Actual expenditure outturns can deviate from the originally approved budget for reasons unrelated to the accuracy of forecasts—for example, a major macroeconomic shock. The calibration accommodates one unusual or “outlier” year and focuses on deviations from the forecast that occur in two or more of the three years covered by the assessment for dimensions 2.1 and 2.2. Dimension 2.3 uses data from all three of the last completed fiscal years.

If there are amounts in suspense accounts at the end of the financial year that could affect the scoring of this indicator if included, it should be noted in the PEFA report narrative. Assessors will need to decide whether the amounts in suspense accounts are sufficient to result in misleading scores based on the amounts allocated to expenditure categories used for this indicator. If the score is likely to be misleading—for example, if the unallocated expenses exceed 10 percent of total annual expenditure—dimensions 2.1 and 2.2, and therefore PI-2 as a whole, should be scored D.

Dimension 2.1 measures the difference between the original, approved budget and end-of-year outturn in expenditure composition, by functional classification, during the last three years, excluding contingency items, and interest on debt. Other expenditures should be included—for example, expenditures incurred as a result of exceptional events such as armed conflict or natural disasters, expenditures financed by windfall revenues including privatization, central government subsidies, transfers, and donor funds reported in the budget.

At the administrative level, differences should be calculated for the main budgetary heads (votes) of budgetary units that are included in the approved budget. If a functional classification based on GFS/COFOG is used, differences should be based on the ten main functions. Where a functional classification not based on GFS/COFOG is used, the measurement of difference should be based on the main heads approved by the legislature. If a program basis is used, the program-based categories should be rated at the same level at which they were voted by the legislature.

The calculations for this indicator include an adjustment to remove the effects of changes in aggregate expenditure. This is achieved by adjusting the budget outturn for each category used by the proportional difference between the total original, approved budget expenditure and the total expenditure outturn. The remaining deviation within each category is based entirely on the absolute value of changes that occurred in and between categories, net of any change assumed to have resulted from aggregate expenditure shifts.

The methodology for calculating this dimension is provided in a spreadsheet on the PEFA website.

Dimension 2.2 measures the difference between the original, approved budget and end-of-year outturn in expenditure composition by economic classification during the last three years including interest on debt but excluding contingency items.

The composition of the budget by economic classification is important for showing the movements between different categories of inputs—for example, capital and recurrent expenditures. The categories of expenditure are the same as for dimension 2.1, with the addition of interest on debt, as this is one of the categories of economic classification. The calculation should use the second level of the GFS classification (2 digits) or similar. If a different classification is used, the level of aggregation should be comparable to the 2-digit GFS.

As with dimension 2.1, the effects of changes in aggregate expenditure between the original, approved budget and outturn are adjusted in the calculations.

The methodology for calculating this dimension is provided in a spreadsheet on the PEFA website.

Dimension 2.3 measures the average amount of expenditure actually charged to a contingency vote over the last three years.

This dimension recognizes that it is prudent to include an amount to allow for unforeseen events in the form of a contingency vote, although this should not be so large as to undermine the credibility of the budget. There may be more than one contingency vote. Assessors should discuss the budgeting and accounting treatment of contingency items in the report narrative. The calibration for this dimension is based on the volume of expenditure recorded against contingency votes, except for transfers to a Disaster Fund or similar reserves, as this represents a deviation from policy-based allocation.