promise to improve living conditions by ridding the country of the dictator and his cronies.


Once in power, harsh budgetary realities and persuasive IFI experts as well as their own permanent secretaries convinced most governments against drastic changes in policy.  Ideational continuities were reinforced by the fact that the political personnel that came to power with the new democratic governments was typically very similar to the outgoing political personnel.  Not only did the government invariably retain virtually all the ministerial permanent secretaries of the previous regime, much of the cabinet had previous ministerial experience in the ancien régime.


Moreover, the dismal state of the civil service through out the continent promoted continuity.  In the transition countries of Eastern Europe, the new democratic governments that emerge have at their disposal a reasonably professional and effective civil service to carry out a set of new policies.  This was much less likely in the regimes in Sub Saharan Africa, where three decades of economic crisis had devastated the professionalism, resources and infrastructure of the state apparatus, sapping its capacity and morale.  In many countries, the state barely functions and is capable of only the most simple administrative tasks.  A significant reorientation of policy may not be possible in the short to medium term.  The main constraints on state action continue to be self-induced, in other words, rather than imposed by external participatory pressures or institutional counterweights to executive dominance.



The Aid Regime  

The final factor favoring continuity has been the relationship with the donors.  It is useful to dismiss the claim that the fledgling democracies have been sustained thanks to extraordinarily generous external financial support.  The pressure to democratize exerted by the western donors and then their support for the ensuing democratic experiments led some to claim that they represented the main motor behind democratization.[66]  These observers could point to countries like Zambia, in which the donors did reward the Chiluba government with an almost doubling of aid in the couple years following the 1991 founding election.  However, the overall numbers on foreign aid tell a different story, with no real difference between the two sets of countries.  According to the data of the OECD Development Assistance Committee, there is virtually no difference between the evolution in aid levels between the New democracies and Non democracies.  The set of New democracies received more aid per capita than the Non democracies, both before and after the transition, but this is largely an artifact of the "small country bias" of aid, resulting in very high totals going to such micro-states as Cape Verde and the Seychelles.  Comparing the average level of aid during the period 1990-1995 with the average for the period 1985-1990 yields an average increase of 37 percent in ODA for the Non democracies and 33 percent for the New democracies.  On the other hand, comparing the shorter periods of 1992-95 and 1988-1991 yields an average increases of 6 percent for the New democracies and a decline of 3 percent for the Non democracies, which suggests a minor advantage to the democracies.[67]


The United States appears to be the only major donor systematically seeking to reward political liberalization in Africa,[68] and it is no longer an important enough donor to shape the overall resource flows to the region.  Other bilateral donors generally do not resort to political criteria for their allocation of aid with any consistency or persistence.  In fact, France, the biggest bilateral donor to the region, clearly favored authoritarian regimes like Cameroon, Côte d'Ivoire and Togo with sharp increases in aid in the early 1990s, seemingly intent on helping vulnerable leaders there survive the democratization wave.[69]


Despite much support for the general principle of democratization, changes in aid volumes during the 1990s appear to have much more to do with the increasing willingness of the leading donors to selectively reward countries that follow their economic policy prescriptions.  Thus, World Bank support to democratic Zambia sharply increased after 1991, but so did its support of authoritarian but pro-policy reform regimes in Uganda and Ghana.  The year 1994 similarly witnessed substantial increase in aid volumes to many Franc Zone countries, as the IFIs (along with France) sought to ensure the success of the devaluation.


Despite donor rhetoric about the importance of democracy, it appears that the quality of the multi-party electoral politics that has emerged in the 1990s has not much mattered to the donors as long as governments play the “presentability” game.  Some donors have sanctioned the particularly egregious manipulation of elections, notably in Togo, where the US, Germany and the European Union have all suspended aid at some point in the 1990s, in response to government repression of the democracy movement.[70] Still, aid to Togo averaged some 13 percent of GDP from 1990-95, slightly above the region’s overall average of 12.7 percent during this period (excluding South Africa and Nigeria), and not much different from the 14 percent it had averaged from 1985-89. In comparison, the liberal democracy and fairly comparable economy,  Benin, received 14.9 percent of GDP in aid.  As with the aggregate numbers, the cases of Togo and Benin thus suggest perhaps a small premium for the new democracies, but hardly a revolutionary change from past patterns.


Not only have the rewards for democratization been small, in addition, the donors have not responded to democratization with any changes in aid delivery modalities.  The aid system that evolved in the 1970s has continued unchanged through the 1990s.[71] A large flow of resources (on average over 15 billion dollars a year during the decade) is given mostly through a system of more or less completely uncoordinated projects by several dozen donors in each country.  Aid has been pretty successful at improving short term individual welfare, notably by promoting projects in the area of social services; it has been considerably less successful at long term institution building.  Indeed, the massive flow of aid for the last two decades appears to have even not prevented an actual decline in state capacity through out the region.    


The wave of democratization that swept Africa in the early 1990s did not appear to change these modalities. We see the same scenario through out the region: the donor-government relationship changes little following a transition.  After a brief honeymoon, and despite perhaps a short spurt of larger aid volumes, the donors go back to their curious brand of toothless conditionality and micro-management, while governments go back to a combination of passive resistance and accommodation to donor pressures in a context of endemic fiscal crisis.



                                                           7. Concluding Remarks

My analysis has pointed to a great degree of continuity in African political economy.  Both domestic and international institutions, and the interests they rest on and sustain can change only slowly.   The wave of democratization that hit Africa in the early 1990s represents a watershed set of events, which will continue to impact upon African politics for years to come.  In the long run, neopatrimonial politics and democracy are almost certainly not compatible, insofar as the latter introduces institutions of horizontal and vertical accountability on the executive.  So, it is not inconceivable that the imperfect democratization of the 1990s will in time represent the beginning of much more fundamental change.  As I have emphasized, however, by itself the introduction of multi-party electoral politics does not create democracy over night, at least in its “liberal” variety.


In some respects, the assessment made in this essay may prove premature.  It was justified, however, by the fact that so many observers argued that democratization would have an immediate and significant impact that the preceding analysis was well worthwhile.  But, the patterns that have emerged in the first few years following the transition and which this essay tracked,  may not be sustained in time.   If it is true that these regimes are hybrid ones, in which rational-legal and patrimonial tendencies coexist, then the recent political reform represents a small defeat for the latter.  All over Africa, the long struggle to consolidate liberal democracy will continue.                                                  

                                  Table 1: Status of African Multi-Party Systems (1998) and

                                                      Legislative elections, 1990-1998



Free                                         Partly Free                                           Not Free


Benin   (2)                                Burkina Faso (2)                                    Cameroon (2)

Botswana* (1)                           C.A.R.  (2)                                            Chad (1)

Cape Verde (2)                          Comoros (3)                                          Congo (B) (2)

Malawi (1)                               Ethiopia (2)                                            Côte d’Ivoire (2)

Mali(2)                                    Gabon   (2)                                            Djibouti (2)

Mauritius* (2)                            Ghana   (2)                                            Equatorial Guinea (1)

Namibia(2)                              Guinea Bissau (1)                                   Gambia* (2)     

Sao Tome (3)                            Lesotho (2)                                           Guinea (1)

South Africa (1)                        Liberia   (1)                                            Kenya (2)

Madagascar (2)                         Mauritania (2)

Mozambique (1)                         Niger (3)

Nigeria  (1)                                            Togo (1)

Senegal* (2)                             

Seychelles (2)                                       

Sierra Leone (1)                                    

Tanzania (1)

Zambia (2)

Zimbabwe (2)




N =                   9                                              18                                                         12


Number of


1990-1998:        16                                             31                                                         21




Countries in Bold and Italics indicate that the incumbent was ousted as part of transition to multi-party rule.  Countries with * indicate that multi-party system has been in existence for an extensive period of time before 1989.  All regimes that are not multi-party regimes are not included in this analysis.  Burundi (Not Free, 1 election in 1993) and Angola (Not Free, 1 election in 1992) were excluded because of the resumption of civil war and end of multi-party rule.  Congo-Kinshasa (Not Free), Eritrea (Partly Free), Rwanda (Not Free), Somalia (Not Free), Sudan (Not Free), Swaziland (Not Free), Uganda (Partly Free) are not included because of the absence of multi-party elections.


Source: Freedom House, 1998/99 Indices



                                                                                               Table 2:

                            Economic Performance, Before and After Transition Election



Period Around Transition

1986-91           1991-95*                1995-98   1986-88   3 Years After


New Democracies

Annual GDP Growthi            2.5                    2.3                           4.1                                           2.3                           3.1

Average inflation rate            22.4                  29.0                         21.2                                         19.3                         31.0

Average Terms of tradeii      120.5                103.0                       99.0                                         129.6                       100.6


Non Democracies

Annual GDP Growthi             3.1                    2.8                           6.5                                           3.2                           3.8

Average inflation rateiii          19.8                  16.5                         11.6                                         24.4                         18.3

Average Terms of tradeii       113.8                101.2                       99.4                                         118.1                       100.8


Old Democracies

Annual GDP Growthi             5.6                    2.7                           4.5                                           6.4                           2.8

Average inflation rate            10.5                  11.9                         9.4                                           11.1                         12.1

Average Terms of tradeii       108.7                102.1                       100.8                                       111.1                       100.4


Liberal Democracies

Annual GDP Growthi                                      3.4                           4.5                                                                           3.6

Average inflation rate                                    17.0                         18.3                                                                         18.8

Average Terms of tradeii                                102.7                       98.6                                                                         104.6



Source: Calculated from the World Bank, African Development Indicators (Washington: the World Bank, 2000); not all of the estimates are based on the same number of data points, as coverage varies by indicator.


i. in Constant US dollars;

ii. 1995 = 100;

New democracies include: Benin, Cape Verde, CAR, Congo, Guinea Bissau, Lesotho, Madagascar, Malawi, Mali, Niger, Sao Tome, Seychelles, Zambia; Non democracies include Burkina Faso, Cameroon, Comoros, Côte d’Ivoire, Eq. Guinea, Gabon, Ghana, Guinea, Kenya, Mauritania, Nigeria, Swaziland, Tanzania, Togo, and Uganda; Old democracies include Botswana, Gambia, Mauritius, Senegal, Zimbabwe; Liberal democracies include Benin, Botswana, Cape Verde, Malawi, Mali, Mauritius, Namibia, Sao Tome and South Africa.



                                                                                               Table 3:

                      Economic Policy Performance Before and After Transition Elections

                                                                                 (As a percentage of GDP)


Period around Transition

1986-911991-95            1995-98                   1986-88   3 Years After

                        (Avg)              (Avg)   


Fiscal Balance (exc. grants)

New democracies:                                   -13.7                -16.3                        -14.0                        -13.7                        -14.9

Non democracies:                                   -9.0                  -8.5                          -4.3                          -9.9                          -8.7

Old democracies:                                     -2.3                  -2.9                          -4.41                        -2.7                          -3.9

Liberal democracies:                                                       -13.4                        -12.9                                                        -12.7


Fiscal Balance (inc. grants)

New democracies:                                   -6.6                  -8.9                          -7.2                          -5.4                          -8.2

Non democracies:                                   -4.5                  -4.5                          -4.0                          -5.3                          -4.6

Old democracies:                                     -1.1                  -1.1                          -2.6                          -1.9                          -4.6

Liberal democracies:                                                       -7.9                          -7.2                                                          -7.0


Government Consumption

New democracies:                                   19.0                 18.8                         17.1                         19.8                         18.0

Non democracies:                                   16.6                 15.6                         13.0                         16.8                         15.5

Old democracies:                                     16.4                 16.5                         15.3                         17.1                         16.5

Liberal democracies:                                                       22.2                         21.4                                                         22.3



New democracies:                                   14.4                 15.6                         12.3                         12.7                         14.7

Non democracies:                                   8.4                   7.5                           6.0                           7.8                           6.3

Old democracies:                                     5.8                   6.2                           7.6                           5.7                           6.3

Liberal democracies:                                                       13.5                         13.5                                                         12.3



Source: Calculated from The World Bank, African Development Indicators, 2000, various tables.

Note: Negative values for Fiscal balance implies a net deficit.

For countries included in each regime category, see Table 2.

                                                                        Table 4:

                                 First and Second Legislative Elections in Africa, 1989-1998



First Election          Second Election

      (n=41)                      (n=26)


# of Parties competing                     13.1                              12.2

# of Parties winning seats                6.3                                            7.6       

# of effective parties                       2.85                              2.4

Share of seats to biggest Party         63.1                              69.7

Turnout (% of reg.voters)               63.9                              61.3


Notes: Data assembled and calculated by author.
































[1] This paper has been adapted from a much larger manuscript, African Economies and the Politics of Permanent Crisis, 1979-1999, (Cambridge University Press, forthcoming, 2001).

[1]. Africa’s wave of democratization is explored in Michael Bratton and Nicolas van de Walle. Democratic Experiments in Africa.  (New York: Cambridge University Press, 1997).  In addition to these states, of course, Botswana, Gambia, Mauritius and Zimbabwe have conducted regular multi-party elections since their independence, and Senegal since 1979.  Gambian democracy suffered a considerable setback, however, when a military coup took power in 1994.

[2]. "Emerging Africa", The Economist, June 14, 1997.  Note that as this essay was being written, Africa once again made the cover of the British news weekly, this time under the title, “The Hopeless Continent,” (May 13, 2000), suggesting a marked change in attitude!

[3]. Célestin Monga, "Eight Problems with African Politics", Journal of Democracy. 8, (1997):156-170; Marina Ottaway,  "African Democratization and The Leninist Option", The Journal of Modern African Studies. 35, (1997): 1-15.  But for a more upbeat assessment,  see Larry Diamond, Prospects for Democratic Development in Africa. Essays in Public Policy, Hoover Institution. (Stanford University: Stanford, 1997).  See also the discussion in John Wiseman, “The Continuing case for Demo-Optimism in Africa”, Democratization, 6, (1999): 128-155.

[4]. A long and distinguished academic tradition studies the long-term relationship between democracy and growth, a slightly different set of issues than those addressed here, and one marked by a far more sanguine view of the relationship.  Good surveys are provided in Adam Przeworski and Fernando Limongi. "Political Regimes and Economic Growth", in Journal of Economic Perspectives. 7,  (1993): 51-69; and Larry Diamond,  "Economic Development and Democracy Reconsidered".  American Behavioral Scientist, 35, (1992): 450-499.

[5]. Henry S. Bienen and Jeffrey Herbst. "The Relationship Between Political and Economic Reform in Africa", Comparative Politics. 29, (1996): 23-42; Thomas Callaghy, "Civil Society, Democracy and Economic Change in Africa: A Dissenting Opinion About Resurgent Societies", in John W. Harbeson, Donald Rothchild and Naomi Chazan (eds.) Civil Society and the State in Africa. (Boulder, Colorado: Lynne Rienner Press, 1994), pp. 231-254; Peter Gibbon, Yusuf Bangura and Arve Ofstad. (eds.) Authoritarianism and Democracy and Adjustment. (Uddevalla, Sweden: Nordiska Afrikainstitutet, 1992); most of the contributions to Jennifer A. Widner. Economic Change and Political Liberalization in Sub‑Saharan Africa. (Baltimore: Johns Hopkins University Press, 1994); and Richard Jeffries. "The State, Structural Adjustment and Good Governance in Africa", in The Journal of Commonwealth and Comparative Politics. 31, (1993): 20-35.  Less pessimistic accounts are provided in John Healey and Mark Robinson, Democracy, Governance and Economic Policy: Sub-Saharan Africa in Comparative Perspective (London: Overseas Development Institute, 1992); and Peter Lewis, "Economic Reform and Political Transition in Africa: The Quest for a Politics of Development" World Politics,  49, (1997): 92-129.

[6].  Stephan Haggard, Stephan and Robert R. Kaufman, The Political Economy of Democratic Transitions. (Princeton: Princeton University Press, 1995),  p. 152.  See also Marc Lindenberg and Shantayanan Devarajan, "Prescribing Strong Medicine: Revisiting the Myths about Structural Adjustment, Democracy and Economic Performance in Developing Countries", Comparative Politics. 25, (1993): 169-183.


[7].  For instance, Haggard and Kaufman, The Political Economy of Democratic Transitions; a good discussion of these issues can be found in Joel S. Hellman, “Winners Take All: The Politics of Partial Reform in Postcommunist Transitions”, World Politics.  50, (1998): 203-234.

[8].  Jeffries, “The State, Structural Adjustment,”  p. 20.

[9]. The interested reader is referred Bratton and van de Walle, Democratic Experiments.  Among the recent major works from different perspectives, see also Richard Joseph (ed.) State, Conflict and Democracy in Africa. (Boulder: Lynne Rienner, 1999); John A. Wiseman, (ed.)  Democracy and Political Change in Sub-Saharan Africa.  (New York: Routledge, 1995);  Jean Pascal Daloz and Patrick Quantin, (eds.) Transitions Democratiques Africaines (Paris: Karthala, 1997);  and Julius O. Ihonvbere and John Mukum Mbaku, (eds.) Multiparty Democracy and Political Change: Constraints to Democratization in Africa (London: Ashgate, 1998).

[10]. On the Côte d’Ivoire, see Ives A. Fauré and Jean François Médard, Etat et Bourgeoisie en Côte D’Ivoire, (Paris: Karthala, 1982). On Nigeria, see Joseph, Democracy and Prebendal Politics.

[11]. Richard Sandbrook, "The State and Economic Stagnation in Tropical Africa", in World Development 14, (1986): 319-332.

[12]. Callaghy, The State-Society Struggle, p. 194.

[13].  See for instance the World Bank, the World Development Report, 1997. (New York: Oxford University Press for the World Bank, 1997).

[14].  Mike Stevens, “Public Expenditure and Civil Service Reform in Tanzania”, in David L. Lindauer and Barbara Nunberg (eds.) Rehabilitating Government: Pay and Employment reforms in Africa, (Washington: the World Bank, 1994), pp. 62-81, especially 66-69.

[15]. On this point, see the perceptive essay by Patrick Quantin, “Les Elites Politiques Face aux Transitions Democratiques”, in l’Afrique Politique, 1995: Le Meilleur, le Pire et l’Incertain, (Paris: Karthala, 1995), pp. 277-85

[16]. Ibid. p. 282.

[17]. See Bratton and van de Walle, Democratic Experiments in Africa. P. 135-6.

[18]. Ibid, p.132-3, 151.

[19]. Ibid.

[20]. The Benin story is well told in Ch. Allen,  "Restructuring an Authoritarian State:  Democratic Renewal in Benin", Review of African Political Economy.  54 (1992): 43-58; and Richard Westebbe. “Structural Adjustment, Rent Seeking and Political Liberalization in Benin”, in Jennifer Widner, (ed.) Economic Change and Political Liberalization in Sub-Saharan Africa.  (Johns Hopkins Press, 1994), pp. 80-100.  On Zambia, see Lise Rakner, Reforms as a Matter of Political Survival. Political and Economic  Liberalisation in Zambia 1991-1996. PhD Dissertation., University of Bergen, Norway, 1998.


[21].  For an examination of French political conditionality, see Richard Banégas and Patrick Quantin, "Orientations et Limites de l'Aide Francaise au Développement Démocratique", Revue Canadienne d'Etudes du Développement, Special Issue, (1996):113-33.

[22]. See my essay, “The Politics of Non-Reform in Cameroon," in Thomas Callaghy and John Ravenhill, (Eds.) Hemmed in: Responses to Africa's Economic Decline. (Columbia University Press, 1993), pp. 357-97; p. 383.

[23]. The African Development Indicators, p.  297.  Thus, Kenya averaged 1.1 billion dollars between 1988-1993, and 640 million dollars between 1994-1997, in constant 1995 US dollars.  Note that there have been further decreases since then.

[24].  See Bratton and van de Walle, Democratic Experiments,  p. 287. This is on a scale of 1 to 7 for both indicators.

[25]. See Fareed Zakaria “The Rise of Illiberal Democracy.”  Foreign Affairs 76, (1997): 22-43; and Larry Diamond.  “Is the Third Wave Over?”  Journal of Democracy, 7 (1996): 20-37.

[26]. Terry Lynn Karl.  "Dilemmas of Democratization in Latin America."  Comparative Politics 22, (1990):1-20.

[27]. Richard Joseph.  “Africa, 1990-1997: From Abertura to Closure.” Journal of Democracy 9, (1998): 3-17.

[28]. Freedom House rates all countries of the world on a seven point scale for both political rights and civil liberties (1 representing the most free and 7 the least free), and then combines the two scores to produce an index.  Freedom House, then, attributes a “Free” rating to countries with average ratings of 1-3, “Partly Free” to those between 3-5.5 and “Not Free” to those averaging between 5.5-7 (Freedom House 1998).

[29]. The term is from Diamond, “Is the Third Wave Over?”

[30]. Ibid.

[31].  Calculated from the Af rican Development Indicators, Table 2-18.

[32]. It should be noted that in previous work (Bratton and van de Walle, Democratic Experiments),  I defined Mozambique as having undertaken a complete political transition.  Nonetheless, the consequences of a decade long civil war of tragic destructiveness continues to powerfully condition economic performance in the country.  Chad was excluded for similar reasons.  For its part, Liberia may or may not have successfully ended its civil war and begun a transition to democratic governance with the elections conducted in July 1997.

[33]. See Bratton and van de Walle, Democratic Experiments, Chapter 3, for a complete justification of this classification.  The countries in each category are listed at the bottom of Table 2.  Note that in all of these countries, except for Guinea Bissau and Seychelles, the incumbent was also ousted during the founding election. The latter two cases are included because of the widespread agreement by domestic and international observers that their founding elections had been free and fair.


[34]. E. Gyimah-Boadi. "Ghana's Encouraging Elections: The Challenges Ahead", in Journal of Democracy. 8, (1997): 78-91; and Terrence Lyons. "Ghana's Encouraging Elections: A Major Step Forward", in Journal of Democracy. 8, (1997): 65-77.

[35]. Caroline Baylies and Morris Szeftel, "The 1996 Zambian Elections: Still Awaiting Democratic Consolidation", in Review of African Political Economy.  71, (1997): 113-128; and Michael Bratton and Daniel Posner. "A First Look at Second Elections in Africa, with Illustration from Zambia", in Joseph State, Conflict and Democracy in Africa, pp 377-408.

[36].If no founding election could be identified, I used the period of data for 1993-1995, since the modal year for founding elections was 1992.

[37]. Jacques Alibert. "Un Bilan de la Dévaluation du Franc CFA", in Afrique Contemporaine. 179 (1996): 16-26; and  Jean A.P. Clément , Johannes Mueller, Stephane Cosse and Jean le Dem. Aftermath of the CFA Franc Devaluation.  Occasional Paper 138. International Monetary Fund.  (Washington: IMF, May 1996).

[38].  On the abysmal quality of African data, see Alexander J. Yeats. "On the Accuracy of Africa Observations: do Sub-Saharan Trade Statistics mean Anything?", in World Bank Economic Review. 4, (1990): 135-56.  The inflation data are particularly problematic as they typically concern only prices in the capital city, and are often based on outdated consumption patterns

[39].  In addition, severe drought through out southern Africa in 1992-93 also hit the economy hard. See Lise Rakner,  Reforms as a Matter of Political Survival.

[40]. I have examined the economic costs of Africa’s democratic transitions at greater length in my essay, “Economic Reform in a Democratizing Africa”, in Comparative Politics, 32, (1999): 21-42.

[41].  Thus, the main data set used (the World Bank, African Development Indicators, 1997. (Washington: the World Bank, 1997b)) does not report data for the period for 4 of the countries in the data set (Equatorial Guinea, Lesotho, Swaziland and Congo/Zaire).  Most annoyingly, there are wide and unexplained divergences between this data set and others that were consulted.  They even vary widely across different publications of the World Bank.  Thus compare the estimates provided in table 3, with those reported in Lawrence Bouton, Christine Jones and Miguel Kiguel. Macroeconomic Reform and Growth in Africa.  World Bank Policy Research Working Paper, no. 1394. (Washington DC: the World Bank, December 1994); or in the World Bank, Adjustment in Africa.

[42].It should be noted in this context that World Bank cross national studies of African economies often exclude states with a GNP inferior to one million US dollars, arguing that aid flows to these states prevent meaningful policy comparisons.  See for example, the World Bank, Adjustment in Africa.

[43]. See for instance  John A.  Wiseman. "Urban Riots in West Africa, 1977-1985", Journal of Modern African Studies. 24 (1986): 509-18.

[44]. For instance William Nordhaus, “the Political Business Cycle” in Review of Economic Studies 42, (1975):169-90; and Michael S. Lewis-Beck, Economics and Elections: the Major Western Democracies. (Ann Arbor: University of Michigan Press, 1990), pp. 137-52.


[45]. For a discussion, see Michael Bratton. “Second Elections in Africa” in Journal of Democracy, 9, (1998): 51-66.

[46]. Nicolas van de Walle, African Economies and the Politics of Permanent Crisis, 1979-1999. (New York: Cambridge University Press, forthcoming, 2001), chapter 1.

[47].  Thomas Callaghy, “Civil Society, Democracy and Economic Change in Africa: A Dissenting Opinion About Resurgent Societies", in John W. Harbeson, Donald Rothchild and Naomi Chazan (eds.) Civil Society and the State in Africa. (Boulder, Colorado: Lynne Rienner Press, 1994b), pp. 231-254.

 1994b; and René Lemarchand.  "Uncivil States and Civil Societies: How Illusion Became reality", in Journal of Modern African Studies.  30, (1992): 177-191; Célestin Monga, “Civil Society and Democratization in Francophone Africa”, in Journal of Modern African Studies, 33, (1995): 359-81; and Peter Lewis,  “Political Transition and the Dilemmas of Civil Society in Africa”, in Journal of International Affairs,  93, (1992): 323-340.

[48].  Georges Courade and Véronique Alary, "Les Planteurs Camerounais ont-ils été réévalués", Politique Africaine, 54 (1994), 74-87, p. 80.

[49]. For instance, “Les Ondes de la Transparence: Avec peu de Moyens, les Radios Privées Freinent les Velléités de Fraude” in Libération.  February 28, 2000.

[50].  For a review and references, see John Healey and Mark Robinson, Democracy, Governance and Economic Policy: Sub-Saharan Africa in Comparative Perspective (London: Overseas Development Institute, 1992).

[51].Bratton and van de Walle, Democratic Experiments in Africa, pp. 167-68.

[52].  On this point see the excellent discussion in Jean‑François Bayart.  L'Etat en Afrique: la Politique du Ventre.  (Paris: Fayard, 1987), especially pp. 193-226; and Donald Rothchild and Michael Foley, “African States and the Politics of Inclusive Coalitions”, in Donald Rothchild and Naomi Chazan, (Eds.) The Precarious Balance: State and Society in Africa. (Boulder: Westview Press, 1988). pp. 149-171.

[53]. See Nicolas van de Walle and Kimberly Butler, “Political Parties and Party Systems in Africa’s Illiberal Democracies”, in Cambridge Review of International Affairs. 13, (1999): 14-28.  The case of Zambia is admirably explored in Daniel Posner,  “Ethnicity and Ethnic Politics in Zambia.”  Unpublished Chapter of Ph.D. Dissertation.  Boston, MA: Harvard University, 1999.

[54]. Recent ethnic politics in Cameroon are well covered in Luc Sindjoun, La Politique d’Affection en Afrique Noire: Société de Parenté, ‘Société D’état” et Libéralization Politique au Cameroun. Occasional Paper Series,  African Association of Political Science.  No. 2, (1998); and Hélène-Laure Menthong, “Vote et Communautarisme au Cameroun: “un Vote de Coeur, de Sang et de Raison”, Politique Africaine  69, (1998): 40-52.  See also Francis Nyamnjoh, “Cameroon: A Country United by Ethnic Ambition and Difference”, in African Affairs, 98 (1999), 101-18.

[55]. Personal communication, Barbara Lewis.  These issues are also discussed from a slightly different perspective by Richard C. Crook, in his essay, “Winning Coalitions and Ethno-Regional Politics: the Failure of the Opposition in the 1990 and 1995 Elections in Côte d’Ivoire”, in African Affairs 96, (1997): 215-242.  Crook argues that ethnic issues were first manipulated by the opposition during the 1990s elections, only to be then adopted by the majority in 1995.  Since then, the government’s anti-northern stance has emerged much more clearly.  See “Fishermen Feel Xenophobia in 'Pure Ivoirian' Pride “, in New York Times, August 8, 1999; and “Turbulences en Côte d’Ivoire” in Le Monde, October 8, 1999 for more recent information on these ethnic issues.

[56].  The ethnic dimension to the 1997 elections are well described in Joel D. Barkan and Njuguna Ng’ethe, “Kenya Tries Again” in Journal of Democracy,  9, (1998): 32-49.

[57]. At the same time, in other work, I have shown that the weaker the largest party, the more likely the democratic gains from the transition to be preserved in the 1990s.  See van de Walle and Butler, “Political Parties and Party Systems”.

[58].  See the illuminating analysis by Bruce Magnusson, "Benin: Legitimating Democracy: New Institutions and the Historical Problem of Economic Crisis", in l'Afrique Politique 1996, Démocratisation: Arrêts sur Images (Paris: Karthala, 1996), pp.33-54.

[59]. Ibid.  See also the interesting work by Richard Banégas, “Marchandisation du Vote, Citoyenneté et Consolidation Démocratique au Bénin, in Politique Africaine  69 (1998): pp. 75-88.

[60].  Oda Van Cranenburgh, "Tanzania's 1995 Multi-Party Elections: the Emerging Party System", in Party Politics.  2, (1996): 535-547, p. 545.

[61]. Caroline Baylies and Morris Szeftel, "The 1996 Zambian Elections: Still Awaiting Democratic Consolidation", in Review of African Political Economy.  71, (1997): 113-128.

[62]. Such constitutional engineering was usually design to manage minority and ethnic divisions. See  Andrew Reynolds.  Electoral Systems and Democratization in Southern Africa.  (New York: Oxford University Press, 1999);  Timothy D. Sisk, and Andrew Reynolds, (eds.), Elections and Conflict Management in Africa. Washington: United States Institute of Peace, 1998); and Harvey Glickman, Harvey (ed.)  Ethnic Conflict and Democratization in Africa. (Atlanta, GA: African Studies Association Press, 1995)

[63].  H. Kwasi Prempeh, “A New Jurisprudence for Africa”, in Journal of Democracy. 10, (1999): 135-49.

[64]. The following account is largely derived from Lise Rakner, Nicolas van de Walle and Dominic Mulaisho, “Aid and Reform in Zambia: A Country Case Study”, Chr. Michelsen Institute, Bergen, Norway.  Unpublished paper for the Aid and Reform in Africa Project, World Bank, September 1999; and Dennis Chiwele and Nicolas van de Walle, "Democratization and Economic Reform in Zambia".  Zambia Democratic Governance Project Monitoring and Evaluations Studies.  Report prepared for the U.S. Agency for International Development, Lusaka, Zambia. (September 1994).

[65]. See Gibbon et al. Authoritarianism, Democracy and Adjustment, for a discussion.

[66]. For instance Samuel Decalo, "The Process, Prospects and Constraints of Democratization in Africa", African Affairs. 91, (1992): 7-35; and Richard .Jeffries. "The State, Structural Adjustment and Good Governance in Africa", in The Journal of Commonwealth and Comparative Politics. 31, (1993):20-35.


[67]. Calculated from Table 12-5, The World Bank, African Development Indicators, 1997.

[68].  The promotion of democracy has been elevated to a central objective of all USAID programs by the Clinton Administration.  For a discussion, see Todd J. Moss. “US Policy and Democratization in Africa: The Limits of Universal Liberalism”, in Journal of Modern African Studies. 33 (1995): 189-209.

[69]. This leads Célestin Monga to argue that the donors have a pro-dictatorship bias in their aid. Only for France, however, does the data clearly bear him out.  See his essay, "Eight Problems with African Politics".  The French case is well discussed in Banégas and Quantin, “Orientations et Limites de l'Aide Francaise”,  and Tony Chafer. "French African Policy: Towards Change", in African Affairs. 91 (1992): 37-51.

[70]. See Kodjo Koffi, “Togo: Les Deux Ruptures de la Coopération” in Afrique Contemporaine.  189, (1999): 63-76.

[71]. See my essay, “Aid’s Crisis of Legitimacy: Current Proposals and Future Prospects” in African Affairs. Vol. 98, (July 1999).