Public financial management reform is often promoted as a means of supporting more accountable and responsive governance. But very little is known about this relationship. There are a number of reasons for this:
- PFM is difficult to measure
- Governance is difficult to measure
- The causal relationship is difficult to disentangle.
Using data from the Public Expenditure and Financial Accountability (PEFA) assessments, our new book, PFM, Good Governance and PEFA, aims to shine some light on this little understood issue.
Although PEFA assessments have some well documented limitations (see Hadley and Miller, 2016), nearly 144 countries have now completed an assessment as of March 2017. These assessments provide a much richer source of information about PFM than existed twenty years ago.
More specifically, we use national level PEFA data to investigate institutional factors that shape PFM performance as well as the effect of the PFM system (as measured by PEFA) on the outcomes we ultimately care about, that is more credible budgets, fiscal discipline, lower corruption, and higher domestic resource mobilisation. Our results are largely reassuring though with some important caveats.
First, do formal political institutions such as the type of government or electoral system matter for PFM performance? Chapter 3 finds little evidence of this for our sample of mostly low and middle income countries.
This is good news for reformers since they are unlikely to have much influence on these factors. However, this does not mean that institutions do not matter at all.
In particular, informal institutions can play a critical role in shaping political incentives for reforms but are beyond the scope of a quantitative study.
Second, do stronger PFM systems produce the desired outcomes even in fragile countries? While Chapter 4 finds that better PFM performance is associated with more credible budgets at the sectoral level in both fragile and non-fragile countries, we do not find that better PFM systems go hand in hand with better fiscal outcomes irrespective of fragility.
This is perhaps not surprising when you consider that fiscal outcomes are influenced by a wide range of factors beyond PFM performance. Budget credibility, on the other hand, is a more intermediate outcome that it is more closely related to PFM performance.
‘Although PEFA is the most comprehensive measure of PFM to date, the PEFA dataset was not designed for rigorous statistical analysis.’
Third, how important is PFM in the fight against corruption? Chapter 5 presents evidence that there is a relationship between better PFM performance and better perceptions of corruption. The relationship is particularly strong for expenditure controls that minimize the discretion of politicians and bureaucrats.
The estimated size of this effect, however, is underwhelming when compared with the magnitude of the relationship between economic growth and perceptions of corruption. Moreover, the causal direction between PFM and corruption is unclear.
Finally, are penalties still relevant for modern tax administrations that are moving to voluntary compliance? Chapter 6 of the book provides evidence that these coercive tools are important. We estimate that countries who score a good practice ‘A’ score on the relevant indicator have tax-to-GDP ratios that are 2.7% higher on average.
In the chapter we discuss how, it may not be the penalties themselves that are important for raising revenues, but rather the political support for revenue collection that they represent.
All of these findings need to be interpreted with caution. Although PEFA is the most comprehensive measure of PFM to date, the PEFA dataset was not designed for rigorous statistical analysis.
Consequently, there are several methodological issues that cannot be fully resolved in this book, or in other papers which apply a similar approach. Another shortcoming stems from whether PEFA and other variables of interest such as corruption and our political/institutional variables actually measure what they are intended to measure. This raises the question of whether comparing two blunt instruments really tells us much?
So what can we do next? First there remains much scope for further research in this area based on more tightly defined individual PFM measures and hypotheses.
There are examples of this in the book, for example our focus on penalties for non-compliance, but more can potentially be done. Second, subnational PEFA assessments are on the rise and can potentially be exploited to investigate outcomes at a more disaggregated level. Finally, making the entire PEFA dataset publicly available can facilitate progress in each of these areas, crowding in research by academics.
But certainly, the PEFA call for research proposals using the PEFA dataset is a step in the right direction.
