This paper comments on various topics relating to the current IMF initiative to 'streamline' itsconditionality, setting this in the context of various other changes occurring simultaneously and of the responses of the World Bank. It undertakes a necessarily preliminary 'reality check' on the basis of brief studies of recent Fund and Bank programmes in six low-income countries:

• Albania
• Ghana
• Kenya
• Mozambique
• Vietnam
• Zambia

The paper additionally draws upon recent documentation relevant to this topic and on conversations in Washington with staff members of the IMF and World Bank.

Some of the principal conclusions include:

• Streamlining is narrowly conceived and supply-driven. It responds to only some of the accumulating reasons for change and not necessarily the most important. It seeks to reverse the proliferation of structural conditions, calls for greater clarity in programme documents about what constitutes Fund conditionality, and seeks to ensure that, in countries where both agencies are operating, there will be a clearer division of responsibility on policy matters between the Fund and the Bank. The evidential basis for expecting streamlining to raise programme effectiveness is rather slight and overall it is conceived as a fairly limited exercise. Streamlining is not being dovetailed with the changes intended to result from the introduction of the PRGF nor with concerns about ways of enhancing ownership. A more 'joined-up' and demand-oriented approach might well help to reduce some of the weaknesses identified here.
• Streamlining is confined exclusively to 'structural' conditionality. There is no comparable re-examination of its traditional quantitative macroeconomic stipulations. While it is true that much of the proliferation occurred in 'structural' conditions, the limited nature of the streamlining exercise missed an opportunity to re-visit the appropriateness of the Fund's conventional macro conditions.
• Results reported so far indicate substantial differences between the aspirations of the Fund=s management and actual changes so far achieved. There has been real streamlining of conditionality in some countries but in others it has largely been business as usual, particular among those with >old= (former ESAF) PRGFs. That streamlining has occurred in some cases suggests that the basic concept and the guidance provided have been practicable, although there remain difficulties about defining what constitutes a 'critical' structural policy condition and in what circumstances streamlining is the most applicable. On average, the number of
• structural conditions in PRGFs has so far been reduced by about a quarter, although the extent of this has been very context-dependent, with large reductions in some programmes and no change in others. We find little progress in the intended clarification of the content of Fund conditionality within programmes.
• Pronouncements and documentation on streamlining make little reference to ownership. Cutting back on the number of detailed structural conditions should be helpful but there may be countervailing factors and it is by no means clear that streamlining will bring any net encouragement to ownership. It was designed for a different purpose and questions of negotiating processes and styles would need also to be addressed if the ownership factor was to be addressed satisfactorily.
• It appears that >new= PRGF programmes - but not >old= ones - are paying more attention to the specifics of social sector spending. It seems also that the content of conditionality in the former is changing to some extent. Detailed structural conditionality is diminishing but there has simultaneously been greater attention to governance and public expenditure management. There also appears to have been some greater accommodation of budget deficits but little movement with respect to consideration of alternative budget scenarios.
• As regards PRSPs, early evidence suggests that this experiment is not yet working fully as planned. It appears so far that PRSPs have tended to be rather general, weakly prioritised and of variable quality. Such defects reduce their value as templates within which IFI programmes can be designed. In some cases,
• there are questions about the degree of government commitment to these strategies and this problem has been aggravated by the perverse effect on incentives of the link between PRSPs and HIPC debt relief.
• Although the World Bank has made statements in favour of streamlining, it appears that, in fact, there is little going on in the Bank comparable with the Fund exercise. There appears to be no institutionalised attempt to further reduce the quantity, or narrow the focus, of the Bank's conditionality. What may happen is that many of the more important structural conditions dropped by the Fund will be taken up in Bank programmes, which may then drop some of its own less central conditions.
• Fund-Bank collaboration appears to be working quite well, although there are apt to be problems of sustainability, there may be difficulties in establishing which is the be the 'lead agency' in particular instances, and it may be more difficult to achieve co-ordination in countries which do not have a PRSP.
• There is a good deal of cross-conditionality between the two institutions at the level of broad performance, probably more so than was in existence before the introduction of streamlining and the new credit facilities. At the same time, both institutions seem clear that cross-conditionality should not be applied to policy specifics within either programme.
• Overall, many many borrowing governments which are also subject to HIPC conditions have probably found themselves in a tighter policy straight-jacket, notwithstanding the possibility of some net reduction in the combined Fund-Bank quantity of conditionality. This trend sits ill with the growing appreciation of the limitations of the modalities of conditionality. For them, it would be very difficult to state with confidence that they now have more freedom of policy choice than they had, say, two years ago, although the evidence does not yet permit any firm conclusions. Reduced policy freedom could scarcely fail to impact adversely on ownership. It would also almost certainly increases the transactions costs to governments of accessing the assistance on offer, a feature worsened by an apparent absence of serious efforts at the co-ordination of PRSPIMF-World Bank-HIPC conditionality and the resulting monitoring and reporting requirements.
• In principle, governments now have, in the PRSP, a vehicle for influencing the framework of conditionality but we have found no examples of governments seeking to use it in that way. A probably more common pattern is that they have drafted their papers to second-guess what they think the IFIs/creditors would like to see, and that papers have so far been too unspecific to make much difference to conditionality. The IMF appears not to have engaged in much bilateral discussion with borrowing governments concerning the design of streamlining.
• Our evidence suggests that neither the small number of PRGF or PRSC programmes so far negotiated have differed markedly in design from their predecessors, although in some PRGFs the extent of structural conditionality has been much reduced. This again suggests that borrower influence on the programmes may be little changed. However, the PRSP experiment is still young and the crucial test will be if a government produces a full PRSP, based on wide consultation, whose policies differ materially from what either IFI would choose.

Policy recommendations are made at the conclusion of the paper.