Jump to ContentJump to Main Navigation
Hunger and Public Action$

Jean Drèze and Amartya Sen

Print publication date: 1991

Print ISBN-13: 9780198283652

Published to Oxford Scholarship Online: November 2003

DOI: 10.1093/0198283652.001.0001

Show Summary Details
Page of

PRINTED FROM OXFORD SCHOLARSHIP ONLINE (oxford.universitypressscholarship.com). (c) Copyright Oxford University Press, 2020. All Rights Reserved. An individual user may print out a PDF of a single chapter of a monograph in OSO for personal use. date: 22 October 2020

Experiences and Lessons

Experiences and Lessons

Chapter:
(p.122) 8 Experiences and Lessons
Source:
Hunger and Public Action
Author(s):

Jean Drèze (Contributor Webpage)

Amartya Sen (Contributor Webpage)

Publisher:
Oxford University Press
DOI:10.1093/0198283652.003.0008

Abstract and Keywords

Some experiences in India and Africa are examined and lessons drawn from them. The authors show how the entitlement protection efforts of India since independence help explain the lack of famines, and then focus on the 1970 Maharashtra drought. They analyse the success stories of Cape Verde, Kenya, Zimbabwe, and Botswana in Africa, referring again to the importance of entitlement protection systems. Special attention is given to the role of early intervention, private trade, diversification of economic activity and public distribution.

Keywords:   Africa, diversification, early intervention, entitlement protection, famine, India, private trade, public distribution

8.1 The Indian Experience

How has India avoided major famines since independence in 1947? It is tempting to attribute her success in this area to a steady improvement in food production. A close look at the facts, however, quickly reveals the inadequacy of this explanation. Indeed, the period during which the frequency of famines started to decline in India (the first half of this century) was actually one of steadily declining food production per head. Since independence, total food output has grown at a substantial rate, but per‐head food production levels have not increased dramatically. They seem to remain, in fact, lower than late nineteenth‐century levels, and also lower than per‐capita food output levels in many countries affected by famines today. Moreover, the increase in per‐capita production has resulted partly in the reduction of imports (and also to some extent the accumulation of large stocks), so that the net aggregate consumption of food per head has remained remarkably constant for the last forty years. Finally, almost every year large and heavily populated parts of India suffer from devastating droughts or floods which, through the disruption of rural livelihoods, remain quite capable of causing large‐scale starvation.1

Nor is it possible to attribute India's success in preventing famines to a significant improvement in the general prosperity of the rural population. The removal of poverty in rural India since independence has, in fact, been shamefully slow.2

The prevention of famines in India cannot be understood without reference to the extensive entitlement protection efforts that have come into play on numerous occasion to sustain the rural population through a crisis. At the risk of oversimplification, it may be said that entitlement protection in India relies on the operation of two complementary forces, viz. (1) an administrative system that is intelligently aimed at recreating lost entitlements (caused by droughts, floods, economic slumps, or whatever), and (2) a political system that acts as the prime mover in getting the administrative system to work as and when required. We shall concentrate first on the administrative aspects, but we (p.123) shall argue later that the administrative structure can be non‐operational and ineffective in the absence of a political triggering mechanism.

The administrative aspects of the system can be traced to ideas championed—and to a limited extent used—in Indian history.3 Kautilya, in the fourth century BC, spoke of employment creation and redistribution to the poor as parts of a sound administrative system to defeat famines. Various Indian rulers (such as Mohammad bin Tughlak in the fourteenth century) made extensive use of work projects and income creation for rebuilding lost entitlements.

As far as systematic exploration and exposition are concerned, the administrative analysis goes back, in many respects, to the detailed recommendations of the Famine Commission of 1880. The policy of governmental inaction that dominated British imperial administration in the early and middle nineteenth century gradually gave way to selective intervention concentrating mainly on the regeneration of cash incomes.

Among the more important of these recommendations of the Famine Commission of 1880 were (1) the framing of region‐specific ‘Famine Codes’ embodying ‘authoritative guidelines’ to the local administration on the measures needed to anticipate and deal with the threat of famine, and (2) a strategy of entitlement protection based on the combination of guaranteed employment at a subsistence wage and unconditional relief (so‐called ‘gratuitous relief’) for the unemployable. The reasoning behind this strategy is explained with great clarity in the Famine Commission Report, which is worth quoting at some length:

. . . we have to consider the manner in which the proper recipients of the public charity can be most effectually ascertained. The problem to be solved is how to avoid the risk of indiscriminate and demoralising profusion on the one hand, and of insufficient and niggardly assistance on the other—how to relieve all who really need relief, and to waste as little public money as possible in the process . . . where limited numbers have to be dealt with, and there is a numerous and efficient staff of officials, it may be possible to ascertain by personal inquiry the circumstances of every applicant for relief sufficiently for the purpose of admitting or rejecting his claim. But in an Indian famine the Government has to deal not with limited numbers, but with millions of people, and the official machinery at its command, however strengthened for the occasion, will inevitably be inadequate to the task of accurately testing the individual necessities of so great a multitude. Nor again is it possible to entrust the administration of public charity to a subordinate agency without providing sufficient checks against dishonesty and neglect on the part of its members. Some safeguards then are essential in the interests of the destitute people no less than of the public treasury, and they are best found in laying down certain broad self‐acting tests by which necessity may be proved, and which may, irrespective of any other rule of selection, entitle to relief the person who submits to them . . . The chief of these tests, and the only one which in our opinion it is ordinarily (p.124) desirable to enforce, is the demand of labour commensurate in each case with the labourer's powers, in return for a wage sufficient for the purposes of maintenance but not more. This system is applicable of course only to those from whom labour can reasonably be required . . . The great bulk of the applicants for relief being thus provided for, we believe that it will be possible for an efficient staff of officers to control with success the grant of relief, on the basis of personal inquiry and knowledge of the individual circumstances of each applicant, among the comparatively small numbers of destitute persons to whom the test of labour cannot be applied.4

Employment in public works was typically remunerated with cash wages. The expectation was that the demands generated by these wage payments would be met by the operation of private trade.

These broad principles are still relevant to the conception of India's entitlement protection system today. In particular the continued power of the strategy of employment generation supplemented by ‘gratuitous relief’ for the unemployable has been apparent in a number of experiences of successful famine prevention since independence.5 At the same time, it would be a mistake to regard this system as a mere legacy of the British Administration. In fact, important advances have been made in famine prevention strategies since independence, even if we confine our attention to the administrative part of the story. One of the most important post‐independence changes relates to the public distribution system.

The British Indian administration considered governmental involvement in food trade or distribution as sacrilegious. This position was grounded on a particular understanding of the teachings of classical economists (especially Adam Smith, John Stuart Mill, and Thomas Malthus), which were sometimes referred to as the ‘Infallible Laws of the Great Masters of Economic Science’.6 Suspicion of government interference with private trade extended also to any kind of public participation in food distribution.

While the government of independent India has, by and large, refrained from directly interfering with the activities of private traders, its own involvement in food trade and distribution has been extensive and important. In order (p.125) to assess correctly the contribution of India's public distribution system to famine prevention, we need to distinguish between its function of price stabilization and income generation.

The stability of food prices in India today is quite remarkable (especially in comparison with the pre‐independence period). For instance, during the most recent drought (that of 1987–8), which led to a considerable decline in food production, foodgrain prices increased by less than 10 per cent.7 This was largely due to large‐scale sales of food through the public distribution system, which held very large stocks at the beginning of the drought.8 Clearly, the contribution of price stabilization measures to the protection of entitlements during crises in India today is a major one.

On the other hand, the income generation aspect of the public distribution of food has often been exaggerated. In per‐capita terms, the subsidies involved in food sales to the rural population in so‐called ‘fair price shops’ are very small indeed, even at times of drought.9 In fact, in most states of India subsidized sales of food have a very limited coverage in rural areas. The really important vehicle of income generation in India during crises is that of large‐scale public works, most frequently for cash wages. The irreplaceable role played by public works in sustaining the purchasing power of the rural poor is illustrated by the case‐study of the Maharashtra drought of 1970–3 in the next section.

The distinction between the price stabilization role and the income generation role of the public distribution system has some bearing on the relevance of the Indian experience for other countries. The possible relevance of this experience for African countries, in particular, has often been dismissed on the grounds that (1) subsidized food sales to the rural population are the cornerstone of entitlement protection in India, and (2) most African countries lack a comparable infrastructure. This assessment is quite misleading. The generation of income through public works is not dependent on subsidized food sales. Adding to the incomes of the victims helps to prevent destitution, even without distributing cheap food. Furthermore, the price stabilization objective of the public distribution system can be pursued with substantial effect through wholesale food operations, and need not be dependent on the public network of ‘fair price shops’.

(p.126) While developments in the administrative aspects of entitlement protection policies since independence have been important, the really crucial changes have taken place in the domain of politics. The existence of the Famine Codes did not, after all, ensure their application, let alone their early and energetic application. The Famine Codes did include very specific instructions on how to recognize and ‘declare’ a famine, and the duty they imposed on the administrative structure made it harder to ignore a threatening crisis. However, this ‘early warning system’ existed within the Famine Codes, and could not guarantee their use in practice. Indeed, the problem of triggering remained an important one, since the Famine Codes did not impose any legal obligation to ‘declare’ a famine. During the Bengal famine of 1943, for instance, the Famine Codes were never invoked and were deliberately ignored, and this fact may well have been responsible for a large part of the extraordinary excess mortality associated with that famine.10

After independence, the political incentives to recognize emergencies, and to take action against the threat of famine, had to assume a new form. The vigour of political opposition has now made it impossible for the government to remain passive without major political risks, and the fear of losing elections reinforces the general sensitivity to political embarrassment in the state assembly and in the central parliament. In the process of making the facts known and forcing the hands of the respective state and central governments, the press too plays a leading role. The affected populations themselves have a much greater ability than in the past to make their demands felt and to galvanize the authorities into action (especially in view of the importance of winning the rural vote). This is one of the positive aspects of Indian democracy. We shall return to this question for fuller assessment in Chapter 11, when we reconsider the respective achievements of India and China in recent decades in combating hunger and famines.11

8.2 A Case‐Study: The Maharashtra Drought of 1970–197312

The state of Maharashtra in Western India, which in the early 1970s had a population of a little over 50 million, is one of the more ‘developed’ among Indian states by a number of conventional aggregate indicators (including (p.127) literacy, urbanization, life expectancy, and average income). However, the divide between urban and rural areas in Maharashtra is very sharp, and the proportion of the rural population below the ‘poverty line’ in this state is among the highest in India.13 Even within the rural sector, there are enormous (p.128) regional inequalities in living standards, and in the semi‐arid parts of the state the precariousness of life is particularly acute.

Table 8.1 District‐Wise Cereal Production in Maharashtra, 1967–1973

District

Index of Cereal Production (1967–8 = 100)

Cereal production per capita, 1972–3 (kg./year)

1969–70

1970–1

1971–2

1972–3

Greater Bombay

77

81

54

31

Thana

88

110

97

42

46

Kolaba

78

101

81

67

131

Ratnagiri

99

117

103

86

85

Nasik

81

107

55

26

32

Dhulia

106

119

74

49

54

Jalgaon

89

74

59

70

72

Ahmednagar

109

80

59

33

47

Poona

90

70

73

43

38

Satara

98

103

91

41

45

Sangli

90

86

90

18

20

Sholapur

92

51

63

18

27

Kolhapur

93

110

115

65

53

Aurangabad

89

74

48

20

31

Parbhani

76

54

42

41

66

Bhir

120

97

54

17

27

Nanded

77

36

48

29

51

Osmanabad

108

54

58

45

61

Buldhana

122

68

82

63

86

Akola

132

55

89

61

64

Amravati

103

61

68

79

62

Yeotmal

131

65

104

85

86

Wardha

97

59

73

68

80

Nagpur

96

71

76

67

49

Bhandara

121

139

114

58

92

Chandrapur

129

109

105

71

118

MAHARASHTRA

99

83

74

47

51

Source: Calculated from the Annual Season and Crop Reports (Government of Maharashtra) of the corresponding years. Per‐capita production figures for 1972–3 (last column) are based on district‐wise population estimates (for 1973) obtained by assuming identical 1973/1971 population ratios for each District; the all‐Maharashtra 1973/1971 population ratio is taken from the Bulletin on Food Statistics, 1982–4. District‐wise population estimates for 1971 are from the Census (as given in Brahme 1983: 13–14.)

At the time of the onset of the terrible drought of 1970–3, rural Maharashtra was facing an alarming problem of environmental degradation, agricultural decline and threatened rural livelihoods, in many ways similar to the crisis faced by a number of African countries today. The sustained downward trend in per‐capita agricultural and food production, which went back at least to the early 1960s, turned into a disastrous crash in the early 1970s, when the larger part of the state was affected by a drought of exceptional intensity for three years in succession. The statistics of food production for that period show a decline which, in terms of rapidity and magnitude, finds few equivalents in the recent history of droughts and famines elsewhere (see Table 8.1, Figure 8.1, and also Table 5.1 in Chapter 5 for an instructive comparison). During the peak year of the drought in 1972–3, the per‐capita production of cereals in the state was as low as 51 kg.—less than one‐third of the average level (itself very low) of per‐capita consumption for India as a whole.

Experiences and Lessons

Fig. 8.1 Production of cereals per capita, 1961–80: India, Maharashtra and Sahel

This acute crisis of food and agricultural production, and of the rural economy in general (with a virtual collapse of private employment and incomes for large sections of the population for a prolonged period), represented a considerable threat of large‐scale entitlement failures. In spite of this, there are good indications that the sufferings caused by the drought, while far from negligible, were remarkably confined. There is, indeed, very little evidence that any of the usual signs of famine developed to a significant extent in this (p.129) event—whether ‘starvation deaths’, increases in mortality rates, nutritional deterioration, land sales, or migration to other states.14

Table 8.2 Earnings from Relief Works and Total Income in Seventy Drought‐Affected Villages, Maharashtra, 1972–1973

Percentage contribution of earnings on relief works to total income

Number of villages

0.0–20.0

7

20.1–40.0

8

40.1–50.0

9

50.1–60.0

10

60.1–70.0

14

70.1–80.0

15

80.1–90.0

6

90.1–100.0

1

TOTAL

70

Source: Brahme (1983: 59). The villages were located in the districts of Poona, Ahmednagar, Sholapur, Aurangabad, Bhir, and Osmanabad (all severely drought‐affected).

How was famine averted? While it would be difficult to give a complete answer to this question, a major part of the story undoubtedly relates to public policies of entitlement protection. The cornerstone of these policies was the generation of employment for cash wages on a large scale, supplemented by ‘gratuitous relief’ for those unable to work and without able‐bodied relatives. At the peak of the crisis, nearly 5 million labourers were employed on public works throughout the state. During the twelve months preceding July 1973 (the peak year of drought), relief works generated nearly one billion persondays of employment.15 In the more severely drought‐affected districts, the contribution of wage income from employment on public works to total income was well above 50 per cent for most villages (see Table 8.2).

One of the effects of this massive programme of income generation in drought‐affected areas was to attract food from other parts of the country into Maharashtra through the channel of private trade. In theory, the drought period was one when severe restrictions were imposed on inter‐state movements of food within India. But large amounts of foodgrains were imported into Maharashtra from neighbouring states in spite of these restrictions. This was an essential part of the mechanism of famine prevention, since the efforts made directly by the government to restore normal levels of food availability in (p.130) Maharashtra (through public sales) fell far short of expectations.16 In fact, had the trade restrictions been effectively enforced with an unchanged amount of food imported through the public distribution system, the aggregate consumption of foodgrains in the state would have been reduced by as much as 40 per cent. It would have been difficult to prevent a decline in average consumption (p.131) of this magnitude (for a population as large as that of Maharashtra) from developing into a famine of major proportions.

In spite of large imports of food on both public and private accounts, a noticeable decline in average food consumption took place in Maharashtra in 1972–3. A striking feature of consumption patterns during that year, however, is that the reduction of aggregate consumption compared to ordinary years was distributed remarkably evenly among different socio‐economic groups (see Table 8.3 and Figure 8.2). This phenomenon reflects the sustained purchasing power of vulnerable groups resulting from large‐scale employment provision, and their ability to battle for their normal share of the available food in spite of the general penury. Correspondingly, the reduction of consumption on the part of relatively privileged groups resulted from the reduction of private money incomes, the increase of food prices, the desire of many asset owners to preserve their productive capital, and their reluctance to join relief works.17

Table 8.3 Cereal Consumption in Maharashtra: 1972–1973 Compared to ‘Normal’ Years

Household class

Cereal consumption per capita (kg./month)

Percentage distribution of households by monthly per capita consumption class (kg./month)

<12

12–15

>15

Total

Large cultivators

1967–8

15.6

25.5

28.7

45.8

100.0

1972–3

12.8

44.7

30.0

25.3

100.0

1973–4

15.3

n/a

n/a

n/a

100.0

Small cultivators

1967–8

13.4

53.0

19.0

28.0

100.0

1972–3

11.1

61.2

25.2

13.6

100.0

1973–4

12.9

n/a

n/a

n/a

100.0

Farm labourers

1967–8

14.5

41.4

19.9

38.6

100.0

1972–3

11.5

60.9

25.3

13.8

100.0

1973–4

13.7

n/a

n/a

n/a

100.0

Industrial workers

1967–8

13.2

42.8

21.7

35.5

100.0

1972–3

12.0

65.3

27.1

7.6

100.0

1973–4

13.3

n/a

n/a

n/a

100.0

Others

1967–8

12.4

n/a

n/a

n/a

100.0

1972–3

10.8

n/a

n/a

n/a

100.0

1973–4

12.1

n/a

n/a

n/a

100.0

All households

1967–8

14.0

38.5

22.5

39.0

100.0

1972–3

11.7

59.9

25.2

15.9

100.0

1973–4

13.9

37.2

26.7

36.1

100.0

Source: Drèze (1988a), Tables 3.7 and 3.8 1967–8 and 1973–4 were years of fairly normal harvest, and were the nearest (to 1972–3) ‘normal’ years for which data are available. The data cover the rural areas of ten drought‐affected districts.

Experiences and Lessons

Fig. 8.2 Cereal consumption in Maharashtra: drought year (1972–3) compared with normal years, by household class

(p.132) As was discussed in the previous chapter, a crucial condition for the success of a redistributive strategy of this type is that preferential support should be given to vulnerable groups. An indication (among others) of the success achieved by the strategy of employment provision in reaching vulnerable groups and areas in this event can be obtained from considering the allocation of relief between different districts within Maharashtra. As can be seen from Figure 8.3, in 1972–3 a striking positive association was observable across districts between the intensity of the drought (as measured by the extent of crop failures) and the extensiveness of entitlement protection measures (as indicated by the proportion of the rural population employed on public works). This association is quite impressive considering the well‐known difficulties that entitlement protection efforts often encounter in reaching the right people in famine situations.18

Experiences and Lessons

Fig. 8.3 Drought intensity and public relief in rural Maharashtra (1972–3) by district

The success of a strategy based on large‐scale cash support also depends, as discussed in Chapter 6, on the ability of the relief system to avoid the danger of large sections of the vulnerable population being left out of the purview of public support and at the same time remaining exposed to the price increases caused by the enhanced purchasing power of the others. This is where the (p.133) principle of guaranteed employment—imperfectly but usefully applied—played a crucial role.

The political factors involved in the response of the government, discussed in general terms in the previous section, are well illustrated by this particular event. The drought situation was the subject of 696 questions in the Maharashtra Legislative Assembly and Council in 1973 alone, and of numerous (often sharply critical) newspaper reports.19 Much of the credit for galvanizing the government into action also belongs to the affected populations themselves, which pressed their demands in numerous ways—including marches, pickets and rallies. As one labourer aptly put it, ‘they would let us die if they thought we would not make a noise about it’.20

8.3 Some African Successes21

In Chapter 5 of this book, we argued that African successes in famine prevention have failed to receive the attention they deserve. Of course, ultimately our concern has to be also with identifying mistakes and failures rather than taking idle comfort in past successes. However, there is a great deal to learn from the way in which various countries have successfully coped with the threat of entitlement failures. While the Indian experience is rich in general lessons, a number of recent successes with famine prevention in Africa are themselves of great interest. This section is devoted to a brief description and analysis of some of these successes. Their implications and lessons will be further discussed in the concluding section to this chapter.

Cape Verde

In his distinguished history of the Cape Verde Islands, Antonio Carreira wrote that ‘everything in these islands combines to impose on man a hard, difficult and wretched way of life’.22 A prominent aspect of the harshness of life in Cape Verde is the recurrence of devastating droughts, which have regularly affected the islands ever since their ‘discovery’ in 1460 by the Portuguese. Many of these droughts were associated with large‐scale famine.23 (p.134)

In fact, it is hard to think of many famines in history that have taken a toll in human life proportionately as high as those which have periodically decimated Cape Verde in the last few centuries. Some of these famines are believed to have killed nearly half of the population (see Table 8.4). Even after assuming some exaggeration in these figures, there are very few parallels of such wholesale mortality in the long and terrible history of famines in the world.

Table 8.4 History of Famine Mortality in Cape Verde, 1750–1950

Mortality attributed to famine (percentage of total population)

1773–6

44

1830–3

42

1863–6

40

1900–3

15

1920–2

16

1940–3

15

1946–8

18

Source: Moran (1982), Table 1. The famines indicated here are only those for which an estimate of famine mortality is provided in that table. For the same period, the author mentions 22 further large‐scale famines for which no mortality estimates are available.

These historical famines went almost entirely unrelieved. When one of the very few exceptions to this pattern occurred in 1825, the governor of the islands was sacked for using Crown taxes to feed the people.24 Left to its own devices, the population had little other refuge than the attempt to emigrate—often encouraged by the colonial authorities. Cape Verde's history of persistent migration is indeed intimately connected with the succession of famines on the islands. However, for most people this option remained a severely limited one, and as recently as the 1940s large‐scale mortality was a predictable feature of prolonged drought.

In recent years, Cape Verde may well have been the worst drought‐affected of all African countries. Indeed, uninterrupted drought crippled the country's economy for almost twenty years between 1968 and 1986—leading to a virtual extinction of domestic food supplies and a near standstill of rural activity.25 (p.135) Half‐way through this prolonged drought in the middle 1970s, the event was already described as ‘the longest and most severe [drought] on record’ for the country.26 In this case, however, not only was famine averted but, even more strikingly, significant improvements in living conditions took place during the drought period. The causation of these improvements is the centre of our interest here.

Famine prevention measures greatly gained in motivation and execution after the independence of Cape Verde in 1975. But even in the first period of the drought preceding independence (1968–75), the Portuguese rulers did make considerable efforts to provide famine relief (in contrast with the experiences earlier on in the colonial period).27 Relief was provided almost exclusively in the form of employment for cash wages in makeshift work (the adequacy of food supplies being ensured separately by food imports). According to one study, as much as 84 per cent of total employment was provided by drought relief programmes in 1970 (this, however, still left 55.5 per cent of the labour force unemployed).28

These preventive measures succeeded to a great extent in averting a severe famine. There were no reports of large‐scale starvation deaths, and the overall increase in mortality seems to have been moderate. The estimated infant mortality rate, for instance, which had shot up to more than 500 per thousand during the famine of 1947–8, was only a little above the 1962–7 average of 93.5 per thousand in the period 1968–75 (Table 8.5). On the other hand, a significant intensification of undernutrition during the same period has been reported in several studies.29

Table 8.5 Infant Mortality in Cape Verde, 1912–1986

Year

Estimated infant mortality rate (deaths per 1000)

(1)

(2)

1912

220.6

1913

174.2

1915

117.9

1920

155.0

1927

217.6

1931

206.7

1937

223.4

1943

317.9

1946

268.7

1947a

542.9

1948a

428.6

1949

203.9

1950

130.7

1962

106.1

1963

109.7

1964

85.3

1965

76.7

1966

83.6

1967

99.9

1968

91.7

1969

123.1

1970

95.0

1971

130.9

1972

90.9

1973

110.6

1974

78.9

1975

103.9

104.9

1980–5

77.0

1985

70.0

1986

65.0

(a) Famine years.

Source: For the period 1912–75 (column 1), Freeman et al.] (1978), Table V.26 (very close estimates are also reported for the 1969–74 period in CILSS 1976, Table VI). For 1975–86 (column 2), World Health Statistics Annual 1985 and UNICEF (1987a, 1988).

Since independence in 1975, Cape Verde has been ruled by a single party with a socialist orientation, viz. the Partido Africano da Independencia da Cabo Verde (PAICV).30 This party, described by the current Prime Minister as ‘reformist, progressivist and nationalist’,31 is flanked by the Popular National Assembly, which is elected every five years by popular ballot (within a single party system). The government of independent Cape Verde has been consistently credited with progressive social reforms and development programmes. Notable areas of improvement have been those of education and health. Drought relief has been among the top political priorities.

Cape Verde's entitlement protection system since independence has (p.136) (p.137) consisted of three integrated components.32 First, a competent and planned use of food aid has ensured an adequate and predictable food supply in spite of the nearly total collapse of domestic production. Food aid is legally bound to be sold wholesale in the open market, and the proceeds accrue to the National Development Fund.33

Second, the resources of the National Development Fund are used for labour‐intensive public works programmes with a ‘development’ orientation. In 1983, 29.2 per cent of the labour force was employed in such programmes.34 The works undertaken include afforestation, soil conservation, irrigation, and road building, and according to a recent evaluation ‘the results of these projects are positive, even on the basis of high standards’.35

Third, unconditional relief is provided to selected and particularly vulnerable groups such as pregnant women, undernourished children, the elderly, and the invalid. This part of the entitlement protection system includes both nutritional intervention (such as school feeding) and cash transfers, and is integrated with related aspects of formal social security measures. In 1983, direct food assistance covered 14 per cent of the population (see van Binsbergen 1986: 10).

The effectiveness of this fairly comprehensive and well‐integrated entitlement protection system is visible from the impact of the drought after 1975.36 Indeed, the adverse effects of the drought on the living conditions of human beings seem to have been remarkably small.37 In addition to the successful prevention of famine, there are indications that the post‐1975 part of the drought period has witnessed: (1) a rapid decline in the infant mortality rate (see Table 8.5); (2) a significant increase in per‐head food intake;38 and (3) a significant improvement in the nutritional status of children (see Table 8.6).39 By any criterion, the success achieved by the government of independent Cape (p.138) Verde in protecting the population from the adverse effects of a drought of unprecedented magnitude must be seen as exemplary.

Table 8.6 Child Undernutrition in Cape Verde, 1977 and 1984

District

Percentage of school children suffering from undernutrition (moderate to serious)

1977a

1984b

Boa Vista

41.8

7.8

Porto Novo

49.2

9.2

Ribeira Grande

54.3

5.8

São Vicente

38.1

10.7

Tarrafal

n/a

7.8

TOTAL

46.4

8.8

(a) Children aged 7–15 years.

(b) Children aged 6–18 years.

Source: van Binsbergen (1986), Table 2. According to the author, the two studies on which this table is based are ‘reasonably comparable’, and ‘although the methodologies used by the different studies were not identical, it is safe to conclude that the nutritional status of school age children has significantly improved since 1977’ (van Binsbergen 1986: 3–4). An independent study carried out in 1973 estimated that 38% of children aged 7–14 suffered from ‘moderate protein‐calorie malnutrition’ (Freeman et al. 1978, Table V.24).

Kenya

The history of Kenya, like that of Cape Verde, has been repeatedly marked by grim episodes of drought and famine.40 As recently as 1980–1, famine struck substantial parts of the population in the wake of a drought of moderate intensity. The government of Kenya has been widely praised, however, for preventing a much more widespread and intense drought from developing into a famine in 1984. This event has been extensively studied elsewhere.41 We shall only recall here the main features of this successful response, and comment briefly on some of its neglected aspects.

Like Cape Verde, Kenya has a single party system and an elected parliament. Since independence in 1963, the country has enjoyed a degree of political stability which compares favourably with many other parts of Africa. The freedom of the national media is limited, but nevertheless more extensive than in most African countries. The country also has a high degree of visibility in the international press.

More than 80 per cent of Kenya's population (around 19 million in 1984) is (p.139) rural, and derives its livelihood largely from agriculture and livestock. Compared to most other parts of Africa, the rural economy is quite diversified, and has experienced relatively rapid growth since the early 1960s. Despite these favourable factors, large parts of the country remain vulnerable to climatic and economic instability, particularly in the largely semi‐arid areas of the Eastern and North‐eastern Provinces.

The strategies adopted by rural households in Kenya to cope with drought or the threat of famine appear to be increasingly geared to the acquisition of food in the market and the diversification of economic activities (partly through wage employment).42 The importance of off‐farm activities in the rural household economy can be seen from the fact that, according to a survey carried out in six districts of the Central and Eastern Provinces in 1985, more than half of smallholder households had at least one member in long‐term wage employment (see Table 8.7 below).

Table 8.7 The 1984 Drought and Smallholder Households in Central and Eastern Kenya

Characteristic

Percentage of households with the specified characteristic, by ecological zone

1

2

3

4

5

All zones

Household member moved during 1984

23

7

21

26

38

25

Has a member in permanent employment

58

47

61

54

54

56

Received cash remittances from relatives or friends during drought

34

28

40

57

46

43

Major food changed during 1984

84

78

76

67

67

73

Received famine relief (from govt or NGOs)

14

35

25

67

77

45

Slaughtered, sold, lost or consumed cattlea

41

45

33

44

32

38

(26)

(35)

(29)

(46)

(51)

(58)

(a) In brackets, the percentage decrease in cattle holding, averaged over all households surveyed in the respective zone.

Source: Anyango et al. (forthcoming), Tables 13.6 and 13.9. Based on survey data collected in January 1985 by the Central Bureau of Statistics on behalf of the National Environment Secretariat. In Kenya, a smallholder is ‘typically defined as a rural landowner with less than 22 hectares’ (Akong'a and Downing 1987: 92). Ecological zones appear in increasing order of drought‐proneness, based on rainfall data.

The 1984 crisis followed a massive failure of the ‘long rains’ in March and April 1984. According to Cohen and Lewis:

It was the worst shortage of rains in the last 100 years. Production of maize, the nation's principal food crop, was approximately 50% below that normally expected for the main rains of March–May. Wheat, the second most important grain, was nearly 70% below normal. Potato production was down by more than 70%. Pastoralists reported losing up to 70% of their stock. The situation had the potential for a famine of major proportions.43

Regional disparities compound the problems reflected in these aggregate statistics. In the Central and Eastern Provinces, maize production for the agricultural year 1984–5 was estimated by the FAO at 14 per cent and 26 per cent (respectively) of the average for the previous six years. In districts such as Kitui and Machakos, maize production was virtually nil both in 1983 and in 1984.44

While in specific areas the drought of 1984 meant the second or even third consecutive crop failure, for most areas the crisis was one of limited duration. The ‘short rains’ of October to December 1984 were above average. However, in terms of intensity and geographical coverage the drought of 1984 was certainly an exceptional one, and distress continued until the harvest of mid‐1985.

The use of formal early warning techniques apparently played little role in precipitating action. The need for action seems to have been detected partly from the visible failure of rains in early 1984 (followed by evident crop failures), and partly from the unusual increase in food purchases from the (p.140) National Cereals and Produce Board later in the year.45 While Cohen and Lewis stress the role of ‘political commitment’ in ensuring an early and adequate response, others comment that ‘the government felt the need to forestall political instability that would result in the event of a widespread famine’.46 The threat of political unrest seems to have been exacerbated by the fact that, somewhat unusually, the drought of 1984 strongly affected a number of politically important and influential areas of the Central and Eastern Provinces, as well as Nairobi.

Active public response to the crisis began in April 1984.47 The first step taken by the government to deal with the threat of famine was to import large amounts of food on a commercial basis. The initial availability of substantial food stocks ensured that the lags involved in the importation of food did not have disastrous consequences. Food aid pledges were also obtained, but with a few minor exceptions their fulfilment occurred only in 1985, several months after the arrival of commercial food imports. The ability of the government to buy large amounts of food on the international market was greatly helped by the availability of foreign exchange reserves and the peak in export earnings resulting largely from high world prices for tea and coffee.

Entitlement protection measures took two different forms. First, the government used food imports to ensure the continued availability of food at reasonable prices through normal commercial channels. In ordinary times, interdistrict food movements are exclusively organized by the National Cereals and Produce Board, which subcontracts the transport and distribution of food to licensed private traders. This arrangement was preserved and intensified during the drought, and most of the imported food was sold through the intermediation of private traders at ‘gazetted’ prices fixed by the government.

Second, direct support was provided to vulnerable households in affected areas. Initially, the government intended to provide such support mainly in the form of employment for cash wages.48 In practice, however, the generation of employment fell far short of target, due to a lack of preparedness and supervisory capacity. On the other hand, the provision of unconditional relief in the form of free food rations (mainly from food aid) assumed considerable importance. In August 1984, nearly 1.4 million people, or 7 per cent of the total (p.141) population, were estimated to be in receipt of free food distribution, and in January 1985 a very similar estimate was reported.49 In drought‐affected areas, the proportion of the population receiving food rations was much larger, and the survey of smallholders in Central and Eastern Kenya mentioned earlier found that over the same period the proportion of households receiving food assistance in the surveyed districts was as high as 45 per cent (see Table 8.7). The size of the rations distributed, however, appears to have been very small prior to the large‐scale arrival of food aid in 1985.50

The allocation of relief to the needy was the responsibility of the provincial administration, which itself relied on local famine relief committees and ‘chiefs’ to identify those in need of support. The precise way in which this system actually worked is far from clear. According to some, the local chiefs ‘knew the needs of their people, and by most reports did an effective, equitable job of distributing the government‐supplied grain’.51 Another account, however, states that ‘moving in the path of least resistance, the GOK [Government of Kenya] would seem to rather divide the available food equally among recipients at the distributions thus defusing potentially uncomfortable situations’.52

It is not implausible that the allocation of food within specific communities was largely indiscriminate, and that ‘targeting’ operated mainly between different villages or regions (the impact of the drought varied greatly between different areas). On the other hand, an important factor facilitating the fair allocation of free food was the fact that most of it consisted of yellow maize, which is generally considered as an ‘inferior’ commodity in Kenya. The element of ‘self‐selection’ involved in distributing yellow maize has been said by a number of commentators to have contributed to an allocation more geared to the most desperate.

Table 8.7 presents some indicators of the impact of the drought on the rural population in different ecological zones of Central and Eastern Kenya, arranged in increasing order of drought‐proneness. The table brings out, inter alia, (1) the role played by wage employment and remittances in sustaining affected households, (2) the responsiveness of consumption patterns to price and income changes (the composition of food consumption changed for most households even in the less affected areas), and (3) the extensive nature of the (p.142) coverage of food distribution in these districts at that time.53 The large cattle losses also confirm the exceptional severity of the drought.54

The overall effect of the drought on the well‐being of the affected populations has not been fully ascertained. Most commentators consider that ‘famine was averted’. The apparent absence of confirmed reports of ‘starvation deaths’, as well as of distress migration on the part of entire families, lends some support to this view. On the other hand, there is clear evidence of widespread hunger as well as rising undernutrition in 1984.55

(p.143)

(p.144)

While the credit which the Kenyan government has been given for averting a severe famine in 1984 appears to be largely deserved, an important query can nevertheless be raised as to whether the strategy it adopted made good use of available opportunities. A particularly relevant issue concerns the balance between income support and price stabilization measures. Considering the small size of food distribution to vulnerable households in per‐capita terms (at least until 1985), it appears that famine prevention measures were geared to work mainly through the level of food prices rather than through the generation of compensating incomes.56 In turn, the stability of prices was pursued through a policy of commercial imports from abroad into the worst affected districts. At the same time, however, government regulations prevented private traders from moving food from surplus to deficit areas within Kenya.

As was mentioned earlier, interdistrict food movements in Kenya are tightly regulated by the National Cereals and Produce Board, which subcontracts food transport to licensed private traders. Several studies have shown that these restrictions on interregional movements have the effect of exacerbating the intensity of local shortages and the disparity of retail food prices between (p.145) regions.57 This phenomenon was clearly visible during the drought year itself: while food prices were sharply rising in drought‐affected districts, they were only sluggishly increasing or even falling in many others (see Figure 8.4 and Table 8.8). Maize prices in different markets varied, at one point, by a factor of nearly ten.58 Even between adjacent districts, price disparities seem to have been exceptionally large (see Table 8.9).

Experiences and Lessons

Fig. 8.4 Minimum and maximum retail price of maize in Kenya, 1983 and 1984

Table 8.8 Retail Prices of Maize in Kenya, 1984

Market

Jan. 1984 (Kshs/kg.)

Nov. 1984 (Kshs/kg.)

Percentage increase

Province

Kalundu

2.50

9.31

272

Eastern

Mwingi

2.50

9.00

260

Eastern

Kiambu

2.14

5.30

148

Central

Machakos

2.40

5.76

140

Eastern

Iciara

2.53

5.94

135

Eastern

Limuru

2.02

4.19

107

Central

Runyenjes

2.57

4.70

83

Eastern

Thika

2.86

4.91

72

Central

Kandara

2.84

4.76

68

Central

Embu Town

2.92

4.89

67

Eastern

Eldoret

2.00

2.92

46

Rift Valley

Kitale

1.51

2.11

40

Rift Valley

Bondo

2.50

3.33

33

Nyanza

Ahero

2.51

3.19

27

Nyanza

Sondu

2.31

2.73

18

Nyanza

Mumias

2.35

2.69

14

Western

Luanda

2.71

2.57

−5

Western

Source: Republic of Kenya, Central Bureau of Statistics, Ministry of Finance and Planning, Market Information Bulletin (Jan.–June 1984 and July–Dec. 1984 issues). The markets in the table are all those for which data are provided in the Bulletin for both months.

Table 8.9 Maize Prices in Central and Eastern Kenya, 1984

Zone

Price of white maize (Kshs/kg.)

Percentage Increase (Jan.–Dec.)

Jan.–Mar.

Oct.–Dec.

1

3.98

5.94

49

2

3.25

5.50

69

3

2.80

6.20

121

4

2.94

7.22

146

5

3.49

10.24

193

Note: The ecological zones are the same as in Table 8.7.

Source: Anyango et al. (forthcoming), Table 13.10.

The detrimental effects of this policy of trade restrictions on the entitlements of vulnerable groups in drought‐affected areas are not difficult to guess. For instance, after stressing the role of food shortages and high prices in undermining the entitlements of poor households in the Samburu District of Northern Kenya, Louise Sperling comments:

[The] problem of local distribution was sufficiently severe to result in the convening of a district‐level meeting as early as 14th June 1984. The District Commissioner called together the eleven or twelve wholesalers to discuss ‘the erratic supply of commodities’. Maize prices are strictly controlled by the state, and the local traders claimed they were losing money on maize sales. The allowed mark‐up could not cover the costs of transport and loading to these more remote areas. Even considerable government pressure to encourage traders to keep their shelves full did not result in an increase in the local availability of maize . . . Again, the poor disproportionately suffered from these shortages. They could not afford to buy grain in bulk when it did arrive. Equally, they did not have the means to purchase alternative, more costly foodstuffs.59

(p.146) To conclude, while the efforts made by the Government of Kenya in 1984 to import food into drought‐affected areas well ahead of large‐scale famine were no doubt remarkable, it may be that certain aspects of the relief programme are yet to receive adequately critical scrutiny. To some extent, the acute need to rush food from abroad into the worst affected regions was a result of the parallel efforts that were made, partly for political reasons, to prevent food exports from surplus areas, or to direct such exports towards Nairobi. It appears that, in some respects, government intervention during the drought was undoing with one hand the harm it had done with the other.

Zimbabwe

The so‐called ‘Zimbabwean miracle’ in food production has received wide attention recently. By contrast, the impressive and largely successful programmes of direct entitlement protection pursued by the Zimbabwean government to prevent the prolonged drought of 1982–4 from precipitating a major famine seem to have been relatively neglected.60 While it is tempting to think that a country with growing food supplies cannot possibly know the threat of famine, the experience of famines all over the world shows how misleading and dangerous this assumption can be. In the case of Zimbabwe too, a closer examination of the facts reveals that the prevention of a famine in 1982–4 must be attributed as much to far‐reaching measures of public support in favour of affected populations as to the growth of food supplies.

The political system in Zimbabwe is that of a limited multi‐party democracy. Since independence in 1980, the country has been ruled by the elected and re‐elected ZANU (Zimbabwe African National Union) party led by Robert Mugabe. A notable feature of ZANU is its very wide and largely rural support base, inherited from the independence struggle. Political debate is intense in Zimbabwe. The press is relatively unconstrained and one of the most active in Africa. The press played, in fact, a conspicuous role in keeping the government on its toes throughout the drought period.61

In spite of the socialist aims of the government, the economy has retained private ownership and market incentives. On the other hand, the government of independent Zimbabwe has carried out a major revolution in the area of (p.147) social services. The great strides made since 1980 in the areas of health and nutrition have, in particular, received wide recognition.62

In comparison with most other African countries, Zimbabwe's economy (including the agricultural sector) is relatively prosperous and diversified. However, the heritage of the colonial period also includes massive economic and social inequalities. The agricultural sector is highly dualistic, the larger part of the more fertile land being cultivated by a small number of commercial farms while peasant production remains the dominant feature of ‘communal areas’. Even within the communal areas, sharp regional contrasts exist both in productive potential and in access to infrastructural support.63 Further divisions exist between racial and class groups as well as between rural and urban areas. As a result, large sections of the population live in acute poverty in (p.148) spite of the relative prosperity of the economy as a whole. At the time of independence, health and nutrition problems in Zimbabwe were extremely serious, even in comparison with other African countries.64

The ‘production miracle’ in Zimbabwe is of relatively recent origin. In fact, from the early 1970s until after the drought period, there was—to say the least—little evidence of any upward trend in food production per capita (see Figure 8.5). On the other hand, it must be remembered that, over the same period, Zimbabwe remained a net exporter of food in most years. The plentiful harvest which immediately preceded the drought ensured that large stocks of maize were available when the country faced the threat of famine.65

Experiences and Lessons

Fig. 8.5. Annual maize production per capita in Zimbabwe, 1970–84

The drought lasted three years, and was of highest intensity in agro‐climatic terms during the second year (i.e. 1983). In the drier parts of the country, the maize crop (Zimbabwe's principal staple crop) was ‘a total failure throughout the drought years’.66 Maize sales to the Grain Marketing Board fell by more than two‐thirds between 1980–1 and 1982–3. Livestock losses between 1978 and 1983 have been estimated at 36 per cent of the communal herd.67

For many rural dwellers, remittances from relatives involved in regular employment or migrant labour were a crucial source of support during the drought. As in the Sahel and in Kenya (discussed earlier), it was found in Zimbabwe that ‘the households most engaged in selling labour to the wider economy . . . are the least susceptible to drought’.68 Many households, however, did not have access to this broader source of sustenance, and for them government relief was often the main or even the only source of food.

The drought relief programme of the government was an ambitious and far‐reaching one. Famine prevention measures were taken early in 1982, and given a high political and financial priority throughout the drought. The main entitlement protection measures were large‐scale food distribution to the adult (p.149) population, and supplementary feeding for children under five. Commenting on the importance of free food distribution for the survival of the poor, one study of drought relief in Southern Zimbabwe comments that ‘for those without access to cash and other entitlements it was their only food intake’.69

It is not easy to assess how large a part of the population benefited from free food distribution. Estimates of 2 to 3 million people being fed in rural areas at the peak of the programme, as against a total rural population of 5.7 million in 1982, have been cited in various studies.70 A survey of 464 households carried out in four communal areas selected for their environmental diversity found that more than 50 per cent of the surveyed households were receiving free maize in 1982–3 and 1983–4.71 Another study, focusing on four villages in one of the most affected districts, reveals a proportion of population in receipt of free food rations ranging from 54 to 68 per cent (see Table 8.10). Very similar findings are reported in a number of further household surveys.72 While the precise extent of food distribution is difficult to ascertain, its scale was undoubtedly impressive. The size of individual food rations—officially 20 kg. of maize per head per month—was also astonishingly large.

Table 8.10 Drought and Drought Relief in Chibi District, Zimbabwe, 1983–1984

Village

Number of households

Percentage of cattle which died in previous twelve months

Percentage of population receiving food rationsa

A

52

28

62

B

36

34

64

C

44

32

68

D

39

47

54

(a) Per capita rations of 20 kg. per month.

Source: Constructed from Leys (1986). According to the author, Chibi District was one of the worst affected districts.

Of course, the task of organizing food distribution on such a gigantic scale was by no means an easy one. While the implementation of relief measures, (p.150) much helped by the popular mobilization and political stability associated with the post‐independence reconstruction efforts, has attracted favourable comments from many observers, frequent complaints on the part of recipients about the delays, uncertainties, and frauds involved in food distribution have also been reported.73 In relieving logistic constraints, the subcontracting of food delivery to the private sector played a major role. In fact, disruptions in food deliveries seem to have intensified after the attempt was made (in September 1983) to substitute government transport for private transport.74

The free distribution of food raised its own problems. The population eligible for food rations seems to have been confined in practice to households without a member in regular employment.75 How fairly this criterion was applied in practice is not easy to ascertain, and conflicting views have been expressed on this question. For instance, one author reports on the basis of extensive field work in Southern Zimbabwe that ‘as far as I could assess, these criteria were applied fairly and, at the sub‐district level, were felt to be fair’, and the evidence presented from four village studies in one of the worst‐affected districts broadly supports this assessment.76 But another author suggests the possibility that ‘in practice, the distribution pattern was indiscriminate; those who were ineligible received relief food, while those who were truly needy may have gone short’.77

The politicization of food distribution during the drought was also apparent in a number of ways—with both negative and positive implications. First, party cadres have played a major role in many places in implementing the distribution of food, and this seems to have led to some favouritism along party lines.78 Second, the coverage of the drought relief programme in Matabeleland, the stronghold of political dissidents, has been described as ‘exceedingly patchy’.79 Third, food distribution was restricted to rural areas—a highly interesting feature of the relief programme given the frequent bias of public distribution systems in favour of urban classes. It is tempting to interpret this (p.151) ‘rural bias’ as a reflection of the politics of ZANU and the predominantly rural character of its support base.

In spite of these reservations, the overall effectiveness of entitlement protection measures during the drought is beyond question. It is not only that ‘starvation deaths’ have been largely and perhaps even entirely prevented.80 The striking effect of the government's far‐reaching relief programme, in combination with the general expansion of health and education facilities since independence, has been a noticeable improvement in the health status of the population of rural Zimbabwe in spite of the severe drought. The most striking aspect of this improvement has been the apparent decline in infant mortality throughout the drought period.81 A significant decline in child morbidity, at least in relation to immunizable diseases, has also been reported, and related to the government's vigorous immunization campaigns.82

The evidence on the nutritional status of the population during the drought is mixed. Concern about sharply rising levels of undernutrition in the early phases of the drought has been expressed in many informal reports.83 There is, however, some evidence of declining undernutrition after the relief programme expanded on a large scale in 1983 (see Table 8.11).84 Taking the drought period as a whole, the available evidence suggests the absence of (p.152) marked change in the nutritional status of the Zimbabwean population.85 This is remarkable enough, given the severity of the initial threat.

Table 8.11 Nutritional Status of Children in Zimbabwe, 1981–2 and 1983

Area

Percentage of children (aged 0 to 5) suffering from second or third degree malnutrition

Weight for age

Weight for height

1981–2

1983

1981–2

1983

Commercial farming area

42

14,20a

16

8, 7a

Communal areab

20

11

13

3

Mine area

22

9

6

4

Urban area

6

4

6

2

(a) These figures refer respectively to (1) farms benefiting from a health project initiated in 1981–2, and (2) farms excluded from the health project.

(b) Resurvey area adjacent to that of baseline survey.

Source: Loewenson (1986), Table 3. Based on a sample survey of nearly 2,000 children in Mashonaland Central. For further details and discussion, see also Loewenson (1984).

Botswana

As a land‐locked, sparsely populated, and drought‐prone country experiencing rapid population growth, massive ecological deterioration, and shrinking food production, Botswana possesses many of the features that are thought to make the Sahelian countries highly vulnerable to famine. There are, of course, also important contrasts between the two regions. One of them arises from the highly democratic nature of Botswana's political regime, and—relatedly perhaps—the comparative efficiency of its administration. Also, while many Sahelian countries have suffered from declining or stagnating per‐capita incomes in recent decades, Botswana has enjoyed a growth rate which is estimated as one of the highest in the world.

Economic growth in Botswana has, however, followed a highly uneven pattern. In fact, much of this rapid growth has to do with the recent expansion of diamond mining, a productive sector of little direct relevance to the rural poor. Against a background of booming earnings in industry, there is some evidence of increasing rural unemployment and falling rural incomes since the early 1970s. One study goes as far as suggesting that rural incomes in Botswana (inclusive of transfers and remittances) declined in real terms at the rate of 5 per cent per year during the period 1974–81.86 The year 1981–2 marked the beginning of a prolonged and severe drought which lasted until 1986–7, and would certainly have been accompanied by an even sharper deterioration of income and employment opportunities in the absence of vigorous public support measures. Fast overall economic growth is no guarantee of protection against famine.

The rural economy, mostly based on livestock, crop production, and derived activities, suffered a predictable recession during the drought. The output of food crops fell to very low levels (Table 8.12). Cattle mortality increased substantially, and the decline of employment opportunities further aggravated the deterioration of rural livelihoods.87 In a socio‐economic survey (p.153) of 284 rural households carried out in 1984, more than half of the respondents reported ‘having no cash income’ (other than relief income).88

Table 8.12 Food Crop Performance in Botswana, 1968–1984

Year

Area planted (000 hectares)

Yield (kg./hect.)

Output (000 tonnes)

1968

200

180

36

1969

240

258

62

1970

202

69

14

1971

246

293

72

1972

251

343

86

1973

139

101

14

1974

255

290

74

1975

250

284

71

1976

261

295

77

1977

255

290

74

1978

260

192

50

1979

160

62

10

1980

268

172

46

1981

274

201

55

1982

193

89

17

1983

226

63

14

1984

197

36

7

Source: Hay et al. (1986), Tables 5 and 6.

By 1981–2, however, Botswana had set up an entitlement protection system exemplary in its scope and integration. This system was, in fact, the outcome of a long process of experimentation, evaluation and learning. Moderately successful but instructive experiments with famine relief in the 1960s and early 1970s were later followed by a series of evaluations and debates which provided the crucial foundation of Botswana's remarkable relief system.89 The drought of 1979–80 played a particularly important role in this respect.

Famine relief during the 1979–80 drought was essentially an experiment in what we have called the ‘strategy of direct delivery’.90 The operation was considerably hampered by logistic difficulties connected with the transportation and distribution of food, though a noticeable improvement occurred after (p.154) the adoption of extensive subcontracting to private truckers. Food deliveries in different parts of the country matched poorly with the extent of distress. The allocation of food within the rural population was largely indiscriminate, partly because the selective distribution of food was found to be ‘socially divisive’.91 While a large‐scale famine was averted, the relief operations did not succeed in preventing increased malnutrition, excess mortality, or even starvation deaths.92

The lessons of this experiment were not lost, however. In fact, the detailed evaluation carried out by Gooch and MacDonald made a crucial contribution to the design of Botswana's entitlement protection system as it exists today. Their recommendations included (1) the issue of a Relief Manual providing clear and coherent advance guidelines to the administration about the provision of drought relief, and (2) the adoption of a famine prevention strategy based on the unlimited provision of employment (for a subsistence wage paid in cash) to the able‐bodied supplemented by unconditional relief for vulnerable groups. These recommendations, while not literally implemented to this day, have provided the basis for a sustained improvement in famine prevention measures.

Careful planning (and buoyant government revenue) would not have gone far enough in the absence of a strong motivation on the part of the government to respond to the threat of famine. Drought relief, however, has consistently been a high political priority in Botswana, and an object of rival promises and actions on the part of competing parties. It is also interesting that, when drought struck the country again in 1981–2, early action was forthcoming in spite of the absence of a formal early warning system.93 As in India, the politics of famine prevention in Botswana are intimately linked with the accountability of the ruling party to the electorate, the activism of opposition parties, the vigilance of the press, and—last but not least—the strong demands for public support on the part of the affected populations.94

The drought of 1982–7 provided a severe test of the country's growing ability to prevent famines. The entitlement protection measures invoked in this event involved three major areas of action: (1) the restoration of adequate food availability, (2) the large‐scale provision of employment for cash wages, and (3) direct food distribution to selected groups.95 (p.155)

Unlike in 1979–80, the restoration of food adequacy in 1982–7 relied on a more varied and discerning strategy than that of direct delivery. While Botswana did receive large amounts of food aid during the drought, the support of incomes through employment generation (financed out of general government revenue) was not tied to the receipt of food aid. Moreover, food aid was substantially complemented by private imports of food from abroad, and it is not implausible that had food aid been interrupted or delayed this alternative source of food supply would have enabled the relief system to operate with no major loss of effectiveness.96

Trade and distribution within the country has been largely ensured by (p.156) Botswana's ‘widespread and highly competitive retail network operating in all but the remoter areas’.97 The effectiveness of this system, and of the process of spatial arbitrage, is visible from the remarkable degree of uniformity in the level of food prices in different parts of the country during the drought (Table 8.13. The contrast with our earlier findings on Kenya is striking.

Table 8.13 Price of Maize Meal at Selected Centres, Botswana 1980–1983

Region

Price in August 1980 (Pula/bag)

Price in April 1983 (Pula/bag)

Percentage increase

Gaborone

3.39

4.51

33

Francistown

3.56

4.41

24

Lobatse

3.34

4.48

34

Selibe‐Pikwe

3.33

4.71

41

Palapye

3.36

4.54

35

Mahalapye

3.48

4.86

40

Mochudi

3.41

4.55

33

Kanye

3.41

4.58

34

Serowe

3.56

4.80

35

Molepolole

3.70

4.81

30

Maun

4.22

5.38

27

Mmadinare

3.73

4.70

26

Tonota

3.59

4.68

30

Shoshong

3.38

4.82

43

Moshupa

3.38

5.02

49

Thamaga

3.35

4.97

48

Ramotswa

3.84

4.73

23

All regions, unweighted average

3.53

4.74

34

Source: Tabor (1983), Table 4.5.

Another important contrast between drought relief in 1982–7 and in 1979–80 has been the much greater reliance, in the former case, on cash‐based employment generation as a vehicle of income generation. The provision of employment has, in fact, fallen short of the vision of ‘employment guarantee’ contemplated by Gooch and MacDonald, and it has been repeatedly observed that the demand for employment has exceeded the number of jobs available.98 Nevertheless, the extent of income support provided to vulnerable households by ‘Labour‐Based Relief Programmes’ has been considerable. In 1985–6 they provided around 3 million person‐days of employment to 74,000 labourers. It has also been estimated that Labour‐Based Relief Programmes ‘replaced’ almost one‐third of rural incomes lost from crop failures between 1983 and 1985.99 Informal evaluations of the productive value of the works undertaken suggest that the contribution of these programmes to national investment has been far from negligible.100

Along with this strategy of employment generation, free food has been distributed on a large scale, mainly in the form of ‘take‐home’ rations. The eligibility conditions for food distribution in various forms in rural areas are very broad, and include not only the destitutes but also other categories such as all pre‐school children, all children in primary school, children aged 6–10 not attending school, and all pregnant or lactating women. As a result, the proportion of the total population in receipt of free food rations was as high as two‐thirds in 1985101

The experience of drought relief in Botswana in 1982–7 amply demonstrates the effectiveness of a famine prevention system based on the combination of adequate political incentives and insightful administrative guidelines. While the drought of 1982–7 was far more prolonged and severe than that of 1979–80, and led to a much greater disruption of the rural economy, the extent of human suffering was comparatively small. There is no significant evidence (p.157) of starvation deaths, or of distress migration on any significant scale.102 The nutritional status of children only deteriorated marginally and temporarily (see Figure 8.6). One study also reports that ‘those who have experienced previous droughts say that the decline in suffering among the disadvantaged is dramatic’.103 Last but not least, drought relief measures in Botswana seem to have met with an impressive measure of success not only in preventing human suffering but also in preserving the productive potential of the rural economy.104

Experiences and Lessons

Fig. 8.6 Incidence of child undernutrition in Botswana, 1980–6

(p.158) There is another aspect of Botswana's experience which deserves special mention here. A number of components of the drought relief programme, such as the distribution of food to certain vulnerable groups, the rehabilitation of undernourished children, and the provision of financial assistance to the destitute, have acquired a permanent status and are now an integral part of Botswana's social security system.105 In the future, therefore, it can be expected that famine prevention measures will perhaps take the form of an intensification of social security measures applying in ordinary times. Such a policy development would be a natural extension of the current reliance on existing infrastructural and institutional arrangements for drought relief purposes. This approach to the protection of entitlements during crises has, in general, much to commend, in terms of administrative flexibility, likelihood of early response, simplification of logistics, and ability to elicit broad political support.

8.4 Lessons from African Successes

The African experiences analysed in the previous section illustrate the rich variety of political, social and economic problems involved in the protection of entitlements in a crisis situation. It would be stupid to attempt to derive from these case studies a mechanical blueprint for famine prevention in Africa. However, a number of commonalities involved in the recent experiences of famine prevention in Botswana, Cape Verde, Kenya and Zimbabwe provide, along with the general analyses of previous chapters, the basis of some useful lessons.

The Importance of Entitlement Protection Systems

There is a tendency, once the dust of an emergency has settled down, to seek the reduction of famine vulnerability primarily in enhanced economic growth, or the revival of the rural economy, or the diversification of economic activities. The potential contribution of greater prosperity, if it involves vulnerable groups, cannot be denied. At the same time, it is important to recognize that, no matter how fast they grow, countries where a large part of the population derive their livelihood from uncertain sources cannot hope to avert famines without specialized entitlement protection mechanisms involving direct public intervention. Rapid growth of the economy in Botswana, or of the agricultural sector in Kenya, or of food production in Zimbabwe, explain at best only a small part of their success in averting recurrent threats of famine. The real achievements of these countries (as well as of Cape Verde) lie in having provided direct public support to their populations in times of crisis.

(p.159) Initiative and Agency

An important feature of recent famine prevention efforts in the four countries studied in the previous section is that, in each case, the initiative and responsibility of entitlement protection efforts rested squarely with the government of the affected country. This is not to say that international agencies played no positive part in such efforts. In fact, their contribution and partnership has, in each case, been helpful. But the essential tasks of coordination and leadership belonged primarily to the government and administration of the affected countries.

The general ‘comparative advantage’ that the governments of affected countries have in managing relief operations can indeed be important. This comparative advantage mainly takes the form of being able to draw at short notice on extensive networks of information, administration, communication, transport and storage. In a long term perspective, a sustainable and efficient system of famine prevention can hardly dispense with the close involvement and leadership of the governments of the concerned countries themselves. This makes it particularly important to see the current contribution of international agencies as cooperation with, rather than replacement of, the efforts of the respective governments.

Early Warning and Early Response

Formal ‘early warning’ techniques have played only a minor role in the famine prevention experiences studied in the preceding section. Early response has been much more a matter of political incentives and motivation than one of informational or predictive wizardry.

As was mentioned earlier, the political systems of the four countries concerned are, in comparison with most other African countries, relatively open and pluralist (e.g. they all have an elected parliament). All except possibly Cape Verde also have an active and largely uncensored press. The role of political opposition, parliamentary debate, public criticism and investigative journalism in galvanizing the national government into action has been central in Botswana as well as in Zimbabwe, and, to a lesser extent, in Kenya.

This does not mean that only countries with highly developed participatory institutions can consistently avert the threat of famine. Even fairly repressive governments are often wary of the prospects of popular discontent in the event of a famine. Political ideology—if it takes the form of a commitment to the more deprived sections of the population—can be another creative force in motivating response. In Cape Verde and Zimbabwe, this influence seems to have been important. As was pointed out in Chapter 5, the attitude of African governments to the threat of famine is not in general one of apathy and callousness.

Food Supply Management

In each of the four countries studied in the previous section, the government took necessary steps to ensure an adequate availability of food. But the exact (p.160) nature of these steps varied a great deal, and appealed to different strategic elements such as government purchases on international markets, private trade, food aid and the depletion of public stocks.

This strategic diversity contrasts with the common belief that food aid is the only appropriate channel to enhance food availability in a famine‐affected country. It is true that three of these four countries (namely, Botswana, Cape Verde and Kenya) have made use of substantial quantities of food aid in their efforts to avert famine, but in no case have their entitlement protection measures been significantly contingent on the timely arrival of food aid. In fact, entitlement protection policies have typically preceded the arrival of food aid pledged in response to the threat of famine. In this respect, entitlement protection measures in these countries have markedly departed from the strategy of ‘direct delivery’ discussed in Chapter 6.

A particularly significant departure from the strategy of direct delivery is the use of cash support to protect the entitlements of vulnerable groups. The general merits of this approach were discussed at some length in Chapter 6. It is worth noting that reliance on cash support, which is sometimes thought to be highly unsuitable in the context of African famines, has been used with excellent effect in two of the four countries concerned (Botswana and Cape Verde).

Private Trade and Public Distribution

Each of the four African countries discussed earlier have induced private trade to supplement the efforts of the public sector in moving food towards vulnerable areas. In Botswana and Cape Verde, this has taken the form of providing cash support to vulnerable groups on a large scale and leaving a substantial part of the task of food delivery to the market mechanism. In Kenya and Zimbabwe, it has taken the form of subcontracting to private traders the transport of food to specific destinations. In each case, private trade could be confidently expected to move food in the right direction, i.e. towards (rather than out of) affected areas.

At the same time, the direct involvement of the public sector in food supply management has also been substantial in each country. The benefits of this involvement were visible not only in terms of its direct effects on the flow of food, but also in the noticeable absence of collusive practices or panic hoarding in the private sector itself. The sharp contrast between the behaviour of food prices in Kenya and Botswana during recent droughts, discussed in sections 6.3 and 8.3, strongly suggests that the positive involvement of the public sector in food supply management is often a far more creative form of intervention than the imposition of negative restrictions on the operation of private trade.

Diversification and Employment

As a final observation, we should note the prominent role played by the diversification of economic activities (notably through wage employment), and (p.161) the acquisition of food on the market, in the survival strategies of vulnerable groups in the countries studied in the previous section. As was discussed in Chapter 5, this observation is in line with a general assessment of survival strategies in Africa. Two of its implications are worth emphasizing.

First, while current problems of famine vulnerability in Africa clearly originate in part from the stagnation or decline of food production in that continent (leading to major losses of income and employment for the rural population), it does not follow that the remedy of this vulnerability must necessarily take the form of reversing that historical trend against all odds. Diversification and exchange have been an important part of the economic opportunities of rural populations in Africa for a long time, and open up alternative avenues of action that also need to be considered.

Second, there are strong reasons to think that the potential of employment provision as a tool of entitlement protection (e.g. in the form of public works programmes) is substantial in large parts of Africa. The general advantages of the strategy of employment provision (notably making possible the use of self‐selection, and also the provision of cash support) were discussed in Chapter 7. The fact that affected populations positively look for work in crisis situations, and do this long before reaching an advanced stage of destitution, strengthens the case for seeing this strategy as a natural avenue of entitlement protection. In many circumstances, the spade is a more powerful tool of famine prevention than the spoon. (p.162)

Notes:

(1) For an examination of the evidence supporting this assessment of post‐independence trends in India, see Drèze (1988a).

(2) This question is further discussed in Chapter 11. On recent trends in the incidence of poverty in rural India, see e.g. Dandekar and Rath (1971), Bhatty (1974), Minhas (1974), Srinivasan and Bardhan (1974, 1988), Ahluwalia (1978), Dutta (1978), Gaiha and Kazmi (1981), Sundaram and Tendulkar (1981), Bardhan (1984), Gaiha (1987, 1988), Minhas et al. (1987), Sagar (1988), Sanyal (1988), and Subbarao (1989).

(3) On the history of famines and famine prevention in India, see Dutt (1900, 1901), Loveday (1914), Bhatia (1967), Srivastava (1968), Ambirajan (1978), Jaiswal (1978), McAlpin (1983a, 1983b), Brennan (1984), Klein (1984), and Drèze (1988a), among others.

(4) Government of India (1880: 36). In practice the British Administration did not, in fact, resist the temptation of providing ‘niggardly assistance’. For instance, the failure of the relief system to prevent dramatic increases in mortality during the two famines which occurred in India at the very end of the 19th century has been partly attributed by several authors to the inadequate nature of relief measures (Bhatia 1967; Klein 1984; Guz 1987; Drèze 1988a). The stinginess of public provision declined markedly after independence, though one may question how far the government of independent India has really departed from the earlier colonial view that ‘while the duty of Government is to save life, it is not bound to maintain the labouring community at its normal level of comfort’ (circular of the Government of India No. 44F, 9th June 1883), however low that ‘level of comfort’ is in the first place.

(5) See Singh (1975), Subramaniam (1975), Desai et al. (1979), McAlpin (1987), Drèze (1988a), and Chen (1989) for some case studies. It must be emphasized that the nature and effectiveness of the famine prevention system vary considerably between different regions of India. Some of these interregional contrasts are studied in Rangasami (1974) and Drèze (1988a).

(6) Etheridge (1868: 3). On the influence exercised by the teachings of Classical economists on many prominent administrators in British India, see Ambirajan (1971, 1978) and Rashid (1980).

(7) On this see Kumar (1988). According to the author, the prices of wheat, rice and ‘all cereals’ increased respectively by 7.3 per cent, 7.4 per cent and 7 per cent between 1986–7 and 1987–8 (p. 26). The decline in per‐capita production of foodgrains during the 1987–8 agricultural year was of the order of 13 per cent compared to the 1984–6 average (calculated from Government of India 1989, Table 1.16). A little more than one century earlier, the Famine Commission of 1880 had boldly pronounced that ‘in time of very great scarcity, prices of food grain rise to three times their ordinary amount’ (Government of India 1880: 27).

(8) Public food stocks amounted to 23.6 million tonnes in January 1987, and declined to 10 million tonnes during the drought period (Kumar 1988: 13).

(9) Also, in spite of the considerable expertise which India has acquired in the field of food logistics and distribution, the public distribution system routinely falls way below target in its supply of food to the rural population during droughts. See e.g. Bhatia (1988), Drèze (1988a), and Harriss (1988b).

(10) See Sen (1981a), Greenough (1982) and Brennan (1988) for further discussion of the non‐declaration of famine in that event.

(11) For further discussion of the role of public pressure in famine prevention in India, see also Sen (1982b, 1982c, 1983a), Ram (1986) and Drèze (1988a).

(12) This section is largely based on Drèze (1988a). The interested reader is referred to that paper for further discussion, as well as for a more detailed examination of the empirical evidence presented here. On the Maharashtra drought of 1970–3, see also Ladejinski (1973), Government of Maharashtra (1973), Krishnamachari et al. (1974), Kulkarni (1974), Mundle (1974), Borkar and Nadkarni (1975), Jodha (1975), Mathur and Bhattacharya (1975), Subramaniam (1975), Oughton (1982), Brahme (1983), and various contributions to the Economic and Political Weekly from 1971 to 1974.

(13) See e.g. Vaidyanathan (1987).

(14) See Drèze (1988a) for further discussion. The evidence on mortality rates is discussed in greater detail in the final version of that paper (to appear in Drèze and Sen, forthcoming), without major changes in the conclusions.

(15) Calculated from official figures given by Subramaniam (1975), Table 11.3 (viii), based on attendance on the last day of each month.

(16) The fact that the purchasing power generated by employment programmes had a dramatic impact on private food movements, and stimulated them in the right direction, does not imply that the public distribution itself played no important role in this episode of averted famine (or that it played a role that could have been easily supplanted by private trade). This should be clear from our earlier discussion of the interaction between private trade and public distribution (Chapter 6).

(17) A decline in average consumption, along with a strikingly even reduction of intake for different socio‐economic groups, is also observed by Marty Chen in her study of the impact of the 1987–8 drought on a village of Gujarat (Chen 1989). In fact, in that study the occupational group which experienced by far the smallest percentage decline in food consumption during the drought is found to be that of ‘labourers’ (Table 39). A completely different pattern has been found in the case of less well handled crises, such as the Bihar ‘near‐famine’ of 1967, when the brunt of hardship was overwhelmingly borne by landless labourers (see Drèze 1988a).

(18) The difficulties and dilemmas involved in the ‘selection problem’ and the advantages of ‘self‐selection’ were discussed in Chapter 7.

(19) A useful guide to some of the newspaper reports on the drought can be found in Luthra and Srinivas (1976).

(20) Cited in Mody (1972: 2483). Many vivid accounts of popular protests during the drought can be found in the columns of Economic and Political Weekly (see e.g. Mody 1972, Anon. 1973, and Patil 1973). The demands involved often went beyond the mere provision of employment. For instance, in May 1973, a strike by 1.5 million labourers working on relief sites forced the government to grant them an increase in wages.

(21) We are extremely grateful to John Borton, Diana Callear, Jane Corbett, Rob Davies, Thomas Downing, Carl Eicher, Charles Harvey, Roger Hay, Judith Heyer, Francis Idachaba, Renée Loewenson, Siddhartha Mitter, S. T. W. Mhiribidi, Richard Morgan, Christopher Murray, David Sanders, Luc Spyckerelle, Samuel Wangwe, and Daniel Weiner for helpful comments, suggestions, and personal communications relating to the case‐studies appearing in this section.

(22) Carreira (1982: 15). According to several analysts the climate of Cape Verde is even harsher and the droughts visiting it more frequent and severe than those of other Sahelian countries. See e.g. Meintel (1984: 56).

(23) For a chronology of droughts and famines in Cape Verde, see Freeman et al.] (1978). For further discussion of famines in the history of Cape Verde, see Cabral (1980), Carreira (1982), Moran (1982), Meintel (1983, 1984), and Legal (1984).

(24) Freeman et al. (1978: 18). A strikingly similar incident occurred during the famine of the early 1940s (Cabral 1980: 150–1). A significant attempt at providing relief was however made during the famine of 1862–5, when employment was provided (with cash wage payments) on road‐building works (see Meintel 1984).

(25) In 1970, 70% of agricultural products consumed in Cape Verde were produced in the country (CILSS 1976: 8). This ratio had fallen to 1.5% by 1973 (CILSS 1976: 8), and only rose marginally thereafter (van Binsbergen 1986). According to one study, ‘during the drought over 70% of the agricultural labour force has been unemployed’ (Economist Intelligence Unit 1984: 38). It is not clear, however, how this calculation treats labour employed on public works programmes (on which more below).

(26) Freeman et al.] (1978: 98).

(27) Several commentators have argued that, in this case, action was motivated by the concern of the Portuguese government for its international image. See e.g. Meintel (1984: 68), CILSS (1976: 4), Davidson (1977: 394), and Cabral (1980: 134).

(28) Calculated from CILSS (1976: 3–4). This was the policy of Apoio or ‘support’, which was later criticized by the government of independent Cape Verde for the unproductive nature of the works undertaken (see CILSS 1976, Legal 1984, and Meintel 1983, 1984).

(29) See e.g. Meintel (1984: 68–9), CILSS (1976: 14), and Freeman et al.] (1978: 149, 203).

(30) In fact, until 1981, Guinea‐Bissau and Cape Verde were jointly ruled by the binational Partido Africano da Independencia da Guine e Cabo Verde, which had earlier led the independence struggle against the Portuguese rulers.

(31) Courier (1988: 27).

(32) For further details, see CILSS (1976), Davidson (1977), Freeman et al. (1978), USAID (1982), Meintel (1983), Legal (1984), Lesourd (1986), and particularly van Binsbergen (1986).

(33) This rule does not apply when the sale of food aid violates the conditions of delivery, e.g. in the case of the comparatively small quantities of food donated to Cape Verde under the World Food Programme. These are used for supplementary feeding.

(34) Economist Intelligence Unit (1984: 38).

(35) See van Binsbergen (1986: 9). See also Courier (1988).

(36) Both in the pre‐independence and the post‐independence periods, remittances from abroad also played an important role in mitigating the effects of the drought.

(37) There is a revealing contrast between this observation and the fact of huge livestock losses, which provide another measure of the intensity of the drought and of the threat of famine. The decline in livestock between 1968 and 1980 has been estimated at 12% for goats, 30% for pigs, 50% for sheep, and 72% for cows (calculated from Economist Intelligence Unit 1983: 43).

(38) On this, see Legal (1984: 12–16), who notes large increases in the consumption of maize, wheat, and rice in the post‐independence period compared to the pre‐drought period. The average consumption of calories, which ‘for the vast majority of the population did not exceed 1500 calories per day’ at the time of independence (CILSS 1976: 8; our translation), is now believed to have ‘moved closer to the required level of 2800 calories per day’ (van Binsbergen 1986: 3).

(39) A USAID study dated 1982 also mentions, without explicitly providing supporting figures, that ‘by providing employment, the Government of Cape Verde's rural work program has had an acknowledged major effect on improving nutritional status’ (USAID 1982: 15).

(40) See e.g. Wisner (1977), O'Leary (1980), Herlehy (1984) and Ambler (1988).

(41) For in‐depth analyses of the 1984 drought and the government's response, see Ray (1984), Deloitte et al. (1986), Cohen and Lewis (1987), Corbett (1987), J. Downing et al. (1987), and T. Downing et al. (forthcoming). A particularly useful and well‐documented account of this event can be found in Borton (1988, forthcoming).

(42) On coping strategies in Kenya, see Wisner (1977), Bertlin (1980), Campbell (1984), Swift (1985), Downing (1988b), Akong'a and Downing (1987), Sperling (1987a, 1987b), Anyango et al. (forthcoming) and Kamau et al. (forthcoming).

(43) Cohen and Lewis (1987: 274). The existence of a serious threat of large‐scale famine in this event is also argued in detail in Corbett (1987). For statistical information on rainfall patterns and crop production during the drought, see Downing et al. (forthcoming).

(44) See Borton (1988), Table 3, and Maganda (forthcoming), Table 9.4.

(45) Cohen and Lewis describe the symptoms of an impending crisis as follows: ‘By April 1984, the situation was obvious. The sun was shining beautifully, when it should have been raining; no early warning system was required’ (Cohen and Lewis 1987: 276). Other authors, however, have also stressed the role of rapidly increasing purchases from the National Cereals and Produce Board in arousing concern for the possibility of a crisis (see e.g. Corbett 1987, and Borton forthcoming).

(46) J. Downing et al. (1987: 266). It appears that the drought enjoyed only limited coverage in the local media, but attracted considerable international attention and concern (Downing 1988a).

(47) For detailed and documented accounts of the famine prevention measures, see Cohen and Lewis (1987), J. Downing et al. (1987), Borton (1988, forthcoming) and T. Downing et al. (forthcoming). This case study concentrates mainly on the government response, which represents the greater part of these measures, though the involvement of non‐government agencies was not insignificant.

(48) The two slogans propounded by the government early on during the crisis were ‘planning, not panic’ and ‘food imports and employment generation’ (Ray 1984).

(49) Deloitte et al. (1986: 12). In 1985, the numbers in receipt of unconditional relief gradually decreased, though the amount of food distributed increased with the enlarged flow of food aid.

(50) The same survey reveals that, between July and December 1984, the median food ration per recipient household varied between 197 and 633 calories per day in different regions (Downing 1988a, Table 5.16; see also Downing 1988b, Table 4.19). For the same period, Anyango et al. (1987) estimate that ‘the food relief averaged 5–10 per cent of individual requirements’ for the recipients (see also Kamau et al. forthcoming). In 1985, the size of food rations was much larger, and did not in fact differ very much on average from the ‘target’ of 10 kg. of maize per person per month (Borton 1988, forthcoming).

(51) Cohen and Lewis (1987: 281).

(52) Ray (1984: 2). Communications from two persons who were involved in the 1984 relief efforts confirm that food distribution centres typically did not discriminate between different groups of people, and provided identical rations to all recipients.

(53) The general significance for famine prevention policies of the phenomena related to the first two observations was discussed in Chapters 5 and 7.

(54) For further details of livestock losses, see Borton (1988), Downing et al. (forthcoming), Chapter 1, Kamau et al. (forthcoming), Anyango et al. (forthcoming) and Mwendwa (forthcoming). The picture presented in other surveys is, if anything, grimmer than that offered by Table 8.7. According to Borton (1988), the drought of 1984 may have depleted the national cattle herd by as much as 50 per cent (p. vii).

(55) See the surveys of Anyango et al. (forthcoming), Neumann et al. (forthcoming) and Kamau et al. (forthcoming). The relative absence of distress migration is discussed in Anyango et al. (forthcoming). Unfortunately, the available data do not permit us to estimate the extent of excess mortality during the drought.

(56) According to Borton (1988), free distribution of food accounted for only 15 per cent of the cereals imported between September 1984 and June 1985 (p. 19).

(57) On this, see particularly Olsen (1984). See also Akong'a and Downing (1987) and Sperling (1987a). Note that the volatility of retail prices is compatible with the control of ‘gazetted prices’ mentioned earlier.

(58) For details of food price patterns during the drought, see the Government of Kenya's Market Information Bulletin, and also Maganda (forthcoming). Careful econometric analysis of time‐series data on food prices in Kenya confirms that the interregional disparity of prices sharply increased in 1984 (Jane Corbett, Food Studies Group, QEH, Oxford, personal communication).

(59) Sperling (1987a: 269).

(60) To the best of our knowledge an in‐depth analysis of these events has not been published to this day. However, see Government of Zimbabwe (1986a, 1986b), Gaidzanwa (1986), Leys (1986), Bratton (1987a), Davies and Sanders (1987a, 1987b), Loewenson (1986), Loewenson and Sanders (1988), Mitter (1988), Tagwireyi (1988), and Weiner (1988), for many valuable insights.

(61) On the extensive coverage of the drought in the Zimbabwean press, see the accounts of Leys (1986), Bratton (1987a), and Mitter (1988). Some of the more widely circulated newspapers, such as the Herald, did not always take a sharply adversarial stance given their generally supportive attitude vis‐à‐vis the ZANU government. However, even they played a role in maintaining a strong sense of urgency by constantly reporting on the prevalence of undernutrition and hardship in the countryside, echoing parliamentary debates on the subject of drought, calling for action against profiteering, and exposing the ‘scandal’ of rural women driven to prostitution by hunger (on this see Herald 1983).

(62) See e.g. Donelan (1983), Government of Zimbabwe (1984), Waterson and Sanders (1984), Mandaza (1986), Davies and Sanders (1987a, 1987b), Loewenson and Sanders (1988), and Tagwireyi (1988). To mention only two important areas of rapid advance, the percentage of children fully immunized in Zimbabwe increased from 27 per cent in 1982 to 85 per cent in 1982 to 85 per cent in 1988 (Tagwireyi 1988: 8), and school enrolment increased at an annual rate of 20 per cent between 1979 and 1985 (Davies and Sanders 1987b: 297).

(63) This point is stressed in Weiner (1987) and Weiner and Moyo (1988). For further discussion of production relations in Zimbabwe's rural economy and their implications for living standards and famine vulnerability, see Bratton (1987a, 1987b), Rukuni and Eicher (1987), Rukuni (1988) and Weiner (1988).

(64) See e.g. Sanders (1982), World Bank (1983b), Loewenson (1984), and Government of Zimbabwe (1984). These studies give a clear picture of the connections between this poor record and the massive inequalities in economic opportunities and access to public services of the colonial period. In the early 1970s, for instance, the life expectancy of a European female was more than twenty years longer than that of an African female (Agere 1986: 359).

(65) It is worth noting that while the Zimbabwean ‘miracle’ has often been attributed to the astonishing power of price incentives (see e.g. The Economist 1985, and Park and Jackson 1985), the expansion of the rural economy since independence has in fact involved a great deal more than a simple ‘price fix’. On the extensive and fruitful involvement of the government of independent Zimbabwe in infrastructural support, agricultural extension, credit provision, support of cooperatives, etc., see Bratton (1986, 1987a), Eicher and Staatz (1986), Eicher (1988a), Rohrbach (1988) and Weiner (1988).

(66) Bratton (1987a: 224). This statement refers, in fact, to the two least fertile among Zimbabwe's five agro‐ecological regions. These two regions account for 64% of Zimbabwe's land, 74% of the ‘communal lands’, and about two‐thirds of the communal area population.

(67) Bratton (1987a: 223–5). Weiner (1988) reports declines in draught stock during the drought period of 47% and 21% respectively in the two agro‐ecological regions mentioned in the last footnote.

(68) Leys (1986: 262). On the importance of wage labour and remittances for the rural economy in general, and for mitigating the impact of the drought on the rural population in particular, see also Bratton (1987a), Weiner (1987) and Weiner and Moyo (1988).

(69) Leys (1986: 270). Similarly, Weiner (1988) states that ‘during the 1982–4 period the government drought relief programme became the primary means of survival for about 2.5. million people’ (p. 71). While the present discussion focuses largely on the food distribution component of the drought relief programme, it is worth noting that (1) policy developments during the drought included an increasingly marked preference for public works programmes (supplemented by unconditional relief for the destitute) as opposed to large‐scale distribution of free food, and (2) the drought relief programme had a number of other important components, such as water supply schemes, cattle protection measures, and inputs provision. On both points, see Government of Zimbabwe (1986a).

(70) See e.g. Government of Zimbabwe (1983: 21), Bratton (1987a: 237), Mitter (1988: 4). The official version seems to be that ‘at the height of the drought about 2.1 million people had to be fed every month’ (Government of Zimbabwe 1986a).

(71) Bratton (1987a), Table 10.8(b).

(72) See especially Weiner (1988), Table 6.4, and Matiza et al. (1988).

(73) See e.g. USAID (1983), Leys (1986), Bratton (1987a), Davies and Sanders (1987b), and Mitter (1988).

(74) Leys (1986: 270).

(75) Bratton (1987a) describes the eligible population as ‘the “needy” . . . , defined as those with insufficient grain in the home granary and without a close family member working for a wage’ (p. 238). Leys (1986), on the other hand, states that free food distribution was intended for ‘the members of households in which the head of the household earned an income under the statutory minimum wage’ (p. 269), but later adds that this involved ‘distinguishing between those households where the head of household held a formal sector job’ and others (p. 270). In practice it is likely that the two sets of criteria described by these authors did not diverge substantially.

(76) Leys (1986: 270). Another author gives credit to the ZANU government for ‘the smooth running of the relief committees’ (Mitter 1988: 5).

(77) Bratton (1987a: 238). The viewpoint expressed by Bratton is based on a personal communication from a colleague at the University of Zimbabwe and could have been less robustly founded than that of Leys.

(78) Daniel Weiner (University of Toledo), personal communication. See also Leys (1986).

(79) Leys (1986: 271). The government put the blame on dissidents for disrupting relief efforts, and at one stage even held them ‘responsible for the drought’ (see Mitter 1988, for further discussion).

(80) According to Bratton, ‘it is safe to say that no person in Zimbabwe died as a direct result of starvation’ (Bratton 1987a: 225). According to Leys, ‘one consequence of the drought for some of the rural African population was hunger, and on occasion and specific places, deaths from starvation’ (Leys 1986: 258).

(81) For detailed discussions of mortality decline in Zimbabwe since independence, see Davies and Sanders (1987a, 1987b), Loewenson and Sanders (1988) and Sanders and Davies (1988). It has been claimed that the extent of infant mortality reduction in Zimbabwe between 1980 and 1985 has been as large as 50% (The Times 1985; Bratton 1987a: 238).

(82) See Loewenson and Sanders (1988).

(83) See e.g. Bratton (1987a: 224), Mitter (1988: 3–4), Moto (1983).

(84) See Loewenson (1984, 1986) for further discussion of these findings.

(85) See Davies and Sanders (1987a, 1987b), Loewenson and Sanders (1988) and Sanders and Davies (1988) for detailed reviews of the evidence. See also the results of regular surveys (carried out by the government of Zimbabwe) presented in Tagwireyi (1988).

(86) See the study of Hay et al. (1986), especially Table 2. A published summary of the main results of this major study of drought relief in Botswana can be found in Hay (1988). On the rural economy of Botswana, see Chernichovsky et al. (1985).

(87) For details, see Hay et al. (1986) and Quinn et al. (1987). It should be mentioned that while livestock losses during the drought were perhaps not dramatic in aggregate terms (according to Morgan 1986, the size of the national cattle herd declined by 22 per cent between 1982 and 1986), it has been frequently noted that these losses disproportionately hit small herds. Cattle deaths for small herds have been estimated at ‘more than 40 per cent for several years’ (Diana Callear, National Food Strategy Coordinator, personal communication). Also, poor households in rural Botswana derive a greater than average part of their total incomes from crops (even harder hit by the drought than livestock). The threat which the drought represented to the entitlements of vulnerable groups was therefore much more serious than aggregate figures about livestock mortality, and about the importance of livestock in the rural economy, would tend to suggest.

(88) Hay et al. (1986: 85).

(89) The Sandford Report (Sandford 1977) provided a useful background investigation of drought in Botswana. The Symposium on Drought in Botswana (Botswana Society 1979) which was convened, quite remarkably, in spite of the then prosperity of agriculture and the economy, was an invaluable forum of discussion on numerous aspects of the problem. Gooch and MacDonald (1981a, 1981b) provided an illuminating evaluation of relief efforts during the 1979–80 drought and far‐reaching recommendations for improvement. For an excellent analysis of the development of Botswana's entitlement protection system, see Borton (1984, 1986).

(90) The following details are based on Morgan (1985), Relief and Development Institute (1985), and especially Gooch and MacDonald (1981a, 1981b).

(91) Gooch and MacDonald (1981b: 11). In other cases, the distribution of food was vulnerable to frank abuses.

(92) On this, see Gooch and MacDonald (1981b: 12–13).

(93) Botswana does have a well developed ‘nutrition surveillance’ system, but this is used mainly for the purposes of monitoring and targeting rather than as a warning device. The decision to launch a major relief operation in 1982 was taken much before the system detected a significant increase in undernutrition (see Borton and York 1987). Nor do other components of Botswana's evolving early warning system seem to have played a major role in triggering the government's response (see e.g. Relief and Development Institute 1985).

(94) For a clear discussion of this issue, and of the ‘political value of drought relief’ in Botswana, see Holm and Morgan (1985) and Holm and Cohen (1988).

(95) It should be mentioned that the drought relief programme as a whole went much beyond these measures of short‐term entitlement protection. Public intervention was also very significant in areas such as the provision of water and the promotion of agricultural recovery. For comprehensive analyses of the drought relief programme, see Tabor (1983), Borton (1984, 1986), Holm and Morgan (1985), Relief and Development Institute (1985), Morgan (1985, 1986, 1988), Hay et al. (1986), Quinn et al. (1987), Hay (1988), Holm and Cohen (1988) and Moremi (1988). See also Government of Botswana (1980, 1985a, 1985b, 1987, 1988).

(96) Botswana belongs to the South African Customs Union (SACU), which inter alia allows the free movement of food between Botswana, South Africa, Lesotho and Swaziland. As a result, domestic variations in agricultural output have little effect on the availability and price of food in Botswana. For an econometric analysis of the effect of SACU membership on food prices and food security in Botswana, see Cathie and Herrmann (1988).

(97) Morgan (1985: 49).

(98) See e.g. Hay et al. (1986) and Quinn et al. (1987). This finding must be interpreted bearing in mind that the level of wages paid is ‘roughly equivalent to the salary earned by maids and security guards in urban areas and considerably more than cattle herders earned on cattle‐posts’ (Quinn et al. 1987: 18).

(99) Quinn et al. (1987: 18, 21). The population of Botswana was a little over one million at the time.

(100) See Hay et al. (1986) for a detailed discussion.

(101) Calculated from Hay et al. (1986), Tables 10 and 11. According to the same source the average size of rations amounted to nearly 60 kg. of food (mainly cereals) per recipient per year. For a helpful account of the various components of Botswana's food distribution programme, see Hay (1988).

(102) According to Morgan (1988), ‘starvation, even among extremely isolated communities, was entirely averted’ (p. 33).

(103) Holm and Morgan (1985: 469). None of the studies cited in this section provide estimates of excess mortality during the drought. According to Borton, ‘mortality estimates are poor in Botswana so it is not possible to estimate whether there has been a significant increase in the death rate’ (Borton 1984: 92). Against the initial increase in undernutrition among children, it must be noted that (1) the incidence of severe undernutrition has been very small (Hay et al. 1986; Holm and Morgan 1985), and (2) seasonal fluctuations in nutritional status have virtually disappeared during the drought (Government of Botswana 1985: Table 5).

(104) On this, see Morgan (1986). The preservation of the productive potential of the rural economy is related partly to the entitlement protection measures discussed here, but also to a wide array of explicit rehabilitation and recovery programmes. Though they are not the focus of our attention in this chapter, the importance of these programmes must not be underestimated.

(105) See Morgan (1986) and Holm and Cohen (1988) for further discussion of the interplay between drought relief and social security measures during the last few decades in Botswana.