ODI paper on Incredible Budgets

[John Short]

In many countries, particularly low-income and fragile states, national budgets are often poor predictors of revenue and expenditure. This paper draws on rational choice theory to set out a framework for explaining what drives budget non-credibility. Three drivers of budget non-credibility are identified –inevitable lack of knowledge as to how the future will unfold (uncertainty); the inability of the head of the executive to fully control his subordinates (unruly agents); and the desire for the head of the executive to gain the support of external stakeholders by publishing a budget that he does not truly intend to undertake (signalling). The latter two drivers represent different variants of the ‘principal-agent’ problem. This framework is applied to a stylised budget execution chain and examples of non-credibility are illustrated with real-world cases from the budgets of Uganda, Tanzania and Liberia. The paper discusses the challenges of using the PEFA framework to consider budget credibility issues and reviews some of the common responses that aim to reduce budget non-credibility. The challenges of appropriately diagnosing budget non-credibility problems and of accessing accurate budget data for further analysis are noted.

[STONE]

Thank you for posting this Interesting paper.  I like the categorisation of uncertainty, unruly agents, and signalling.   

One aspect of signalling that is not brought out is related to mistakes and errors, and ignorance and incompetence. 

In one country I know, there has been much noise from development partners about underspending of the budget - particularly with respect to development expenditure. There was even loose talk about making improved expenditure out-turn a General Budget Support condition.

Significant amounts of development expenditure in the budget are DP financed. A detailed investigation of why only 14 % of one ministry's development budget was spent in a particular year revealed that most of the apparent underspend was due what was known by government staff, who had earlier done a preliminary investigation, as 'over-budgeting'.  The term sounds like one aspect of what Simson and Welham call ‘signalling’ and so attracts suspicion of intent.  If only.

When adressing the issue of why the government does not spend the money DPs provide for the budget (at the request of one DP), a question asked was how development expenditure estimates entered the budget. The answer was the aid management staff asked DPs to indicate their 'budgets' for the next year. These requests were typically passed to project administrators who provided information on the cash flow set out in the original DP Project Document or in several cases the total project amount.  The numbers were not checked for ‘credibility’ by the DP task managers (we never found a DP Task Manager who had read the budget presentation of their projects), nor were they checked by the Ministry staff.  This did not change much even after the awful shaming evidence was presented in all its horrible detail to an ‘event’ for the ministry management and the DPs.

One DP ‘believed’ that its project did not appear in the Budget. It could not appear in the budget because expenditure did not go through the Single Treasury Account. Evidence to the contrary (the relevant page of the published budget book) was met with ‘incredulity’.  That ‘belief’ also results in ‘under-reporting’ of expenditure.  The DP was informed of the procedure for DP reporting of such expenditure to MoF in the interest of fiscal transparency comparison but no action was taken in the following year: and the budget continued to show the budget but related expenditure was not reported. This was not an isolated case. Incredible but true!

Now, ‘over-budgeting’ and ‘under-reporting’ did account for all of the difference between budget and actual, but it did account for a ‘material’ amount.

I find it ‘hard-to-believe’ the extent of continuing DP ignorance of budget preparation and expenditure reporting processes. There seemed to be no concern: neither about the risk of revealed DP incompetence nor about the effects on the reputation of the government partner.