PUBLIC ADMINISTRATION AND DEVELOPMENT, Vol.

7, 383-397 (1987)

The implementation of decentralization: a case study

of district planning in Agra District, Uttar Pradesh,
India
MUKUL SANWAL

SUMMARY
Though the literature on decentralization has clarified some issues, it has left unsettled the
question of the administrative structure for its implementation: who should take what
decisions, and at what level? There are limits to the conventional methods of classification.
where answers have been sought in terms of Western experience, participation and manage-
ment. and a legal perspective has been adopted with decentralization seen in terms of fully
elected local governments. semi-autonomous local governments. or coordinating commit-
tees. Rather than stress procedures of participation. the substantive content of decision-
making needs to he analysed. A classification of rural development programmes provides
an alternative model for the reallocation of functions within the system of government to
meet development needs. The case of district planning in Agra District. India. illustrates the
process of decentralization. underlines its composite character. and shows the importance of
the budget in initiating reform in the political-administrative system. Since decentralization
is now motivated by different factors than in the past. the requirements of the anti-poverty
programmes will move the decision-making process closer to the cutting edge of the village
level, with even greater responsiveness to demands from below.

Decentralization is one of the many strategies for rural development. That decen-
tralization is a major administrative reform for improved implementation of rural
development is widely recognized. Whether decentralization is an end in itself. or
a process to reach other goals, is debated. According to the dominant tradition
decentralization can be a hierarchical or spatial distribution of authority initiated
by the central government or by pressure from below (Conyers, 1984). The
structure dictates relations to local organizations: interventions organized by pro-
gramme content tend to be more responsive to central decision-making while
those organized by area tend to be more responsive to local needs and interests.
Decentralization is generally considered to take four major structural
formsdeconcentration within the central government, devolution to locally
elected units outside the central government, delegation to agencies distinct from
the central government, or privatization (Hicks, 1961; Maddick. 1963). In the
process of decentralization three major decisions are involved: what to decentral-
ize, to whom, and how to coordinate implementation. Since administration and
organization is essentially a matter of people, decentralization involves factors of

Mukul Sanwal is Director of the Administrative Training Institute, Nainital. Uttar Pradesh. India
0271-2075/87/040383-15%07.50
0 1987 by John Wiley & Sons, Ltd.
384 M . Sanwal
power distribution, areal jurisdictions and a relating of the parts to the whole-a
major restructuring of the political and administrative system. This explains why
it has been so difficult to define the term and to operationalize it.
The central argument in this paper is that the unimodal strategy of rural
development (Johnston and Clarke, 1982) requires development administration
to establish new and different relationships-in society, the political system and
bureaucracy-as a precondition for effective local organizations t o develop, and
eventually for the development of policies that will give a greater role to, and
promote the generation of income for, the poor. While rural development is a
political task (Huntington and Nelson, 1976) its organization has to be adminis-
tered, and the organizational form adopted will largely affect outcomes. The
design characteristics of rural development programmes, rather than western
models, should determine the institutional arrangements for decentralization-who
has to take what decision and at what level.
The paper is in three parts. In the first part the issues in the study of decentraliz-
ation are discussed. In the second part a specific case of decentralization is used
to illustrate the processes of implementation. In the third part lessons for policy
are drawn.

ISSUES IN THE IMPLEMENTATION OF DECENTRALIZATION

There are two philosophically related trends in public administration, and both
have affected the study of decentralization. The first trend concerns greater
participation of citizens in government decision-making and implementation. The
second trend relates to changes in attitudes, values and structures of public
organizations so as to adapt them better to changes in the environment and in
the demands placed upon them. These trends explain the two major characteristics
emphasized by writers on decentralization. First, administrative devolution, involv-
ing a local organization and participation, is considered desirable (Cohen and
Uphoff, 1977). Second, administrative deconcentration with linkages between the
central and local units for the two-way flow of resources, influence and information
to sustain the process, is considered necessary (Leonard and Marshall, 1982). A
point that has not been appreciated in the literature needs to be stressed; in intent
decentralization concerns relations between the government and the masses. but
in its implementation decentralization involves relations wirhin the national govern-
ment. We need to consider alternative ways in which the choice of institutional
arrangements can be approached. The centre will need to give up some of its
power and there will have to be a powerful political incentive for this, a critical
factor that is generally ignored.
The driving force for decentralization and the form it takes are largely influenced
by the needs of the political and social environment. Thus the decentralization
programmes of the 1950s and 1960s in the developing countries were modelled
on western systems of local government, with the object of encouraging political
education and involvement at local levels, for extending the reach of the political
system rather than sharing power (UN, 1962). The present trend towards decentra-
lization is a way of increasing the effectiveness of rural development programmes
by making them more relevant and responsive to local needs and conditions,
 The Implementation of Decentralization in lndia 385

allowing greater flexibility in their implementation and providing a means of

coordinating the various agencies involved at the local level (Rondinelli, 1983;
Cheema and Rondinelli, 1983). Since western models are used, evaluations com-
pare these efforts to western experience; the earlier decentralization was evaluated
in terms of participation just as the present decentralization is being evaluated
in terms of management.
Parallels from the experience of developed countries are inappropriate for
development administration. There are three major differences in the conditions,
which are usually ignored. First, in developing conditions, even with decentraliz-
ation, the central bureaucracy continues to take decisions at local levels, rather
than locally employed staff. Second, financial resources continue to be provided
by the centre, and local decision-making involves primarily allocation and location,
rather than resource mobilization and control over financial affairs. Third, the
requirements of coordination are seen in terms of complex management practices
and structural adjustments, heirarchical control, optimal allocation of responsibilit-
ies and the monitoring of system performance, ignoring the fact that such linkages
have limited validity in developing countries where more than one agency is
involved and administrative influence is not limited to regulation but involves a
substantial element of politics. In developing countries interdependence of politics
and administration, with mutually respected jurisdictions, is a part of the organiz-
ation structure (Somjee, 1982). Organizations, especially successful ones. also
build up their own constituency and coalitions of support.
There is a relationship between the administrative and political roles of field
administrators. Colonial government was highly decentralized, with the delegation
of administrative, judicial and political authority to field officials responsible for
law, order and good government. The most important task of the District Officers
was to maintain political stability and the existing system of government. After
independence this apparatus has been retained in almost all the developing coun-
tries for preserving the power base of the new regime, with two major differences.
Firstly, in colonial government field administrators substituted for political
structures-interest articulation, aggregation, political communication. As society
becomes politically stable and accommodates the political process within differen-
tiated structures this role diminishes. Secondly, field administration by initiating
economic and social change has an impact on the political system; economic
development brings in new social and political values. Since the District Officer
maintains public order, it is easy for the institution to be labelled authoritarian
and undemocratic. Functional decentralization to departments with Ministers of
their own actually leads to more centralization. Since all decisions tend to gravitate
to the political power centre implementation suffers.
It is only now being recognized that programmes for human development that
governments seek to implement require spatial decentralization and coordination.
Providing irrigation through canals can be done effectively by engineers as long
as the concern is protection against drought, but with the Green Revolution the
timeliness and requirement of a minimum quantity of water brings on questions
of access, use and farmer organization. The electricity board can lay transmission
lines very efficiently but rural electrification requires the planning of these lines
according to village concentrations, and an economic return on the investment is
dependent on power connections taken for village industries. Health clinics await
 Table 1. Design strategy for rural development programmes

Mainly infrastructure Mainly Welfare Mainly anti-poverty programmes

programmes programmes (3)
(1) (2)

1. Goals Physical output, Human development, Social change, redistribution

construction distribution, extension of
technology
2. Activities Expansion of services, Health, education, Economic diversification of households for
buildings, roads, nutrition, family planning generating incomes; assets transfer; securing
irrigation, rural rights in land reforms and common property
electrification, housing, resources; house sites and houses; individual
drinking water economic activity; wage income and minimum
wages; social security. pensions, legal aid
3. Beneficiaries General public, mainly in Geographical coverage of Personal well-being of households in special
less developed areas target groups of the poor, need-including agricultural labourers, groups
including women and in backward regions, tribes, women, aged
children widows and handicapped
4. Constraints Finance, information Organization. supply, Social institutions. erosion of traditional rights.
access lack of opportunities, accountability and
motivation of bureaucracy
5. Resources Capital-intensive, Management-intensive, Labour-intensive. local initiatives and
technology, investment mix of services, synergy organization, appropriate technology, credit,
subsidy, legal aid
6. Nature of decisions Pre-programme/project, Recurrent, maintenance- Continuous, decentralized selection of
investment-oriented at the oriented coordination at beneficiaries and projects, sustenance-oriented
regional or state level the district level at the village level
7. Formulation Designing, primarily Providing, primarily social Enabling, primarily group processes
engineering services
8. Implementation Construction project Top-down service delivery Bottom-up initiatives for use
9. Organization Government, Government, Implementer-client village level committee,
supplemented by private supplemented by local paraprofessionals, voluntary organizations
representatives
10. Evaluation Input, cost-effectiveness, Output, quantitative Outcomes, qualitative changes, ability to meet
simple, with frequent changes, complex, with changing beneficiary needs, extremely
redesign, short time path sporadic redesign, long complex, irregular redesign, short time path 3
for costs and benefits time path for costs, for costs, and benefits to accrue immediately, s
benefits apparently early basic changes take time to accrue g.
%
388 M. Sanwal
the patient, but family planning requires medical personnel to go out to the
people, and approach the village leaders. Decisions that are frequently but not
routinely made, that require local knowledge, flexibility and beneficiary response,
need to be decentralized to the local level (Montgomery, 1980).
In the political economy of developing countries decentralization should be seen
as a process, and in terms of the preponderance of central or local influences or
accountability in decision-making. The essential characteristics are political will to
shift the locus of decision-making for planning and implementation closer to the
field, with a locally based set of persons, political and administrative, jointly
taking decisions, with continuing financial, administrative and political supportive
linkages with the centre. Since these characteristics are both cognitive and struc-
tural, decentralization will be an incremental process.
Thus, for operationalizing decentralization, in addition to political will. the
critical factors are internal to the administration, or the organization of the public
service, and not, as is commonly assumed, external to the central administration
in the creation of a separate legal unit, or the nature and form of participation.
The process has two important aspects. Firstly, the distinction between functions
and resources must be recognized. Decentralization of functions can be a way of
reducing social responsibility (Guibert and Lanvin, 1983). decentralization of the
budget effectively decentralizes decision-making. Secondly. since human develop-
ment programmes (basic human needs) predominate an intermediate level between
the centre and local organizations-the district-will be the key determinant in
decentralization. When anti-poverty programmes. which specifically target benefits
to the poor, are taken up a further decentralization of decision-making is required,
and the village becomes the key implementation level. The objectives and the
form of decentralization in developing countries should meet the requirements of
rural development programmes (see Table 1). Generalizations made from case
studies of experiences in the developing countries (Rondinelli. 1983) do not draw
a distinction between the administrative arrangements required for programmes
and arrangements for.the structure of government; programme requirements will
be influenced by government, but the two are quite distinct. The structure of the
national government is the critical variable.

DISTRICT PLANNING IN UTTAR PRADESH

The merit of the decentralization scheme implemented in Uttar Pradesh lies in

recognizing these political and administrative constraints and opportunities. in its
three important features. First, the existing structural arrangements were modified,
blending functional and spatial decentralization, and a separate agency, legal unit,
or national organization. has not been created. Second, there is political rather
than administrative domination. Third. decision-making is regulated by a well
defined set of norms rather than by persons, blending participation and manage-
ment (Table 2). These features have ccmbined to sustain the process for over 2
years, and have made the change irrevocable.
District planning in Uttar Pradesh recognizes the interrelationships between
policy formulation and implementation, and the budget process has been decentra-
lized. The development budget has been divided into the State and District sectors,
 The Implementation of Decentralization in India 389
Table 2. District planning framework
~~ ~

State Level District Level Local Level

1. Actors Chief Minister, District Minister, Village leaders, field-

Ministers, State- District-level level administrators,
level administrators political leaders. public
District-level
administrators
2. Issues Goals, allocation Locational decisions, Assistance. feedback
decisions, control consultation,
and support, feedback
redesign
3. Structural Design (and Implementation Implementation, local
Features redesign) (and design), organization
vertical-horizontal
integration,
linkages with other
agencies,
coordination-control
relations
4. Process roles Political support Administrative Demands, informal
through meetings commitment, meetings. field visits.
and visits. meetings, consultation. feedback
consultat ion consultation.
consensus
5 . Effects (+) Commitment to Speed and quality of Public initiatives
development of decisions. changes in
backward areas and roles and attitudes
poor in bureaucracy,
political leaders
(-) Reform Mediating leadership Village factions lead
dependent on role of District to some delays
continued support Minister and District
of Chief Minister Officer crucial to
success

in the ratio 70 : 30. Major programmes and projects whose benefits cover a
number of districts are retained in the State sector-such as major canals, while
the District sector comprises programmes of local utility-tube wells and minor
irrigation works. Simple guidelines have been laid down for the formulation of
District Plans: committed expenditure is to be covered; the priorities of the
National Plan are to be adhered to, and the Plan is to provide 60 per cent to
productive schemes-providing employment through agriculture, irrigation. etc. ;
and 40 per cent to non-productive schemes-roads, electricity, education, medical
and health, etc. Allocation between Districts is done on the basis of a simple
criterion of backwardness-taking total population, population of scheduled
castes (who tend to be the landless and thus an index of poverty) and development
of infrastructure (UP, 1981) in the District (roads and electrification). In this
390 M. Sanwal
arrangement backward Districts get more than in the past, and the measure will
reduce regional disparity.
Unit costs and time schedules for completion of activities provide for a realistic
planning and implementation exercise. At the District level the requirements are
kept simple so that they can be met by the available technical staff. Essentially,
only two decisions have to be taken: sectoral allocation of the surplus funds which
requires an assessment of the physical infrastructure, land use, commercial activity
and potential and identification of a predominant activity of each block, or sub-
unit; and location of the infrastructure facilities, which requires a cartographic
depiction of unreserved areas. The decisions are to be taken by a Planning and
Monitoring Committee, chaired by a Minister (not from the local area), and
comprising all elected members of the State legislature and the National Parliament
and the District Officer. The final approval is given by the State Planning Depart-
ment, and the proposals are then incorporated in the State budget. In cases where
the administrative departments oppose certain allocations they can appeal to
the Planning Department, which as a rule agrees with the Districts, if the Plan is
in accordance with the norms.
How the reform was formulated was as important as what is being done. Political
acceptability was obtained by the participation of those who could sabotage it-a
process which took 6 months. A committee of ministers toured the country and
put up their proposals to the Council of Ministers. The reform thus had, at least.
the tacit consent of those who stood to lose the most power. The members of the
legislative party stood to gain, and welcomed the reform. Even then, the idea was
sold by the Chief Minister to the legislative party of the government in a 2-day
session. He stressed the importance of objective criteria, in taking allocation and
location decisions; that the planning exercise would be a buffer for things not
happening, and would demonstrate a positive relation with the electorate. It would
also underline the importance of discussing resources and equitable regional
development rather than just demands. N o public announcement was made, but
a meeting of all elected representatives of the State legislature, Heads of Depart-
ments and District Officers was chaired by the Chief Minister, to explain the
reform and to obtain their views on it, a level of participation consistent with
expectations. The Chief Minister also toured major districts and personally
exhorted the District Officers and elected representatives to make the reform a
success; he regarded this as the most important measure of his government.
The reform itself was simple, unambiguous and carefully planned, effecting a
balance between competing interests. A statutory body was not created; flexibility
in design was thus retained and adverse political repercussions avoided. A second
tier would adversely affect the political position of members elected to the State
legislature. The chairmanship of the Minister stressed the political and participa-
tory nature of the Committee, and being from outside the area he was expected
to take an objective view.The presence of the Minister also raised the status of
the Committee, enabling the District to have an idea of the thinking and policy
of government; the aspirations and needs of the District could also be conveyed
to government. The presence of the District Officer in the Committee fixed the
responsibility for implementation of the decisions on the bureaucrats.
Since relations between officials and non-officials are still adversarial in the
State, bringing them together in implementation has been avoided. The planning
 The Implementation of Decentralization in India 391
and monitoring committee includes all elected members of the legislature from
the District, thus coopting all the interest groups. A separate implementation
committee of all departmental officers has been formed under the District Officer
for affecting the necessary coordination, without politicization. Thus allocation
and locational decisions in the District sector have been given to a committee
guided by a set of norms, so that no area (or elected representative) can unduly
benefit at the expense of others. This proved a politically acceptable proposition:
if it is not possible to favour friends, the system ensures that opponents are not
favoured either.
Success, however, depended on the transactions during implementation, which
in certain aspects extended the definition of clients and implementers from
District-level politicians and District-level officers to include all parties, field-level
bureaucrats, and village-level leaders. In Agra District three modifications were
deliberately made to assist implementation, which were subsequently rec-
ommended by the State government to all the Districts. Firstly, since all elected
representatives of the government party and and the opposition were members of
the District Committee certain ground rules or conventions needed to be accepted
by all, at least informally, to avoid clashes in the meetings. At the instigation of
the District Minister and the District Officer, an acceptable principle emerged-not
everyone could get tube wells or roads, but each elected representative could get
an approximately equal share of the allocations which would be used to bring the
infrastructure facilities in each development block up to the level of the District
and State average. Most important, they could choose the location from amongst
the unserved areas. A perfectly rational allocation and locational decision (if that
is possible) would have been politically unacceptable and could have created a
deadlock; this happened in some Districts. It would also have been unnecessary:
locational decisions confer benefits and so are political and not administrative in
nature. The important point (of the ethics of administration) is that implementation
requires bargaining, which can be carried out within the prescribed norms without
affecting efficiency, effectiveness and equity; by doing this implementation can
also change the environment to its advantage.
Secondly, the gram sabhas (village assemblies), though outside the purview of
the design, were involved in the decision-making and evaluation process through
a series of meetings held prior to finalization of the District Plan. In these
meetings, open to all local villagers, members of parliament and the legislative
assemblies, as well as district and field-level bureaucrats from that area, were
requested to be present. It was seen that open debates bring out balanced argu-
ments. This made it more difficult for the elites to get away with heavy-handed
political manipulations that could sabotage the District Planning process.
Thirdly, a simple bar chart prepared by the District Planning office compared
the level of infrastructure in the development blocks, and cartographic depiction
of the existing infrastructure facilities highlighted the gaps in the distance to
services and the distribution of existing services. CPM techniques also helped in
decision-making and did not require sophisticated analysis or data collection
requirements on technical staff which was not available in the District. This helped
in decision-making as well as in coordination among the agencies involved in
implementation.
In the process of implementation new relationships were established, which
392 M. Sanwal
provided for continuous negotiation and dispute settlement. The norms for allo-
cation and locational decisions simplified roles and duties, leading to immediate
and visible results. Local capacity was built up in identifying problems, finding
solutions, resolving conflicts, and in management. Decisions were more rational
than when taken in the capital city, since there was less scope for patronage. The
vertical relations within departments changed, with more emphasis on meeting
time schedules, timely availability of materials and making field officers seek early
decisions. The District Committee also introduced a new monitoring system; the
ministers provided government and the departmental ministers with an additional
channel of information on the actual state of affairs, leading to a better appreci-
ation of the local problems. Ministers and Secretaries at the central level had more
time for policy issues and major programmes and projects, as decentralization of
locational decisions greatly reduced the number of visitors coming to influence
decisions.
The adversary relations between non-officials and officials were also affected
with the improved lateral communications. The elected representatives became
more aware of the problems and constraints of implementation, and consensual
decision-making modes emerged. The cooperative partnership was most visible in
jointly pressurizing government for the provision of adequate funds and equipment
for the District. Involvement in decision-making also brought responsibility on
the elected representatives for settling subsequent local disputes, which even
infrastructure projects require. Rural electrification is financed through the Rural
Electrification Corporation, which allots funds for village electrification provided
there is a minimum return from domestic connections and the power-load of small
rice, oil or flour mills, etc. Progress of this important scheme had been poor
because the State Electricity Board did its planning and implementation in iso-
lation. In the new scheme of things the responsibility for collecting the required
number of applications was placed on the elected representatives. They were able
to do this within 1 month, showing the crucial role local involvement can play in
extending the outreach of administration.
The most noticeable change was in the commitment of the officials. Since
decisions were now taken at the District level, the officials were committed to
achieving their targets. A shortage of equipment was no longer used as a ready
excuse. The tube well engineer borrowed rigs from a World Bank project in the
neighbouring district where site selection had not been done and the rigs were
idle. The Medical Officer was even able to obtain laparoscope teams from the
neighbouring region to keep up the momentum of the family planning programme.
Rural electrification had suffered in the past because individual villagers found it
difficult to spare the time, and lacked the confidence to go to the district head-
quarters. The electricity engineer agreed to complete the formalities at the village
level. The anti-poverty programmes, for small farmers and landless labour,
improved both in terms of meeting targets and quality. Field-level implementers
(Block Development Officers) took bank managers with them to villages, dates
were fixed in banks for groups of villages in order to expedite procedures in
obtaining loans, and villagers were taken in groups to village markets to select
the best milch animals, carts and implements. Improved transactions among the
implementers, and between them and clients, meant that procedures usually cited
as the major obstacles to implementation melted away.
 The Implementation of Decentralization in India 393
Horizontal relations within the bureaucracy also changed. In place of failure to
coordinate activities, inter-departmental rivalries and conflicts during implemen-
tation, the 'ego-involvement' of District-level officers in decision-making motivated
them towards better task performance: targets were met, the budget utilized, and
implementation speed and quality improved. This required no additional monetary
incentives, staff or equipment but only a sharing of tasks and responsibilities.
Interdependency of the team or group was also stressed in order collectively to
face the elected representatives better. The process of planning in the Districts
and the proposals for the budget, were initiated 6 months before the budget was
to be presented to the legislature. Sanctions were issued within 1 months of the
beginning of the fiscal year, and with standardization in the engineering depart-
ments implementation started immediately-a major administrative reform by
itself.
For the generalist administrator the workload increased. The more enterprising
departmental officers began using the District Officer directly to move govern-
ment, by-passing the normal department channels for quicker decisions. The
District Officer was also able to draw the attention of the government and heads
of departments towards vacant posts, problems of transportation and mobility.
and timely release of funds. The District Officer had a delicate role to play.
Essentially he had to win the cooperation of his 'team' through persuasion rather
than control; the department officers were wary of his possible 'domination'.
Good personal relations were important; in this joint touring helped. At times.
however, threats had also to be administered. This did not sour relations because
everyone understood the underlying purpose, and regular monthly meetings helped
to keep the lines of communication open and avoid misunderstandings. The new
system required more initiative and anticipation of problems rather than the usual
administrative style of merely reacting to events. It was participatory and improved
management and put implementer discretion to good use. The result was that
unexpected problems, common in development administration. got a quick
response, and there were fewer delays in implementation. A new organization
culture developed and the quality of administration improved.
Once the villagers, the principal clients, had got over the initial surprise of
seeing the bureaucrats and political leaders coming to them. rather than their
going to the offices to wait upon the 'bosses', they soon get into the spirit of
things. They started pressurising the officials for implementation, though the
means and the demands varied. Some groups of villagers sought to establish good
relations with selected officials to get the benefits earlier. This was not corruption
but khushamad (pleading). The locational decisions for tube wells were taken by
the District Committee, but the tube well engineer decided in which order these
tube wells would be constructed. Villages which got tube wells early could use
the water for an extra crop. Some groups sought to force decisions-they would
construct school buildings and then press for the posting of a teacher. Some groups
took direct action to press their claim. and stopped traffic on the highways till the
officials came and repaired their electricity transformer, changed the location of
a school to a more central point or provided public standposts according to the
size of the population in the water supply scheme. Greater pressure. through the
voicing of complaints, was also felt in State sector activities-cleaning of canals,
maintenance of roads, upkeep of hospitals, and attendance of teachers in schools,
394 M. Sanwal
to name a few activities of government that remain neglected in the absence of
pressure from below.
Similar results were achieved in another critical administrative function, the
maintenance of projects. In social forestry, site selection within the village grazing
lands requires sufficient to be left in the right places for the animals of the village.
When this locational decision was left to the village, villagers came up with offers
of land, where previously there had been continuous disputes. It became possible
to develop the existing resources-ground water potential was exploited in difficult
areas which the departments had avoided since villagers helped in transporting
heavy machinery over difficult terrain. The positive response of the villagers, and
the gratitude they expressed to officials, acted as a source of motivation for the
bureaucracy, and the existing patron-client relationships started changing to one
of coequal partnership in implementation.
Along with efficiency and effectiveness, equity considerations were also met to a
greater degree than before. The Development Plan now gave priority to backward
Districts. The allocations for sectors like agriculture, social welfare and education,
which affect the poorer areas as well as the poorer people, increased substantially.
A closer look at the problems of the District brought attention to neglected areas.
Officers did more touring, and so visited more areas and came in direct contact
with the poor, rather than dealing only through intermediaries. Programmes that
benefited the poor and neglected areas were given priority: for example, primary
education, the least costly and most useful infrastructure activity, could be pro-
moted, and in 1982 98 schools were opened in Agra District, while the average
of the previous 5 years had been seven. There was greater responsiveness to the
needs of the poor.
The most significant results were achieved in the implementation of programmes
which depended for their success on a positive client response. Family planning
had been taboo in the State since 1977. The transactions involved in the implemen-
tation of District planning changed the environment in the District, and the field-
level implementers as well as the village heads could be exhorted to motivate
eligible couples. The District performance shot up and government mustered the
courage to quote this example to other Districts. Decentralization strengthened
local organizations.
The costs of too much participation also emerged. It was difficult to determine
in all cases whether the headman reflected village opinion, or the opinion of a
faction. Disputes arose especially over location of schools. It was also seen that
the interests of the landless generally went unrepresented. On the whole the
initiative of involving the village in decisions which directly concern them helped
in the speed as well as quality of implementation.
District planning had its problems, costs and inadequacies, some of which
have been subsequently rectified. No training for officials and non-officials was
organized at the District level. The release of funds continued to be through the
departments and there were instances of diversions from one District to another;
separate budget-heads for District sector schemes and direct allocation to districts
from government had been suggested in 1983-84. Flexibility in implementation
also required that for District sector schemes the powers of reappropriation of
the head of department be delegated to the District level, to complete the
decentralization of the budget process.
 The Implementation of Decentralization in India 395
Departments also tried to push schemes whose benefits were not limited to a
District, and schemes involving recurring expenditure, into the District Sector in
order to gain more funds for the State sector schemes. Categorization of District
sector schemes into productive and non-productive has also been disputed; e.g.
some feel that education is a productive scheme. There are also examples of
staff-oriented schemes being sanctioned; and of schemes which provided for
disproportionate subsidies which could have an adverse impact on other pro-
grammes. These were teething troubles, and on the whole the reform worked
well. The Planning Department was quick to acknowledge difficulties and sugges-
tions, and had an open mind in affecting improvements. Most important, the
interest of the Chief Minister, and the role of the Planning Department, in
mediating the processes in favour of the District level was crucial in institutionaliz-
ing the reform.

CONCLUSION

Writers on decentralization generally conclude that it has not lived up to its

expectations. The problems are divided into three broad categories: the degree
of decentralization has been limited, decentralization has done little to improve
planning and implementation, and power has been transferred to the elites so that
there is no meaningful increase in the participation of the masses (UNCRD, 1981).
Clarity of objectives, and careful design and management, will reduce these
problems. Evaluation criteria should consider the programme characteristics. the
specific decisions that are decentralized, and the actors involved in these decisions.
Legal questions about the nature of the authority are important, but it is more
important to see how decentralization is actually operating (GOI, 1984). Sufficient
evidence is available to show that allocation and locational decisions, as well as
implementation decisions, e.g. choice of beneficiary and project for activities that
affect local people, when taken at the local level improve government efficiency.
effectiveness and equity.
Policy prescriptions must take a realistic view of both the constraints and the
opportunities of the development process. Decentralization combines elements of
politics (who benefits), organization theory (structural changes) and bureaucratic
reorientation (change in tasks, roles, attitudes and behaviour patterns), and the
concept cannot be explained by merely looking at one of these elements. The
process will be politically dominated, especially since at the implementation level
the interplay of politics and administration creates tensions. Efforts to institutional-
ize these changes require norms to order relationships within the bureaucracy and
between the bureaucracy and the elected representatives. The collection and
presentation of data in an easily understood format facilitates decision-making.
The process, in the initial stages, has also to be overseen from outside the
bureaucratic stream in the Planning Department or Chief Ministers office.
For development administration the budget is more important than the plan.
Decentralization of the budget achieves the desired results with a minimum of
bureaucratic and political dislocation, and helps to build up local capacity. Budget-
ary reform is also useful and acceptable to central governments chronically short
of resources. Decentralization of the budget is easy to operationalize because of
396 M. Sanwal
its clarity and simplicity, and because it specifies the goal and the scope of
decentralization, taking into account local capacities and political constraints.
Secondly, decentralization of the budget modifies government structures and
central procedures, and the scope can and should be expanded as political support
and administrative reorientation increases and conflict decreases. Thirdly, decen-
tralization of the budget and the new bargaining relationships require new roles
in interactions within the bureaucracy to create appropriate supportive linkages, as
well as between the bureaucracy and the poor for mobilizing support for change.
Decentralization is a tool to meet the requirements of development programmes
rather than a form of Government. The essential requirement is the development
of local capacity, or consensus-building and commitment at the local level. The
emphasis has to be on incremental changes in the political and administrative
process, if the reform is to be sustained. Though decentralization will be initiated
from outside it cannot be sustained by external interventions.

ACKNOWLEDGEMENTS

The case study is based on the author's experience as District Officer in Agra
District, when District Planning was introduced. Academic work for this paper
was done at Harvard University. The author is grateful to John Montgomery,
John Cohen, Donald Warwick, and David Leonard for their comments. The views
expressed and the errors are, of course, the author's.

REFERENCES

Cheema, G. S. and Rondinelli, D. A. (1983). Implementing Decentralization Programme

in Asia: Local Capacity for Rural Development (UNCRD), Nagoya, Japan.
Cohen. John and Uphoff, N. T. (1977). Rural Development Participation: Concepts and
Measures for Project Design, Implementation and Evaluation, RDC. Cornell University,
Ithaca, New York.
Conyers. Diana (1984). 'Decentralization and development: a review of the literature',
Public Administration and Development. 4 (2), 187-197.
GO1 (1984). Report of the Task Force on District Planning, Planning Commission, Govern-
ment of India. New Delhi.
Guibert, G. and Lanvin. B. (1983). 'Decentralisation in government: the United States
and France compared', Journal of Policy Analysis and Management, 3 ( 2 ) .
Hicks, U. (1961). Development from Below: Local Government and Finance in Developing
Countries of the Commonwealth, Clarendon Press, Oxford.
Huntington, Samuel P. and Nelson, J. M. (1976). No Easy Choice: Political Participation
in Developing Countries, Harvard University Press, Cambridge, Mass.
Johnston, Bruce F. and Clark, W. C. (1982). Redesigning Rural Development. Johns
Hopkins, Baltimore.
Leonard, David K. and Marshall, Dale Rogers (eds) (1982). Institutions for Rural Develop-
.
ment: Decentralisation and Organisational Linkages, ISS, University of Berkely Califor-
nia.
Maddick, H. (1963). Democracy, Decentralisation and Development, Asia Publishing
House, Bombay.
Montgomery, J. D., (1980). 'Administering to the Poor (or, if we cannot help rich dictators.
what can we do for the poor)', Public Administration Review, 40 ( 5 ) .
 The Implementation of Decentralization in India 397
Rondinelli, Dennis A . (1983). 'Implementing decentralisation programmes in Asia: a
comparative analysis', Public Administration and Development, 3 (3), 181-208.
Somjee, A. H . (1982). 'The technomanagerial and politico-managerial classes in a milk
cooperative', Journal of Asian and African Studies, 17.
U N (1962). Decentralisation for National and Local Development, New York.
UNCRD (United Nations Centre for Regional Development) (1981). Implementing Decen-
tralisation Policies and Programmes, Report of the UNCRD Workshop, Nagoya, Japan.
UP (1981). Guidelines for the Formulation of District Plans, Planning Department, Govern-
ment of Uttar Pradesh.