The dialectics of decentralisation

Few experiments in decentralisation have been so thoroughly considered as the People's Campaign in Kerala, which establishes the conclusions of scholarly research that in cases of 'good government', relations between central and local government s are much more complex than proposed by the simple 'decentralisation logic'.

JOHN HARRISS

KERALA'S remarkable experiment with democratic decentralisation over the last four years deserves to be much more widely known, both because it shows just what can be achieved through the devolution of resources and of decision-making to local bodies, an d, at the same time, why the common prescription of 'decentralisation' will often do more harm than good. Decentralisation is one of the magic bullets offered in the current consensus (the 'Washington Consensus') on what is required to bring about sustai nable development that will deliver higher living standards to more people. It seems so obviously right that the more governmental authority is devolved to local bodies, the better informed government will be about the specifics of local circumstances, t he more accountable it will be because it will be more vulnerable to citizen pressures - partly because it is easier for citizens, in these circumstances, to become well-informed about what government is doing - and the more effective government will be because its actions are tailored more to local needs and conditions. Decentralisation, therefore, seems to go hand-in-hand with participation at community level, and with democratisation, as well as with more efficient government. I believe that these sy nergistic effects are being realised in Kerala. Certainly, over years of visiting many different parts of rural India, I have never seen such demonstrably active and well-organised local offices as the panchayat offices which I saw in Malabar recently.

But the pious hopes which surround decentralisation have often lent themselves to abuse. They also serve the neo-liberal interest of 'cutting back the state', and can easily be made to serve the interests of power-holding elites at both the central and l ocal levels. International comparative research has recently shown, indeed, that the outcomes of decentralisation are crucially influenced by the political relationships between 'centre' and 'locality', and by configurations of local power, which mean th at very similar decentralisation schemes can have different purposes and outcomes - sometimes serving to extend central power downwards through patronage, or to break potential sources of opposition. Research has also shown that 'good government' actuall y involves much more complex interactions across the levels of central and local government, and of civic action than is supposed in the stylised notion of decentralisation as the unidirectional transfer of power and funding from 'centre' to 'locality'. There is in fact an 'ironic paradox of decentralisation': strengthening the capacity of local government may actually mean that the government at the centre has to play a stronger role in certain critical respects. What has been happening in Kerala shows that the architects of the People's Campaign anticipated these conclusions of scholarly research.

Richard Crook of the Institute of Development Studies at the University of Sussex was particularly interested in the outcomes of decentralisation in the 12 cases he studied - which included both West Bengal and Karnataka in India, as well as Bangladesh a nd a number of countries in Africa and Latin America - in regard to poverty reduction. His strongest conclusion was that in West Bengal and Brazil, which were the most clearly successful instances, "pro-poor outcomes were a product of the synergy between local and central factors (in which) centrally funded, poverty relevant programmes (were) implemented in cooperation with local governments, and given a strong ideological and organisational impetus at the local level from the commitment of local employ ees and political activists. To ensure this it was necessary to have either a strong political party which mobilised an electoral coalition in favour of such policies (as in West Bengal), or (as in Brazil) elites who, in the process of competition, saw t he policies as a way of achieving popularity." In both cases the 'centre' (the State government) supported local people in their struggles against local power-holders - in West Bengal, of course, notably through agrarian reform. Decentralisation in Karna taka has had positive effects, no doubt, but it has not made government more responsive to poor people, because it has served - by contrast with the successful cases - to enhance the clout of local power-holders.

Crook's findings are amplified in studies, by Judith Tendler of MIT, of the government in the State of Ceara in northeast Brazil. From being one of the States with the poorest records of all in relation to human development criteria, it was turned around in a remarkably short space of time under the aegis of two young reformist Governors. What was involved was a transformation - not universally, or without resistance - in the performance of government employees. Where programmes have been made to work m uch more effectively than before it has been because of changes in work organisation and in the commitment of government workers. In many ways their patterns of work have come to resemble those which are associated with the most successful private-sector enterprises: they are flexible, involve team-work, and a client-centred, problem-solving approach (rather than the delivery of centrally determined 'products'). These in turn have been made possible through the creation by the state of an imagery of 'ca lling' around public service, in the publicity which has been given to public employees and the public celebration of their achievements. This same publicity has a flip-side, for it is also an instrument for monitoring the workers' performance because it has shown the public just what should be expected of government employees. Publicity has been given, too, to the importance of complaints from ordinary citizens about failures of local government performance.

DID all this happen because of 'decentralisation and participation'? The answer is, emphatically, 'No'. It involved, rather, a three-way dynamic between local government, civil society and an active central (that is, here, State) government. Most strikin gly the centre actually took away some powers from municipalities in order, paradoxically, to strengthen local government. The success of a health sector programme, for example, depended upon the fact that the central government took away powers of makin g appointments of local health workers from Mayors - so that community workers were no longer the dependent clients of local power-holders. And while it certainly was the case in Ceara that institutions in civil society played an important role in the im provement of governmental performance, their impact on central government was at least as important as at the local level. A professional association of health workers, notably, exercised a very strong influence on the government's health policies. Furth er, some of the significant institutions of civil society were established at the instance of government; and the state's publicity and information campaigns "had awareness-raising impacts on the public that were similar to those of the (no doubt) more ' purely' motivated messages of this nature from independent advocacy NGOs".

So in important ways the central government intervened, in a sense to support and protect citizens against local government, in the context of these three-way relationships, and - says Tendler - there is a "magic that was worked by distrust between centr al and local governments". Her point, of course, is not to advocate that central government should return to its old centralised ways, but to show, very much like Crook, that the relations between central and local government in cases of 'good government ' are actually much more complex than simple 'decentralisation logic' proposes.

The way in which devolution to local bodies has worked in Kerala shows up these 'dialectics of decentralisation' in sharp relief. Together with the bold step of devolving responsibility for a large share of Plan expenditure to local bodies, the State gov ernment has also intervened actively to assist local people in identifying opportunities and constraints, often through encouraging the activities of volunteers like those - notably - of the Kerala Sastra Sahitya Parishad (KSSP, the People's Science Move ment) who have been involved in facilitating local resource mapping, and in providing training in project planning. There is a three-way dynamic, therefore, between local and central government, and civil society. The presence of members of the KSSP on t he State Planning Board shows the possibility of the influence of a civil society organisation at the central as well as at the local level, exactly as Tendler found in Ceara; while non-governmental organisations (NGOs) provide inputs into locally concei ved plans (though not invariably so - some of the presentations that I heard at the International Conference suggested that NGOs in some instances have been setting up parallel structures and projects, which may in the end weaken the efforts of the local bodies).

There is an element, too, of the 'magic' worked by 'distrust' between the central and local government in the procedures which have been instituted for subjecting local plans to expert as well as to democratic scrutiny, albeit in a way which involves loc al volunteers and not just staff from the line departments of the State government. And the whole experiment is underlain - it cannot be stressed too much - by the prior accomplishment of land reform in Kerala. Inequalities persist in local society, of c ourse, but local power has been contained through land reform in Kerala as it has been nowhere else in the country. It is true, no doubt, as one panchayat member who described himself as an "independent, CPI(M) sympathiser" said to me, that where the pan chayat is controlled by the Communist Party of India (Marxist), there is a tendency to favour party members and sympathisers, but then the regular conduct of open and fair local elections, as well as the transparent procedures which have been put into pl ace for beneficiary identification, will make for accountability. And, of course, not all of the successful panchayats, as we found in Malabar, are under the control of the ruling party at the centre. There are few instances, anywhere, of so thoroughly c onsidered an experiment in decentralisation as this one. But in the absence of the kinds of conditions and procedures that characterise the Kerala case, and which entail, paradoxical though it may seem, a strong role for the (State) government at the cen tre, decentralisation is likely to involve neither the deepening of democracy nor positive developmental outcomes.

(The research studies referred to in this article are by Richard Crook with Alan Sverrisson, 1999, 'To what extent can decentralised forms of government enhance the development of pro-poor policies and improve poverty-alleviation outcomes?', Institute of Development Studies, University of Sussex; and by Judith Tendler, 1997, 'Good Government in the Tropics', Baltimore and London: Johns Hopkins University Press.)

John Harriss is Reader in Development Studies at the London School of Economics and Political Science.