Our Programmes



Sign up to our newsletter.

Follow ODI

Suspending disbelief - true love and budget credibility

Written by Bryn Welham


​Question: what’s the similarity between true love; and a pile of budget documents from low-income countries?  The answer?  As anyone who has worked on public financial management or been in love will know, sometimes you have to suspend disbelief in order to reconcile reality with expectation.

The issue of incredibility – defined in the dictionary as something that is ‘impossible or very difficult to achieve’ – matters hugely for understanding how public money is used in developing country budgets. But as much as it matters to donors and institutions like the IMF who work closely with governments in low-income countries, budget credibility also is a major contributor to building trust in accountable public institutions. So why, in spite of this considerable effort to strengthen budget credibility, do many governments still struggle to execute their budgets to plan? Last week’s announcement that the Public Expenditure and Accountability Program (PEFA) is now open for public consultation merits a closer look at why budget management in developing countries often goes wrong.

Firstly – and most obviously – it’s difficult to plan expenditure in an uncertain world.  This is particularly true for the poorest countries that are prone to economic shocks. Beyond this unavoidable fact of life, there are also other reason why plans may not match execution.  Short of expensive and time-consuming systems for the President to monitor every last piece of activity in government, there will always be occasions when mice will play when the cat’s away. This means that unruly civil servants can undermine, subvert and amend budgets during execution to meet their own ends – good or bad.  Drawing on a rational choice framework, we can say that the President (as the principal) cannot always control his subordinates (the agents). 

Stakeholders such as citizens, other politicians and (in many countries) donors will also have an interest in public spending choices.  To keep them on board, each of these important constituencies must be at least minimally satisfied with the budget.  The budget, a public document, can therefore also act as a signalling tool that enables important stakeholders to see what they want to see.  In practice, budget managers often know that they don’t intend to follow the plan but are aware that these stakeholders are equally unlikely to overcome the high costs that come with exhaustively verifying budget implementation.  In rational choice terminology, this turns the principal-agent dilemma around, and the President (the agent) is under the imperfect supervision of his stakeholders (the principals).

Having reviewed instances in Uganda, Liberia and Tanzania as part of my working paper ‘Incredible Budgets – Budget credibility in theory and practice’ however, not executing budgets to plan matters for two reasons. Firstly, it makes the development of accountable public institutions all the more difficult: if the budget is not credible it undermines any form of ‘contract’ between those who provide funds (taxpayers and donors) and those entrusted to administer them in accordance with the agreed plan (government).  Secondly, if the budget is not credible, one of its most important functions will be reduced: it can’t be used as a management tool for coordinating action to achieve government objectives.

But making theoretical distinctions between drivers of incredibility is much easier in theory than in practice.  For example: Were the cost overruns on a construction project merely human error (uncertainty), an effort by road engineers to dupe the central government into a larger contract than they would otherwise have accepted (unruly agents) or an attempt by the government to gain parliamentary support for a high profile project by understating its cost (signalling)? Is it endemic public sector corruption that appears to go unpunished a principal-agent problem (unruly agents), or is it tacitly ‘allowed’ by the President as a means of augmenting civil service pay against the wishes of the donor community (a signalling issue)?

Donors also may be well aware that the proposed budget will not be implemented as approved by Parliament.  However, openly challenging the detail of budget execution brings its own costs in terms of reputational risk and additional work for the donor leading to a ‘Garden of Eden’ approach to budget credibility: eating from the tree of knowledge (i.e. facing up to the reality of how budgets are actually executed) means a great deal more hassle than remaining intentionally ignorant in budget Paradise. 

A recent roundtable event discussed the conclusions of my budget credibility paper with practitioners that deal with managing developing country budget management on daily basis. These had no shortage of examples from their own experience of where budgets had gone off track.  Some noted that too often budget execution simply means controlling expenditure to stop overspending rather than managing expenditure to actively follow the budget. Participants agreed that there were few easily available indicators to track credibility below aggregate level, although the World Bank outlined some initial work into reviewing what indicators it has used in the past to do so.

Budget credibility remains an important, yet difficult area to study – particularly because of the lack of detailed data available for researchers.  Although new data as part of the World Bank ‘BOOST’ initiative will allow for more countries to be examined, those working with developing country budgets need to be more rigorous in how they understand non-credibility and which of the drivers is at play.  Only by correctly diagnosing the underlying drivers can those who support more credible budgets design and implement the correct response.