What are the implications of this assessment? Is the PRSP idea still viable? I believe this needs to be answered in two parts. First, the observation that donor-driven policies do not work has lost none of its validity.

In fact, it has been powerfully reinforced by the reform experience of the last few years. In this respect, if PRSPs did not exist, we would have to invent them: there is no question of going back to what existed before 1999/2000, because none of the problems which that initiative was trying to address have gone away.

Second, the implicit theory about political change that underlay the concept and its operationalisation has been proven naïve.

The theory says that the more participatory reform processes are, the more likely they are to be effective.

The claim is that, having committed themselves to doing things in a dialogue with other stakeholders, governments will be more likely to be called to account for their actions and their results. They will therefore take their responsibilities more seriously.

The theory is not easy to refute directly. It has a history. The belief that participation improves effectiveness gained considerable support from studies and evaluations of project and programme management over several decades.

Indeed, 1999 was the occasion when this accumulated wisdom seemed to many people to be ripe for scaling up to national policy making. This was also a culminating moment in the long flirtation between the big official agencies and the development NGOs, for whom the inclusion of civil society is a litmus test of sound policy making.

The idea had, and still has, the additional attraction of suggesting that all good things go together. It is ideologically congenial for donors based in rich developed countries to think that inclusive and democratic practices are good for development, as well as good in themselves.

It is not surprising, therefore, if the theory continues to influence donor thinking in spite of the limited benefits that seem to have been obtained as a result of citizen participation in PRSP processes.

A typical view today is that PRSP processes may have had important weaknesses as exercises in participatory planning, but the solution is to do this better … for example doing more to involve parliamentarians and the private sector, and giving civil society organizations more time to prepare.

In other words, the results of the PRSP experiment have not yet dented very seriously the belief that involving a wider public in discussions about how to reduce poverty is a powerful means of building both commitment and accountability.

This is a belief that has real attractions. It is nonetheless naïve when applied without heavy qualification in the typical aid-dependent poor country.

I believe the PRSP experience has helped to show that this is the case, although it may not be news to political scientists and anthropologists who follow the situation on the ground, or indeed to informed citizens of the same countries.

The following are a few bold and unqualified generalizations about PRSP (that is, poor and highly indebted) countries.

They may well contain some exaggeration, and not apply well in particular cases; but if even half of what they say is true, the theory of political change that underlies the PRSP idea is mistaken.

A cogent argumnt is that in the politics of reform all good things do not go together. Many or most of the key decisions are made informally, by small groups of politicians linked together by networks of clientelism and patronage. Governments are not unified and well-coordinated actors.

To a wide extent there is no effective accountability, it often operates through the patronage system, and follows criteria other than those formally agreed.

Formal decision-making processes, even including the national budget and certainly including the PRSP, are from a political point of view window dressing, or largely theatre. Whereas in some middle income countries, there is an informed public opinion and a degree of effective parliamentary scrutiny, in many areas the needed preconditions (e.g., a large local newspaper readership and an effective press) do not exist.

The reasons are partly sociological: to function well, the institutions of formal, universal franchise democracy require a level of urbanization, literacy and de-peasantisation that is unusual at this level of economic development.

The easy notion that donors only have to back off from demanding accountability for aid funds in order for domestic accountability to flower in its place is unrealistic and unhistorical.

NGOs have been and will continue to be a useful means of promoting progressive measures that donors cannot promote directly. However, they cannot substitute for the absence of representative forms of civil and political society.

Could Development Cooperation do Better? Many of the observations above are widely recognized by donor staffs. The problem is not that they are personally naïve but that it is not clear to anybody exactly what else might be done to be supportive of country-owned policies under these circumstances.

The topic is under active consideration much of the time in several bilateral and multilateral agencies. Some proposals currently being discussed are certainly unwise if anything like the above is generally true.

Others make good sense but are insufficient. It would not be the right solution to abandon wholesale what the main group of donors in the developed countries currently consider to be the best approach.

Rather, the argument I want to advance is about the importance of going beyond the PRSP framework as currently conceived and addressing some of the missing links in the chains of causation needed to harness politics to the goals of development and poverty reduction in PRSP countries.

Two obvious focuses for discussion here are the aid harmonization and alignment agenda of the developed world donors,  and the arguments around general budget support as a preferred aid modality in a PRSP context.

Both are areas where there is no doubt donors could do better, but where the conventional concepts are insufficient.

The harmonization and alignment commitments made by high-level representatives of the donors in Rome in 2003 and Paris in 2005 are important landmarks.

This is so in spite of and partly because of the points made above about constraints at the country level. There are severe limitations to the ability of donors to engineer progressive change and create country ownership of policies.

However, there is not much doubt about the ability of the aid business to do harm. Both the multiplicity of competing donor agencies and the donor tendency to take over the policy-making function from country authorities (especially when this is done in aggressive or adversarial ways) have done, and continue to do, damage to whatever exists by way of country policy-making systems.

For that reason, it is right that the Rome and Paris Declarations should be regarded as priorities for implementation by all the signatory countries.

The qualification that needs to be added is that, even if fully implemented, these measures would not be a sufficient solution to the problem of aid misalignment, important parts of which are about the recipient side of the relationship.

I would argue that the same applies to the promotion of general budget support as a preferred method of achieving the desired alignment with national policies and systems.

In some poor countries, projects with their own reporting and accounting arrangements are the only thing that is feasible, because the minimum conditions for using government channels do not exist.

However, from the point of view of institutional development, this remains very much a second- or third-best option, one that is very likely to reinforce the characteristics of the state that make it the only option in the first place. General budget support and pooled sector funds are both preferable (in that order) whenever they are possible.

This follows the general principle that systems cannot be expected to improve unless they are being used.

On the other hand, the new aid modalities are not a panacea. The key thing in the cases of both budget and sector support is to go in with eyes open, bearing firmly in mind the learning about conditionality that led up to the PRSP initiative.

That is, programme aid plus conditionality does not work for complex reforms, unless there is a significant amount of real commitment within the country.

This means that conditions attached to budget support programmes should normally be limited to policy actions to which the government is in some real sense committed. Disbursement should be based on past progress and not on promises to do things better in the future.

In practice this distinction is often hard to make but the principle that buying reforms does not work is well established and a good general guide.

The most important conditions be supported are those that surround the decision to embark on budget support in the first place, as it becomes difficult to withdraw later on. Those are most intelligently based on the observation of the country’s trajectory of institutional development over a suitable recent period, and not just on the observation of a basic minimum of fiduciary good practices.

If there are no significant domestic pressures leading to improvements in the budget process, it is doubtful whether any of the benefits from having donors make use of government systems will materialize.

The evaluation framework here advocated is that a general budget support incorporates an assumption in the theoretical model of the conditions under which budget support will be effective that political competition in the country is moving away from use of state resources for patronage and towards a focus on results, and that there are domestic constituencies and pressures for higher standards of accountability.

Once again, however, the main worry about programmatic forms of aid (sector and budget support) is that they come to be regarded as self-sufficient means of addressing the weaknesses that PRSPs have helped to show up.

If the PRSP process has not generated real buy-in to poverty reduction by a country’s politicians, a budget support programme will not solve the problem. The conditions for reforms to work are essentially the same as the conditions for PRSPs and the associated reforms to deliver what they promise.

The donor conditionality or performance assessment has a modest role to play, mainly as a marker for policy dialogue. But if conditionalities are limited to policy actions that the government is showing some inclination to undertake anyway, something else is needed to get other and more difficult issues onto the development agenda.

(To be continued…)