Introduction

Regrettably, the discipline of Development Studies has often had a tendency of displaying less interest in critically testing assertions about the merits of democratic institutions than it has in exposing the adverse consequences of more authoritarian political structures. In few places is this more evident than in the field of famine research. Amartya Sen, Nobel Laureate in Economics, has made two influential contributions to modern famine research. His first contribution, The Entitlement Approach, is an analytical famine framework founded on the simple yet groundbreaking idea of focusing on the lack of access to food rather than a lack of availability of food (Sen, 1981). By showing how some of the worst famines of the twentieth century took place with virtually no (or a miniscule) decline in food availability, Sen rejected the hitherto dominant causal famine theory, the Food Availability Decline (FAD) theory, emphasizing the availability of food as the key determinant of famine. Instead, Sen proposed his entitlement approach highlighting the distributional impacts of famine. His second major contribution is the assertion that democratic institutions together with a free press provide effective protection from famine. The rationale is that electoral political dynamics (the competitive struggle to win over public opinion and crucial votes) will induce the incumbent government to take timely and adequate action in the face of a famine disaster.

What is remarkable is that while Sen's entitlement approach has sparked much controversy in academic circles (see Bowbrick (1986); Kula (1987); De Waal (1990); Osmani (1995); Devereux (2001); Elahi (2006); and Rubin (2009a)), his causal theory about the merits of a particular political system in reducing the risk of famine has been allowed to stand largely uncontested. This despite the fact that Sen devoted a whole monograph to theoretically developing and empirically underpinning the entitlement approach while his thoughts on democracy and famine were initially merely advanced at the now-famous Coromandel Lecture with limited references to empirical evidence.

This article will seek to remedy this bias of academic attention by scrutinizing whether indeed democracies effectively provide protection from famine. Rather than focusing on the authoritarian political dynamics that have exposed populations to famine (with more than one million people believed to have perished during the 1990s due to starvation, North Korea would be an obvious such candidate), attention will here be devoted to identify democratic dynamics that appear to have short-circuited effective protection in famine situations. These famine situations will first and foremost relate to India after independence, the empirical cornerstone of Sen's hypothesis of the merits of democracy in famine protection, but recent famine situations in Malawi and Niger will also be briefly addressed. The article will conclude by theorizing as to the deficiencies within pluralist institutions that can account for inadequate famine protection. First, however, Sen's thoughts on the link between democracy and famine protection will briefly be laid out.

Sen's Theory of Democracy and Famine

It was not until the beginning of the 1980s that the true atrocities of the Chinese famine of 1958–1961 began to surface in the western world. With an estimated 30 million deaths, the Chinese famine appears to have inspired Sen to develop the hypothesis that democracy provides effective protection from famine. The fact that such a terrible famine could unfold despite noteworthy Chinese achievements in reducing chronic undernourishment and child mortality led Sen to complement his entitlement approach with thoughts about dynamics at the political level. At the 1982 Coromandel Lecture, Sen elaborated at some length: This particular aspect of the Chinese famine – its linkage with the lack of democracy in China – fits into a more general pattern of association between democracy and successful prevention of famines, or – seen the other way – between the absence of democracy and the lack of any guarantee that serious attempts to avert famines will be undertaken. Sen (1982)

Sen thus developed a causal theory stressing the importance of a pluralist political system combined with a free press in famine protection. In a democracy, the government is faced with opposition parties, a free press and regular elections. An elected government is therefore more responsive to popular demand, because voters have the power to replace it after the next election. Formulated more vividly by Sen: 'Democracy and an uncensored press can spread the penalties of famine from the destitute to those in authority. There is no surer way of making the government responsible to the suffering of famine victims' (Sen, 1990, p. 12). Hence, what triggers government responsiveness is not linked to the benevolence of a democratic regime vis-á-vis an authoritarian regime, but to the fact that governments – fighting for survival (reelection) – will go great lengths to avoid a humanitarian catastrophe. This is why democracy must be combined with a free press. The free press is important as a mediator of information and ensures that a famine cannot be concealed from the greater public. Knowledge is obviously a prerequisite for making a government accountable. Sen notes that 'newspaper reports and public protests carry not only information that authorities can use, but also elements of pressure that may make it politically compelling to respond to these danger signals' (Drèze and Sen, 1990, p. 263). In a democracy, the free press thus plays two important roles in famine protection: one as a mediator of information and the other by holding the government accountable.

Whereas China provided the empirical inspiration for his thoughts regarding the perils of authoritarian regimes in famine protection, India provided evidence of the other side of the equation. Until the formation of a pluralist political system following independence in 1947, the Indian continent had regularly been tormented by famines. Since then, according to Sen, democratic India has successfully combated major famines.

Sen's basic idea about the merits of democracy in famine protection has persisted since it was first conceived and forwarded publicly in 1982. In one of his latest books, The Argumentative Indian, published in 2005, Sen repeats (Sen, 2005, p. 51) that 'famines do not occur in democracies.' Furthermore, he once again draws attention to his own 'prominent observation that major famines do not occur in democracies, even when they are very poor' (Sen, 2005, p. 188). Again, he refers to the fact that famines 'disappeared abruptly with the establishment of a multiparty democracy in India. In contrast, China had the largest famine in recorded history 1958–61, when nearly 30 million people, it is estimated, died' (Sen, 2005, p. 188).

Writing for the New York Times, Massing (2003) notes that 'just as Adam Smith is associated with the phrase "invisible hand" and Joseph Schumpeter with "creative destruction," Mr Sen is famous for his assertion that famines do not occur in democracies.'

The few academics that have critically addressed Amartya Sen's democracy hypothesis (De Waal, 1997, 2000; Banik, 2007 and Neumayer and Plümper, 2009) used his hypothesis as a stepping stone for forwarding their own slightly revised single-factor explanations for famine. Rather than democracy, De Waal emphasized anti-famine contracts as the determining variable in famine protection; Banik argued that freedom ought to replace democracy as the explanatory variable; and Neumayer and Plümper hypothesized that the size of the electorate was what mattered. Thus, one causal theory replaced another one while retaining the basic premise of having one factor alone responsible for certain famine outcomes. The mono-causal approach itself was new questioned. Sen's democracy theory still remains far and away the most dominant famine theory at the political level. Indeed, Sen's democracy hypothesis has received wide recognition among famine theorists and has been somewhat uncritically adopted as a proven fact that democracy must lead to increased famine protection (D'Souza, 1994; Varshney, 2000; Jana, 2002; Martin Paldam, 2006; Barnett and Adger, 2007; Rummel, 2007; Diamond, 2008). The question is whether it is a valid one.

Approach for Empirically Assessing the Democracy Hypothesis

The assessment of the democracy hypothesis draws on empirical evidence from two levels of analysis (i) the aggregate level where focus is on the relationship between the two key variables: democracy and famine; and (ii) the disaggregate level where focus is on the actual electoral political dynamics by which democracies are pressured to protect against famine.

Testing Aggregate Deterministic and Probabilistic Relationships

Although the strength of the aggregate correlation in Sen's democracy hypothesis is open to some interpretation, the disaggregated process itself is very clearly defined. At the aggregate level, some of Sen's statements (including the ones reproduced above) could indicate that the relationship between democracy and famine is a deterministic one – famines are unable to occur in democracies. Myhrvold-Hanssen (2003, p. 1) finds this deterministic interpretation 'reasonable,' although one interpretation among many. This deterministic interpretation can be evaluated empirically via falsification; the preferred method would therefore be to confront the proposition with case studies in which both famines and a democratic system can be identified. Due to the fact that democracy and famine are both contested concepts, to increase analytical robustness, the study relies on a cumulative research design based on five cases where famine appears to have coincided with a democratic regime. Three of these cases are from India (Bihar, 1966; Maharashtra, 1970–73 and Orissa, 1990s) while two are more contemporary cases from Africa (Malawi, 2002 and Niger, 2005). The cases from India constitute critical cases e.g. cases of strategic relevance to the hypothesis (Flyvbjerg, 2006). India is repeatedly referred to by Amartya Sen as an empirical verification of his thesis stressing the famine-protective effects of democracy. If famine has nevertheless occurred in India despite its democratic institutions, or if an equivalent level of democracy can be found in other famine-stricken countries, then the deterministic nature of the hypothesis becomes questionable.

The aggregate relationship between democracy and famine could also be interpreted more leniently as meaning that democracies are less likely to experience famine because the incumbent government would go to great lengths to prevent such a disaster; such probabilistic causality would not rule out the possibility of famine in a democracy altogether. While a cumulative research design, through methods of falsification, can assess the deterministic propensities of the relationship, cross-country quantitative methods can be used to obtain measures on the strength of the causality between democracy and famine.

Assessing the Hypothesized Electoral Political Processes

Regardless of whether the aggregate correlation is deterministic or probabilistic, the actual channel through which democracies provide effective famine protection is carved in stone; the electoral political dynamics (with a free press and opposition parties) pressure the government to timely and adequate interventions. Theoretically, many other dynamics could be responsible for a correlation between democracy and fewer famines; most democracies have not existed for very long making famine less likely; democratization in many developing countries coincided with the Green Revolution; many democracies are so wealthy that a widespread entitlement collapse is unrealistic; and democracies might have better access to relief aid from likeminded donors. All are factors that could lower the risk of famine in democracies independent of the electoral political dynamics per se. The clear identification of one specific process by which democracies are hypothesized to deliver effective famine protection makes case studies a beneficial method for validity testing: are the hypothesized political processes (that pressures from a vocal opposition together with a vigilant free press compel an incumbent government to act) actually prominent in famine situations or do the political dynamics in fact differ from what has been envisaged? Here the article will mainly focus on the political processes during the Bihar famine due to its status as a critical case, and to a lesser extent on the dynamics during famines in Malawi and Niger.

Critical Case: The Indian Experience Revisited

Despite Sen's use of India and China as the yin and yang of his thesis on democracy and famine, the empirical evidence indicates that famines did not appear to 'abruptly' end in India following the establishment of a pluralist political system in 1947. Instead, much empirical evidence points to the continued existence of famines in India following independence; in particular, the states of Bihar (1966), Maharashtra (1970–1973), and perhaps even Orissa (1990s), experienced famine-like conditions. However, in the 50 years following the great Chinese famine of 1958–1961, large-scale famines do appear to have abruptly ended in China despite the continued existence of a totalitarian regime. Hence, the superiority of democracies vis-à-vis other regimes in providing effective famine protection might not be as obvious as Sen took it to be. The key exponent for the merits of democracy in famine protection, India since Independence, will therefore be revisited with a special focus on the adverse pluralist dynamics during the Bihar 1966 famine. The Indian experience calls into question both the deterministic relationship between democracy and famine as well as the hypothesized electoral political dynamics that should impel the incumbent government to provide protection from famine.

The Bihar 1966 Famine

Following independence in 1947, India was established as a federal republic, and has successfully held 14 national elections together with countless state elections. Such long traditions of democratic rule and press freedom are indeed an accomplishment in a populous, diverse and poor country such as India. Just preceding the 1967 national elections, India' second most populous state, the state of Bihar, appears to have experienced a major famine. Before 1967, Bihar suffered less than normal food grain production for 2 years in a row due to a disastrous combination of drought in the south and floods in the north. More than half of the districts in Bihar produced less than 50 per cent of the normal output, and more than a quarter of the population had to survive on less than 1800 calories per day (Myhrvold-Hanssen, 2003, p. 3). The impact on mortality was clearly discernable. Based on the Annual Report on Vital Statistics of Bihar, the mortality rate peaked in 1967 with a rate that was 22 per cent higher than in 1966 and 34 per cent higher than in 1968. For some of the worst affected districts, the Gaya district for instance, the mortality rate in 1967 was in fact three times the mortality rate the following year. If 1966 and 1968 are taken to be normal years, then the excess mortality rate in 1967 was in the order of 3.5/1000/year for the whole of Bihar. According to the Government of Bihar, 34 million persons were reportedly affected by the famine, of which 13.4 million were in the famine area (Brass, 1986, p. 247). In other words, the casualties of the famine are likely to be counted in tens of thousands. This is supported by reports of 'eye-witness accounts of people eating wild leaves and roots, picking pieces of grain from the dust around railway sidings, undergoing appalling skeletonization, and even starving to death' (Drèze, 1999, p. 111). The famine is noteworthy because it appears to have evolved under democratic rule. Furthermore, I will argue that the crisis did not occur despite political action to remedy the famine, but that the lack of an adequate response was caused at least in part by democratic institutions that exacerbated the situation.

With the appointment of a Famine Commission in 1878 and the drafting of Famine Codes, India had taken the first, and in many ways unique, step toward an institutionalized response to the threats of famine. The Codes laid out authoritative guidelines of famine relief but placed the responsibility for drawing up and implementing the policies at state level. In the Bihar case, these policies were compiled in the Bihar Famine and Food Relief Code of 1957 stating that if existing relief measures are inadequate to cope with distress in a crisis situation, the vulnerable districts should be declared 'famine affected' in order to trigger increased relief intervention from the national level. In the summer of 1966, demands to declare famine started flowing in from various drought-stricken districts in Bihar to the State Government. The State Government was initially reluctant to acknowledge a potential crisis situation; after several visits by prominent ministers to the famine stricken districts; however, the State Government openly talked about the need for food aid to save people from starvation in September/October 1966, and the State Government requested support from the Central Government. The Central Government, however, refused for some time to accept the claims of an impending disaster in Bihar.

The Government of India distrusted and discounted the reports coming from Bihar and treated them as being politically motivated. At the time of the famine, both the Central and State Governments were being run by the Congress Party. At the state level, the Congress Party was weak and fragmented by major internal divisions, and it suffered from widespread membership defections before 1967 elections. The famine coincided with the process of selecting State Congress candidates for the imminent general elections; preparing for the election diverted focus and resources away from the drought, and a dispute between the Central and State Governments over the selection of candidates increased tension between the Center and the State. Furthermore, Bihar Chief Minister Krishana Sahay had only been a lukewarm supporter of the Prime Minister, Indira Gandhi, to succeed the late Lal Shastri as Prime Minister, and he had been unable to procure a Bihar vote for Gandhi in the caucuses. On the administrative side, the Bihar bureaucracy had received a regular pounding by the press for its handling of various high profile issues. As such, the Central Government had little confidence in the capacity of the Bihar administration to cope efficiently and impartially with a major crisis such as a drought.

The Central Government did not engage directly in the crisis until November following Indira Gandhi's emotional speech to the nation in which she called for a 'political truce,' and made the case that 'survival should remain above politics' (quoted in Brass, 1986, p. 259). A relevant question is how the Central Government for months could get around the existing institutional arrangement that should secure provision of famine relief. The main reason appears to be the State Government's reluctance to declare famine. Declaring a famine immediately before the election could give the impression of an incapable State Government – in particular if the declaration of famine was not followed up with the necessary resources; and those resources largely rested with the Central Government. The declaration of famine first came about after the State Congress government lost the election and was replaced with a six-party coalition government, which declared the famine for political reasons (having demanded the declaration of famine while in opposition, they had to follow suit when in government).

Both the State and Central Governments were worried about the potential impacts an impending famine would have on the election results. The State Government attempted to gain ground politically by appealing for Central support rather than engaging in widespread redistributional policies. The Chief Minister, Sahay, told reporters that without Central assistance, it would be difficult for the State Government to save the people from starvation (Brass, 1986, p. 257). The Central Government also saw a political 'hot potato' in the impending famine. The Central Government needed to strike a difficult balance between acting responsively but at the same time avoiding responsibility, which had to be placed at state level. Several members of the Congress Party openly criticized the Bihar Government for delaying relief (Brass, 1986, p. 258). The Central Government also abstained from major redistributional policies, relying instead on strengthening Bihar's administrative apparatus and allocating US food aid to Bihar. Brass (1986, p. 259) notes that almost all of the grain provided to Bihar from the Central Government came from the US. This reliance on US grain took place in a context where India as a whole actually increased staple food production by 2 million tons in the period 1966–1967.

In summing up, the Bihar famine 1966–1967 – like all famines – had several causes. Because of India's high degree of federalism, the Bihar famine had two strong government players: the Bihar State Government and the Central Government. The fact that both governments were from the same party did not prevent political divisions and mistrust. Brass concluded his analysis by stating that the 'Bihar Famine crisis was not only politicized from its onset, but it was democratized.' Democratized refers to the fact that many different voices in the democratic process – the free press, the citizens and the opposition parties – all had a say in defining the situation. The many actors were not speaking with one voice – not due to a lack of coordination, but because the politicization of famine fostered the incentive to assume political stances relative to other actors. Questions such as whether or not there was a famine; who should intervene to prevent the famine; and who should bear the responsibility for the famine were the focus of a political struggle ultimately leading to a suboptimal response. The 1967 elections, together with the strained relationship between the Central and State Governments, appear to have created incentives to dodge responsibility. The decision not to implement redistributive policies to mitigate the famine might have root in the imminent election; the State Government asked the Central Government for external support, whereas the Central Government in turn sent requests for support to the US. Redistributive policies moving resources from the majority to the minority could prove dangerous in an election year, and the Congress Party might have considered the redistribution of grain from surplus districts or states to have adverse political implications (Brass, 1986, p. 259). Drèze (1999, p. 118) also highlights the clear political aspect of the famine by noting the delay in the declaration of famine due to the general election in February 1967 as well as the Center–State intrigues.

A counterargument could be that without the 1967 election, the situation might have been even more dire, as the incentives to intervene might have been less pressing; after all, the Central Government did act in November 1966 with the likely agenda of preventing a famine from developing just outside the voting booths. Nevertheless, the argument here is not to advocate that without democratic institutions Bihar famine relief might have had greater success. Rather, the Bihar 1966–1967 case shows that democratic institutions can be a double-edged sword in the fight against famine. In Sen's democracy hypothesis, the politicization of famine is a key driver for government action, but the politicization also creates other dynamics that might be counterproductive in a famine situation. Mistrust between different public institutions, incentives to 'pass the buck', 'electoral tunnel vision' where issues not deemed essential for the election vanish out of sight until after the election, and restrained famine response in the light of the impending election were all factors that hampered effective famine relief in democratic Bihar.

The Maharashtra 1973 Famine

Less than a decade later, yet another major famine took place in India under a democratic government – the Maharashtra famine of 1973. Dyson (1993, p. 25) notes that the famine mortality data for Maharashtra is much better than the Bihar data, and he arrives at a rough estimate of 130 000 excess deaths in 1972–1973. This figure is also reproduced in Devereux's list of twentieth century famines (Devereux, 2000, p. 6). Despite the estimated 130 000 excess deaths, Drèze uses the Maharashtra food crisis as an empirical example of a successful application of the correct sets of policies within the optimal political setup. Unlike his investigation of the Bihar case, Drèze does not address the issue of excess deaths at all in the Maharashtra case. Were it not for other scholars (Dyson and Maharatna, 1992; Menken and Campbell, 1992; Dyson, 1993), one could easily be left with the misleading impression that no excess deaths had occurred. However, as Menken and Campbell (1992, p. 111) write, the conclusion rather appears to be that 'even in what was a relatively well-managed disaster, there were probably around 130 000 excess deaths.' This indicates that the sheer scale of a food crisis can overwhelm even a responsive democratic regime with a well-functioning bureaucracy.

The Orissa 1990s Hunger Crisis

In more recent times, widespread starvation deaths occurred in the state of Orissa during the 1990s. Although the hunger situation in Orissa throughout the 1990s might not be of famine proportions, the hundreds of reported starvation deaths do indicate that India's typical levels of chronic malnutrition turned into a more acute starvation epidemic in Orissa. Through extensive fieldwork in the starvation-stricken western parts of Orissa (in the Kalahandi and Naupada districts), Currie (1998, 2000) reports how the response of successive State Governments to adversarial pressures from the press and opposition parties was not to initiate more effective famine protection but instead to use time and energy to deny any wrongdoing and to deflect blame on the opposition. In other words, there is no guarantee that critique from the press and political opposition only fosters the most constructive type of government response, which is to provide more effective famine relief aid. An alternative strategy – the strategy chosen in fact by governments in Orissa – is to inflate own achievements while blaming the preceding governments as well as the public for not taking advantage of the opportunities available to them. Very similar political dynamics were found by Dan Banik (1998, 2007) in his investigation of the Kalahandi starvation deaths. Although the starvation deaths were used by the opposition to display the incompetence and callousness of the incumbent government, the consequence of this politicization was not always constructive for famine protection policies. The accusations quickly turned into a struggle of general liability and credibility rather than spurring increased attention to the implementation of famine policies. Even the free press could not exert sufficient pressure on the government because the information was 'contested by the ruling party instead of being reacted upon' (Banik, 2007). Moreover, when the opposition parties came to power, 'they were equally quick to deny allegations of starvation deaths similar to the ones they had made only a few months before coming to office' (Banik, 2007, p. 190). Other than disputes between the government and the opposition parties, Banik also describes how 'in virtually all drought years, there is a set pattern in Center–State disputes – the Center insists that it has sent the money while states claim not to have received any' (Banik, 2007, p. 134). In effect, drought relief has become an arena for playing a political blame game.

More Counterevidence: Malawi, Niger and Cross-Country

India is not the only pluralist regime to have suffered from famine. Just in this millennium, two constitutional democracies both with substantial press freedom and a credible opposition experienced famine – Malawi in 2002 and Niger in 2005.

Malawi 2002

In February and March 2001, floods in Central and southern Malawi cut maize production by around 30 per cent. As maize is the most important staple food covering more than 60 per cent of Malawian daily calorie intake, this production shock generated the large food deficit and food price increase that triggered the famine in 2002 (Devereux, 2000; Kerr, 2005, p. 54; Devereux & Tiba, 2007). Two health assessments produced jointly by different UN agencies (WHO, 2002a, 2002b), from the peak of the famine point to excess mortality rates in the most affected areas. There have been several different estimates of the famine mortality stretching from Howe and Devereux (2004) estimate of 300–500 deaths over Devereux's (2002) guesstimate of 1000–3000 deaths to Devereux and Tiba's (2007) estimate of more than 10 000 deaths. As with most famines, there were also a multitude of factors involved in the Malawi 2002 famine; however, here, the focus will be to identify the deficiencies of the democratic dynamics.

Together with much of the rest of Africa in the beginning of the 1990s, Malawi legalized opposition parties and held competitive, multiparty elections. At the time of the 2002 famine, most scholars found Malawi to be a liberal democracy; indeed, one of the few democratic successes in Africa (Wiseman, 1995; Schmidt, 2001; Van de Walle, 2001; Devereux, 2002; Devereux and Tiba, 2007). Larry Diamond (2000), a seasoned expert in African democracy, despite being generally critical of the second wave of democratization that swept over Africa in the period, argued that a rare combination of civil society coalitions and external pressure had indeed succeeded in generating a transition to democracy in Benin, Mali and Malawi (p. 6).

The situation in Malawi in 2002 appears to offer a case in point for the Malawian saying, Sungadye demokalase, which loosely translated means that you cannot eat democracy. Devereux and Tiba (2007, pp. 167–171) even argue that the political deficiencies should be found in the political system itself; a famine situation in 1992 was averted exactly because Malawi was a benevolent authoritarian regime, whereas the 2002 famine was perpetuated by the democratic institutions, in which interest group politics led to the marginalization of the poor and powerless. Two interrelated political factors could account for the failure of Malawi's democratic institutions in preventing the famine – a straitjacketed pluralistic government and strained donor relations (Rubin, 2008b).

Malawi is a poor and highly indebted developing country with both a population and a government short on cash: 65 per cent of the population survives on less than a dollar a day and more than 40 per cent of Malawi's budget (83 per cent of Malawi's development budget) is dependent on international aid. This has two implications. The first is that a large proportion of the population will have no other option than to rely on public transfers during a crisis situation, and the second is that the government too will have to turn to external funding to provide the needed public transfers. In such a context, donors have substantial influence over policy decisions – especially over policy decisions in the field of poverty alleviation and food security. The Malawi famine shows that if a large proportion of the state budget comes from external sources over which the elected government has little control, then the de facto power lies with the external agencies – not with the elected government. Hence, an elected government without sufficient state capacity could end up being nothing more than an empty shell of democratic practices.

Malawi's strong dependence on foreign exchange inflows and budgetary support proved to be critical in the emergency situation. Because of allegations of Malawian mismanagement of donor funds, relations with donors had cooled down just before the famine, instead mistrust and disrespect prevailed. Kydd and Dorward et al (2002, p. 3) argue, 'this was a major factor in delayed recognition of and response to emerging evidence of a famine crisis.' The IMF withheld 47 million USD that had been promised to the Malawian government due to government overspending. By the end of 2001, the EU not only suspended 13 million USD in aid but it also reclaimed seven million USD that had been given to Malawi due to what the auditing EU delegation tactfully termed as 'anomalies'; the US also diverted six million USD that were meant for Malawi to another country; and Denmark suspended 87 million USD earmarked aid intended to cover a four year period citing a weak administration and a misuse of donor money. Apparently, the USAID even followed a policy dictating that if the recipient government had exported food (as Malawi had done), then the US were prohibited from bringing in emergency food in the same year (Devereux, 2002, p. 15). Of course this does not exonerate the Malawi government from mishandling of the famine but it is indicative of much more complex political dynamics (involving international players) than can be contained in the democracy hypothesis with its focus on national parliamentary dynamics.

Niger 2005

Despite a drop in cereal production of 11 per cent from a 5-year average in 2005 that could have been covered by a mere 3 per cent increase in Niger's normal cereal imports, Niger experienced a food crisis with an excess mortality in the tens of thousands according to mortality estimates from the ground-based NGOs Médecins Sans Frontières and World Vision Canada (Carr, 2005; Moonen and Steffen, 2005; Tectonidis, 2006; detailed mortality calculations in Rubin, 2009b). The famine appears to have been caused by a surge in cereal food prices combined with declining livestock prices that left livestock-dependent households with just a quarter of their purchasing power in 2005 compared to 2004. An MSF survey revealed that nearly 90 per cent of the families in the vulnerable regions reported that they had no remaining food reserves (MSF 13 September).

The two municipal and the presidential elections preceding the 2005 famine were described as generally free and fair. The UN characterized Niger as a 'model democratic state' having held three peaceful elections in a region prone to political conflicts and complex emergencies (OCHA, 2005, p. 3). DFID (2007) appeared to second that description by profiling Niger as a stable country with a well-established multiparty system. Larry Diamond (2003, 2008) also categorized Niger in the group of electoral democracies together with Sen's own exponent for the validity of the democracy theory, India.

The question remains why the famine was not halted in time by the responsiveness of a democratic regime backed with the resources and expertise of the international community. There are many mitigating circumstances surrounding the government's lack of response and flawed interventions. Some of the policies forwarded by the IMF might have perpetuated the situation; the international community reacted belatedly as well; and not only is Niger extremely poor, but the government is highly dependent on external funding in order to intervene.

However, pluralist processes also failed to work according to the democracy hypothesis forwarded by Sen. There is ample evidence that the government did not want to acknowledge the severity of the situation. In a widely quoted and criticized statement, Nigerien president Mamadou Tandja denied that Niger was experiencing famine because, as he told some BBC reporters, 'the people of Niger look well-fed' (BBC, 2005). Other comments signaled apathy and distrust toward the international community. In September 2005, the Prime Minister, Hama Amadou, justified the interruption of food aid out of concern for Nigerien dignity, stating, 'our dignity suffered. And we've seen how people exploit images to pledge aid that never arrives to those who really need it' (quoted in Mousseau and Mittal, 2006, p. 14). These actions and statements do not bear witness to a responsive government that is concerned with the welfare of its citizens (if for no other reason than to be re-elected), but that is forced to passivity, straitjacketed by the international community. In the face of government inaction, the role of watchdog and strong advocate for extensive interventions to mitigate the famine was neither assumed by the opposition nor the domestic media; instead, it was left to MSF to break the silence about the famine. The formal request to intervene more directly to mitigate the impact of the famine came from the UN, not from the government. IRINnews (2005) describes how the UN had to ask the government in July 2005 'to immediately begin the free distribution of food to vulnerable groups to ensure their survival.' According to the democracy hypothesis, the appeal to initiate the distribution of food aid ought to flow from the democratically elected government to the more autonomous humanitarian organization, not the other way around.

Cross-Country

These case studies appear to reject the deterministic proposition of Sen's democracy hypothesis because they strongly suggest that famines have occurred in what could be termed electoral democracies. They also pinpoint some process dynamics that appear to run directly counter to what is expected in the democracy hypothesis; not only has the incumbent government displayed a high degree of passivity but even when pressured to action by opposition parties and a free press, there was no guarantee that the action would actually promote effective famine mitigation. The case studies cannot, however, reject the notion that democracies are better than authoritarian regimes in protecting against famines that relates back to the probabilistic causality. Such proposition can be tested quantitatively using cross-country data. In 2008, I conducted a study addressing the probabilistic causality between democracy and famine by conducting various regressions using different model specifications and proxies (Rubin, 2008a).

In general, there was no robust support for the hypothesis that the political system has a significant effect on the occurrence of famine, as the results varied depending on the proxy for famine and the analytical model applied. My own logistical and Ordinary Least Squares (OLS) regressions using years of famine (based on data from Von Braun et al (1998) and the EM-DAT disaster database, respectively) as well as the number of famine reports from NGOs as the dependent variables found only ambiguous evidence of an impact of democracy (proxied by Polity IV and Political Rights Index). Neumayer and Plümper's (2007) analysis, based on zero-inflated negative binomial models, with famine mortality from EM-DAT as the dependent variable, on the contrary, suggested that democracy did have an impact on famine morality. What the analyses had in common was the difficulty of producing a robust proxy for famine. There are major discrepancies in famine mortalities from one estimate to the next and even inconsistencies with regards to whether or not to count an occurrence as a famine. The EM-DAT database, for instance, surprisingly does not contain data on major famines such as the Chinese 1958–1961 famine or the Bangladesh famine in 1974.

Keeping the data problem in mind some of the more robust results from my own analyses were the following. There was generally a strong correlation between income and famine incidents after controlling for other socioeconomic variables. With substantial multicollinearity, one ought to be cautious about drawing strong conclusions, but it does appear as though being poor matters more than the political system. One of the most robust results from the quantitative analyses was also a strong significance of the African dummy. This tells us that after the effect of other included control variables (proxy for war, GDP per capita and population density), there is still an independent effect just from being an African county.

In two of the model specifications, some significant results involving the political system variable were produced but let me reiterate that the quantitative exercises as a whole did not appear to produce robust correlations between proxies of democracy and famine. Including an interaction term between the political system and educational level (proxying for some sort of democracy deepening) did, however, render the income level insignificant and suggested that democracy combined with a certain level of human capital could lower famine incidents (based on EM-DAT). In another model specification, I restricted the sample to developing countries with a large domestic staple food production, and only included observations in extreme scarcity years (defined as observations in which the domestic production deviates negatively by more than a standard deviation from the mean). Testing for the socioeconomic determinants of variations in imports of staple food could provide indications as to whether democratic governments are more likely to divert fiscal resources to imports of essential foods during times of distress. It turned out that income per capita had a positive and significant effect on food imports. Thus, in cases of extreme food production declines, wealthy countries import a greater share of food to make up for the deficit. The political system (having reversed the Freedom House Index) also had a significant positive impact on the imports in such extreme situations even after controlling for wealth; the more democratic, the higher the food imports in scarcity situations. If indeed the imports of staple food in a food scarcity situation can make up for a crude proxy for government involvement, then this last result provides some support for the hypothesis that democracies could have positive effect on famine mitigation measures.

Theoretical Caveats with the Democracy – Hypothesis

The case studies above have forcefully displayed two limitations with the hypothesis: (i) famines have taken place in electoral democracies; (ii) and in those cases, some of the political dynamics ran counter to the hypothesized process. Even the empirical support for the probabilistic interpretation, that the political system matters in famine protection, was not persuasive. In the following section, the article will briefly point to a few (the list is far from exhaustive) deficiencies with the democracy hypothesis that could account for the lack of empirical support.

Democracies are Diverse

Democracies are not homogenous political entities. With respect to India, the democratic political dynamics have changed substantially from single-party hegemony (by the secular Congress Party), over 21 months of authoritarian emergency rule (imposed by Indira Gandhi) to a bipolar party system (with the Bharatiya Janata Party emerging as a serious contender for power). In addition to changes over time, there are also huge differences in democratic depth from state to state. The state of West Bengal, for instance, has had a communist led government for more than 30 years, whereas neighboring Orissa has had four changes in government coalitions (1980, 1990, 1995 and 2000). Beer and Mitchell (2006) have tabulated a democracy index across the Indian states based on two dimensions, namely electoral competition and electoral participation. The index displays considerable variance across different states with a factor difference of 2.7 between the lowest (Jammu and Kashmir) and the highest scores (Manipur).

Within and between democratic systems – or authoritarian systems for that matter – there are also substantial differences in institutional setup that could have implications for pressure that can be mounted against the incumbent government. A presidential democratic system, for instance, contains more authoritarian traits than a parliamentary democratic system, which, according to Linz and Valenzuela (1994), could increase the risk of conflicts and deadlocks in a crisis situation. Malawi has a full presidential system of government while it is semi-presidential in Niger. The specific electoral dynamics could also play a role. Chhibber and Nooruddin (2004, p. 24) reveal that states with multiparty competition in elections provide less public goods than do states with bipolar competition. Thus, more attention could be directed toward the concrete political arrangements in place in the individual countries. An even better response would be to abandon the focus on systems (whether democratic/authoritarian, presidential/parliamentarian or proportional representation/first-past-the-pole) and examine the political processes.

The Importance of the Domestic Electorate and Press is Eroding

In contemporary famine situations, the political pressure exercised by the electorate might be eroding due to a globalization of the pressure to intervene. The international community is increasingly interfering in matters within sovereign states, and food policies are often shaped by donor influence rather than the internal competition for votes in a pluralistic political system. Both Malawi and Niger are highly dependent on donor resources, meaning that donor ideologies and policies largely set the political agenda. The political debate often assumes the form of a bipolar debate between the international donors and elected governments. Such political dynamic has two consequences: the citizens are cut off the policy loop and the incumbent government can play the role of the watchdog aggressor vis-à-vis the international community. Both undermine the hypothesized mechanism of electoral and opposition pressures.

Famines are not Necessarily Politicized in Democracies

According to Sen, disasters and catastrophes such as famines are easily portrayed in the media. They have news value and provide the reporter with good, strong pictures. This media exposure will in turn create pressure on the government from the public or other political movements to act. From a political economy perspective it appears paradoxical that a government should be made more responsible for their handling of catastrophes, disasters and other extraordinary occurrences, than for their handling of important issues in normal times. This is because even though catastrophes are easier to portray in the media, they could also more easily be blamed on 'force majeure.' Catastrophes might even possibly cement alliances between opposition and government rather than raise criticism and accusations in the media. Furthermore, even if opposition parties did criticize the government for their handling of a famine, the effect of such criticism on voters should not be overestimated. The opposition almost always argues that they could have attained far better policy results than the government in almost any field. Why should they succeed in persuading the voters in this particular case? Like Amartya Sen, Bardhan (2005) also argues that the Indian electorate 'does not in general punish the politicians for the continual scourge of poverty, unemployment, disease and illiteracy which afflict the lives of common people.' What is interesting about Bardhan's observation is that the exception to this general rule is not famine. Rather, it is inflation. According to Bardhan (2005, p. 22), the poor hold the government directly responsible for inflation, and he argues that it will be politically intolerable to have a double-digit inflation rate. Nevertheless, inflation does not possess any of the disastrous traits of a famine, nor does it lend itself well to headlines in the media.

The Strength of Democratic Pressure and Accountability Varies

Accountability is an essential component for democratic pressure. However, it is not always straightforward to identify accountable decision makers in a democratic political system. Such a system can very well consist of a complex web of democratic institutions, many of which might have mutual/overlapping powers of authority. A multilayer democratic system (through a federal structure and/or extensive decentralization) can dilute accountability, thus fostering a political blame game. In India, all three levels of elected authorities had a hand in food security policies: the Central Government, through the major national schemes, for increasing food security; the State Government for the state schemes and distribution within the state; and the local government for implementation and monitoring. The institutions can avoid responsibility by 'passing the hot potato' and blaming other democratic institutions. This is even more the case if – as is increasingly the case in India – competing political parties are in power at the Central and State levels. Even when the same political party controls of both levels, the Bihar case showed how tension could nevertheless compromise effective famine protection. All these factors blur and weaken the relationship, thus reducing the opportunities to hold those responsible accountable.

Politicization Does Not Entail Effective Famine Protection

Even if – for the sake of argument – it is assumed that famines are very high on the political agenda and voters will punish governments failing to prevent famines, there might be many countervailing forces in a democracy. Within the democratic system, therefore, other dynamics must also be taken into account that could be thought to counteract the mechanism described by Sen, thus rendering the net impact of a democratic system on famine protection an empirical question. These factors could include everything from direct lobbying conducted by outside groups to 'log rolling' and 'vote trading' from factions within the political system. Famine relief could be further prevented by 'pork barrel politics' or the corresponding 'not in my backyard' attitude, where locally elected politicians place the interests of their home constituents ahead of the emergency in working for or against initiatives depending on the interests of their own constituents. These 'dark' sides of politicization were evident in both the Bihar and the Orissa food crises where it appeared as though the politicization of famine created many perverse incentives that countered the envisaged mechanism, such as infighting between the State and Central Governments, as well as reluctance to undertake the necessary redistributive measures in an election year.

Even making yet another leap of faith by assuming that the pressure mechanism described by Sen would dominate all other democratic processes that might counter the mechanisms still does not imply that a democratic government should be preferred over an authoritarian regime. Pointing to democratic mechanisms with a positive effect on famine protection does not exclude the possibility that others, even more effective, can be identified under authoritarian rule. The argument about the merits of democracy in famine protection has clear roots in cost-effectiveness reasoning (given the assumed superiority of the democratic political system, what are the processes that could account for effective famine protection?) when one really ought to rely on cost–benefit reasoning (under different political rules, which political processes foster the most effective famine protection?). From the perspective of short-term famine relief, it is not difficult to present arguments that could favor a more authoritarian political system. Some of the counterproductive mechanisms described above (log rolling, vote trading, pork barrel politics, not in my backyard and the political blame game) would be largely absent or assume a different form under authoritarian regimes. Famine relief could be marginalized in democracies if organized interest groups have strong influence on government decisions, and are capable of 'capturing' the state making it serve the interests of urban elites rather than famine vulnerable groups. Mesquita et al (2003) document the tendency for democracies to prefer broad-based policies over targeted schemes, which might hinder effective famine protection as it depends to a large extent on a fast deployment of targeted schemes. It is also possible that authoritarian regimes could manage a much prompter and more extensive mobilization of resources for famine prevention when needed. An elected government might have to engage in compromises and negotiations with other political parties, which might not only slow down the process, but also avert resources to other political purposes through log rolling.

Theoretically, therefore, the democracy hypothesis is not convincing. The mechanism has problems pertaining to both its strength and causality in relation to famine protection. Within the democratic system, other factors must also be taken into account that could be thought to counteract the mechanism described by Sen, thus making the net impact of a democratic system on famine protection an empirical question. And it is a question for which the empirical evidence has not been able to produce an authoritative answer. The question appears to be best addressed on a case-to-case basis.

Conclusion

There is little doubt that Amartya Sen is a true champion of democracy. He declares the struggle for democracy around the world to be 'the most profound challenge of our times' (Sen, 2003). In an article in Journal of Democracy, he refers to the rise of democracy as the most important development in the twentieth century (Sen, 1999).

However, this article has shown that democracy is no panacea against famine; a more holistic approach is desirable to capture the political dynamics of famine. To fully grasp the complexities of famine, the insights stemming from both acknowledging and understanding the political aspects of famine beyond simple dichotomized thinking is of paramount importance. Even assuming that a probabilistic relationship exists between democracy and increased famine protection, the implications of focusing on the democratic/authoritarian dichotomy are the emergence of tunnel vision that ignores other vital political dynamics. Sen rightly argued against the causal theory of FAD, not because FAD was unrelated to famines but because focusing exclusively on FAD would (i) neglect situations in which famine occurs without a FAD as well as (ii) leave out important distributional dynamics in the cases where FAD did occur. When it comes to causal hypotheses at the political level, the argument remains the same – democracies might be related to famine protection (although the empirical evidence is murky at best); but by focusing on the political system, we will (i) overlook situations where famines do take place in democracies; (ii) be ill-equipped to analyze the complex political dynamics involved in famine disasters. In my view, it is somewhat ironic that Amartya Sen, so successful in dismissing the previous causal theories of famines and replacing them with a general analytical framework at the micro level, goes on to forward a deterministic (or at least probabilistic) causal theory of famine on the political macrolevel. Improving our understanding of the political dynamics of famine would entail replacing the democratic-centered causal theory with more general tools for context-specific political analysis. Simple qualitative tools such as Stakeholder Analysis, Institutional Analysis and Political Mapping would be forceful tools in uncovering key stakeholder and agency characteristics, their interests, and the nature and degree of their influence on famine formation and famine prevention policies.

References

1. Banik, D. (1998) India's freedom from famine: the case of Kalahandi. Contemporary South Asia 7(3): 265–281. | Article
2. Banik, D. (2007) Starvation and India's Democracy. London: Routledge.
3. Bardhan, P. (2005) Democracy and Distributive Politics in India. Berkeley Paper, http://emlab.berkeley.edu/users/webfac/bardhan/papers/BardhDemoDist.pdf.
4. Barnett, J. and Adger, W. (2007) Climate change, human security and violent conflict. Political Geography 26(6): 639–655. | Article
5. BBC. (2005) Niger leader denies hunger claims, http://news.bbc.co.uk/2/hi/africa/4133374.stm, accessed 9 August 2005.
6. Beer, C. and Mitchell, N.J. (2006) Comparing nations and states – human rights and democracy in India. Comparative Political studies 39(8): 996–1018. | Article
7. Bowbrick, P. (1986) The causes of famine – a refutation of professor Sen's theory. Food Policy 11(2): 105–124. | Article
8. Brass, P.R. (1986) The political uses of crisis – the Bihar famine of 1966–1967. Journal of Asian Studies 45(2): 245–267. | Article
9. Carr, S. (2005) Assessment of the crude and under-5 mortality rates in two world vision Niger intervention zones: Maradi and Zinder. World Vision, Canada, http://www.humanitarianinfo.org/niger/uploads/w3/nutrition/WV_Maradi.doc.
10. Chhibber, P. and Nooruddin, I. (2004) Do party systems count? The number of parties and government performance in the Indian states. Comparative Political Studies 37(2): 152–187. | Article
11. Currie, B. (1998) Public action and its limits: Re-examining the politics of hunger alleviation in Eastern India. Third World Quarterly 19(5): 873–892. | Article
12. Currie, B. (2000) The Politics of Hunger in India: A Study of Democracy, Governance and Kalahandi's Poverty. Basingstoke: Macmillan.
13. De Mesquita, B., Smith, A., Siverson, M. and Morrow, J. (2003) The Logic of Political Survival. Cambridge, MA: MIT.
14. De Waal, A. (1990) A re-assessment of entitlement theory in the light of recent famines in Africa. Development and Change 21(3): 469–490. | Article
15. Devereux, S. (2000) Famine in the Twentieth Century. Brighton, UK: Institute of Development Studies at the University of Sussex. ISD Working Paper 105, http://www.ids.ac.uk/ids/bookshop/wp/wp105.pdf.
16. Devereux, S. (2001) Sen's entitlement approach: Critiques and counter-critiques. Oxford Development Studies 29(3): 245–263. | Article
17. Devereux, S. (2002) State of disaster – causes, consequences & policy lessons from Malawi. ActionAidMalawi, http://www.actionaid.org.uk/_content/documents/malawifamine.pdf.
18. Devereux, S. and Tiba, Z. (2007) Malawi's first famine, 2001–2002. In: S. Devereux (ed.) The New Famines – Why Famines Persist in an Era of Globalization. Abingdon: Routledge, pp. 143–177.
19. DFID. (2007) Niger Fact Sheet, January, London: DFID, http://www2.dfid.gov.uk/countries/africa/niger-fact-sheet.pdf.
20. Diamond, L. (2000) Developing democracy in Africa: African and international imperatives. Cambridge Review of International Affairs 14(1): 191–213.
21. Diamond, L. (2003) Can the whole world become democratic? Democracy, development, and international policies. Irvine, CA: Center for the Study of Democracy, University of California. Working Paper 0305, http://repositories.cdlib.org/03-05.
22. Diamond, L. (2008) The Spirit of Democracy: The Struggle to Build Free Societies Throughout the World. New York: Times Book.
23. Drèze, J. (1999) Chapter 3 – famine prevention in India. In: J. Drèze, A. Sen and A. Hussain (eds.) The Political Economy of Hunger. Oxford: Oxford University Press.
24. Drèze, J. and Sen, A. (1990) The Political Economy of Hunger. Oxford: Clarendon Press.
25. D'Souza, F. (1994) Democracy as a cure for famine. Journal of Peace Research 31(4): 369–373. | Article
26. Dyson, T. (1993) Demographic responses to famines in South Asia. IDS Bulletin 24(4): 17–31. | Article | PubMed | ChemPort |
27. Dyson, T. and Maharatna, A. (1992) Bihar famine, 1966–67 and Maharashtra drought, 1970–73 – the demographic consequences. Economic and Political Weekly 27(26): 1325–1332.
28. Elahi, K.Q.I. (2006) Entitlement failure and deprivation: A critique of Sen's famine philosophy. Journal of Development Studies 42(4): 541–558. | Article
29. EM-DAT (2009) The International Emergency Disasters Database. WHO Collaborating Centre for Research on the Epidemiology of Disasters, http://www.emdat.be/Database/terms.html.
30. Flyvbjerg, B. (2006) Five misuuderstandings about case-study research. Qualitative Inquiry 12(2): 219–245. | Article
31. Freedom House. (2008) Freedom in the World Country Rankings 1972 – 2006 http://www.freedomhouse.org/uploads/fiw/FIWAllScores.xls.
32. Howe, P. and Devereux, S. (2004) Famine intensity and magnitude scales: A proposal for an instrumental definition of famine. Disasters 28(4): 353–372. | Article | PubMed
33. IRINnews. (2005) Niger: Free food needed now as millions teeter on the brick of famine, UN says, http://www.irinnews.org/report.asp?ReportID=48128 accessed 14 July 2005.
34. Jana, J. (2002) Poverty rules in all polities. The Telegraph, http://www.telegraphindia.com/1021017/asp/opinion/story_1280576.asp, accessed 17 October 2002.
35. Kerr, R.B. (2005) Food security in Northern Malawi: Gender, kinship relations and entitlements in historical context. Journal of Southern African Studies 31(1): 53–74. | Article
36. Kula, E. (1987) The inadequacy of the entitlement approach to explain and remedy famines. The Journal of Development Studies 25(1): 112–116.
37. Kydd, J., Dorward, A. and Vaughan, M. (2002) The Humanitarian Crisis in Southern Africa: Malawi. Submission to the International Development Committee, http://www.sarpn.org.za/documents/d0000392/P346_IDC_Malawi.pdf.
38. Linz, J.J. and Valenzuela, A. (1994) The Failure of Presidential Democracy. Baltimore; London: Johns Hopkins University Press.
39. Massing, M. (2003) Does democracy avert famine?. The New York Times, http://query.nytimes.com/gst/fullpage.html?res=9800E5DD103CF932A35750C0A9659C8B63, accessed 1 March 2003.
40. Menken, J. and Campbell, C. (1992) Age patterns of famine-related mortality increase: Implications for long term population growth. Health Transition Review 2(1): 91–101.
41. Moonen, B. and Steffen, C. (2005) Statut nutritionnel et mortalité retrospective. Geneva: MSF.
42. Mousseau, F. and Mittal, A. (2006) Sahel: A Prisoner of Starvation? A Case Study of the 2005 Food Crisis in Niger. Report from The Oakland Institute. California, http://www.oaklandinstitute.org/pdfs/sahel.pdf.
43. Myhrvold-Hanssen, T. (2003) Democracy, News Media, and Famine Prevention: Amartya Sen and the Bihar Famine of 1966–67. Paper 1–10, http://www.disasterdiplomacy.org/MyhrvoldHanssenBiharFamine.rtf.
44. Neumayer, E. and Plümper, T. (2007) Famine Mortality, Rational Political Inactivity, and International Food Aid. LSE PSPE Working Paper no. 2, http://www.lse.ac.uk/collections/government/PSPE/pdf/PSPE_WP2_07.pdf.
45. Newport, F. (2006) Little impact of Katrina on Bush's overall job ratings. Gallup News Service, http://www.gallup.com/poll/24283/Little-Impact-Katrina-Bushs-Overall-Job-Ratings.aspx.
46. OCHA. (2005) Niger 2005 – Flash Appeal. Geneva: OCHA, http://www.reliefweb.int/rw/RWB.NSF/db900SID/SKAR-6CJAQP?OpenDocument, accessed 23 May 2005.
47. Osmani, S. (1995) The entitlement approach to famine: An assessment. In: K. Basu, P. K. Pattanaik and K. Suzumura (eds.) Choice, Welfare, and Development: A Festschrift in Honour of Amartya K. Sen. Oxford: Oxford University Press.
48. Paldam, M. (2006) P1 radio interview. 1 November (In Danish), http://www.dr.dk/P1/Krause/Udsendelser/2006/10/25115519.htm.
49. Rubin, O. (2008a) Contemporary famine analysis – from causal theories to general frameworks. PhD dissertation 2008/6. Department of Political Science, University of Copenhagen.
50. Rubin, O. (2008b) The Malawi 2002 famine – destitution, democracy and donors. Nordic Journal of African Studies 17(1): 47–65.
51. Rubin, O. (2009a) The entitlement approach – a case for framework development rather than demolition. Journal of Development Studies 45(4): 621–640. | Article
52. Rubin, O. (2009b) The Niger famine – a collapse of entitlements and democratic responsiveness. Journal of Asian and African Studies 44(3): 279–298. | Article
53. Rummel, J. (2007) The Blue Book of Freedom. Nashville, TN: Cumberland House Publishing.
54. Schmidt, S. (2001) Is small beautiful? The significance of decentralization and local self-government in the context of recent processes of democratization in Africa, http://www.kas.za/Publications/SeminarReports/Traditionalparticipation/SCHMIDT.pdf.
55. Sen, A. (1981) Poverty and Famines: An Essay on Entitlements and Deprivation. Oxford: Oxford University Press.
56. Sen, A. (1982) Food battles: Conflicts in the access to food. The 12th Coromandel Lecture; December, New Delhi.
57. Sen, A. (1990) Public action to remedy hunger. Tanco Memorial Lecture: 1–15, http://www.globalpolicy.org/socecon/hunger/general/1990/0802public.htm.
58. Sen, A. (1999) Democracy as a universal value. Journal of Democracy 10(3): 3–17. | Article
59. Sen, A. (2003) Democracy and its global roots. New Republic 229(14): 28–35, http://www.countercurrents.org/eco-sen041003.htm.
60. Sen, A. (2005) The Argumentative Indian: Writings on Indian History, Culture and Identity. London: Allen Lane.
61. Tectonidis, M. (2006) Crisis in Niger – Outpatient Care for Severe Acute Malnutrition. Paris: MSF, http://fieldresearch.msf.org/msf/bitstream/10144/20833/2/243
    _Crisis_in_Niger_outpatient_care_NEJM_2006.pdf, accessed 19 January 2006.
62. Van de Walle, N. (2001) The impact of multi-party politics. Forum for Development Studies 1: 5–42.
63. Varshney, A. (2000) Why have poor democracies not eliminated poverty? Asian Survey 40(5): 718–736. | Article
64. Von Braun, J., Teklu, T. and Webb, P. (1998) Famine in Africa. Baltimore: The Johns Hopkins University Press.
65. Wiseman, J.A. (1995) Democracy and Political Change in Sub-Saharan Africa. London: Routledge.
66. WHO. (2002a) Health assessment mission in Malawi (19 April–10 May 2002), http://www.who.int/disasters/repo/7958.doc.
67. WHO. (2002b) Report on Update of Health Assessment, UNFPA, UNICEF, Ministry of Health and WHO, Malwai September 2002, http://www.who/int/disasters/repo/8416.doc.

Acknowledgements

I wish to extend my gratitude to Erich Kofmel, Managing Director at the Sussex Centre for the Individual and Society (SCIS), for convening the DSA 2008 Panel on Anti-Democratic Development. This article is highly inspired by his somewhat provocative idea of an anti-democratic bias in much of the Development Studies discipline.