By Paolo de Renzio, Global Economic Governance Programme, Oxford University
Nancy Birdsall’s response on this blog to the issues raised about cash-on-delivery aid in the 2010 Education for All Global Monitoring Report reflect the amount of work that has gone into addressing some potential pitfalls of the new approach to aid delivery championed by her organization, the Center for Global Development. However, it still does not fully address some of the more fundamental criticisms that I raised in a note written with Ngaire Woods of Oxford University at the time when the COD aid initiative was taking shape and which Nancy Birdsall alludes to in her blog post.
Cash-on-delivery aid is based on problematic assumptions about aid relationships that are not backed up by any extensive empirical evidence. It is premised on the assumption that tying aid disbursements to concrete development results will provide a strong incentive for recipient governments to focus their efforts on achieving those results. It also assumes that the low-income countries that cash-on-delivery aid is likely to target have not only the willingness, but also the capacity to improve the effectiveness of service delivery in key areas of public intervention such as education.
Now, if that were the case, then the incentive would not really be necessary, as these countries would already be attempting to use all available resources as efficiently and effectively as possible. What cash-on-delivery aid might do, in these cases, is provide some marginal additional resources that have fewer of the problems – such as unpredictability and donor micro-management – that other aid processes suffer from. In this sense, it could well be a welcome addition.
The problem is that there is ample evidence to show that in most of the countries that cash-on-delivery aid is likely to target, the assumptions it is based on simply do not hold. Governments have not only insufficient capacity, but also limited political interest in using available resources to maximize development impact. Of course, politicians in these countries (like in all others) do consider better health and education for poor people as a priority. But in practice they have to balance this objective with many others. Evidence has shown that political realities in low-income countries may dictate a different, short-term focus, linked to political survival, patronage networks, conflict and other factors.
For cash-on-delivery aid to work, the incentive it provides needs to be large enough to induce the kind of behavioural (and probably institutional) change that leads to a reorientation of government priorities towards better public service delivery. If instead, as envisaged, cash-on-delivery aid is to be another, additional process, representing a small percentage of overall aid flows, not only is its incentive effect likely to be too small to make any difference, but its existence is likely to put additional strains on overstretched bureaucracies, which would face conflicting and competing incentives linked to different aid processes, with little space for the kind of innovation and policy experimentation that cash-on-delivery aid aims to promote.
After nearly 40 years of working with over 35 developing countries to improve their educational systems, I could not agree more that an external stimulus is needed upfront to provoke and support change (hopefully improvement) in any nation’s educational system. The pressures are simply too great and the resources too small to move away from the status quo without help.
Paolo makes some really key points on cash-on-delivery, which are all the more important in the light of the growing emphasis on showing value-for-money. This potentially places in danger the progress that some donors have been making by taking risks to support conflict-affected countries – which has had high pay-offs, but would not show the short-term type of results that proposed cash-on-delivery schemes demand.
Incentivising needs to go further than national education ministries for quality education for all to be delivered. It also needs to filter down to practitioner level. A colleague today from Cusco, Peru, was telling me how 40% of posts remain unfilled in the public education sector in the Cusco department, principally of course in the poorest, most isolated regions where children have the greatest needs.
We appreciated Paolo and Ngaire’s paper and inputs to developing the COD aid concept. I’m somewhat surprised, though, by this particular critique of COD Aid coming up again at this time because the point is so focused on whether COD Aid will induce changes in developing country government’s behavior when the political economy we have to consider is also how to change the behavior of developed countries that are funding these programs. My interpretation of the evidence is that current aid modalities do little to induce uninterested governments to make faster progress. Rather, they create vested interests to keep money flowing even when progress is lacking. COD aid could untangle this mess. I address this more fully on the CGD site at: http://blogs.cgdev.org/globaldevelopment/2010/07/what-is-the-counterfactual-for-cod-aid.php
Very Interesting post, would you mind if I translate it into Catalan and link your blog?
I’m strating a new blog on educational issues Interalia and I really would linclude your interestig posts.
Best regards
Marga
LA ENTREGA DE DINERO EN EFECTIVO PARA LA MEJORA EN EDUCACIÓN ES VÁLIDO SIEMPRE CUANDO LOS ORGANISMOS QUE DAN Y QUIEN LO RECIBE DEBE DE EXISTIR UN CONTROL PERMANENTE DE LOS GASTOS EN LOS PROYECTOS PREVISTOS EN MEJORA, REALMENTE DE LAS COMUNIDADES POBRES DONDE ES NECESARIO MEJORAR EL SERVICIO EDUCATIVO DE LOS NIÑOS, NIÑAS Y ADOLESCENTES.
My commentary is on the necessity of contextualisation, targeting, and incentives when considering cash-for-results. In their absence, it might be tempting for practitioners to see cash-for-results as a simple solution. It’s not. We can see this if we think of cash-for-results as a management-by-numbers approach to public service performance management. That approach has been tested in the UK public service. It has been shown to be problematical. The issue is that it has been inadequately contextualised, targeted, and incentivised. In its its rudimentary form, the UK variety of cash-for-results neglects the importance of combining with other components of a more comprehensive system of performance management. Simple dependency on value-added assessments is shown to have actually hindered efforts to carry out comprehensive service improvements.
Christopher Hood’s research understands this. The Quality and Outcomes Framework introduced for UK’s general medical practitioners in 2004 produced payable results. In fact results were far in excess of targets established for the measurable treatments incentivised. But results produced no inevitable performance improvement over the long-run. Sustainability did not materialise. At times, the results generated a one-event improvement, then subsequently flattened-out. Sometimes the results led to distorted outputs. For instance, Guthrie et.al show that medical interest fell off non-incentivised treatments. Holistic patient care received lower priority. Doctors were content to focus on incentivised treatments. Treatment was neglected that did not fall under the prescribed focus. The successful meeting of targets led to a major overspend on the 2005 National Health Service budget (Checkland et al. 2009). In other cases, cash-for-results in the UK health sector led to gaming by providers to boost rankings or to achieve targets. Bevan & Hamblin’s 2009 study of the English ambulance service strongly suggests misreporting or altering of response times, in order to meet targets. Hood points out that similar issues surround cash-for-results experimentation associated with literacy and numeracy targets for English primary schools. Educator’s prioritized literacy and numeracy at the expense of learning in the broader sense.
The key implication I draw attention to is that the extent to which cash-for-results can lead to better public service performance depends on how well the indicators are designed.
How effectively do the indicators capture the full range of performance elements that policy-makers want providers to focus on? What does the country in question want to focus on? Inequality? Geographical disparities in accessing the same good quality services? How to get professionals to relocate to distant areas?
Cash-for-results experimentation relies heavily on specificities. These are directly linked to capacity within the recipient country. Thus, improvement in the donor-recipient relationship, expected to flow from cash-for-results experimentation, suggests anything but a hands-off approach. The notion of reliance on independent auditing is a further problem. It fits too readily into the concept of a one-size-fits-all template. This variety of cash-for-results would take us backwards.
Instead, these issues suggest the importance of a thorough-going reappraisal of donor collaboration. The kind that leads towards targeted, contextualised identification of results, and includes sophisticated incentivisation – both monetary and non-monetary.