Full text of DOI:10.1111/1467-7717.00129

Disasters, 2000, 24(1):18^37
Cyclone Mitigation, ResourceAllocation and Post-
disaster Reconstruction in South India: Lessons
fromTwo Decades of Research
Peter Winchester
Flood Hazard Research Centre
Middlesex University
This paper opens with a history of development and disaster-prevention strategies in a
cyclone-prone area of the east coast of India and traces the evolution in the area of
British and Indian governments’ programmes and policy over a century. Research
over the last 20 years has shown however that the programmes and policies have
failed to balance economic growth with safety. Resources intended for the benefit of
all have been diverted by alliances of powerful people to a small minority, and recent
developments have reduced the physical protection of the area. The result is that
increasing numbers of people are vulnerable to the effects of cyclones and floods. The
findings suggest that the best way to reduce vulnerability is to improve the socio-
economic standing of the most vulnerable and for this to happen these people must
have an assured income based on assets that will enable them to acquire social and
economic credit-worthiness within the local economy. This paper presents evidence
that suggests that non-governmental organisation (NGO)-supported co-operatives are
the best way to achieve this through self-help and self-employment schemes. It also
suggests that NGOs should be encouraged to take up environmentally and
ecologically beneficial activities involving the poorest groups in the communities, in
this way combining sustained self-employment with environmental protection.
Keywords: Vulnerability, cyclones, poverty alleviation, power relations, NGOs, co-
operatives.
Introduction
Since the introduction of canal irrigation into the cyclone and flood-prone deltas of
India’s eastern seaboard, and the subsequent population explosion, the major problem
for those engaged in coastal management has been to achieve a balance between
economic growth and safety (Government of Madras, 1878, 1908). Successive
governments have discouraged people from migrating to the overcrowded cities from
coastal areas by opening up the land and taking steps to protect the populations from
flooding and cyclones (Andhra Pradesh Government, 1977a, 1980a, 1980b, 1981a;
Government of Madras, 1938). When protection schemes fail, however, people
ß Overseas Development Institute, 2000.
Published by Blackwell Publishers,Oxford OX41JF, UK and 350 Main Street, Malden, MA 02148,USA.

Cyclone Mitigation, Resource Allocation and Post-disaster Reconstruction
19
increase their exposure to physical and economic risks, especially the poor
Winchester, 1986, 1992b; Brammer, 1990; Thompson and Penning-Rowsell, 1994;
(
Thompson and Sultana, 1996).
Research has shown that physical risks presented by topography and climate are,
for poor people, of secondary importance compared with the economic risks which are
principally determined by the power relations formed by alliances of the people who
control resources (Wade, 1982, 1984; Winchester, 1986, 1992a; Blaikie et al., 1994).
People with resources can protect themselves economically and physically from
cyclones and flooding and the speed of their recovery is relative to the size of their
asset base (Winchester,1992a). Unfortunately, the poor are vulnerable at every stage
and government measures have been ineffective at strengthening their socio-economic
resilience, resilience being quantified by the speed of recovery from disastrous events.
This paper focuses on poverty alleviation and makes the case that this should be the
central aim of all disaster reconstruction and development programmes. Field
evidence from a case study in Divi Seema, an island at the mouth of the River Krishna
in the Krishna delta, supports and emphasises views expressed elsewhere (Lipton,
1977; Chambers and Leach, 1989, Chambers et al., 1989a, 1989b) that poverty can
only be alleviated by improvements in the socio-economic resilience of the poor. It is
the contention of this paper that, just as engineers have seen the relative or absolute
failure of many ‘hard’ engineering structures in hazard mitigation and coastal
management, moving in the 1990s towards more natural approaches via ‘soft’
engineering solutions working with the natural morphological and ecological
characteristics of coastal areas, so post-hazard development programmes should
move away from focusing on whole populations and target instead the poorest and
most vulnerable through programmes based on self-help. It also argues that NGOs
could play major roles in carrying out large-scale development programmes in coastal
areas (Pearce and Turner, 1990; Zazweta, 1995; Krishna et al., 1997).
This paper sets out the evolution of disaster-prevention strategies in Divi Seema and
reveals why different strategies have succeeded or failed. The first part describes Divi
Seema and its historical development from 1907 to 1998. It provides an explanation of
why cyclone mitigation and disaster management programmes have had so little success
in alleviating vulnerability in this area. The second part presents the data and an analysis
of 20 years’ field-work (1978–98) in Divi Seema that supports this explanation and
reveals why some households survive and prosper while others do not. The third part
examines the issue of equitable resource allocation and looks at the work of some NGOs
working in Divi Seema and elsewhere on the sub-continent. It focuses on the scale of
operations and shows why size is the major issue for NGOs, defining which NGOs fail
and which continue to operate despite the potentially destructive socio-political power
relationship between government and alliances of powerful people. Finally, the paper
summarises some of the problems and makes some recommendations.
Cyclone mitigation
Location: Divi Seema
The island of Divi Seema lies at the eastern tip of the Krishna delta; its two distinct
topographical areas are the inland plain and the coastal strip, with the land rising

20
Peter Winchester
Figure1 Map of a part of the Krishna River delta showing the island of Divi Seema, the
Mission and Trust villages, irrigation canals and principal land and vegetation zones
gently to three metres above sea level at its highest point (see Figure 1). The delta is
very fertile being mostly composed of rich alluvial deposits. However, near the coast
there are extensive areas of low-lying land where soil salinity levels are high due to
seasonal marine flooding; in addition these areas are also frequently waterlogged by
watercourse flooding from excessive rainfall during the monsoon (Andhra Pradesh
Government, 1977b). The Krishna delta is the second most cyclone- and flood-prone
stretch of the eastern seaboard of the Indian sub-continent after the Ganges delta in
Bangladesh. Cyclonic storms, usually accompanied by heavy rainfall, strike the delta
on average every five years reducing standing crops by up to 20 per cent and every 10
years by up to 50 per cent (Winchester, 1986). In November 1977 a cyclone and storm

Cyclone Mitigation,Resource Allocation and Post-disaster Reconstruction
21
surge struck Divi Seema killing 10,000 people in one night and devastating hundreds
of square kilometres of crops and property (Andhra Pradesh Government, 1977a;
Raghavulu and Cohen, 1979). This cyclone was the initial focus for the author’s
interest and research into the subject (Winchester,1979). Since 1977 there have been
further cyclones in the delta, in 1979, 1981, 1990 and 1994.
History1907--1977
Canal irrigation was introduced into Divi Seema in 1907, 20 years after the rest of the
Krishna delta had been irrigated due to its remoteness and liability to sea flooding
(Government of Madras, 1908). As a result of irrigation, large landowners came from
the mainland and took over the land as absentee landlords pushing the previous dry
land cultivators towards the coast. Between 1907 and 1937 the irrigation system was
extended, further marginalising the indigenous inhabitants and pushing them into the
dangerous coastal strip where only fishermen had lived previously. Mangrove forests
had once protected the coastal strip, acting as windbreaks to cyclones and a
deceleration zone for the storm surges accentuated by the delta’s extensive offshore
shallows (Nageswara Rao and Vaidyanaham, 1978; Dube et al., 1981). By the mid-
1930s most of the mangroves had been cut down and by the early 1940s over-
population of the coastal strip had led to their total destruction (Figure 1). The British,
aware of the increasing physical threats of cyclones and flooding on the expanding
population, and seeing the problem in terms of physical safety, built the first coastal
earthen bund in the late 1930s to protect property further inland and save the repetitive
costs of replacing infrastructure (Government of Madras, 1938). At the same time they
built embankments (lev e´ es) on both sides of the river Krishna together with land
drainage and associated works. None of these works prevented the low-lying areas
from periodic waterlogging.
In the late 1950s, the state government, in accordance with the government of
India’s First and Second Five Year Plans (1947–60), set up Primary Agricultural Co-
operative Societies and transferred 20,000 acres of government land in the coastal strip
to fishermen, cultivators and landless labourers. However, by the mid-1960s the land
had been taken over by large landlords because the intended beneficiaries could not
get credit from the banks for land development and the agricultural inputs had been
appropriated by the large landlords and their allies who, working closely together with
the merchants and traders (henceforth collectively termed the Alliances) control the
agrarian economy of Divi Seema. The Alliances have connections with local
politicians, irrigation engineers and banking officials ensuring that nearly all
government resources go to them. The two most important resources are water
supply (Wade, 1982, 1984) and credit (Winchester, 1986, 1992a). Little development
was carried out in Divi Seema after the collapse of the co-operatives and it remained
cut off during the rainy seasons, retaining its reputation as a wild place populated by
undisciplined people (Washbrook, 1976; Andhra Pradesh Government, 1977a).
Between 1907 and 1977, the population on Divi Seema doubled but the economy
remained based on an uneven distribution of wealth, principally in the form of land
ownership (Andhra Pradesh Government, 1977b). Ninety per cent of the population
were dependent for their staple food and employment on a handful of large
landowners. The banks lent only to those who owned land or other major assets as
collateral and tended to write-off loans at election times — a practice that is

22
Peter Winchester
widespread all over India, but irrelevant to the poor (who have no bank loans). The
poor had no choice but to continue to go to the moneylenders (mainly the large
landlords, shopkeepers and traders) whose extortionate rates kept most of them in debt
for life. The informal market had many other destructive features; one being the
system of interlocked contracts in which lenders set conditions on the borrowers to
provide their labour and/or produce at disadvantageous terms to the borrowers;
another was the arrangement whereby those with rice surpluses offered rice loans
preceding the harvest on condition that they were repaid immediately after harvest at
the rate of two bags for one (Winchester, 1986).
The economic vulnerability of the majority of the population and the acute physical
vulnerability of the area made Divi Seema a disaster in waiting. The disaster duly
occurred in 1977, bringing tragedy to thousands but also focusing attention on this
neglected area (Raghavulu and Cohen, 1979; Winchester, 1979). The result of the
worldwide publicity was a huge aid response and intensive development programmes,
all of which altered the surface appearance of the economy, but did little to affect its
underlying structure as the findings of 20 years of research have shown.
The1977 cyclone and subsequent cyclone mitigation programmes
The 1977 cyclone was the turning-point in government cyclone-mitigation policy.
There had been previous attempts to produce a cohesive cyclone mitigation policy,
notably the recommendations by the Cyclone Distress Mitigation Committee in 1971
(Government of Andhra Pradesh, 1971), but very few recommendations had actually
been enacted by the end of the decade. After the devastation of the 1977 cyclone the
government accelerated measures to save and protect the coastal populations. The
government (as the British had before them) saw the problem of vulnerability as
principally attributable to exposure, topography and the ‘savagery of nature’
(
Winchester, 1986: 6). The government also thought that the poor always made
difficult situations worse by their intransigence at not taking notice of warnings
Andhra Pradesh Government, 1977a, 1979b; Winchester, 1986, 1992a). This view
(
that still prevails today, particularly in the press.
In response to the calamity of 1977 the government embarked on relief and
rehabilitation programmes and a public-works programme (Government of Andhra
Pradesh, 1977a, 1977b). The earthen seawall was rebuilt as well as the roads; the
irrigation and drainage networks were extended and electricity was extended to all
parts of the island. A community cyclone shelter and house-building programme was
undertaken, using village labour. The public works provided short-term employment
for the poor, but since the roads and irrigation works went to land owned or controlled
by the Alliances and their friends in the villages (the elders and elected officials) and
since the programmes were targeted at relief rather than development, there have been
few long-lasting benefits for the poorest (Winchester, 1986, 1992a). Shelter belts had
been planned for a strip 20km wide along the entire coastline but were never planted
(Andhra Pradesh Government, 1977a, 1979, 1981a). Many of the cyclone shelters are
in use today as ‘village halls’ for nursery schools, clinics and meeting places, and can
be judged a success. However, the 15,000 brick and concrete ‘pukka’ houses built in
conjunction with the aid agencies have been of limited value (Winchester 1986,
1992a). Another devastating cyclone and storm surge hit the Krishna delta in 1979 but
the death tolls were less than 10 per cent of those in 1977 due principally to improved

Cyclone Mitigation,Resource Allocation and Post-disaster Reconstruction
23
warning systems and evacuation procedures (Andhra Pradesh Government, 1979).
Since 1979 there have been cyclones and flooding in the Krishna delta in 1983, 1987,
1990 and most recently in 1995 but in each case there were fewer than a hundred
human fatalities.
Changes since1977
The greatest change since 1977 to the physical and economic landscape of Divi Seema
has been the introduction of aquaculture (prawn and fish farming ) into areas
previously under rice cultivation, particularly on irrigated land at the farthest end of
the coastal strip. To make use of poor-quality agricultural land the government
promoted aquaculture in the late 1980s and encouraged the banks to give larger loans
for aquaculture than for rice cultivation. The result was that the area under agricultural
cultivation in the coastal strip of Divi Seema since 1977 has decreased by more than
80 per cent, while the number of noncultivators — namely, landless labourers —
increased by more than 50 per cent (Government of India, 1981, 1991). Rice
production thus halved with a commensurate decline in agricultural employment. This
has produced higher rates of migration for agricultural work to other dangerous coastal
areas than ever before, land for grazing has become scarce and animal husbandry — as
a supplementary occupation for agricultural labourers — has been severely curtailed
as it is limited by fodder availability (personal communication with the president of a
coastal village, 1998). In the same period nearly all the trees and thorn bushes growing
on the bunds and between the paddy fields in the coastal strip have been cleared, and
inland areas, once protected to some extent by the flora of the coastal strip, are now
fully exposed to wind damage and flooding. Half the area is now permanently under
water, with serious implications for health through the spread of water-borne diseases
(Bradley, 1977) — one of these the increasingly prevalent malaria.
In the mid-1990s a virus almost destroyed the fish and prawn stocks and the
industry nearly collapsed. Only those with connections to the Alliances who enabled
them to write off their aquaculture loans survived to continue with aquaculture.
Agricultural labourers suffered most. They lost their traditional access to
encroachment lands (low-lying common land) because many cultivators, with small
holdings who had sold their land to repay the aquaculture loans, took them over (noted
in Table 3). This deprived labourers of one of their traditional insurance ‘buffers’
against bad times (see also Jodha, 1991), as well as further increasing their
vulnerability to wage and price fluctuations. Although agricultural wages and the costs
of basic foodstuffs have remained at roughly equivalent since 1977, the costs of other
necessities such as building materials and agricultural inputs have escalated, and so
too have costs for medical and veterinary services and bribes.
Elsewhere in the world, with similar problems to Divi Seema, some problems are
being addressed through benign coastal-management programmes. These programmes
use the natural morphological and ecological characteristics of coastal areas to
increase protection and incorporate environmentally appropriate land uses as
employment generators (Baan et al., 1997; Green et al., 1994; Klein et al., 1998;
Turner et al., 1998). Land uses include the planting and management of shelter-belt
schemes with mangrove and other trees using social-forestry techniques (CIDA, 1995;
Saxena, 1995). Institution of shelter-belt and employment schemes could be of great
benefit to the people of Divi Seema in the future.

24
Peter Winchester
There is no doubt that the provisions of the Cyclone Contingency Plan of Action
(Andhra Pradesh Government, 1981a), particularly the provisions concerned with
improved warning systems and evacuation procedures, have saved thousands of lives.
The specific anti-cyclone measures may have been successful in Divi Seema in regard
to loss of life, but the empirical, visual and anecdotal evidence suggests that
government anti-poverty and development programmes aimed at strengthening the
economic base of the most vulnerable sections of the population, carried out in tandem
with the anti-cyclone programmes, have not succeeded as much as was predicted.
Empirical evidence to support this statement is presented below.
Research1978--1998
Sample surveys
Research in Divi Seema has been carried out over 20 years, starting in 1987
(Winchester, 1986, 1992a); the central findings are summarised in Tables 1–4. In the
first stage of the field-work programme, methodology was designed to find the links
between vulnerability, cyclone and flood impact and recovery with the object of
disentangling conflicting views on the subject. Vulnerability in this paper is defined as
the relative ability of an individual, household or community to withstand and recover
from a shock such as a cyclone or flood (Winchester, 1986: 122, 123, 1992a: 44, 46,
59), or in other words: resilience. By including recovery within this definition,
vulnerability is released from the strict confines determined by exposure and aids the
understanding that the major forces governing (cyclone) vulnerability are not
necessarily physical (see also Blaikie et al., 1994).
The view held by government and technocrats was that recurrent exposure to
cyclones and associated flooding increased physical and economic vulnerability and
that this could be reduced only if the impact of cyclones and flooding were reduced by
physical means (Andhra Pradesh Government, 1977a; Winchester 1992a: 31). The
alternative view held by some social scientists (Westgate and O’Keefe, 1976; Baker,
1981; Blaikie, 1981 among others) was that the degree of peoples’ vulnerability to and
ability to recover from cyclones and flooding were directly related to the forces that
governed their everyday lives, that is, the historical development of the area and the
underlying power relations within the local political economy, so that peoples’
vulnerability was only accentuated by the incidence of cyclones and flooding and not
principally caused by them (Winchester, 1992a: 44).
Methodology
Data to test both hypotheses had to distinguish, initially, between variables that have
an apparent effect on vulnerability to cyclones and flooding, such as topography and
location, from other variables that would appear to have a marginal effect. As a result,
topography was chosen as the major variable with caste distribution and occupations
as secondaries. My sample area was located in an area where the Society of Jesus (the
Jesuits) had carried out a socioeconomic survey after the 1977 cyclone (DSSSS,
1978). My area included 12 villages and, since topography was the major variable,
one-half of the sample frame of 202 households came from the coastal strip and one-

Cyclone Mitigation,Resource Allocation and Post-disaster Reconstruction
25
half from the inland areas. I also used the Jesuit sample area as the frame for my ‘site
and condition surveys of the concrete houses’ carried out in 1983, 1988 (Winchester,
1992a) and 1997. Diagrammatic village plans were drawn up for the interview
programme based on a random sample of every 10th house, but the sample was
stratified according to house type since I was initially interested in the role of housing
relative to cyclone and flooding vulnerability.
The pilot study of 51 households revealed that housing was considered by all but
the wealthier households as having a very low priority relative to reducing
vulnerability to cyclones and flooding. It became apparent that the majority of
households in the pilot study considered that economic factors affecting social
standing determined their survival and recovery from cyclones and flooding to a far
greater extent than their incidence. The focus of the field-work programme was
consequently altered to take account of this. The results of the surveys of 1981–3 and
1988 produced ‘indicators of vulnerability’ (Winchester, 1992a) and these became the
basis for the 1997/8 surveys for identifying vulnerable households.
Results
Table 1 shows the distribution of losses in 1977 as displayed in government data,
according to topography and location. It shows that total human losses are six-times
higher in the unprotected coastal areas than in the protected inland areas and total asset
losses in the coastal areas are at least twice that of those inland. Table 1 also shows that a
very high percentage of animals were lost on the inland plain. Table 2 displays
occupations and widens the government classifications of agricultural occupations —
cultivators, cultivator/labourers, labourers (Government of India, 1981) by including a
fourth classification: ‘others’ (whose principal occupation is not in agriculture but
involves specific caste trades — herders, toddy-tappers, potters — traders, small
shopowners or minor government officials). This table shows that cultivators, who have
two or three income-earning occupations, sustained fewer losses than (landless)
labourers who only have one income-earning occupation. While the cultivators had the
lowest percentage losses both in the coastal strip and the inland areas, the labourers had
Table1 Losses in1977 as totals and percentages of previous household totals according to
topographyand location
a
Households
111
Deaths
Assets
Animals
Coastal
strip
115
86
163
(23%)
(93%)
(93%)
Inland
plain
91
18
(5%)
34
(70%)
93
(74%)
Totals
202
133
120
256
as a percentage
(14%)
(66%)
(85%)
of previous totals
Source: Field-work: 1981–2 (Winchester, 1986: 166)
a
Assets are plough teams and carts
Original 202 household sample survey 1981–2
b

26
Peter Winchester
Table 2 Losses in1977 as a percentage of the previous group totals stratified according to
topography, location and occupations
Occupation
H/holds
Deaths
%
Assets
%
Animals
%
a
(n)
Coastal
Cultivator only
(6)
3
35
26
25
38
75
90
—
35
88
90
92
Cult/ labourer
(66)
(26)
(16)
b
Others
Labourer only
Inland
Cultivator only
Cult/labourer
Others
(12)
(42)
(12)
(22)
—
8
7
17
50
70
—
18
60
70
92
Labourer only
10
Sources: Winchester 1986: 180–2, 1992: 104.
a
Original 202 household survey 1981–2
b
Others are: fishermen, government officials, petty shop keepers and tradesman, some of whom own
land and some of whom (the majority) work as agricultural labourers
the highest percentage of human deaths and animal losses. Village surveys three years
later (Winchester, 1986) showed that 90 per cent of the houses and animal sheds owned
by cultivators in the coastal area had survived the five-metre storm surge that had
flattened everything else and showed that those with sufficient resources could protect
themselves adequately despite the absence or failure of defence systems (Winchester,
op. cit.). Table 3 presents data on recovery over the 1977–88 period and reveals that
although minor cyclones and floods in 1979, 1983 and 1987 played their part in delaying
or reversing the recovery of the labourers, they did not affect the progressive recovery
overall of the cultivators (Winchester, 1992a). Table 4 shows that during the 20-year
period 1977–97, the cultivators increased the size and extent of all their assets. Other
groups also increased their assets, except for land. Cultivator/labourers, petty traders
and minor government officials were in 1998 only slightly better off than they were
before the 1977 cyclone, but the landless were worse off (Table 4).
A survey of 202 households is a small sample from which to deduce the success or
failure of development programmes in reaching the poorest in a population of 70,000.
However, the findings set out in the tables give substance to the overall pattern
previously discerned and reinforce the view that the historical development of an area,
its local political economy and underlying power relations are fundamental to any
understanding of the causes of vulnerability.
Compared with pre-1977 in 1998 everyone looked better off; for instance, people
wore better clothes; there were tea shops in every village hamlet; there were doctors,
clinics and schools in every large village; and there were more shops in the main
towns than previously. Appearances, however, masked the reality for the poorest
(Winchester, 1990). The survey showed that changes in the physical and economic
landscape introduced principally by aquaculture affected the poor in two ways:
•
agricultural labourers’ debts to moneylenders are no less than they were 20 years
ago, and

Cyclone Mitigation,Resource Allocation and Post-disaster Reconstruction
27
Table 3 Mean distribution of land ownership and other assets per household, classified by
occupations, and relative change over the period1977 ^1988
Occupation
Time
Land
ownership
Other assets
(plough teams
& carts)
Animals
(acres)
a
Cultivator
1977
1981
1988
4.9
5.1
6.2
3.4
3.9
5.4
2.5
4.1
5.0
only
b
c
(n = 12)
Relative change
1977–88
1.3
2.0
2.5
a
Cultivator/
1977
1981
1988
1.5
1.3
1.6
0.5
0.6
1.2
1.3
1.2
1.8
labourer
b
c
(n = 51)
Relative change
1977–88
0.1
0.7
0.6
a
Others
1977
1981
0.4
0.5
0.6
0.1
0.1
0.2
0.6
1.3
1.3
b
(n = 16)
c
1988
Relative change
1977–88
0.2
0.1
0.7
a
d
Labourer
1977
1981
0.0
0.3
0.2
0.0
0.0
0.0
0.1
1.3
1.0
b
(n = 26)
c
1988
e
Relative change
1977–88
0.2
0.0
0.9
Source: Field-work 1981: 82, 83, 88
a
Based on anecdotal evidence from the 202 households and verified as far as possible during the
981–2 field-work programme
Based on total of 105 households
1
b
c
Re-interviews with 39 of the 42 case study households of the 1983 survey plus re-interviews with 66
households from the original 202 sample (January–March 1988)
d
Landless households own no freehold land. These figures refer to the extent of encroachment land
that they cultivate
e
This figure indicates the effect of the Jesuit programmes which peaked in 1985
•
migration for work to other dangerous coastal areas and inland have increased,
especially in the last 10 years (Winchester, 1992a).
No matter where they lived, however, the cultivators have prospered since 1977
Tables 2, 3, 4). In my conversations with them the cultivators informed me that they
(
attributed their survival and progressive prosperity since 1977 to two factors, despite
intervening cyclones, floods and poor harvests. These were:
•
•
They already had an established income base from assets (land, animals,
equipment) on which to build sufficiently strong houses and sheds to protect
themselves and their assets regardless of what the government did before the
cyclone.
As a consequence, they had sufficient social standing and economic credit-
worthiness with the Alliances who controlled the distribution of government

28
Peter Winchester
Table 4 Mean distribution of land ownership and other assets, per household classified by
occupations, and relative change over the period1977--1997
Occupation
Time
Land
ownership
Other assets
(plough teams
& carts)
Animals
(acres)
1
977
4.9
6.2
8.2
3.4
5.4
5.6
2.5
5.0
6.2
a
c
Cultivator
only (n=16)
1988
1997
b
Change
1977–97
3.3
2.2
3.7
1
977
1.5
1.6
1.2
0.5
1.2
1.2
1.3
1.8
2.8
a
c
Cultivator/
labourer (n=45)
1988
1997
b
Change
1977–97 Minus 0.3
0.7
1.5
1
977
0.4
0.6
0.2
0.1
0.2
0.1
0.6
1.3
1.8
a
c
Others
1988
1997
b
(n =15)
Change
1977–97 Minus 0.2
0.0
1.2
1
1988
1997
977
0.0
0.2
0.0
0.0
0.0
0.0
0.1
1.0
0.2
a
c
Labourer
(n = 34)
Change
1977–97
0.0
0.0
0.1
Source: Fieldwork 1988, 1997–8
a
b
c
c
As for note in Table 3
Based on total of 110 households
Two surveys were carried out in March and August 1997 in which 37 of the 42 case study
households of the 1983 survey were revisited together with a further 73 households from the original
2
02 household survey
resources to ensure that they gained access to those resources and so recovered
quickly after the cyclone. Having done so, connections with the Alliances have
been maintained ever since.
Summary
The research showed that an established income and credit-worthiness are the two key
attributes for reducing vulnerability, while location is of less significance. People live
in dangerous areas because either they can make money there and the risks they
choose to take are offset by the gains they hope to make, or their caste occupation (for
instance fishermen) enables them to weigh up risks and make choices. But, most
people live in dangerous areas like Divi Seema because they have no choice. Since it
is unrealistic to consider moving the populations of the coastal areas, the problem
becomes how to reduce the vulnerability of those people most at risk from indifferent
or hostile political and economic circumstances.

Cyclone Mitigation,Resource Allocation and Post-disaster Reconstruction
Resource allocation
NGO projects on the Indian subcontinent
29
The historical records and the evidence brought forward here suggest that the solution
to the problem lies in access to and distribution of resources. Clearly the way
government resources have been allocated is ineffective. This has been remedied
historically by non-governmental organisations (NGOs) which have offered an
alternative approach to resource allocation and distribution. NGOs have some
advantage over government agencies by being nominally outside socio-political power
relationships. Even though their scales of operations are far smaller than government
ones, it is worth examining how the successful NGOs cope with hostile political and
economic circumstances.
NGOs face numerous problems on the Indian subcontinent; one is the sheer scale of
the poverty that surrounds them, but another is the size of their resources compared
with the government’s. Many social scientists and economists recognise that poverty is
caused by forces governing access to and distribution of resources, enshrined in power
relationships (Mitra, 1977; Lipton, 1977; Chambers et al., 1981) which the poor are
unable to alter (Chambers, 1982, 1983, 1988). Some Indian NGOs in the early 1980s
promoted social activism and civil rights as an alternative to providing economic
resources (Kothari, 1983; Sheth, 1983, 1987; Lewis, 1991) but most NGOs kept away
from overtly political issues and concentrated on social welfare, health-care and the
provision of cheap credit (Hashemi et al., 1991; Black, 1992; Hulme and Montgomery,
1994; Lipton, 1996; Yunus, 1998). One prominent NGO (Action India) after 10 years’
work concluded that the most useful activity for an NGO to engage in was to organise
the poor to get government resources themselves instead of the NGO providing them
(Lewis, 1991).
One problem facing NGOs is to find the most cost-effective size to mobilise the
resources they command. The problems connected with poverty are so complex and so
massive that large agencies with many areas of expertise are needed. This involves the
development of complex management structures that bring with them some of the
more negative aspects of bureaucracies. Small NGOs are limited to one or two areas of
expertise (health-care, agriculture, credit and so on), but their size makes them
adaptable in changing circumstances more quickly and effectively than larger, more
cumbersome agencies. However, small NGOs cannot hope to solve the causes of
poverty, at best they are palliative and at worst their presence provides an excuse for
the authorities to do nothing. One solution is for small NGOs to join together, or form
co-operatives with other groups.
Some examples are presented below of NGOs that have been successful in avoiding
some of the perils that often befall government programmes; these examples show that
the scale of operations is crucial in determining whether organisations achieve their
goals. Even though NGO programmes cannot be diverted in the same way as the
government’s, they can be damaged in other ways, principally by falling foul of the
people who control the local political economies (Winchester, 1981, 1986, 1992a,
1992b). Thus, their success depends largely on their relationship with local power
structures. This relationship governs whether they survive or not (see also Cassen et
al., 1994). For example, the largest NGOs whose scales of operations encompass
hundreds of thousands of people over large areas can survive and prosper within

30
Peter Winchester
Figure 2 DSWSG structure for one or more groups, showing content and time scale of
programmes
initially hostile political economic systems by eventually becoming part of the
commercial establishments of those political economies. The intermediate-size NGOs,
involving several thousand people can alter some aspects of prevailing socio-
economic systems as long as they remain non-confrontational, but — if they leave the
field — the former socio-economic status quo will re-establish itself. The smallest
village-level NGO cannot hope to change the prevailing political and economic
systems, but it may survive if it does not confront them (Hartmann and Boyce, 1983;
Winchester, 1981, 1992a; Beck, 1994).

Cyclone Mitigation,Resource Allocation and Post-disaster Reconstruction
Large-scale operations involving tens of thousands of people
31
One of the best-known examples of a large co-operative NGO consisting of many
groups of varying sizes is the Grameen Bank of Bangladesh (Yunus, 1998) which
pioneered a micro-credit loan system based on the savings of poor people without
financial collateral (see also Johnson and Regally, 1997). These people are
incorporated into a structure which provides access to a wide range of social and
financial support services with the services forming the backbone for a wide range of
multi-sectoral enterprises (see Figure 2) that increase their members’ incomes and
commensurately reduce their poverty (Khandker et al., 1995; Jain, 1996; Yunus,
1998). Another large well-known NGO supported in its early days by Oxfam, is the
Self-Employed Womens’ Association (SEWA) of Ahmedabad, Gujarat, India which is
a ‘parent’ organisation geared to supporting a series of worker co-operatives of urban
and rural multi-sectoral commercial activities (SEWA, 1981). The combination of
many ‘independent’ commercially oriented groups under the umbrella of SEWA has
protected hundreds of small groups of women workers from exploitation and
destruction by the people who control the local economies. As a result of their success
SEWA is accepted as a respectable commercial force by the powerful urban and rural
alliances (Oxfam, 1995).
Intermediate-scale operations for thousands of people
Two intermediate-scale NGOs operating in the Krishna delta have joined forces with
larger NGOs to carry out projects which neither could do on their own while
remaining independent. The Divi Seema Social Service Society (1977–85), run by the
Jesuits, and the Salvation Army (1977–82) came to Divi Seema immediately after the
1977 cyclone and worked in the areas with the highest death tolls on the coastal strip
and low-lying inland areas where approximately 25,000 people lived (or one-third of
the area’s population). They departed after friction between their respective head
offices and their field operatives. Both organisations established village credit unions,
membership of which gave access to soft loans and subsidised agricultural inputs; their
other programmes covered health-care, adult and nursery education and skills training
for a wide range of self-employment activities. Membership of credit unions was open
to everyone, but the better-off found the prospect of regular payments and compulsory
attendance at meetings too tedious and the loans and subsidies too small. Nevertheless
they did not feel excluded and consequently did nothing to undermine the
programmes. Both NGOs worked with Oxfam and other international aid agencies
on large projects (affecting up to 15,000 people). One of these was to take safe
drinking-water to remote villages and extend the irrigation network in the tail-end
lands. Neither the Jesuits nor the Salvation Army could have accomplished these
projects alone. Despite it being their stated intention, neither organised the villagers to
take over the programmes when they left so that when they did leave, the banks
withdrew their loans, asset accumulation ceased and the tolerable relations between
poor villagers and the Alliances evaporated. The collapse of these NGO programmes
mirrors the fate of many government projects when a senior officer such as a collector
sets up special small-scale anti-poverty projects that cease abruptly once he or she is
transferred.

32
Peter Winchester
Small-scale operations involving hundreds of people
Three small-scale NGOs now operate in the Krishna delta. From 1985 onwards, two
small NGOs have worked continuously in Divi Seema: The Sisters of St Ann’s (St
Ann’s Sisters) at Avanigadda and the Sisters of St Joseph of Cluny at Nagayalanka
(
(
Figure 1). Between them they reach perhaps 7,500 people in a population of 75,000
Government of India, 1991). They have encouraged the formation of women’s
savings groups, eligibility for which is based on economic criteria and restricted to
women who come from families with no assets apart from their own labour, widows,
families where the female is the main earner and from families most at risk from the
effects of cyclones and floods. Membership in a savings group opens the door to
social support and economic programmes more or less immediately. Their
programmes tackle the symptoms of poverty — low income, poor nutrition, poor
health, lack of education and so on, without significantly altering their socio-
economic conditions.
Recently a third NGO has established itself: the Divi Seema Women’s Savings
Groups (DSWSG) with the explicit aim of tackling the causes of poverty. DSWSG is a
typical small NGO made up of social and a financial support arms (Figure 3) capable
of providing support to no more than 10 groups of around 30 women each (300
families). Its structure is almost identical to the that of the other two NGOs. In the case
of DSWSG, members have to save regularly for six months before becoming eligible
for loans although during this time ‘community development’ services are provided;
after six months they become eligible for social and economic support (Figure 2).
During its initial phase this NGO has limited its activities to establishing group and
loan structures. A second phase of expansion is planned to enable it to achieve its
lending aim.
Lessons learned from the successes of the larger NGOs suggest that the most
effective survival route for small NGOs like DSWSG would be to join up with other
NGOs such as the St Ann’s Sisters, under the umbrella of a larger ‘parent’ NGO. The
parent NGO could be a federation of several independent varying-sized groups or it
could be a co-operative (like SEWA). The principal function of a parent would be to
link the small NGOs with scarce resources through its own nation-wide links to
institutions that provide skills expertise, technology and money, of such variety and
scale normally unavailable to small NGOs. Figure 3 presents a model for the future
structure of the organisation showing the relationship between DSWSG and a parent
NGO. A revolving fund is a central feature of the structure; this is a fund into which
the women put their own savings and from which they take out loans, matched at
regular intervals by grants from DSWSG — and which are the seed-corn capital for
the multi-sectoral enterprises.
Conclusions and recommendations
The NGO examples in this paper indicate that the size of an NGO is the crucial factor
in its effectiveness in tackling the causes of poverty. The small and intermediate
NGOs in the area have had little or no effect on reducing the causes of poverty,
although the Jesuits and the Salvation Army had been achieving this on a limited scale
before their activities were curtailed. Despite their continuity, the Sisters of St Ann’s

Cyclone Mitigation,Resource Allocation and Post-disaster Reconstruction
33
Figure 3 Proposed structure of DSWSG linked with St Ann's and parent NGO
Note to Figure 3: A revolving fund is a joint bank account in the names of all group members, managed by the
*
women and the coordinator, into which the women put their monthly savings (and any other contributions, all of
whichattract acommercialrate of interest)andfromwhich theyalso takeout loans,inturn,atcommercial ratesof
interest. Every18 months the fund receives a matching grant from DSWSG funds or from government sources.
and the Sisters of St Joseph of Cluny have been unable to reduce the rise in poverty
and the DSWSG has yet to achieve its stated aim. By contrast, the track records of
Grameen Bank and SEWA show that large co-operative structures can provide a
supportive framework which enables poor people to help themselves. The evidence
suggests that such structures succeed for three reasons:
•
•
•
They are big enough to attract sufficient resources to enable their ‘beneficiaries’ to
compete and take their place in the local markets.
They have a policy and a structure, supplying training and education that enables
their ‘beneficiaries’ to take responsibility for running their programmes.
They provide a cohesive social network able to withstand individual failures, the
corrosive socio-economic forces of local politics and the pressures of the political
economy.

34
Peter Winchester
The Indian government now faces the same problem at the end of the century as the
British government faced at the beginning: that of striking a balance between safety
and economic growth. Successive government policies to reduce vulnerability and
poverty in the coastal areas have succeeded in reducing loss of life but have failed to
reduce poverty due, not to failures in the delivery strategy, but to failures in the
delivery structure. Chronically, resources allocated at state level for poverty
alleviation get diverted at local level (see also Rao and Erappa, 1987). For decades
government civil servants have tried to solve this problem but their efforts have
largely been dissipated in a web of local alliances; they have also been hindered by an
historically uneasy relationship with local politicians.
The research findings, both historical and empirical, lead to the conclusion that the
problems in Divi Seema and other similarly remote areas could be resolved if NGOs
could do the following:
•
combine together to be of sufficient size and credibility to be able to receive and
handle aid directly from the large international financial institutions, such as World
Bank and IMF;
•
To work, in partnership with government, on huge-scale problems such as coastal
management along the eastern coastline of India, a coastline that is continually
under threat from cyclones.
Successive governments have been unable to alter local political economies and
thereby change significantly the socio-economic and political position of the poorest
people, in which case a partnership seems to be the most positive way to break the
cycle. The governments would continue to be responsible for the infrastructural
elements of coastal management through public-works programmes and other public
services. The co-operative NGOs would be responsible for carrying out the economic
aspects of the anti-poverty programmes, which have the possibility of reaching
millions of people. These programmes are normally well beyond their means.
Together these programmes should ensure both protection and reduced vulnerability
as well as bringing lasting economic benefits in cyclone and flood-prone coastal areas.
Acknowledgements
The author would like to thank Edmund Penning-Rowsell of Middlesex University for
helpful comments at an early stage, Vanessa Winchester of Oxford University for her
invaluable comments on numerous drafts of this paper and to the editor for her helpful
comments on the final draft. The author is grateful for financial support and would like
to thank the Royal Institute of British Architects, the universities of East Anglia and
Middlesex, in this respect. In India the author has received a large amount of help from
the villagers and village and other officials up to the highest levels of government.
Senior members of the Indian Administrative Service provided valuable background
documentation, particularly A.V.S. Reddy, I.A.S (Collector, Krishna district 1980–83)
and Mrs Chitra Ramanchandram I.A.S. (Collector, Krishna district 1996–98).
Notwithstanding these acknowledgements the author wishes to stress that he alone
is responsible for the views, analyses and conclusions in this paper.

Cyclone Mitigation,Resource Allocation and Post-disaster Reconstruction
References
35
Andhra Pradesh, Government of (1971) Final Report of the Cyclone Distress Mitigation
Committee. Office of the Chief Engineer, Major Irrigation, Hyderabad.
—
— (1977a) Statement on Cyclone and Tidal Wave on 19th November 1977 by Sri
Vengala Rao, Chief Minister of Andhra Pradesh. Hyderabad: Government Secretariat
Press.
—
—
—
—
— (1977b) Andhra Pradesh District Gazetteer, Krishna. Hyderabad: Government
Secretariat Press.
— (1979) Note on the Recent Cyclone by Dr Channa Reddy,Chief Minister of Andhra
Pradesh. Government Secretariat Press, Hyderabad.
— (1980a) Community Cyclone Construction Programme for Cyclone Prone Areas of
Andhra Pradesh. Revenue Department, Hyderabad.
— (1980b) Rehabilitation and Construction of Permanent Houses in Cyclone-affected
Areas of Krishna District. Masulipatnam: Office of the District Collector, Krishna.
— (1981a) Cyclone Contingency Plan of Action. Hyderabad: Revenue Department.
— (1981b) The Andhra Pradesh Natural Calamities (Relief and Rehabilitation) Act
—
—
1981.
Baan, P.J., A. Hulsbergen, C.H. and M. Marchand (1997) Veerkracht van de kust-
ontwikkeling en operationalisering van een ‘veerkrachtmeter’. Publ. No. Z2136,
Waterloopkundig Laboratorium, Delft.
Baker, R.P. (1981) Land Degradation in Kenya: Economic or Social Crisis. School of
Development Studies Discussion Paper No. 82, School of Development Studies,
University of East Anglia, Norwich.
Beck, T. (1994) The Experience of Poverty: Fighting for Respect and Resources in Village
India. IT Publications, London.
Black, M. (1992) A Cause for Our Times — Oxfam the First Fifty Years. Oxfam and
Oxford University Press, Oxford.
Blaikie, P.M. (1981) Class, Land Use and Soil Erosion. In Overseas Development Institute
Review (2): 52–66.
—
—, T. Cannon, I. Davis and B. Wisner (1994) At Risk: Natural Hazards, People’s
Vulnerability and Disasters. Routledge, London.
Bradley, D. (1977) The Health Implications of Irrigation Schemes and Man-made Lakes in
Tropical Environments. In R. Feachem, M. McGarry and D. Mara (eds.) Water, Waste
and Health in Hot Climates. John Wiley and Sons, Chichester.
Brammer, H. (1990) Floods in Bangladesh: II. Flood Mitigation and Environmental
Aspects. Geographical Journal 156(2): 158–65.
Canadian International Development Agency (CIDA) (1995) Andhra Pradesh Social
Forestry Project. CIDA. New Delhi.
Cassen, R. and Associates (1994) Does Aid Work? Clarendon Press, Oxford.
Chambers, R. (1982) Health, Agriculture and Rural Poverty: Why Seasons Matter. Journal
of Development Studies 18(2): 218–38.
—
—
— (1983) Rural Development: Putting the Last First. Longman, London.
— (1988) Poverty in India: Concepts, Research and Reality. IDS Discussion Paper No.
241, Institute of Development Studies, Brighton.
—
—
—
—
—, R. Longhurst and A. Pacey (1981) Seasonal Dimensions to Rural Poverty. Frances
Pinter, London.
— and M. Leach (1989) Trees as Savings and Security for the Rural Poor. World
Development 17(3): 329–42.
—, A. Pacey and L. Thrupp (eds.) (1989a) Farmer First: Farmer Innovation and
Agricultural Research. IT Publications, London.
—, N.C. Saxena and S. Shah (1989b) To the Hands of the Poor: Water and Trees. IT
Publications, London.
Divi Seema Social Service Society (DSSSS) (1978) Socio-economic Survey Conducted
May 1978: Six Months After the Tidal Wave Tragedy of November 1977.

36
Peter Winchester
Nagayalanka, mimeo. held in the Indian Institute Library, New Delhi.
Dube, S.K., P.C. Sinha and A.D. Rao (1981) The Response of Different Wind-Stress
Forcings on the Surges along the East Coast of India. Mausam 32: 315–20.
Government of India (1981) District Census Handbook, Andhra Pradesh, Krishna District.
Village and Town Directory. Village and Town Primary Abstract. Series 2. Parts X. A
&
B. Hyderabad. Government Secretariat Press.
—— (1991) Krishna District. Village and Town Directory. Village and Town Primary
Abstract. Series 2, Parts X, A and B, Government Secretariat Press, Hyderabad.
Government of Madras (1878) Annual Progress Reports of the Irrigation Branch of the
Public Works Department in the Madras Presidency. Government Gazette Press,
Madras.
—
— (1908) Administrative Report of the Irrigation Branch of the Public Works
Department in the Madras Presidency for the Year 1907/8. Government Press, Madras.
— (1938) Administrative Report of the Irrigation Branch of the Public Works
Department in the Madras Presidency for the Year 1937/8. Government Press, Madras.
—
Green, C., A. van der Veen, E. Wierstra and E. Penning-Rowsell (1994) Vulnerability
Refined: Analysing Full Flood Impacts. In E. Penning-Rowsell and M. Fordham (eds.)
Floods Across Europe: Flood Hazard Assessment, Modelling and Management,
Middlesex University Press, London.
Hartmann, B. and J.K. Boyce (1983) A Quiet Violence: A View from a Bangladesh Village.
Zed Press, London.
Hashemi, S.M., S.R. Schuler and A. Riley (1996) Rural Credit Programs and Women’s
Empowerment in Bangladesh. World Development 24(4): 25–39.
Hulme, D. and R. Montgomery (1994) Cooperatives, Credit and the Poor: Private Interest,
Public Choice and Collective Action in Sri Lanka. Savings and Development 18(3):
xxx
Jain, P. (1996) Managing Credit for the Rural Poor: Lessons from the Grameen Bank.
World Development 24(1): 79–89.
Jodha, N. (1991) Rural Common Property Resources: A Growing Crisis. Gatekeeper Series
No. 24. International Institute for Environment and Development, London.
Johnson, S. and B. Rogally (1997) Micro Finance and Poverty Reduction. Oxfam, Oxford.
Khandker, S.R., B. Khalily and Z. Khan (1995) Grameen Bank: Performance and
Sustainability. World Bank Discussion Paper 306, World Bank, Washington.
Klein, R.J.T., M.J. Smit, H. Goosen and C.H. Hulsbergen (1998) Resilience and
Vulnerability: Coastal Dynamics or Dutch Dikes. Geographical Journal 164(3): 258–
68.
Kothari, R. (1983) The Party and State in our times; The Rise of Non-Political Formations.
Alternatives 9: 595–618.
Krishna, A., N. Uphoff and M.T. Esman (1997) Environmental Entitlements: Reasons for
Hope; Instructive Experiments in Rural Development. Kumarian Press, West Hartford.
Legislative Assembly Bill 38 (1981) Andhra Pradesh Gazette. Part IV A. Extraordinary.
Government Central Press, Hyderabad.
Lewis, P. (1991) Social Action and the Labouring Poor. Vistar Publications, Delhi.
Lipton, M. (1977) Why Poor People Stay Poor. Temple-Smith, London.
—— (1996) Successes in Anti-poverty. International Institute of Labour Studies, Geneva.
Mitra, A. (1977) Terms of Trade and Class Relations. Frank Cass, London.
Nageswara Rao, K. and R. Vaidyanaham (1978) Geographic Features in the Krishna Delta
and Its Evolution. Proceedings of Morphological Evolution of Land Forms. University
of Delhi.
Oxfam (1995) Report on SEWA. Oxfam internal report. Oxford.
Pearce, D. and R.K. Turner (1990) Economics of Natural Resources and the Environment.
Johns Hopkins University Press, Baltimore.
Raghavulu, C.V. and S. Cohen (1979) The Andhra Cyclone: A Study in Mass Death. Vikas,
Bombay.

Cyclone Mitigation,Resource Allocation and Post-disaster Reconstruction
37
Rao, U.M. and S. Erappa (1987) IRDP and Rural Diversification. Economic and Political
Weekly XII(52): xxx.
Saxena, N.C. (1995) Forests, People and Profit: New Equations and Sustainability. Natraj
Publications, Dehra Dun.
SEWA (1981) We, the Self Employed: Banking with Poor Women. Internal Report, Oxfam,
Oxford.
Sheth, D.L. (1983) Grass-roots Stirrings and the Future of the Future of Politics.
Alternatives 9: 1–24.
— (1987) Alternative Developments as Political Practice. Alternatives 12: 155–71.
—
Thompson, P. and E. Penning-Rowsell (1994) Socioeconomic Impacts of Floods and Flood
Protection: A Bangladesh Case Study. In A.Varley (ed.) Disasters, Development and
the Environment. Belhaven Press, London.
—
— and P. Sultana (1996) Distributional and Social Impacts of Flood Control in
Bangladesh. Geographical Journal 162(1): xxx
Turner, R.K., I. Lorenzoni, N. Beaumont, I.J. Bateman, I.H. Langford and A.L. McDonald
1998) Coastal Management for Sustainable Development: Analysing Environmental
(
and Socio-economic Changes on the UK Coast. Geographical Journal 164(3): 269–81.
Wade, R. (1982) The System of Administrative and Political Corruption: Canal Irrigation
in South India. Journal of Development Studies 18(3): 287–327.
—
— (1984) The Market for Public Office: Why the Indian State is not Better at
Development. Discussion Paper No.1, Institute of Development Studies, University of
Sussex.
Washbrook, D.A. (1976) The Emergence of Provincial Politics: The Madras Presidency
880–1920. Cambridge University Press, Cambridge.
1
Westgate, K.N. and P. O’Keefe (1976) Some Definitions of Disaster. Disaster Research
Unit, Occasional Paper No. 4, Department of Geography, University of Bradford,
Bradford.
Winchester, P.J. (1979) Report on Conference After 1977 cyclone. Disasters 3(2): 173–7.
—
—
— (1981) From Disaster to Development: Notes from India. Disasters 5(2): 154–63.
— (1986) Cyclone Vulnerability and Housing Policy in the Krishna Delta, South India,
1977–83. Unpublished thesis. School of Development Studies, University of East
Anglia, Norwich.
—
—
—
— (1990) Economic Power and Response to Risk: A Case Study from India. In J.
Handmer and E. Penning-Rowsell (eds.) Hazards and the xxx of Risk. Gower, London.
— (1992a) Power, Choice and Vulnerability. A Case Study of Disaster Mismanagement
in South India. James and James Science Publishers, London.
— (1992b) Recurring Vulnerability in Disaster Prone Areas: Implications for Outside
Intervention. Working Paper No. 5, Centre for Resource and Environmental Studies,
Australian National University, Canberra.
Yunus, M. (1998) Banker to the Poor. Aurum Press, London.
Zazweta, A. (1995) Policy Hits the Ground: Participation and Equity in Environmental
Policy Making. World Resource Institute, Washington.
Address for correspondence: Flood Hazard Research Centre. Middlesex University,
Queensway, EN3 4SF E.mail address: FHRC1@mdx.ac.uk