Chapter

IV Ghana

Author(s):
Andrew Berg, Mumtaz Hussain, Shaun Roache, Amber Mahone, Tokhir Mirzoev, and Shekhar Aiyar
Published Date:
March 2007
Share
  • ShareShare
Show Summary Details
Author(s)
Shaun Roache

Ghana’s economy has performed well in recent years compared with its regional peers. Per capita GDP growth averaged 1.6 percent annually in the 1990s. After a period of economic volatility around the turn of the century, Ghana has pursued economic policies that have delivered a degree of fiscal consolidation, lower inflation, and steadily increasing real GDP growth. Starting in 2000–01, international support jumped by several percentage points of GDP. Net aid averaged 7.3 percent of GDP during 2001–03 compared with 2.8 percent during 1996–2000 (Table 4.1).

Table 4.1.Ghana: Net Aid Inflows and Selected Economic Indicators(Percent of GDP)
19971998199920002001200220032004
Gross aid inflows8.88.77.58.814.95.99.57.4
Project aid7.86.95.65.09.33.44.43.3
Program aid1.01.81.93.85.62.55.14.1
Debt service15.35.64.79.04.23.42.4–0.6
Net aid inflows3.53.22.8–0.310.62.57.18.0
Private inflows214.06.512.914.812.713.913.78.5
Memorandum items:
GDP (real percent change)4.24.74.43.74.24.55.25.2
Inflation (percent change)18.416.313.139.323.514.124.0
Cedi per U.S. dollar (average)2,0502,3142,6695,4557,1707,9328,6779,004
Percent change–20.1–11.4–13.3–51.1–23.9–9.6–8.6–3.6
Real exchange rate vs. U.S. dollar (percent change)–7.61.5–4.1–34.0–8.61.510.8
Real effective exchange rate (percent change)6.08.20.5–35.50.6–0.61.40.0

This case study looks at Ghana’s experience during 1996–2003, a period when the country was almost continuously engaged in economic adjustment programs supported by the International Monetary Fund’s Enhanced Structural Adjustment Facility (ESAF) and its successor, the Poverty Reduction and Growth Facility (PRGF).1

Pattern of Aid Inflows

The pattern of aid inflows is critical to understanding Ghana’s policy response to aid during 2000–03 and the resulting economic outcomes. Net aid jumped in 2001, collapsed in 2002, and jumped again in 2003. This volatility was unexpected, and thus required large and rapid policy adjustments (Table 4.2). Most of these changes were driven by changes in gross aid flows. Debt relief under the Heavily Indebted Poor Countries (HIPC) Initiative was also a major, and much smoother, contributing factor. Ghana reached the HIPC Initiative “decision point” in 2002, and the value of the associated debt service relief was about half that of loans and grants in 2002–03. Private flows, mostly transfers but including unidentified items as well, also rose sharply, amplifying the effects of aid shocks.

Table 4.2.Ghana: Aid Shocks(In percent of GDP)
199819992000200120022003
Project grants–0.1–0.5–0.71.8–1.0–0.2
Program grants0.30.2–0.62.0–0.10.3
HIPC assistance0.00.00.00.00.00.2
Project loans1.91.1–0.7–0.1–1.80.2
Program loans0.2–0.3–3.91.0–1.20.6
Gross aid shock2.30.6–6.04.7–4.11.8
Net aid shock0.50.0–6.14.9–4.51.3
Note: HIPC = Heavily Indebted Poor Countries Initiative. “Aid shocks” are defined as aid received in a given year less aid expected just prior to that year, as reflected in IMF staff projections. A negative number thus implies a shortfall in actual aid compared to the expected level of aid.

Was Aid Absorbed?

During the entire 2001–03 period when aid soared, Ghana’s non-aid current account deficit narrowed. In other words, the aid was not used to increase net imports, or more generally to raise investment relative to domestic savings. Rather, inflows were fully accumulated as reserves. The narrower non-aid current account deficit reflected stable or declining import volumes (Table 4.3).

Table 4.3.Ghana: Balance of Payments(In percent of GDP)
1997199819992000200120022003
Levels
Aid and reserves6.46.24.94.46.9–2.2–2.5
Net aid15.16.22.43.38.42.47.5
Change in gross reserves (decrease –)1.30.02.51.1–1.5–4.6–10.0
Non-aid balance of payments–6.4–6.2–4.9–4.4–6.92.22.7
Non-aid current account–15.2–8.4–11.7–9.3–8.3–1.1–1.0
Non-aid capital account28.82.26.84.91.43.23.6
Changes
Aid and reserves (changes)–0.2–1.3–0.52.5–9.1–0.3
Change in net aid11.2–3.80.95.1–6.05.1
Change in gross reserves (decrease –)–1.32.5–1.4–2.6–3.1–5.4
Non-aid balance of payments (changes)0.21.30.5–2.59.10.5
Change in non-aid current account6.8–3.42.51.07.30.1
Change in non-aid capital account2–6.64.7–2.0–3.51.80.4
Memorandum items (percent change)
Real GDP growth4.24.74.43.74.24.55.2
Gross reserves (millions of U.S. dollars)508.0508.0317.0264.0344.0635.01,427.0
Cedi per U.S. dollar (average)–20.1–11.4–13.3–51.1–23.9–9.6–8.6
Real effective exchange rate (percent, average)6.08.20.5–35.50.6–0.61.4
Trade balance (percent of GDP)–17.9–10.8–16.0–16.5–18.2–10.8–12.0
Exports f.o.b.0.015.6–4.1–3.5–3.610.220.1
Export volume–0.616.3–2.82.2–1.3–2.1–6.8
Nontraditional export volumes2.6–1.3–18.25.86.3–2.4–32.9
Imports f.o.b.32.4–4.712.3–15.22.6–4.120.1
Import volumes14.424.510.4–24.910.0–6.86.9
Non-oil import volumes15.823.610.3–30.38.6–7.69.0
Terms of trade–0.713.7–8.7–16.64.89.414.8
Import cover (months)2.11.91.91.11.52.34.7
Note: f.o.b. = free on board.

Consistent with Ghana saving the aid as higher reserves, there was little evidence of Dutch disease. The real effective exchange rate changed by less than 1 percent during the period and exhibited little volatility.2 In view of the absence of real appreciation, it would seem unnecessary to examine the performance of exports for Dutch disease symptoms. Nonetheless, the decline in nontraditional export volumes during the period of higher aid is puzzling (Table 4.3).

This picture for 2001–03 masks three distinct episodes that mirrored the fluctuations in aid:

  • About half of the 2001 aid jump (equivalent to 5.1 percent of GDP) remained in reserves, while the rest financed a deterioration in the capital account (indeed, more than the rest, as the non-aid current account deficit shrank).

  • The 2002 aid collapse of 5.9 percent of GDP was more than outweighed by a decline of the non-aid current account deficit.3 This and some capital inflows permitted a further reserve accumulation equivalent to 3.3 percent of GDP.

  • Finally, the 2003 aid jump (equal to 6.5 percent of GDP) and a further decline of the non-aid current account deficit (of 1 percent of GDP) fed a further reserve build-up (to 7.3 percent of GDP).4

GDP growth and the real exchange rate were remarkably steady through the entire period. The nominal exchange rate for the cedi depreciated by nearly 50 percent in 2000, but the use of aid in 2001 helped arrest the decline. Reserve accumulation after 2001 contributed to ongoing nominal depreciation, a policy choice discussed below.

Was Aid Spent?

On a cumulative basis, none of the aid Ghana received was spent in the sense of allowing a widening fiscal deficit (net of aid). This does not mean that aid money itself was not spent, but that if it was, other spending was reduced correspondingly.

Aid going to the budget followed the same pattern as aid measured through the balance of payments (Table 4.4). During 2001–02, fiscal policy was sensitive to aid flows. In 2001, the fiscal deficit widened by half the increase in aid. The next year’s large and unexpected aid shortfall triggered a fiscal consolidation that was large but still not enough to close the shortfall. However, fiscal policymakers responded cautiously to the aid surge in 2003; the fiscal balance did not react at all. Although spending rose, it was fully financed by higher nongrant revenues. On a cumulative basis, aid had little fiscal impact during 2001–03.

Table 4.4.Ghana: Net Aid Flows and the Fiscal Response(In percent of GDP)
1997199819992000200120022003
Change in financing−0.7−0.50.7−0.84.5−5.61.3
Change in net aid−1.4−0.4−0.3−3.110.9−8.16.1
Of which: unexpected0.50.0−6.14.9−4.42.3
Change in domestic financing10.7−0.11.02.3−6.42.5−4.8
Change in balance (before grants)1.00.90.4−0.2−4.66.5−0.1
Change in expenditure−1.30.2−2.41.55.0−6.62.9
Change in revenue excluding grants−0.31.1−2.01.30.4−0.12.8
Memorandum items:
Revenue and grants19.220.518.019.825.021.125.5
Revenue17.318.416.417.718.118.020.8
Grants1.92.21.72.16.93.14.7
Expenditure28.428.626.227.732.726.129.0
Recurrent expenditures (excluding interest)210.310.310.811.112.113.813.8
Wages and salaries5.35.55.65.26.18.58.4
Capital expenditure11.611.39.89.212.86.18.9
Poverty expenditure4.54.86.5
Overall balance−9.2−8.1−8.2−7.9−7.7−5.0−3.5
Excluding grants−11.1−10.2−9.8−10.0−14.6−8.1−8.2
Debt service (percent of revenue excluding grants)37.638.034.142.443.133.929.8
Debt service (percent of exports)324.725.021.719.422.119.919.8
Interest payments6.57.05.67.57.86.16.2

The volatility of the aid flows seems to have complicated expenditure management and perhaps undermined the efficiency of spending (Figure 4.1). The increase in spending associated with the 2001 aid jump came mainly in the form of higher public capital expenditure, but also recurrent expenditure (Table 4.4). When aid declined by more than 8 percent of GDP in 2002, capital expenditure fell by nearly 7 percentage points of GDP, as current expenditures continued to rise. In 2003, capital expenditures rose with higher domestic revenues, but to levels well below the average of the four pre-aid boom years.

Figure 4.1.Ghana: Aid Flows and Public Expenditure Patterns

(In percent of GDP)

The volatile but increasing trend in net aid did not lead to any apparent reduction in Ghana’s revenue efforts. Revenues excluding grants have increased steadily as a share of GDP since 1999.

Why did Ghana not spend the aid on a cumulative basis? Three motivations seem to have been at work:

  • A desire to resolve underlying fiscal problems and achieve disinflation;

  • IMF conditionality on fiscal policy and its interaction with volatility; and

  • An underlying concern for the implications of aid volatility.

Throughout 2001–03, the IMF-supported program targets implied that a large part of the aid increment—between 2 and 4 percentage points of GDP—was not to be spent, in order to reduce the large stock of domestic public debt, high domestic interest rates, and the resulting large share of interest payments in expenditures.

Given the (unpredicted) volatility of aid to Ghana, the interaction between the surprise component of aid flows and the fiscal performance criteria is an important part of the story. The criteria were subject to asymmetric adjustors to account for deviations between expected and realized aid flows. Positive aid shocks were to be saved and were not to be used to increase spending. Negative aid shocks were to be partially dealt with through a reduction in spending and, hence, a narrower deficit (before grants) (Box 4.1).

As noted earlier, roughly half of each aid jump during 2001–03 was unexpected. In 2001, the increase in domestic credit roughly matched the expected component of the aid inflow. Correspondingly, the actual size of the primary deficit and domestic financing of the budget were close to, though somewhat below, the adjusted program targets. This small degree of outperformance implies that the program was almost binding; there was little scope to boost spending and this led the authorities to save almost all of the aid surprise. Figure 4.2 illustrates this point for domestic financing, which implies the same constraints applied to the fiscal deficit.

Figure 4.2.Ghana: Performance Against Domestic Financing Targets

(In billions of cedis)

Note: Shaded areas: positive aid surprise; nonshaded areas: negative aid surprise.

In 2002, Ghana’s IMF-supported program again called for a large reduction in domestic financing. This was because of the need to begin reducing the large stock of domestic debt—which had reached 20 percent of GDP—and because of high real interest rates. But the substantial fiscal consolidation (equivalent to 6.5 percent of GDP) did not fully compensate for the huge aid shortfall. The adjustors allowed for some rise in domestic financing, but the program’s targets were still breached.

In 2003, however, even the expected component of the surge was not spent. In this case, the targets were clearly not a binding constraint on fiscal policy. This is illustrated by the large degree of outperformance against the domestic borrowing criteria in Figure 4.1, indicating that the authorities had room to increase spending under the program. Such increases in government expenditure as were observed were largely financed by terms-of-trade-related increases in tax revenues.

Ghana’s choice not to spend the aid surge of 2003 requires further explanation. One obvious inference is that the largely unexpected aid volatility of the previous few years dictated caution, particularly in view of the impact of aid and, hence, fiscal volatility on capital expenditures.

Ghana’s fiscal caution in 2003 is consistent with its policy on reserve accumulation. The authorities saved the entire 2003 aid jump (and more) in reserves. Spending the aid in the face of reserve accumulation would have been equivalent to a domestically-financed fiscal expansion. Thus, the reserve accumulation policy made the fiscal savings more advisable. The reserve accumulation itself may have been driven partly by fiscal policymakers’ desire to save the aid jump. In any case, the two policies together served to save the aid inflows.

Finally, the lack of fiscal expansion in 2003 despite the surge in aid afforded the authorities scope to use the aid inflows to stabilize the exchange rate and inflation, which had increased sharply in late 2002 and early 2003. In order to understand this better, we now turn to Ghana’s monetary policy response to the aid inflows.

Monetary Policy Response

The pattern of absorption and spending in Ghana—essentially the don’t-absorb-and-don’t-spend strategy on a cumulative basis—implies that the aid flows in and of themselves had no overall impact on Ghana’s monetary policy. In effect, the aid left the country in the form of higher reserves, and the government compensated for any aid-related expenditure increase by cutting spending elsewhere. This contrasts with the strategy outlined in Ghana’s PRGF-supported program, which was to mostly absorb and partly spend the expected aid increments in order to reduce the burden of domestic debt on the economy.

Again, this big picture masks some interesting year-by-year variations. Ghana experienced two major surges of inflation during the period—in 2000–01 and again in 2002–03. Each followed sharp and unexpected aid declines, and in the first case a major terms-of-trade decline as well. Each was also accompanied by a loosening of domestic monetary and fiscal policy. And each was followed by an aid surge. In the first case, the authorities used aid to help stabilize the economy; in the second, they did not.

The first aid surge came when the Ghanaian economy was still reeling from a terms-of-trade decline of 25 percent during 1999–2000 (Box 4.2).5 In response to this shock, and with foreign reserves equal to one month of imports at the end of 2000, the authorities failed to tighten fiscal policy sufficiently; rather, they increased domestic borrowing and ran up domestic arrears to plug the shortfall. Reserve money growth took off and the result was a 50 percent loss in the currency’s nominal effective value and a sharp rise in inflation, which rose to an annual average of about 40 percent (Figure 4.3).

Figure 4.3.Ghana: Exchange Rates, Inflation, and Aid Flows

(In percent annual change)

Note: NEER = nominal effective exchange rate.

The aid surge in 2001 was partly sold into the foreign exchange market, strengthening the cedi and helping reduce money growth. Given the worsening capital account, it is likely that in the absence of the aid surge the nominal exchange rate declines in 2001 would have been much larger than the modest 5 percent experienced.

Box 4.1.Ghana: Program Adjustors for Aid Surprises

The criteria set out in the program supported by the Poverty Reduction and Growth Facility for Ghana were subject to asymmetric adjustors to account for deviations between expected and realized aid flows. Positive aid shocks were to be saved in higher net international reserves (NIR) and lower net domestic assets (NDA), while negative aid shocks were to be partially dealt with through a reduction in NIR and some increase in NDA. Some cushion was built in, so that the adjustment was some proportion of the aid shortfall. But if the criteria were binding, shortfalls still implied a tightening of policy.

IMF-Supported Program Aid Flow: Relevant Performance Criteria
Positive Aid Shock (actual higher than programmed)Negative Aid Shock (actual lower than programmed)
Performance criteria
Net domestic financing of government (ceiling)Lowered by full amount of the aid shockRaised by some proportion of the aid shock1
Net domestic assets of the central bank (ceiling)2Lowered by full amount of the aid shockRaised by some proportion of the aid shock1
Net international reserves (floor)Raised by full amount of the aid shockLowered by some proportion of the aid shock3
Indicative targets
Domestic primary fiscal balance (floor)4No effectImplicitly raised by some proportion of the aid shock
Reserve money (stock)2No effectImplicitly lowered by some proportion of the aid shock

In 2001, IMF staff supported the Ghanaian monetary authorities’ strategy, advising that aid flows be used to support the currency as a way to curb inflation. The IMF-supported program specified a floor for the accumulation of net international reserves (NIR) and a ceiling for central bank net domestic assets (NDA). Both targets were generally met in 2001, except that NDA targets were waived as money demand increased more than what was forecast, while inflation eased.

As the aid shortfall hit in 2002, monetary policy was eased. The fiscal contraction was less than the aid decline, and the government borrowed directly from the central bank to partly cover the resulting financing gap. There was little sterilization of these liquidity injections during this period, either domestically or in terms of foreign exchange, as reserves continued to be accumulated. In part, reserve accumulation was programmed, because import cover at the start of 2002 remained fairly low (at just 2.3 months). The nominal exchange rate depreciation accelerated and inflation began to pick up by the end of 2002.6

The monetary and fiscal targets under the IMF-supported program were overshot by the end of 2002. These slippages prevented the IMF’s Executive Board from completing the final review of the 2002 PRGF-supported program.7 Reserve floors under the program were exceeded, however. Thus, the program left substantial room to sell reserves; indeed, reserves ended the year $80 million above the adjusted program floor. Selling some reserves would have reduced monetary expansion without requiring further contraction of domestic credit by the central bank.8

In 2003, the authorities faced a need to stabilize. One approach might have been to sell some reserves to reduce money growth and stabilize the exchange rate. The authorities in fact steadily reduced the rate of nominal exchange rate depreciation, contributing to the inflation stabilization. However, the authorities could have made more aggressive use of foreign exchange sales. Instead, as discussed above, the authorities continued to accumulate more reserves. One motivation, emphasized earlier, was the need to create a buffer for volatile aid flows. A desire to keep the real exchange rate from appreciating may have also played a role, however, as the authorities actually bought reserves in the foreign exchange market and accumulated more than would be implied by the aid jump.

To reduce inflation, the authorities conducted policy through domestic monetary operations, selling treasury bills for local currency and increasing reserve requirements for domestic banks to reduce money supply growth and raise interest rates. In the event, inflation fell to less than 5 percent on a six-month annualized basis by the end of 2003.

IMF staff urged the authorities to avoid further reserve accumulation and once again use the aid to allow the currency to adjust, in order to slow the pace of monetary expansion, restrain inflation expectations, and avoid the crowding-out effects of domestic sterilization. Reserve accumulation exceeded the program floor by $310 million, approximately $200 million more than the positive aid surprise. Despite this, high-powered money was just 7 percent above its indicative target as a result of the aggressive domestic monetary tightening.

The potential costs of domestic sterilization were twofold: the possible effect on private investment and the quasi-fiscal costs of the higher domestic debt level. Excluding the early part of the year, when petrol prices affected inflation, real interest rates in Ghana tended to rise from already high levels in 2003 (Figure 4.4). Private investment remained stable, at about 14 percent of GDP. Meanwhile, domestic debt remained high, at 20 percent of GDP, and with interest rates high, domestic debt service continued to absorb 5 percent of GDP, or 17 percent of total public expenditure.

Figure 4.4.Ghana: Real Treasury Bill Rates Using Annualized Quarterly Inflation

(In percent)

Note: Shaded area denotes the aid-surge period.

One aim of policy over this period was to reduce the domestic debt burden in order to lower real rates, stimulate private investment, and reduce the burden of domestic interest payments on the budget. A number of factors supported higher private investment after 2001, including improving terms of trade, recovering GDP growth, and fiscal consolidation.9 It is possible that the domestic sterilization policy may have kept real interest rates higher, the domestic debt burden higher, and private investment lower than otherwise would have been the case (Table 4.5). While some of this may have been appropriate in view of the need to reduce inflation, a more aggressive use of aid—via sales of foreign exchange—may have mitigated some of these problems.

Table 4.5.Ghana: Monetary Conditions(In millions of cedis, except where indicated)
1997199819992000200120022003
Change in reserve money3013155841,1451,3332,1161,882
Change in net foreign assets111294−550−9301,4051,7526,115
Change in net domestic assets189211,1332,076−72364−4,233
Memorandum items:
Net aid flows (millions of U.S. dollars)243236221−13563157566
Change in gross reserves (millions of U.S. dollars)−910−62−18280291792
Percent change in nominal effective exchange rate (average)−16−5−21−49−5−18−16
Percent change in real effective exchange rate (average)681−361−11
Inflation, percent change (average)18161339231424
Treasury bills (average nominal rate)43273041332622
Treasury bills (average real rate)251117210122
Reserve money (percent change)33263855414628
Broad money (percent change)44172554324934
Velocity5444443
Private investment (percent of GDP)12121315141414
Private investment (percent of investment)49525862526960

Box 4.2.Ghana: Terms-of-Trade Shocks and Aid Inflows

The table below illustrates the relative importance of net aid flows, private flows, and the trade balance for overall foreign exchange flows to Ghana between 1998 and 2003. Three commodities dominate the terms of trade: gold and cocoa exports, and oil imports. Price volatility in these commodities can greatly influence the flow of foreign exchange into the economy.

In 1996, two commodities, gold and cocoa, accounted for roughly equal shares of 74 percent of exports; by 2003, this had declined modestly to roughly equal shares of 64 percent (i.e., one-third each). Price independence provides some diversification (the correlation of annual price changes of the two commodities since 1996 is not significantly different from zero), but overall exposure to prices remains high. Between 10 and 20 percent of imports are accounted for by oil.

In 2000, Ghana suffered a large negative terms-of-trade shock that amplified the impact of the aid decline. During 2001–03, terms-of-trade effects were not as large as aid movements but were still important. In 2002, they offset some of the decline, and in 2003 they moved in the same direction as the large jump in aid, driven largely by export price increases.

Net Aid and Terms-of-Trade Effects(In millions of U.S. dollars)
199819992000200120022003
Exports (higher export values +)282−86−70−69190414
Of which: price effects−11−28−112−44234594
Imports (lower import values +)144−355493−72117−545
Of which: price effects714−49−422185−81−335
Trade balance (narrower deficit +)426−441423−141307−131
Of which: terms-of-trade effect703−77−534141153259
Net aid (higher net aid flows +)−7−15−233576−406409
Private capital (higher inflows +)−350405−30513210285
Total69−51−115567−89563
Memorandum items:
Change in gross reserves (millions of U.S. dollars)−62−62−18280291792
Terms of trade (annual percent change)14−9−175915

Conclusions

Ghana avoided real exchange rate appreciation and Dutch disease during a three-year period when aid flows were 4.4 percentage points of GDP higher than during the previous five years. This cannot be attributed to offsetting exogenous forces, as Ghana’s terms of trade improved and private inflows rose significantly during the same period. Rather, the fact that Ghana effectively saved all of the aid and prevented it from having a cumulative impact on the economy explains the lack of a real exchange rate effect.

The aid inflows and Ghana’s policy response differed in each of the three years:

  • In 2001, aid inflows jumped unexpectedly. Fiscal policy was restrained, in keeping with the IMF-supported program, which did not allow spending of aid surprises. Much of the extra aid was absorbed in that it was sold in the foreign exchange market, putting this episode in the category of aid mostly absorbed but not spent. The sale of foreign exchange and associated policies allowed the exchange rate to stabilize and inflation to fall.

  • In 2002, aid inflows fell just as unexpectedly. Fiscal policy contracted, but not proportionally, leaving a net increase in domestic financing of the deficit. Reserves were further accumulated, in excess of the actual aid flows. The combined impact of the less-than-full fiscal adjustment and the accumulation of reserves was a loose monetary policy that contributed to an increase in inflation by the end of 2002.

  • In 2003, aid again surged unexpectedly. This time, the authorities more than accumulated the jump in reserves and avoided any increase in the fiscal deficit before grants. Aid was neither absorbed nor spent. Inflation was stabilized, in part through the sale of government paper that contracted the money supply, which contributed to high interest rates and a large stock of domestic public debt. The authorities did not take the opportunity to stabilize through more aggressive foreign exchange sales, a more appreciated real exchange rate, and a reduction in the level of domestic public debt.

A number of conclusions emerge. First, why did policymakers limit the impact of aid on the economy? The answer does not seem to lie in the strictures of the IMF-supported program. Most notably, the lack of fiscal and exchange rate reaction to the 2003 aid inflow expansion cannot be ascribed to program fiscal and reserve accumulation targets.

Part of the reason for the reserve accumulation was presumably Ghana’s desire to rebuild reserves from low levels. And part of the reason for the lack of proportionate increase in the fiscal deficit was a desire to “crowd in,” that is, to reduce the large stock of domestic public debt that had led to high real interest rates and incurred large interest costs for the budget.

However, these explanations are perhaps not the full story. The clearest indication of this is that IMF floors for reserve accumulation and ceilings on domestic financing of the deficit were substantially exceeded in 2003. In addition to the above motivations, the policy may have reflected a desire to avoid real exchange rate appreciation. However, there is little direct evidence on this point.

Aid volatility—particularly unexpected and large swings—has contributed to Ghana’s policy difficulties and, most likely, to the authorities’ caution in spending the 2003 aid surge. The volatility of aid carried other costs as well. It is surely more difficult to revert to a given level of expenditure after an increase than to simply maintain the current level of spending. In this sense, the volatility in 2001–02 contributed to the fiscal, and then monetary, policy relaxation, and to high inflation in 2002. Moreover, recurrent expenditures appear to have been harder to restrain than capital spending during the 2002 downturn. This implies that spending the aid as it arrives could help reduce the share of capital expenditures over time. Recent improvements in donor coordination in Ghana should help limit the volatility of aid flows in the future.10

IMF-supported program targets, with their asymmetric adjustors, may significantly affect how aid surprises are absorbed. The system of adjustors in Ghana was, in part, designed to reduce the risks of excessive public domestic borrowing and facilitate crowding in. With these objectives in mind, although the asymmetric adjustors may have appeared excessively tight ex ante, in the event, they turned out to be useful. Had the aid jump of 2001 been fully spent, the collapse of 2002 would have been even harder to manage.11 Had the effect of the aid collapse on policy been fully smoothed by domestic borrowing, the public debt stock would have risen from already high levels, while the emerging inflation problem of 2003 would have been worse.

An earlier version of this chapter appears in Isard and others (2006).

The real exchange rate versus the U.S. dollar appreciated by roughly 10 percent, mainly when the dollar began to broadly depreciate.

This was driven by a $307 million narrowing in the trade balance, plus an $80 million rise in private transfers, largely made up of remittances.

In one sense, some of the aid could be said to have been absorbed, with the terms-of-trade-related increase in export proceeds going into reserves. However, the terms-of-trade effect simply means that it would have taken a larger real exchange rate appreciation to absorb the aid than otherwise.

Compounding the effect of the terms-of-trade shock was a decline in net aid inflows of 0.3 percent of GDP, as against IMF staff expectations of a 5.8 percent of GDP increase in aid flows in 2000 (Table 4.1).

In early 2003, inflation pressures were exacerbated by one-off petroleum price hikes linked to the removal of subsidies.

It would seem unlikely that this event contributed much to the shortfall in aid, which had already mostly emerged by then.

The monetary performance criterion was changed from reserve money to net domestic assets in June 2001. Reserve money became an indicative target.

Private investment is only an estimate, constructed by IMF staff, and therefore any inference about its level or change needs to be treated with appropriate caution. Ghana’s Statistics Service does not produce expenditure components of GDP.

The government takes the overall lead in coordinating external assistance and has introduced a “mini-Consultative Group” process, in which it meets with external partners on a quarterly basis. The World Bank has stepped back from its traditional role to leave scope for the government to lead external partner coordination.

The assumption is that the aid collapse of 2002 was not a result of donor or recipient dissatisfaction with the partial failure to spend the aid in 2001.

    Other Resources Citing This Publication