Does donor support to public financial management reforms in developing countries work? An analytical study of quantitative cross-country evidence

Working and discussion papers
April 2011
Paolo de Renzio, Matt Andrews and Zac Mills

This study is part of a broader evaluation of donor support to public financial management (PFM) reforms in developing countries. It brings together available quantitative evidence on the quality of PFM systems, to assess the factors that are associated with and may have determined cross-country differences and variations over time, with a particular focus on the impact of donor support to PFM reforms. The bulk of the analysis draws on data from Public Expenditure and Financial Accountability (PEFA) assessments in 100 countries, data on donor support to PFM reforms collected directly from some of the donor agencies most active in this area and a large dataset on other economic/social, political/institutional and aid-related variables identified as relevant from previous research.

A number of findings from the cross-country econometric analysis are relevant for the evaluation, and for broader donor approaches and policies on PFM reforms. They can be summarised as follows:

  • Economic factors are most important in explaining differences in the quality of PFM systems. Aid-related factors, on the other hand, have more limited explanatory power. As a consequence, PFM systems are more likely to improve in response to changing economic circumstances rather than to donor efforts.
  • More specifically, countries with higher levels of per capita income, with larger populations and better recent economic growth records, are characterised by better quality PFM systems. On the other hand, state fragility, defined as being in a conflict or post-conflict situation, has a negative effect on the quality of PFM systems.
  • Interestingly for the purposes of the evaluation, donor PFM support is also positively and significantly associated with quality of PFM systems. On average, countries that received more PFM-related technical assistance have better PFM systems. However, the association is very weak: an additional $40-50 million per year would correspond to a half-point increase in the average PEFA score (equivalent to, say, a change from ‘C’ to ‘C+’ in PEFA indicators).
  • These results remained consistent through a number of robustness checks and model changes. Interesting additional results come from using more recent data or focusing on low-income countries only. In these cases, the share of total aid provided as general budget support (GBS) is also positively and significantly associated with better PFM quality. In other words, aid modalities, and not just direct support to PFM reforms, contribute to explaining differences in the quality of PFM systems in some of the poorer countries where most donor efforts are concentrated.
  • Finally, different aspects of donor support differ in their relationship with more specific PFM processes. A longer period of donor engagement, for example, is associated with better performance in upstream, de jure and concentrated processes. This may owe to donors’ historical tendency to pay more attention to these simpler reform areas, but could also reflect the fact that downstream, de facto and de-concentrated processes take longer to improve.
  • The level of donor PFM support is also associated more strongly with scores for de jure and concentrated PFM processes, again highlighting how donor PFM support seems to focus more on rules, procedures and specific actors within government. Results are reversed when it comes to upstream vs. downstream processes. Here, the association is stronger with downstream processes, possibly highlighting the large amounts of funding devoted to Integrated Financial Management Information System (IFMIS) projects, a typical downstream PFM reform.

At the same time, these results suffer from a number of serious limitations and challenges, including the following:
 

  • Data quality remains an issue, especially when it comes to information about donor PFM support. Given limitations in the information provided by donors, we focused on yearly disbursements for PFM-related activities. This gives undue weight to large projects such as IFMIS introduction, at the expense of ‘softer’ interventions. We also focused on data post-2002, for which availability is much greater. This means we cannot capture earlier donor PFM support, from a time when the foundations for PFM reforms were laid in some of the countries included in our sample.
  • While the positive and significant relationship between donor PFM support (and GBS as a share of total aid in certain cases) on the one hand and quality of PFM systems on the other is particularly encouraging for the purposes of the evaluation, it clearly cannot be interpreted as causal, given the nature of the data. It could merely reflect the fact that donors tend to provide more PFM-related assistance (and more GBS) to countries that have already achieved certain success in improving the quality of their PFM systems. Despite various attempts to tackle this issue, we could not prove the direction of causality.
  • Assessing the impact of donor support on PFM reforms requires tracking the quality of PFM systems over time. Given the lack of sufficient time series data, the analysis assumes that a higher PEFA score today is a valid proxy for past reform success. However, its findings are only partly confirmed by evidence from a smaller dataset looking at changes in PFM systems over the past decade in 19 African countries.

These limitations and challenges point to the need to interpret the results of the analysis presented in this paper with a great deal of caution. Moreover, they highlight the need to complement these quantitative findings with in-depth qualitative research at country level, explaining not only if and when donor PFM support has had an impact on PFM systems, but also why and how it did. Nevertheless, case study countries can and should be selected taking into account some of the insights provided in this paper.

Read the research
Documentpdf0.99M