In this chapter we compare actual fiscal performance against targets projected in programs. We first examine the large cross section of programs to assess the frequency and nature of fiscal shortfalls relative to targets, the sources of these shortfalls, and the extent to which programs have been flexible in revising targets as the program unfolds. We then examine in some depth the 15 sample programs to study the composition of the fiscal adjustment and some qualitative dimensions of that adjustment that cannot be detected in the cross-country analysis.
Cross-Country Analysis
Table 4.1 compares actual with envisaged (average) changes in the current account and fiscal balances over a two-year horizon for the sample as a whole and also for the different subgroups. The following broad patterns emerge:
-
Whether we look at the overall balance or the primary balance, fiscal balances improved by half the projected amounts for the sample as a whole. Shortfalls relative to projections were about ¾ of 1 percent of GDP in both cases. However, there were important differences across subgroups. Fiscal targets were met in the transition countries but not in the other subgroups. Concessional arrangements and SBA/EFFs in nontransition countries experienced a shortfall in the primary balance equal to half the targets, although these shortfalls were small as a percentage of GDP (0.4 percent and 1 percent, respectively).
-
The composition of the fiscal adjustment also shows significant variation across types of programs. In the case of SBA/EFF arrangements in nontransition countries, the shortfall is largely due to the expenditure side: while programs projected on average a reduction in public expenditures to the tune of 0.7 percent of GDP, expenditures actually increased by 0.6 percent of GDP. In the case of SBA/EFF arrangements in transition countries, both revenue and expenditure declined more or less in line with projected values.
-
While fiscal balances improved much less than projected, the current account balances on average improved slightly more than projected, though the pattern varied across subgroups. In concessional programs, the actual developments turned out to be equal to projections but in the SBA/EFF arrangements, the current account position adjusted more than projected in both subgroups (nontransition countries and transition countries). In the nontransition countries the average current account adjustment exceeded projections by more than 1 percent of GDP, a statistically significant change.
Changes in External and Fiscal Balances from (T−1) to (T+1)1
(In percent of GDP)
Figures subject to rounding errors.
Based on a sample of 115 arrangements.
Changes in External and Fiscal Balances from (T−1) to (T+1)1
(In percent of GDP)
SBA/EFF | |||||
---|---|---|---|---|---|
All Arrangements | ESAF/PRGF | Transition countries |
Nontransition countries |
||
Envisaged | |||||
Current account | −0.3 | 0.1 | −2.0 | −0.2 | |
Government balance | 1.7 | 1.6 | 1.1 | 2.0 | |
Primary balance2 | 1.4 | 1.0 | 0.4 | 2.0 | |
Government revenues | 0.4 | 0.4 | −1.7 | 1.3 | |
Government expenditures | −1.2 | −1.2 | −2.8 | −0.7 | |
Actual | |||||
Current account | 0.3 | 0.1 | −1.3 | 1.1 | |
Government balance | 0.8 | 1.0 | 1.8 | 0.2 | |
Primary balance2 | 0.7 | 0.6 | 0.3 | 1.0 | |
Government revenues | 0.2 | 0.1 | −1.4 | 0.9 | |
Government expenditures | −0.7 | −1.0 | −3.2 | 0.6 | |
Count | 133 | 60 | 21 | 52 |
Figures subject to rounding errors.
Based on a sample of 115 arrangements.
Changes in External and Fiscal Balances from (T−1) to (T+1)1
(In percent of GDP)
SBA/EFF | |||||
---|---|---|---|---|---|
All Arrangements | ESAF/PRGF | Transition countries |
Nontransition countries |
||
Envisaged | |||||
Current account | −0.3 | 0.1 | −2.0 | −0.2 | |
Government balance | 1.7 | 1.6 | 1.1 | 2.0 | |
Primary balance2 | 1.4 | 1.0 | 0.4 | 2.0 | |
Government revenues | 0.4 | 0.4 | −1.7 | 1.3 | |
Government expenditures | −1.2 | −1.2 | −2.8 | −0.7 | |
Actual | |||||
Current account | 0.3 | 0.1 | −1.3 | 1.1 | |
Government balance | 0.8 | 1.0 | 1.8 | 0.2 | |
Primary balance2 | 0.7 | 0.6 | 0.3 | 1.0 | |
Government revenues | 0.2 | 0.1 | −1.4 | 0.9 | |
Government expenditures | −0.7 | −1.0 | −3.2 | 0.6 | |
Count | 133 | 60 | 21 | 52 |
Figures subject to rounding errors.
Based on a sample of 115 arrangements.
As noted in other comparisons, the group averages mask considerable variation across countries and this is shown in Table 4.2, which presents the distribution of programs according to their fiscal performance relative to program targets by the second year of the program.
-
About 58 percent of programs had a shortfall with respect to targets in the overall fiscal balance, and the percentage in the case of primary balances was 66 percent. The mean fiscal shortfall in this group was 2.8 percent of GDP for the overall balance and 2.2 percent of GDP for the primary balance.
-
The incidence was largest in SBA/EFF arrangements in nontransition countries, with about three-fourths of these programs having fiscal shortfalls. It is followed by ESAF/PRGF programs, where about half the arrangements had shortfalls. The mean shortfalls for both these groups are similar, about 3 percent of GDP.
-
In contrast, SBA/EFF arrangements in the transition countries had the lowest incidence of shortfalls in overall balances (40 percent). However, the picture changes when the incidence of shortfalls refers to primary balances (60 percent). It is clear that in these arrangements, programs have frequently underestimated favorable developments in interest payments providing relief to the budget.
Differences Between Actual and Projected Changes in Fiscal Balances1
(From T–1 to T+1)
This difference between actual and envisaged adjustment is statistically significant at the 99 percent or better confidence level.
Values in parentheses show the results when overperformance and underperformance are defined in terms of primary balances.
Differences Between Actual and Projected Changes in Fiscal Balances1
(From T–1 to T+1)
Distribution of Programs (In percent) |
||||||
---|---|---|---|---|---|---|
SBA/EFF | ||||||
Differences (In percent of GDP) |
All arrangements | ESAF/PRGF | Transition countries |
Nontransition countries |
||
Positive differences (“overperformance”) | ||||||
Larger than 4 | 4.5 | 3.3 | 13.6 | 2.0 | ||
Between 3 and 4 | 3.7 | 5.0 | 4.5 | 2.0 | ||
Between 2 and 3 | 5.2 | 6.6 | 0.0 | 5.9 | ||
Between 1 and 2 | 10.5 | 13.3 | 18.2 | 4.0 | ||
Between 0 and 1 | 18.2 | 20.2 | 22.8 | 13.8 | ||
Subtotal | 42.1(33.8) | 48.4 (43.1) | 59.1 (41.2) | 27.7 (22.4) | ||
Mean | 1.9 (2.1) | 1.9 (2.2) | 2.2 (1.8) | 1.7 (2.1) | ||
Negative differences ("underperformance") | ||||||
Between 0 and −1 | 12.8 | 11.7 | 22.8 | 9.8 | ||
Between −1 and −2 | 18.1 | 13.3 | 9.1 | 27.7 | ||
Between −2 and −3 | 9.0 | 10.0 | 4.5 | 9.8 | ||
Between −3 and −4 | 5.2 | 3.3 | 4.5 | 7.8 | ||
Smaller than −4 | 12.8 | 13.3 | 0.0 | 17.1 | ||
Subtotal | 57.9 (66.2) | 51.6 (56.9) | 40.9 (58.8) | 72.2 (77.6) | ||
Mean | −2.8 (−2.2) | −2.9 (−3.0) | −1.4 (−1.5) | −3.1 (−2.5) | ||
Total | 100.0 | 100.0 | 100.0 | 100.0 | ||
Count | 133 | 60 | 21 | 52 | ||
Overall mean | −0.8 * (−0.6)* | −0.6 (−0.4) | 0.7 (−0.2) | −1.8* (−0.9)* | ||
Std | 3.5 | 3.4 | 2.6 | 3.7 |
This difference between actual and envisaged adjustment is statistically significant at the 99 percent or better confidence level.
Values in parentheses show the results when overperformance and underperformance are defined in terms of primary balances.
Differences Between Actual and Projected Changes in Fiscal Balances1
(From T–1 to T+1)
Distribution of Programs (In percent) |
||||||
---|---|---|---|---|---|---|
SBA/EFF | ||||||
Differences (In percent of GDP) |
All arrangements | ESAF/PRGF | Transition countries |
Nontransition countries |
||
Positive differences (“overperformance”) | ||||||
Larger than 4 | 4.5 | 3.3 | 13.6 | 2.0 | ||
Between 3 and 4 | 3.7 | 5.0 | 4.5 | 2.0 | ||
Between 2 and 3 | 5.2 | 6.6 | 0.0 | 5.9 | ||
Between 1 and 2 | 10.5 | 13.3 | 18.2 | 4.0 | ||
Between 0 and 1 | 18.2 | 20.2 | 22.8 | 13.8 | ||
Subtotal | 42.1(33.8) | 48.4 (43.1) | 59.1 (41.2) | 27.7 (22.4) | ||
Mean | 1.9 (2.1) | 1.9 (2.2) | 2.2 (1.8) | 1.7 (2.1) | ||
Negative differences ("underperformance") | ||||||
Between 0 and −1 | 12.8 | 11.7 | 22.8 | 9.8 | ||
Between −1 and −2 | 18.1 | 13.3 | 9.1 | 27.7 | ||
Between −2 and −3 | 9.0 | 10.0 | 4.5 | 9.8 | ||
Between −3 and −4 | 5.2 | 3.3 | 4.5 | 7.8 | ||
Smaller than −4 | 12.8 | 13.3 | 0.0 | 17.1 | ||
Subtotal | 57.9 (66.2) | 51.6 (56.9) | 40.9 (58.8) | 72.2 (77.6) | ||
Mean | −2.8 (−2.2) | −2.9 (−3.0) | −1.4 (−1.5) | −3.1 (−2.5) | ||
Total | 100.0 | 100.0 | 100.0 | 100.0 | ||
Count | 133 | 60 | 21 | 52 | ||
Overall mean | −0.8 * (−0.6)* | −0.6 (−0.4) | 0.7 (−0.2) | −1.8* (−0.9)* | ||
Std | 3.5 | 3.4 | 2.6 | 3.7 |
This difference between actual and envisaged adjustment is statistically significant at the 99 percent or better confidence level.
Values in parentheses show the results when overperformance and underperformance are defined in terms of primary balances.
The pace of the adjustment
The pace of the fiscal adjustment during the first two years of the program provides some interesting insights. Table 4.3 compares projected and actual changes in fiscal balances between the preprogram year T−1 and the program years T and T+1.
-
Almost all fiscal adjustment on average takes place during the first year of the program. Except in the transition countries, programs were unable to achieve further fiscal gains in the second year of the program in spite of more ambitious fiscal targets.
-
In SBA/EFF arrangements in nontransition countries, revenue ratios did not increase beyond the gain of 1 percentage point of GDP achieved during the first year of the program and expenditure ratios could not be reduced.
-
Concessional programs exhibited similar features, except that these programs were able to reduce expenditure ratios by the second year of the program. It is possible that this is because financing for these countries was more of a binding constraint than for the other cases.
The Dynamics of Fiscal Adjustment
(In percent of GDP)
Difference refers to the actual minus envisaged magnitudes. Hence, negative values show underperformance.
Difference statistically significant at the 95 percent confidence level. The other differences in means are not statistically significant.
The Dynamics of Fiscal Adjustment
(In percent of GDP)
(T−1) to T | (T−1) to (T+1) | |||||||
---|---|---|---|---|---|---|---|---|
N | Envisaged | Actual | Difference1 | Envisaged | Actual | Difference1 | ||
Changes in fiscal balances | ||||||||
All arrangements | 133 | 1.0 | 0.8 | −0.2 | 1.7 | 0.8 | −0.92 | |
SBA/EFF | ||||||||
Transition countries | 21 | 1.0 | 1.3 | 0.3 | 1.1 | 1.8 | 0.7 | |
Nontransition countries | 52 | 1.0 | 0.5 | −0.5 | 2.0 | 0.2 | −1.82 | |
ESAF/PRGF | 60 | 1.0 | 0.9 | −0.1 | 1.6 | 1.0 | −0.6 | |
Changes in revenues | ||||||||
All arrangements | 133 | 0.8 | 0.4 | −0.4 | 0.4 | 0.2 | −0.2 | |
SBA/EFF | ||||||||
Transition countries | 21 | −1.2 | −1.2 | 0.0 | −1.7 | −1.4 | 0.3 | |
Nontransition countries | 52 | 1.7 | 1.1 | −0.6 | 1.3 | 0.9 | −0.4 | |
ESAF/PRGF | 60 | 0.6 | 0.3 | −0.3 | 0.4 | 0.1 | −0.3 | |
Changes in expenditure | ||||||||
All arrangements | 133 | −0.3 | −0.4 | 0.1 | −1.2 | −0.7 | −0.5 | |
SBA/EFF | ||||||||
Transition countries | 21 | −2.2 | −2.4 | 0.2 | −2.8 | −3.2 | 0.4 | |
Nontransition countries | 52 | 0.6 | 0.6 | 0.0 | −0.7 | 0.6 | −1.32 | |
ESAF/PRGF | 60 | −0.4 | −0.6 | 0.2 | −1.2 | −1.0 | −0.2 |
Difference refers to the actual minus envisaged magnitudes. Hence, negative values show underperformance.
Difference statistically significant at the 95 percent confidence level. The other differences in means are not statistically significant.
The Dynamics of Fiscal Adjustment
(In percent of GDP)
(T−1) to T | (T−1) to (T+1) | |||||||
---|---|---|---|---|---|---|---|---|
N | Envisaged | Actual | Difference1 | Envisaged | Actual | Difference1 | ||
Changes in fiscal balances | ||||||||
All arrangements | 133 | 1.0 | 0.8 | −0.2 | 1.7 | 0.8 | −0.92 | |
SBA/EFF | ||||||||
Transition countries | 21 | 1.0 | 1.3 | 0.3 | 1.1 | 1.8 | 0.7 | |
Nontransition countries | 52 | 1.0 | 0.5 | −0.5 | 2.0 | 0.2 | −1.82 | |
ESAF/PRGF | 60 | 1.0 | 0.9 | −0.1 | 1.6 | 1.0 | −0.6 | |
Changes in revenues | ||||||||
All arrangements | 133 | 0.8 | 0.4 | −0.4 | 0.4 | 0.2 | −0.2 | |
SBA/EFF | ||||||||
Transition countries | 21 | −1.2 | −1.2 | 0.0 | −1.7 | −1.4 | 0.3 | |
Nontransition countries | 52 | 1.7 | 1.1 | −0.6 | 1.3 | 0.9 | −0.4 | |
ESAF/PRGF | 60 | 0.6 | 0.3 | −0.3 | 0.4 | 0.1 | −0.3 | |
Changes in expenditure | ||||||||
All arrangements | 133 | −0.3 | −0.4 | 0.1 | −1.2 | −0.7 | −0.5 | |
SBA/EFF | ||||||||
Transition countries | 21 | −2.2 | −2.4 | 0.2 | −2.8 | −3.2 | 0.4 | |
Nontransition countries | 52 | 0.6 | 0.6 | 0.0 | −0.7 | 0.6 | −1.32 | |
ESAF/PRGF | 60 | −0.4 | −0.6 | 0.2 | −1.2 | −1.0 | −0.2 |
Difference refers to the actual minus envisaged magnitudes. Hence, negative values show underperformance.
Difference statistically significant at the 95 percent confidence level. The other differences in means are not statistically significant.
Composition of the adjustment in programs with fiscal shortfalls
We now turn to examine the anatomy of programs with fiscal shortfalls, namely whether fiscal shortfalls are primarily due to revenues (as a share of GDP) being below target or expenditures (as a share of GDP) above target. The relevant data are presented in Table 4.4.
Percentage Distribution of Programs with Fiscal Shortfalls
(Shortfalls expressed as a share of GDP)
The shortfall is the difference between actual and projected values. For example, if the fiscal deficit is 3 percentage points of GDP higher than programmed, and spending is 1 percentage point of GDP higher than envisaged while revenue is 2 percentage points of GDP lower than projected, spending accounts for one-third of the fiscal shortfall and revenue for two-thirds.
Percentage Distribution of Programs with Fiscal Shortfalls
(Shortfalls expressed as a share of GDP)
SBA/EFF | ||||||
---|---|---|---|---|---|---|
All Arrangements | ESAF/PRGF | Transition countries | Nontransition countries | Large Episodes of Envisaged Fiscal Adjustment | ||
Programs where at least half the fiscal shortfall is due to:1 | ||||||
Expenditure shortfalls | 72.0 | 84.0 | 50.0 | 67.0 | 29.0 | |
Revenue shortfalls | 28.0 | 16.0 | 50.0 | 33.0 | 71.0 | |
Total | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 |
The shortfall is the difference between actual and projected values. For example, if the fiscal deficit is 3 percentage points of GDP higher than programmed, and spending is 1 percentage point of GDP higher than envisaged while revenue is 2 percentage points of GDP lower than projected, spending accounts for one-third of the fiscal shortfall and revenue for two-thirds.
Percentage Distribution of Programs with Fiscal Shortfalls
(Shortfalls expressed as a share of GDP)
SBA/EFF | ||||||
---|---|---|---|---|---|---|
All Arrangements | ESAF/PRGF | Transition countries | Nontransition countries | Large Episodes of Envisaged Fiscal Adjustment | ||
Programs where at least half the fiscal shortfall is due to:1 | ||||||
Expenditure shortfalls | 72.0 | 84.0 | 50.0 | 67.0 | 29.0 | |
Revenue shortfalls | 28.0 | 16.0 | 50.0 | 33.0 | 71.0 | |
Total | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 |
The shortfall is the difference between actual and projected values. For example, if the fiscal deficit is 3 percentage points of GDP higher than programmed, and spending is 1 percentage point of GDP higher than envisaged while revenue is 2 percentage points of GDP lower than projected, spending accounts for one-third of the fiscal shortfall and revenue for two-thirds.
Except for programs in transition countries, a much larger proportion of programs reflects situations where excess expenditure as a share of GDP (relative to targets) is the dominant source of the fiscal shortfall. This is particularly important in the case of ESAF/PRGF arrangements.
It is relevant to ask whether the shortfalls are the result of very ambitious fiscal targets (on either the revenue or expenditure side) or the result of moderate targets combined with very little progress in the actual adjustment. Table 4.5 presents a comparison of fiscal targets and actual achievements for the group of programs showing a shortfall.
-
The fiscal shortfall is largest for the group of nontransition SBA/EFF arrangements, where the fiscal deficit, far from showing an improvement by T+1, actually shows a deterioration. However, the volume of adjustment proposed in this group was not larger than for others. In fact, it is the subgroup of transition economies that shows the highest proposed improvement and this group also had the best compliance record.
-
In the case of SBA/EFF in nontransition countries, about two-thirds of the adjustment was expected to come from the revenue side and one-third from expenditure. In fact, revenues increased much less than expected while expenditure increased in relation to GDP, instead of declining as programmed.
-
In the ESAF/PRGF programs, both revenues and expenditures moved in the opposite direction compared with projections. Revenues declined originally instead of increasing, and expenditures increased instead of declining as programmed and the latter effect explains most of the shortfalls.
-
In the transition countries, the fiscal shortfall is due to expenditure shortfalls, in spite of programs achieving expenditure reductions equivalent to 5.3 percent of GDP. One may argue that this shortfall is to be expected given the significant expenditure cuts being programmed, equal to 7 percent of GDP.
The Composition of Fiscal Adjustment in Programs with Fiscal Underperformance
(In percent of GDP; values in parentheses refer to primary balances or primary expenditures1)
Values for primary balances or primary expenditures are presented only if they significantly differ from overall fiscal balances or total expenditures.
Difference refers to the actual minus envisaged magnitudes. Hence, negative values show underperformance.
The Composition of Fiscal Adjustment in Programs with Fiscal Underperformance
(In percent of GDP; values in parentheses refer to primary balances or primary expenditures1)
(T−1) to (T+1) | |||||
---|---|---|---|---|---|
N | Envisaged | Actual | Difference2 | ||
Changes in fiscal balances | |||||
All arrangements | 77 | 2.3 | −0.5 | −2.8 | |
SBA/EFF | |||||
Transition countries | 8 | 4.3 | 2.9 | −1.4 | |
Nontransition countries | 38 | 2.0 | −1.0 (0.0) | −3.0 (–2.0) | |
ESAF/PRGF | 31 | 2.1 | −0.8 | −2.9 | |
Changes in revenues | |||||
All arrangements | 77 | 0.5 | −0.1 | −0.6 | |
SBA/EFF | |||||
Transition countries | 8 | −2.7 | −2.4 | −0.3 | |
Nontransition countries | 38 | 1.4 | 0.5 | −0.9 | |
ESAF/PRGF | 31 | 0.3 | −0.3 | −0.6 | |
Changes in expenditures | |||||
All arrangements | 77 | −1.8 | 0.4 | −2.2 | |
SBA/EFF | |||||
Transition countries | 8 | −7.1 | −5.3 | −1.8 | |
Nontransition countries | 38 | −0.6 | 1.6 (0.6) | −2.2 (–1.2) | |
ESAF/PRGF | 31 | −1.8 | 0.5 | −2.3 |
Values for primary balances or primary expenditures are presented only if they significantly differ from overall fiscal balances or total expenditures.
Difference refers to the actual minus envisaged magnitudes. Hence, negative values show underperformance.
The Composition of Fiscal Adjustment in Programs with Fiscal Underperformance
(In percent of GDP; values in parentheses refer to primary balances or primary expenditures1)
(T−1) to (T+1) | |||||
---|---|---|---|---|---|
N | Envisaged | Actual | Difference2 | ||
Changes in fiscal balances | |||||
All arrangements | 77 | 2.3 | −0.5 | −2.8 | |
SBA/EFF | |||||
Transition countries | 8 | 4.3 | 2.9 | −1.4 | |
Nontransition countries | 38 | 2.0 | −1.0 (0.0) | −3.0 (–2.0) | |
ESAF/PRGF | 31 | 2.1 | −0.8 | −2.9 | |
Changes in revenues | |||||
All arrangements | 77 | 0.5 | −0.1 | −0.6 | |
SBA/EFF | |||||
Transition countries | 8 | −2.7 | −2.4 | −0.3 | |
Nontransition countries | 38 | 1.4 | 0.5 | −0.9 | |
ESAF/PRGF | 31 | 0.3 | −0.3 | −0.6 | |
Changes in expenditures | |||||
All arrangements | 77 | −1.8 | 0.4 | −2.2 | |
SBA/EFF | |||||
Transition countries | 8 | −7.1 | −5.3 | −1.8 | |
Nontransition countries | 38 | −0.6 | 1.6 (0.6) | −2.2 (–1.2) | |
ESAF/PRGF | 31 | −1.8 | 0.5 | −2.3 |
Values for primary balances or primary expenditures are presented only if they significantly differ from overall fiscal balances or total expenditures.
Difference refers to the actual minus envisaged magnitudes. Hence, negative values show underperformance.
Large episodes of envisaged adjustment
The case of large envisaged fiscal adjustment (defined as larger than 3 percent of GDP between T−1 and T+1) is of special interest because the results described above are reversed; revenue shortfalls account for most of the cases.1 The results are shown in Table 4.6.
-
The average targeted fiscal adjustment in this subgroup is 5.7 percent of GDP (Table 4.6) compared with only 1.7 percent of GDP for the 133 arrangements taken together (Table 4.3). The initial fiscal deficit in the preprogram year in this group is also higher, at 7.8 percent of GDP compared with 4.1 percent for the overall sample.
-
Programs achieve half of the envisaged adjustment; substantial adjustment was undertaken notwithstanding the shortfalls. Much of this adjustment was on the expenditure side. In fact, expenditure reductions for this group were significantly stronger than in milder cases of fiscal adjustment.
-
Revenue shortfalls remain significant in these cases, in spite of the higher requirements for revenue increases stemming from the need to reduce a more severe initial fiscal imbalance.
Changes in Government Balances in Large Episodes of Envisaged Adjustment from (T−1) to (T+1)
(In percent of GDP)
Difference statistically significant at the 95 percent confidence level.
Changes in Government Balances in Large Episodes of Envisaged Adjustment from (T−1) to (T+1)
(In percent of GDP)
Government Balance | Revenues | Expenditures | |||||||||
---|---|---|---|---|---|---|---|---|---|---|---|
N | Envisaged | Actual | Shortfall | Envisaged | Actual | Shortfall | Envisaged | Actual | Shortfall | ||
Total | 39 | 5.7 | 3.6 | −2.11 | 1.7 | −0.5 | −2.21 | −4.0 | −4.1 | 0.1 | |
SBA/EFF | |||||||||||
Transition countries | 6 | 6.5 | 5.2 | −1.3 | −2.1 | −5.7 | −3.6 | −8.6 | −10.8 | 2.2 | |
Nontransition countries | 17 | 5.4 | 3.4 | −2.0 | 3.0 | 1.5 | −1.5 | −2.4 | −2.0 | −0.4 | |
ESAF/PRGF | 16 | 5.7 | 3.1 | −2.61 | 1.7 | −0.8 | −2.51 | −3.9 | −3.9 | 0.0 | |
Difference statistically significant at the 95 percent confidence level.
Changes in Government Balances in Large Episodes of Envisaged Adjustment from (T−1) to (T+1)
(In percent of GDP)
Government Balance | Revenues | Expenditures | |||||||||
---|---|---|---|---|---|---|---|---|---|---|---|
N | Envisaged | Actual | Shortfall | Envisaged | Actual | Shortfall | Envisaged | Actual | Shortfall | ||
Total | 39 | 5.7 | 3.6 | −2.11 | 1.7 | −0.5 | −2.21 | −4.0 | −4.1 | 0.1 | |
SBA/EFF | |||||||||||
Transition countries | 6 | 6.5 | 5.2 | −1.3 | −2.1 | −5.7 | −3.6 | −8.6 | −10.8 | 2.2 | |
Nontransition countries | 17 | 5.4 | 3.4 | −2.0 | 3.0 | 1.5 | −1.5 | −2.4 | −2.0 | −0.4 | |
ESAF/PRGF | 16 | 5.7 | 3.1 | −2.61 | 1.7 | −0.8 | −2.51 | −3.9 | −3.9 | 0.0 | |
Difference statistically significant at the 95 percent confidence level.
A possible explanation for the heavier reliance on expenditure cuts in programs with very large initial fiscal imbalances is simply that the deficits could not be financed and large expenditure cuts became unavoidable when revenue measures did not yield results quickly enough.
This conclusion has special applicability in SBA/EFF arrangements in nontransition countries. In spite of the relatively higher level of development in these countries, programs were not able to raise more than 1.5 percent of GDP in extra revenues by the second year of the program, irrespective of the severity of the initial fiscal deficit and the size of the targeted fiscal deficit. Expenditures were then cut residually.
Determinants of fiscal shortfalls
Regression analysis is one way of identifying possible determinants of the fiscal shortfalls and our results are presented in detail in Appendix 1, Table A1.2. The most significant variable was the difference between the envisaged and actual rate of growth for T+1.2 Lower-than-envisaged GDP growth was associated with less fiscal adjustment than envisaged; a shortfall in the growth rate with respect to projections equal to 1 percentage point of GDP was associated with a fiscal shortfall compared to programmed levels of 0.3 percentage point of GDP.
It is relevant to ask whether this effect operates via the expenditure or revenue side. To explore these channels, separate regressions were run to explain revenue and expenditure ratio shortfalls. Interestingly, growth optimism proved to be significant in explaining optimism in reducing expenditures as a share of GDP (also with a coefficient equal to 0.3 but negative) but not in explaining optimism in forecasting revenues as a share of GDP. In fact, the elasticities of projected and actual revenues with respect to GDP happen to be similar and close to one, so that shortfalls in GDP growth lead to proportional shortfalls in revenue without much effect on the revenue/GDP ratio.3 On the other hand, optimism in growth generates optimism in projecting declines in the expenditure/GDP ratios which are not realized in the end because nominal expenditure levels are usually less sensitive to growth.
The low explanatory power of the regressions—they explained only about 22 percent of the variation in the differences between envisaged and actual fiscal adjustment—suggests the omission of important explanatory variables that may also influence revenue ratio shortfalls. This is particularly true in the episodes of large envisaged fiscal adjustment, where programs may have overestimated the speed at which tax policy and tax administration measures could be implemented, or the extent by which these policies could quickly yield revenue increases. The overall role of optimism regarding the progress of structural reforms in the fiscal area is examined later on in Chapter 7.
Flexibility of Fiscal Targets During Program Reviews
In view of the persistent shortfall in fiscal performance compared with targets, it is relevant to consider how programs are adjusted to take account of shortfalls. This section first looks at the magnitude and direction of revisions in fiscal targets in a large sample of programs in the 1993–2001 period. The results are then complemented by a qualitative analysis of the 15 program case studies. We examine three interrelated questions associated with the extent of fiscal flexibility in programs. First, in what direction are fiscal targets being revised? Second, are the revisions linked to changes in other projected variables such as GDP growth? Finally, do program documents provide a good rationale for the revision? To answer these questions we have analyzed how the first and second reviews modified fiscal targets for year T+1, namely, the second year of the program.4
Patterns in the revision of fiscal targets
Revision in fiscal targets (and indeed in other key macroeconomic targets) is common in the course of program review. Figure 4.1 presents the distribution of differences between the original fiscal target and the revised target in the first review (Panel A) and between the original target and the second review (Panel B).
Distribution of Programs According to Differences in Fiscal Adjustment Between Original Targets and Reviews
(In percent of GDP)
Sources: MONA and WEO databases.Note: Negative (positive) values correspond to situations in which the review has lowered (raised) the targeted improvement in the fiscal balance.Panel A shows that at the first review 55 percent of the programs experienced minor revisions (between plus/minus 0.5 percentage point of GDP) and the few cases of large revisions are more or less distributed symmetrically in both directions. These results are not surprising given that the first review is relatively close to the date of the original program request. However, by the time of the second review (Panel B), the center of the distribution shifts to the left suggesting that by the second year many more targets are relaxed (targeting a lower improvement in the fiscal balance) than tightened (targeting a greater improvement in the fiscal balance).
Since fiscal outcomes are affected by growth outcomes, it is relevant to consider whether revisions in fiscal targets in the course of programs reflect revisions in growth projections (Table 4.7). The data suggest that both distributions are indeed related.5 However, it is interesting to note that growth revisions do not seem to operate in a symmetric way: when growth projections are revised downward, fiscal balance targets are adjusted downward in twothirds of the cases. By contrast, when growth projections are revised upward, fiscal targets are adjusted upward in only half of the cases.
Distribution of Programs According to Revisions in Growth and Fiscal Balances
(First Review)
Distribution of Programs According to Revisions in Growth and Fiscal Balances
(First Review)
Revisions in Fiscal Balances |
||||
---|---|---|---|---|
Revisions in GDP Growth |
No change | Upward Revision | Downward Revision | Row Total |
No change | 14 | 10 | 6 | 30 |
Upward revision | 1 | 14 | 14 | 29 |
Downward revision | 4 | 13 | 21 | 38 |
Column total | 19 | 37 | 41 | 97 |
Distribution of Programs According to Revisions in Growth and Fiscal Balances
(First Review)
Revisions in Fiscal Balances |
||||
---|---|---|---|---|
Revisions in GDP Growth |
No change | Upward Revision | Downward Revision | Row Total |
No change | 14 | 10 | 6 | 30 |
Upward revision | 1 | 14 | 14 | 29 |
Downward revision | 4 | 13 | 21 | 38 |
Column total | 19 | 37 | 41 | 97 |
This asymmetry was also captured when we ran separate regression equations for the cases where GDP growth was adjusted downward and for those where it was adjusted upward. The results show that when growth is revised downward by 1 percentage point, fiscal targets (as percent of GDP) are on average revised downward by ⅓ of 1 percent of GDP. However, no statistically significant impact was found on fiscal targets when growth was revised upward.6
These results suggest that program targets do respond to changes in growth outlooks as the program unfolds but the response tends to be asymmetric. Fiscal targets are revised downward when growth is below expectations, but they are less often revised upward when growth turns out to be higher than originally projected.
Rationale for revisions in fiscal targets
The rationale for the mid-course revisions of fiscal targets was further examined on the basis of the 15 case studies used in this evaluation. Table 4.8 shows the revision in targets for these case studies. We selected 11 cases in which the fiscal deficit target was adjusted upward or downward by more than 1 percentage point of GDP between the original program request and the first review (Algeria, Ecuador, the Philippines, Romania, Uruguay, and Venezuela programs), or between the first review and the second review (Bulgaria, Jordan, the Philippines, Senegal, and Tanzania programs).7 Three of these cases (Ecuador, Bulgaria, and Venezuela) were examples where the fiscal target was actually tightened, whereas in the other eight it was relaxed.
Revisions in Fiscal Balance Targets for T+1
(In percent of GDP)
Cases of revision between the original program target and the first review or between the original program and the second review, more than 1 percent of GDP.
For Ecuador and Venezuela we examined original program projections and reviews for year T due to insufficient data for year T+1. For all other programs, figures refer to changes in the government fiscal balance from T−1 to T+1.
Fiscal balance excluding grants.
Revisions in Fiscal Balance Targets for T+1
(In percent of GDP)
Fiscal Balance T−1 | Original Program | First Review | Second Review | |
---|---|---|---|---|
Algeria SBA 19941 | −8.6 | 3.3 | 1.2 | N/A |
Bulgaria EFF 19981 | −2.5 | −2.0 | −2.8 | −1.5 |
Costa Rica SBA 1995 | −5.1 | −1.1 | −1.1 | N/A |
Ecuador SBA 20002 | −5.8 | −3.9 | −2.8 | N/A |
Egypt SBA 1996 | −1.3 | 0.2 | −0.6 | −0.9 |
Jordan EFF 19993 | −9.5 | −5.5 | −5.5 | −7.5 |
Pakistan SBA 2000 | −6.0 | −5.3 | −5.3 | −5.3 |
Peru EFF 1996 | −2.8 | −0.7 | −0.7 | −0.8 |
Philippines SBA 19981 | −0.7 | −0.1 | −1.6 | −2.8 |
Romania SBA 19991 | −5.0 | −1.1 | −3.4 | N/A |
Senegal PRGF 19981, 3 | −2.0 | −1.0 | −1.0 | −4.4 |
Tanzania ESAF 19961 | −4.4 | −2.0 | −2.0 | −3.4 |
Ukraine EFF 1998 | −5.4 | −1.0 | −1.0 | −1.3 |
Uruguay SBA 20001 | −4.2 | −1.2 | −2.6 | −3.3 |
Venezuela SBA 19962 | −3.2 | −3.3 | −0.4 | N/A |
Cases of revision between the original program target and the first review or between the original program and the second review, more than 1 percent of GDP.
For Ecuador and Venezuela we examined original program projections and reviews for year T due to insufficient data for year T+1. For all other programs, figures refer to changes in the government fiscal balance from T−1 to T+1.
Fiscal balance excluding grants.
Revisions in Fiscal Balance Targets for T+1
(In percent of GDP)
Fiscal Balance T−1 | Original Program | First Review | Second Review | |
---|---|---|---|---|
Algeria SBA 19941 | −8.6 | 3.3 | 1.2 | N/A |
Bulgaria EFF 19981 | −2.5 | −2.0 | −2.8 | −1.5 |
Costa Rica SBA 1995 | −5.1 | −1.1 | −1.1 | N/A |
Ecuador SBA 20002 | −5.8 | −3.9 | −2.8 | N/A |
Egypt SBA 1996 | −1.3 | 0.2 | −0.6 | −0.9 |
Jordan EFF 19993 | −9.5 | −5.5 | −5.5 | −7.5 |
Pakistan SBA 2000 | −6.0 | −5.3 | −5.3 | −5.3 |
Peru EFF 1996 | −2.8 | −0.7 | −0.7 | −0.8 |
Philippines SBA 19981 | −0.7 | −0.1 | −1.6 | −2.8 |
Romania SBA 19991 | −5.0 | −1.1 | −3.4 | N/A |
Senegal PRGF 19981, 3 | −2.0 | −1.0 | −1.0 | −4.4 |
Tanzania ESAF 19961 | −4.4 | −2.0 | −2.0 | −3.4 |
Ukraine EFF 1998 | −5.4 | −1.0 | −1.0 | −1.3 |
Uruguay SBA 20001 | −4.2 | −1.2 | −2.6 | −3.3 |
Venezuela SBA 19962 | −3.2 | −3.3 | −0.4 | N/A |
Cases of revision between the original program target and the first review or between the original program and the second review, more than 1 percent of GDP.
For Ecuador and Venezuela we examined original program projections and reviews for year T due to insufficient data for year T+1. For all other programs, figures refer to changes in the government fiscal balance from T−1 to T+1.
Fiscal balance excluding grants.
The reasons given in the review documents for the revision in fiscal targets are summarized in Table 4.9 classified into two groups: those with little or no explanation, and those that provide some justification for the revisions. We find that out of the 11 cases in which the fiscal target was revised by more than 1 percent of GDP, program documents provided some justification for the new target in 7 cases with little or no justification in the other 4.
Summary of Reasons for the Revised Fiscal Balance Target
Summary of Reasons for the Revised Fiscal Balance Target
Little or No Discussion | Relatively Clear Discussion | |
---|---|---|
Algeria SBA 1994 (Firstreview) | No explanation for the downward revision in the fiscal target for year T+1. | |
Ecuador SBA 2000 (First review) | Higher growth and higher oil prices than expected led to better revenue performance. Lower expenditures due to cuts in investment spending. | |
Jordan EFF 1999 (Second review) | Prudent fiscal stance is central to achieving macroeconomic objectives. Need to diminish high public debt burden. Mindful that unduly rapid fiscal adjustment would have a negative effect on growth and employment. | |
Philippines SBA 1998 (First review) | Loosening of the fiscal target to accommodate the effects of sharply lower GDP growth and somewhat higher social spending than originally envisaged. | |
Romania SBA 1999 (First review) | Review document contains a comprehensive analysis of fiscal policy but does not provide a good sense of why the fiscal deficit target was revised downward. | |
Uruguay SBA 2000 (First review) | Loosening of fiscal target owing to expected revenue shortfalls associated with a large output gap relative to potential GDP. Adhering to the original fiscal deficit target, the review argues, would imply large tax increases that would make the resumption of growth more difficult. | |
Venezuela SBA 1996 (First review) | Review document explains that higher-than-expected oil prices and lower-than-expected interest payments justify the much larger improvement in the fiscal balance than originally expected. | |
Bulgaria EFF 1998 (Second review) | Fiscal target is tightened to anticipate a worse-than-expected external position. Stronger revenues than initially expected also played a key role. | |
Philippines SBA 1998 (Second review) | Fiscal targets are relaxed to accommodate the effects of weaker economic conditions and also to permit higher social expenditures. | |
Senegal PRGF 1998 (Second review) | No explanation for the revision in the fiscal target. | |
Tanzania ESAF 1996 (Second review) | No explanation for the revision in the fiscal target. |
Summary of Reasons for the Revised Fiscal Balance Target
Little or No Discussion | Relatively Clear Discussion | |
---|---|---|
Algeria SBA 1994 (Firstreview) | No explanation for the downward revision in the fiscal target for year T+1. | |
Ecuador SBA 2000 (First review) | Higher growth and higher oil prices than expected led to better revenue performance. Lower expenditures due to cuts in investment spending. | |
Jordan EFF 1999 (Second review) | Prudent fiscal stance is central to achieving macroeconomic objectives. Need to diminish high public debt burden. Mindful that unduly rapid fiscal adjustment would have a negative effect on growth and employment. | |
Philippines SBA 1998 (First review) | Loosening of the fiscal target to accommodate the effects of sharply lower GDP growth and somewhat higher social spending than originally envisaged. | |
Romania SBA 1999 (First review) | Review document contains a comprehensive analysis of fiscal policy but does not provide a good sense of why the fiscal deficit target was revised downward. | |
Uruguay SBA 2000 (First review) | Loosening of fiscal target owing to expected revenue shortfalls associated with a large output gap relative to potential GDP. Adhering to the original fiscal deficit target, the review argues, would imply large tax increases that would make the resumption of growth more difficult. | |
Venezuela SBA 1996 (First review) | Review document explains that higher-than-expected oil prices and lower-than-expected interest payments justify the much larger improvement in the fiscal balance than originally expected. | |
Bulgaria EFF 1998 (Second review) | Fiscal target is tightened to anticipate a worse-than-expected external position. Stronger revenues than initially expected also played a key role. | |
Philippines SBA 1998 (Second review) | Fiscal targets are relaxed to accommodate the effects of weaker economic conditions and also to permit higher social expenditures. | |
Senegal PRGF 1998 (Second review) | No explanation for the revision in the fiscal target. | |
Tanzania ESAF 1996 (Second review) | No explanation for the revision in the fiscal target. |
Two patterns emerge that are worth reporting:
(1) When fiscal performance by the time of the review was weaker than projected, program documents did not clearly analyze and try to separate what part of this weaker performance was due to weak policy implementation and what part to factors outside the control of governments—for example, lower growth than expected, higher interest payments, terms of trade shocks, etc. Furthermore, there is a general tendency to emphasize the role of factors outside the government’s control. However, as we show later, insufficient progress in structural reforms in the fiscal area is an important factor behind shortfalls in fiscal adjustment in IMF-supported programs.
(2) Review documents tend to be more back-ward-looking than forward-looking; they typically elaborate why past fiscal developments call for revisions, but not much is said why the new targets are appropriate given the overall objectives of the program.8 Only the Philippines program, and to a lesser extent the Uruguay program, provided sufficient forward-looking analysis or the need for revising fiscal targets. In the case of the Philippines program, the review documents for the first and second reviews contain a comprehensive analysis of the fiscal stance (EBS/98/172). They provide an assessment of the past and an analysis of the new fiscal balance target for T+1. The report justifies the relaxation of the fiscal target in terms of lower growth than originally envisaged and the need to accommodate higher social spending. In the case of Uruguay the review document argues that, given the weak revenue performance by the time of the review, sticking to the original fiscal target would have a contractionary impact on output.
What Accounts for Large Fiscal Underperformance? Evidence from the Smaller Sample
The analysis of differences between envisaged and actual fiscal adjustment in the large sample of programs can be supplemented by evidence from the smaller sample of 15 from which we extract the 9 cases of fiscal underperformance (Table 4.10). Specifically, we focus on the 7 cases of large fiscal shortfall, that is, where actual adjustment was 2 or more percentage points of GDP less than envisaged. These 7 cases of large fiscal shortfall comprise 2 (Algeria SBA 1994 and Uruguay SBA 2000) dominated by expenditure overruns and 5 (Costa Rica SBA 1995, Jordan EFF 1999, the Philippines SBA 1998, Romania SBA 1999, and Tanzania ESAF 1996) mainly caused by revenue shortfalls.
Envisaged and Actual Fiscal Adjustment in Nine IMF-Supported Programs
(Ranked by the magnitude of the fiscal shortfall)
The percentage contribution of revenue plus expenditure shortfalls add up to the 100 percent shortfall in the fiscal adjustment. Negative (positive) values show that revenues or expenditures adjusted less (or more) than was projected. When the revenue component adds up to more than minus 100 percent, it means revenue shortfalls were larger than the total fiscal adjustment shortfall—expenditures then adjusting more than projected.
Envisaged and Actual Fiscal Adjustment in Nine IMF-Supported Programs
(Ranked by the magnitude of the fiscal shortfall)
Initial Conditions |
Fiscal Adjustment (T−1 to T+1) | Decomposition of the Fiscal Shortfall1 |
|||||
---|---|---|---|---|---|---|---|
Program | Fiscal balance (T−1) |
Envisaged | Actual | Difference | Revenue component |
Expenditure component |
|
Cases of fiscal underperformance | (In percent of GDP) | (In percent of the overall adjustment) | |||||
Algeria | −8.6 | 11.9 | 7.2 | −4.7 | −25 | −75 | |
Philippines | −0.7 | 0.8 | −3.5 | −4.3 | −93 | −7 | |
Tanzania | −4.4 | 2.7 | −0.7 | −2.0 | −90 | −10 | |
Romania | −5.0 | 3.9 | 1.0 | −2.9 | −175 | +75 | |
Costa Rica | −5.1 | 4.0 | 1.1 | −2.9 | −155 | +55 | |
Uruguay | −4.2 | 3.0 | 0.0 | −3.0 | −7 | −93 | |
Jordan | −9.5 | 4.0 | 0.6 | −3.4 | −166 | +66 | |
Ukraine | −5.4 | 4.4 | 3.1 | −1.3 | −115 | +15 | |
Egypt | −1.3 | 1.5 | 0.3 | −1.2 | −75 | −25 |
The percentage contribution of revenue plus expenditure shortfalls add up to the 100 percent shortfall in the fiscal adjustment. Negative (positive) values show that revenues or expenditures adjusted less (or more) than was projected. When the revenue component adds up to more than minus 100 percent, it means revenue shortfalls were larger than the total fiscal adjustment shortfall—expenditures then adjusting more than projected.
Envisaged and Actual Fiscal Adjustment in Nine IMF-Supported Programs
(Ranked by the magnitude of the fiscal shortfall)
Initial Conditions |
Fiscal Adjustment (T−1 to T+1) | Decomposition of the Fiscal Shortfall1 |
|||||
---|---|---|---|---|---|---|---|
Program | Fiscal balance (T−1) |
Envisaged | Actual | Difference | Revenue component |
Expenditure component |
|
Cases of fiscal underperformance | (In percent of GDP) | (In percent of the overall adjustment) | |||||
Algeria | −8.6 | 11.9 | 7.2 | −4.7 | −25 | −75 | |
Philippines | −0.7 | 0.8 | −3.5 | −4.3 | −93 | −7 | |
Tanzania | −4.4 | 2.7 | −0.7 | −2.0 | −90 | −10 | |
Romania | −5.0 | 3.9 | 1.0 | −2.9 | −175 | +75 | |
Costa Rica | −5.1 | 4.0 | 1.1 | −2.9 | −155 | +55 | |
Uruguay | −4.2 | 3.0 | 0.0 | −3.0 | −7 | −93 | |
Jordan | −9.5 | 4.0 | 0.6 | −3.4 | −166 | +66 | |
Ukraine | −5.4 | 4.4 | 3.1 | −1.3 | −115 | +15 | |
Egypt | −1.3 | 1.5 | 0.3 | −1.2 | −75 | −25 |
The percentage contribution of revenue plus expenditure shortfalls add up to the 100 percent shortfall in the fiscal adjustment. Negative (positive) values show that revenues or expenditures adjusted less (or more) than was projected. When the revenue component adds up to more than minus 100 percent, it means revenue shortfalls were larger than the total fiscal adjustment shortfall—expenditures then adjusting more than projected.
In the case of the Uruguay and Algeria programs, shortfalls in the expenditure/GDP ratio relative to targets cannot be attributed to weaknesses in implementing the program. In the case of Uruguay, nominal spending was in fact within the agreed ceiling, with the shortfall reflecting a significantly lower growth performance than projected. (In fact, the program projected recovery of growth while in reality growth was negative.) In the Algeria program, nominal spending was indeed higher than envisaged. However this reflected unexpected shocks: specifically, spending in the wake of an earthquake, and higher-than-expected outlays to protect public safety in response to heightened security concerns.
In contrast, the revenue shortfalls seem to be associated with weak implementation as outlined below. Table 4.11 compares the shortfall in envisaged GDP growth to the revenue shortfall in the five large revenue underperformers.
Comparing Growth and Revenue Underperformance
Difference between the actual average growth in T and T+1 and the equivalent projected value.
Difference between the actual revenue over GDP in T+1 and the equivalent projected value.
Consistency of data may be compromised by data revisions in the GDP series after the original program request. The revenue shortfall after taking these revisions into account is still substantially large.
Comparing Growth and Revenue Underperformance
Shortfall with Respect to Program Targets | |||||
---|---|---|---|---|---|
Growth (In percent) |
∆Revenues (In percent of GDP) |
||||
Program | T | T+1 | T−1 to T+1 | GDP growth1 (Percentage points) |
Revenue2 (In percent of GDP in T+1) |
Costa Rica | 2.4 | 0.7 | −2.3 | −1.2 | −4.53 |
Jordan | 3.1 | 4.0 | −2.4 | 1.3 | −3.0 |
Philippines | −0.6 | 3.4 | −3.6 | −2.1 | −3.4 |
Romania | −2.3 | 1.6 | −1.3 | 0.1 | −5.23 |
Tanzania | 3.7 | 3.7 | −0.8 | −1.8 | −4.5 |
Average | 1.3 | 2.7 | −2.1 | −0.8 | −4.1 |
Difference between the actual average growth in T and T+1 and the equivalent projected value.
Difference between the actual revenue over GDP in T+1 and the equivalent projected value.
Consistency of data may be compromised by data revisions in the GDP series after the original program request. The revenue shortfall after taking these revisions into account is still substantially large.
Comparing Growth and Revenue Underperformance
Shortfall with Respect to Program Targets | |||||
---|---|---|---|---|---|
Growth (In percent) |
∆Revenues (In percent of GDP) |
||||
Program | T | T+1 | T−1 to T+1 | GDP growth1 (Percentage points) |
Revenue2 (In percent of GDP in T+1) |
Costa Rica | 2.4 | 0.7 | −2.3 | −1.2 | −4.53 |
Jordan | 3.1 | 4.0 | −2.4 | 1.3 | −3.0 |
Philippines | −0.6 | 3.4 | −3.6 | −2.1 | −3.4 |
Romania | −2.3 | 1.6 | −1.3 | 0.1 | −5.23 |
Tanzania | 3.7 | 3.7 | −0.8 | −1.8 | −4.5 |
Average | 1.3 | 2.7 | −2.1 | −0.8 | −4.1 |
Difference between the actual average growth in T and T+1 and the equivalent projected value.
Difference between the actual revenue over GDP in T+1 and the equivalent projected value.
Consistency of data may be compromised by data revisions in the GDP series after the original program request. The revenue shortfall after taking these revisions into account is still substantially large.
In most cases (except Costa Rica) growth improved during the program period. Despite this acceleration in growth, revenue ratios declined. This suggests that GDP growth played a limited role in accounting for the poor revenue performance. Neither could the large shortfall in revenue performance with respect to targets be explained by the observed shortfall in growth performance in the program period relative to projections. Indeed, revenue under-performance is about four times the growth under-performance. These magnitudes cannot be explained by typical revenue-GDP elasticities, which are normally around one.9
We can summarize the above findings as follows:
-
For the programs with expenditure shortfalls, it is either unexpected shocks or the optimism in the envisaged GDP growth that explains expenditure overruns as a share of GDP.
-
In programs with significant revenue shortfalls with respect to targets, neither actual growth performance nor growth optimism can explain these shortfalls.
Optimism in growth projections cannot, therefore, explain the large underperformance of revenue. Instead, underperformance must be related to other factors such as structural reforms. Either reforms were implemented less rapidly than envisaged or staff over-estimated the impact of these reforms. Indeed, Table 4.12 suggests that underperformance was mainly the result of insufficient progress in revenue-enhancing structural reforms. Cases of large revenue shortfalls were mostly the result of poor implementation of structural reforms envisaged in the program (e.g., the Costa Rica, Jordan, Philippines, and Romania programs), or the implementation of reforms likely to have an impact on revenue only over the medium term (e.g., the Tanzania program that envisaged preparatory steps for the implementation of VAT and parliamentary approval of associated legislation).
Revenue-Related Structural Reform Measures in Selected Programs with Large Revenue Shortfalls
Revenue-Related Structural Reform Measures in Selected Programs with Large Revenue Shortfalls
Program | Implementation | |
---|---|---|
Philippines |
|
Partial progress. Partial/poor progress. Slow progress. Partial progress. |
Tanzania |
|
Done. Done with delay. |
Romania |
|
Partial progress. Done. Implemented with delay. Not done. |
Jordan |
|
Done. |
Costa Rica | No structural benchmarks related to revenue but there was a PC on the net borrowing requirements of the nonfinancial public sector that incorporated an anticipated 3 percentage point of GDP increase in taxes from:
|
A waiver was required for the PC due, inter alia, to delays in adopting tax measures in 1995. |
Revenue-Related Structural Reform Measures in Selected Programs with Large Revenue Shortfalls
Program | Implementation | |
---|---|---|
Philippines |
|
Partial progress. Partial/poor progress. Slow progress. Partial progress. |
Tanzania |
|
Done. Done with delay. |
Romania |
|
Partial progress. Done. Implemented with delay. Not done. |
Jordan |
|
Done. |
Costa Rica | No structural benchmarks related to revenue but there was a PC on the net borrowing requirements of the nonfinancial public sector that incorporated an anticipated 3 percentage point of GDP increase in taxes from:
|
A waiver was required for the PC due, inter alia, to delays in adopting tax measures in 1995. |
Three extreme cases of major expenditure collapse were excluded from this group: Armenia SBA 1995; Equatorial Guinea ESAF 1993; and Malawi ESAF 1995. In these programs, public expenditures collapsed between 18 and 26 percent of GDP.
It may be surprising that deviations in growth projections from actuals do explain deviations in fiscal adjustment while growth projections did not seem to have influenced a program’s fiscal adjustment projections (as found earlier in Chapter 2, “Factors Determing the Scale and Nature of Fiscal Adjustment”). This apparent puzzle is explained by the fact that actual fiscal adjustment was indeed found to be associated with actual growth (Appendix 1, Table A1.1). In fact, the coefficient of the growth variable in the equations for deviations is similar to the one in the equation for actuals. The fact that growth is not significant in the fiscal projection equations does not mean that errors in growth projections do not influence shortfalls in fiscal adjustment. This effect will persist as long as actual fiscal adjustment is influenced by actual growth rates.
Regressions to explain both envisaged and actual changes in revenue/GDP confirm that forecast and actual revenue-GDP elasticities are similar and close to one.
We chose T+1 rather than T for two reasons. First, very few programs have more than one revision during the first year of the program. Hence, looking only at one revision would provide a partial and incomplete picture of the review process. Second, we are interested in separating situations where revisions in targets reflect an ex post rationalization (i.e., some months into the program actual data for part of the year become available and targets may be simply revised to conform with actual developments) from situations where revisions represent a genuine forward-looking policy response in the face of changing economic circumstances. Revisions at year T are often too close to the end of the first program year to be able to separate the two effects.
A Chi-square test rejects the null hypothesis of no association between the two distributions. Chi-square = 23.08; df = 0.4; p-value = 0.0001.
This result is based on the following regression framework:
Where
We focus on cases of relatively large revision in the initial fiscal target where the need to explain why the new target is needed is presumably more relevant.
This is certainly understandable in those cases in which the review takes place very close to the end of year T+1 (as there is little time to change course in fiscal policy). But when the distance between the time of the review and the end of T+1 is sufficiently large (e.g., more than six months), review documents should explain why the new target is consistent with the overall objectives of the program.
In some cases, growth may have been concentrated in lightly taxed sectors (e.g., agriculture and exports in the Philippines), a factor that may not have been anticipated in the original revenue projections. Nevertheless, if the shift to lightly taxed sectors was permanent, relatively painless policy action should have been feasible to restore the traditional share of taxes by taxing part of the unexpected growth, for example by rolling back exemptions.