III Ethiopia

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

The Derg regime, a political dictatorship employing socialist central planning, held sway in Ethiopia from 1974 until its overthrow in 1991. After the totalitarian regime fell, the Ethiopian authorities implemented an economic reform program to revive the war-torn economy, and to liberalize many of the central planning policies of the previous government. The authorities made great strides in consolidating the gains of these reforms in subsequent years, with generally favorable results for the economy.

With the new transitional government in place, the country then experienced a surge in aid inflows as it undertook a wide range of structural reforms. These aid flows decreased from 1997 to 2000, when Ethiopia and Eritrea engaged in a violent border conflict. The country’s Enhanced Structural Adjustment Facility (ESAF) program with the International Monetary Fund ended in October 1999, with the full amount of programmed funds undrawn. The border conflict was partly resolved in 2000. Shortly thereafter, in March 2001, Ethiopia began a Poverty Reduction and Growth Facility (PRGF) program that accompanied a second period of high aid inflows. It is this second period of increased aid inflows, from 2001 to 2003, that is the subject of this case study.1

Since 1992, the country has had nearly consecutive Structural Adjustment Facility (SAF), ESAF, and PRGF programs. Under the most recent PRGF program, Ethiopia achieved a fairly stable macroeconomic position, although growth remains heavily reliant on the success or failure of the annual rains and the agricultural sector.2 Growth in GDP per capita has averaged 1.4 percent since 2001, even withstanding the country’s recession in 2002 and 2003 due to a severe drought.3 These growth rates, while inadequate, compare favorably to the average GDP per capita growth rate of only 0.04 percent over the previous 10 years.

Ethiopia remains one of the poorest countries in the world, with a per capita income of around $150. The country’s 2002 Poverty Reduction Strategy Paper reported a poverty head count of 45.5 percent in 1996, and of 44.2 percent in 2000. The country rates near the bottom of the Human Development Index, at 170th out of 177 countries. As the second most populous country in Africa, with approximately 67 million people, Ethiopia plays a critical role in determining Africa’s success in reducing poverty and meeting the Millennium Development Goals (MDGs).

The next section examines the patterns of aid inflows during the study period more closely. To put the subsequent discussions of policy choices in context, this chapter then reviews the behavior of exchange rates and prices, and examines the potential influence of the terms of trade. The core of the chapter is the discussion of the overall pattern of absorption and the fiscal policy response. The final section breaks down the effects of the aid surge on a year-by-year basis in order to understand the authorities’ monetary policy responses.

Pattern of Aid Inflows

The year 2001 marked the beginning of the turning point for Ethiopia, with aid flows, (defined as net public inflows) increasing to 8.6 percent of GDP, up from 5 percent of GDP the previous year. The U.S. dollar amount of net public inflows increased by 47 percent from 2000 to 2001. In 2002, aid inflows increased even more dramatically, to 16.1 percent of GDP, an increase of 70 percent over U.S. dollar amounts in 2001. In 2003, aid inflows leveled off, though remaining high, at levels near 15 percent of GDP.

Ethiopia received its largest surge of gross public aid in 1997, with levels of nearly 75 percent of GDP. However, most of this assistance was used for arrears clearance, resulting in net public aid inflows of only 2 percent of GDP that year. Ethiopia received a rescheduling of debt from the Paris Club in 1998 as well as debt relief under the enhanced Heavily Indebted Poor Countries (HIPC) Initiative.4

Aid inflows are generally difficult to predict, and Ethiopia is no exception. When actual aid flows through the budget are compared with IMF aid forecasts, it can be seen that the initial surge in aid inflows, in 2001 was far less than predicted (Figure 3.1). In 2002 and 2003, predicted flows were largely on target with actual outcomes.

Figure 3.1.Ethiopia: Programmed vs. Actual Levels of Net Budget Aid

(In percent of GDP)

Note: PRGF = Poverty Reduction and Growth Facility.

Net private inflows to Ethiopia are an important factor in the economy as well. They have been relatively stable, however, averaging 7 percent of GDP in the most recent three years, compared to an average of 6 percent between 1996–2000, thus allowing a focus on the macroeconomic impact of public inflows (Table 3.1).

Table 3.1.Ethiopia: Aid and Other Inflows(Percent of GDP)
Net public inflows1.
Net private inflows5.
Gross public inflows74.613.111.78.824.318.117.5
Debt relief168.
Change in arrears2––
Memorandum item:
GDP (real annual percent change)5.1––3.9
Note: Figures in bold represent the aid-surge period.

Includes a Paris Club rescheduling agreement in 1997/98 and traditional and Heavily Indebted Poor Countries (HIPC) Initiative debt relief.

Negative values indicate a clearance of arrears.

Note: Figures in bold represent the aid-surge period.

Includes a Paris Club rescheduling agreement in 1997/98 and traditional and Heavily Indebted Poor Countries (HIPC) Initiative debt relief.

Negative values indicate a clearance of arrears.

Exchange Rate, Prices, and Terms of Trade

The Ethiopian authorities maintained a close crawling peg to the U.S. dollar over the course of the aid surge. The real effective exchange rate depreciated slightly over the study period, as did the nominal effective exchange rate and the nominal exchange rate against the U.S. dollar (Figure 3.2).

Figure 3.2.Ethiopia: Exchange Rate and Price Developments

Sources: IMF staff estimates and country authorities.

Note: REER = real effective exchange rate; NEER = nominal effective exchange rate; CPI = consumer price index; PRGF = Poverty Reduction and Growth Facility.

One possible explanation for the stability of the exchange rate despite the aid surge is a decline in food prices of approximately 10 percent in 2001 and 13 percent in 2002. The food surplus experienced in these years was due to a bumper harvest and large amounts of food aid.

A second potential explanatory factor is that Ethiopia’s terms of trade fell sharply during the sample period, due to the collapse in coffee prices, the country’s primary export (Table 3.2). Nonetheless, an analysis of the relative impact of both the aid inflows and the declining terms of trade shows that, in the first two years of surging aid inflows, the dollar gain from net aid inflows far outweighed the loss due to declining coffee exports (Table 3.3). The fact that 2001 and 2002 still saw a depreciation in the real exchange rate, despite the overall positive net impact of aid and trade, suggests that the Ethiopian authorities may have pursued policies to avoid an appreciation of the real exchange rate.

Table 3.2.Ethiopia: Trade Indicators
Terms of trade index100.
Coffee price (U.S. dollars per thousand metric tons)
Annual change in coffee value (millions of U.S. dollars)–138.8–19.2–80.0–18.82.0
Annual change in coffee volume (millions of U.S. dollars)–18.815.4–21.014.715.8
Total imports, c.i.f. (millions of U.S. dollars)1,558.01,611.01,557.01,696.01,856.0
As percent of GDP24.124.723.927.927.9
Total exports f.o.b. (millions of U.S. dollars)484.0486.0463.0452.0483.0
As percent of GDP7.
Noncoffee exports (millions of U.S. dollars)203.0224.0281.0289.0318.0
As percent of GDP3.
As percent of total exports41.946.160.763.965.8
Exports nonfactor services (millions of U.S. dollars)430.0498.0516.0530.0657.0
Percent change–
Notes: c.i.f. = cost, insurance, and freight; f.o.b. = free on board. Figures in bold represent the aid-surge period. To estimate the dollar size of the terms-of-trade shock, exports and imports are calculated based on what they would have been had prices remained what they were in the previous year.
Notes: c.i.f. = cost, insurance, and freight; f.o.b. = free on board. Figures in bold represent the aid-surge period. To estimate the dollar size of the terms-of-trade shock, exports and imports are calculated based on what they would have been had prices remained what they were in the previous year.
Table 3.3.Ethiopia: Terms-of-Trade Shocks and Aid Inflows(In millions of U.S. dollars)
Trade Balance
Effect of terms-of-trade shock3–54–484–13–36–59
Aid effect–152423541721
Net effect–69–460222381–38
Note: Figures in bold represent the aid-surge period.

Calculated as the value of exports (imports) from the previous year, multiplied by the annual percentage change of the export (import) volume index for the current year.

Calculated as the difference between the counterfactual exports and the counterfactual imports.

Calculated as the difference between the actual trade balance and the counterfactual trade balance.

Note: Figures in bold represent the aid-surge period.

Calculated as the value of exports (imports) from the previous year, multiplied by the annual percentage change of the export (import) volume index for the current year.

Calculated as the difference between the counterfactual exports and the counterfactual imports.

Calculated as the difference between the actual trade balance and the counterfactual trade balance.

On average, across the 2001–03 aid-surge period, Ethiopia absorbed only about one-fifth of the 8 percentage point surge in aid flows (Table 3.4). Most of the increment in aid was channeled into increasing official reserves, which rose from 5.2 percent of GDP in 2001 to 14 percent by 2003. The import cover of gross official reserves also improved substantially over the aid-surge period, from 2 to 3.8 months of imports.

Table 3.4.Ethiopia: Was Aid Absorbed?(Annual averages in percent of GDP)
Pre-Aid-Surge Average, 1999–2000Aid-Surge Average, 2001–03DifferenceIncremental Aid Absorbed?
Net aid inflows5.313.38.0
Non-aid current account balance–9.2–10.8–1.6
Non-aid capital account balance2.01.3–0.7
Change in reserves (increase –)1.9–3.8–5.7Partly absorbed

Spending Out of Aid

The average fiscal response to the inflows during the three-year period of high aid inflows from 2001 to 2003 was very cautious. Ethiopia saw an increase in aid to the budget of approximately 6 percent of GDP during the aid surge.5 The average overall fiscal balance (before aid) shows a small improvement, indicating that overall, the government did not spend any increment of the increased aid (Table 3.5). However, these averages mask significant year-to-year variations, which are taken up in the next section.6

Table 3.5.Ethiopia: Was Aid Spent?(Annual averages in percent of GDP)
Pre-Aid-Surge Average, 1999–2000Aid-Surge Average, 2001–03DifferenceAid Spent or Not?
Net fiscal aid inflows5.311.25.9
Revenue (excluding grants)
Expenditure (excluding external interest)31.832.50.7
Overall fiscal balance before aid–13.8–13.00.8Not spent

The composition of government expenditure showed major improvements over the period in question. Before the aid surge, expenditure was dominated by the war with Eritrea, and defense expenditure reached nearly 13 percent of GDP in 2000. In 2001, the war subsided, and there followed a sharp decline in military defense expenditure. Total social spending (both current and capital), which remained low during the war years, has increased steadily since 2001.

Revenue collection in Ethiopia remained fairly steady, increasing by slightly more than 1 percentage point of GDP. During the aid-surge period, revenue collection fell short of PRGF projections on average by 1.7 percent of GDP. This shortfall is partly explained by the drought years of 2002 and 2003, when tax and nontax revenue collection was lower than expected, largely due to the heavy reliance on the agricultural sector. Overall, the differences between revenue collection before and during the aid-surge period are marginal, and there is no clear evidence that higher volumes of aid inflows reduced the country’s revenue collection efforts.

Monetary Policy Response

Reflecting the overall pattern of absorption and spending, the 2001–03 aid surge presented no major monetary policy challenges on a cumulative basis. By largely saving the aid in the form of higher reserves and reduced fiscal deficits (after aid), the authorities avoided real exchange rate appreciation and the need to manage aid-related increases in liquidity. Thus, on the whole, Ethiopia fits into the don’t-absorb-don’t-spend category during the period covered. However, this overall picture hides considerable and economically meaningful variation over different subperiods of the aid surge.

Monetary policy was anchored by a close, crawling peg to the U.S. dollar over the period of high aid inflows. The authorities maintained a crawling peg at about 7 percent depreciation annually between 1998 and 2000. With the surge in aid inflows, the authorities moved to a much slower rate of crawl starting in 2001, with an annual rate of depreciation of about 1.6 percent between 2001 and 2003 (Figure 3.3).

Figure 3.3.Ethiopia: Exchange Rate and Treasury Bill Rates

Sources: IMF staff estimates and country authorities.

Note: REER = real effective exchange rate, 1990 = 100; PRGF = Poverty Reduction and Growth Facility.

The capital account in Ethiopia is largely closed, leaving some scope for monetary policy even given the peg. The Ethiopian banking sector remains underdeveloped, dominated by the National Bank of Ethiopia and the Commercial Bank of Ethiopia. Interest rates remain low and controlled. Moreover, throughout the study period, the banking system contained excess liquidity. All this makes monetary policy somewhat difficult to observe.

Before examining monetary policy during the aid surge, it is useful to investigate the period prior to the aid surge, which was characterized by loose monetary policy, downward pressure on the exchange rate, and a loss of reserves. The 1998–2000 war with Eritrea presented a series of challenges, in particular, escalating defense expenditure. The authorities engaged in a fiscal expansion of 7 percent of GDP, well above the PRGF indicative target for expenditure. In the context of the crawling peg exchange rate, one result was an increase in the current account deficit and a collapse of reserves (Figure 3.4). The fiscal expansion, loose monetary policy, and exchange rate depreciation also began to push up nonfood inflation, which increased from levels near zero in 1999 to 5 percent by December 2000.7

Figure 3.4.Ethiopia: Gross Official Reserves

Source: IMF staff reports.

Note: e = estimate.

During the aid surge, monetary policy was largely driven by how the country addressed the question of spending the aid inflow. Table 3.6 shows the change in the annual averages of the balance of payments and fiscal identity presented previously as period averages. The aid-surge period can be broadly divided into two: 2001: Fiscal contraction in the first year of the aid surge, when aid inflows were saved while none of the aid was absorbed; and 2002–03: The late aid-surge period, when spending of aid inflows exceeded absorption.

Table 3.6.Ethiopia: Year-on-Year Changes in Absorption and Spending of Aid Inflows(Change in annual averages in percent of GDP)
Balance of payments identity
Net balance of payment aid inflows––1.1
Non-aid current account balance–5.71.5–0.2–3.40.3
Non-aid capital account balance5.7––0.5
Change in reserves (increase –)0.23.1–3.3–6.61.3
Fiscal identity
Net fiscal aid inflows2.4––0.1
Revenue excluding grants0.0––0.5
Expenditure excluding external interest6.61.3–
Change in overall balance before aid–6.6–1.43.9–2.5–2.2
Note: Figures in bold represent the aid-surge period.
Note: Figures in bold represent the aid-surge period.

2001: No Absorption or Spending: Stabilization via Fiscal Contraction

Policy changed substantially in 2001 for two main reasons. First, reduced tensions with Eritrea allowed for reduced military spending and an improvement in the overall fiscal balance. Second, aid inflows surged. The authorities seized this opportunity to undertake a sharp fiscal contraction and increase international reserves. By reducing the fiscal deficit net of aid, even as large amounts of aid began to flow into the economy, the authorities were able to accumulate foreign reserves while eliminating the depreciating pressures on the pegged exchange rate resulting from previous monetization of the fiscal deficit.

In the first year of high aid inflows, Ethiopia cut spending, and so did not spend any of the aid surge through the budget. The government absorbed a negligible portion of total aid inflows, putting most of the aid into international reserves.

The fiscal contraction, of some 4 percent of GDP before aid, allowed a stabilization of the macroeconomic situation. The year 2001 was characterized by a sharp fall in net domestic assets and slowly rising net foreign assets, which reversed their decline of the previous period (Table 3.7). The supply of reserve money decreased as spending was cut. The authorities also were able to retire some domestic debt in 2001 as a part of the stabilization program (Table 3.8).

Table 3.7.Ethiopia: Selected Monetary Indicators
Reserve money
Percent change (12-month)65.438.7–27.0–9.020.734.827.515.1
Percent change (6-month)–4.473.0–19.8–
Velocity (GDP/broad money)
Excess reserves (percent deposits)6.324.
Broad money
Percent change (12-month)13.715.913.
Percent change (6-month)–
Note: Figures in bold represent the aid-surge period.
Note: Figures in bold represent the aid-surge period.
Table 3.8.Ethiopia: Domestic Debt and Debt Service Indicators(Percent of GDP)
Domestic debt31.232.136.639.037.741.438.2
Interest payments on domestic debt8.
Note: Figures in bold represent the aid-surge period.
Note: Figures in bold represent the aid-surge period.

The increased level of official reserves held by the central bank may also have had a positive effect on expectations, thereby indirectly boosting the authorities’ efforts to stabilize the depreciating currency.

Given that Ethiopia was beginning the PRGF-supported program from high levels of spending and such a depleted base of international reserves, holding reserves equivalent to only 2.2 months of imports in 2001, the program encouraged saving, and not absorbing, most of the aid inflows in 2001. In the end, the Ethiopian authorities ended up reducing the fiscal deficit by even more than targeted by the PRGF program, and put most of the increment of aid inflows into reserves, as targeted.

In a country with a liberalized financial market, one would assume that a fiscal contraction and reduced domestic financing would be reflected in falling real interest rates. However, in Ethiopia, as mentioned previously, the financial sector is not liberalized. Interest rates are largely controlled by the government. In addition, the banking system in Ethiopia consistently contained high excess liquidity, reflecting the underdeveloped financial markets in the country. These factors tend to sever the normal relationship between monetary policy and interest rates. Table 3.9 demonstrates that, in 2001, treasury bill rates rose and deposit rates remained fixed, even as Ethiopia cut domestic financing and expenditure.

Table 3.9.Ethiopia: Investment and Interest Rates(In percent)
Treasury bill rate4.
Deposit rate6.
Private investment/GDP8.910.
Note: Figures in bold represent the aid-surge period. Central bank interest rates shown as 6-month backward averages.
Note: Figures in bold represent the aid-surge period. Central bank interest rates shown as 6-month backward averages.

Perhaps surprisingly, private investment rates showed no increase during this period. One might have expected to see lower interest rates crowding in the private sector. However, the policy objectives in 2001 were to stabilize the post-war economy and increase international reserves. It is questionable whether, in the absence of large-scale banking sector reform, the government could have stimulated private sector investment. Even as real interest rates remained low and the banking sector accumulated more excess liquidity, private sector investment rates remained steady.

2002–2003: Spending Exceeds Absorption

In 2002 and 2003, net aid inflows doubled from the already increased levels of the previous year, hitting 16 percent of GDP in 2002 and 15 percent in 2003. The authorities put about half of the aid increase in reserves, bringing their reserves up to the equivalent of 3.3 months of imports. The rest of the aid increment was absorbed, with the non-aid current account deficit increasing by about 3 percent of GDP. At the same time, the authorities spent most of the second jump in aid, increasing the fiscal deficit cumulatively by about 5 percent of GDP over 2002 and 2003. Domestic debt levels began to rise again, after their brief fall in 2001, as the authorities began increasing spending in 2003, especially on poverty reduction.

The pace of the central bank’s sale of foreign exchange was largely dictated by the need to support the crawling peg exchange rate regime. In this context, the degree of absorption of the aid reflects the import component of aid-related spending and possible second-round effects from the fiscal stimulus. It does not reflect the effects of an appreciated exchange rate, as there was none: inflation matched the rate of crawl fairly closely. Because spending exceeded absorption, the fiscal expansion implied a potentially inflationary increase in the money supply (for example, by 20 percent in 2002). However, the authorities did not attempt to sterilize through domestic monetary operations in order to achieve money targets and reduce demand.

The increased money supply put some pressure on inflation, but overall, nonfood inflation remained relatively low. Nonfood inflation increased from levels near zero in December 2002 to approximately 5 percent by the start of 2004. It is somewhat surprising that inflation did not increase further still, given that the authorities maintained the currency’s nominal peg to the dollar and also allowed the money supply to increase.

There are several (possibly complementary) explanations for the moderate rise in inflation and lack of a real exchange rate appreciation, despite the fiscal and monetary expansion, which could be expected to increase the price of nontraded domestic goods and thus the real exchange rate. First, a high import content of aid could help avoid appreciation pressures. Second, the income elasticity of import demand could have been very high. Indeed, imports increased by an average of nearly 5 percent of GDP from the pre-aid-surge period (1997–2000) to the aid-surge period (2001–03). Since net aid increased by 9.3 percent of GDP over the same period, the increase in imports represents a sizable fraction of the increase in aid flows. Third, the recession and negative terms-of-trade shocks in 2002 and 2003 may have had a dampening effect on prices and hence the real exchange rate. In 2003, the negative terms-of-trade shock actually outweighed the size of aid inflows. Finally, the banking sector’s excess reserves continued to increase as a percentage of deposits, restricting the growth in broad money and thus containing pressures for inflation and a real exchange rate appreciation.

The IMF-supported program did not directly constrain the pace of absorption over 2002–03. The Ethiopian authorities accumulated more reserves than required by the program. During this period, the country authorities exceeded the program floor for net foreign assets (NFA) and net foreign reserves accumulation (Figure 3.5).

Figure 3.5.Ethiopia: Selected Program Targets and Outcomes

(In millions of birr, flow values, unless otherwise specified)

Source: IMF staff reports.

Note: NFA = net foreign assets; NDA = net domestic assets; NFR = net foreign liquid reserves.

Either fiscal policy or monetary policy would have had to have been looser in order to increase absorption further.8 Some loosening would have been possible under the program. The authorities did not reduce the fiscal deficit (after grants) by as much as expected under the PRGF program in 2002 and 2003. However, the program targeted net domestic assets (NDA) (and net central bank credit to the government), not the deficit itself. The higher-than-expected NFA accumulation kept NDA lower than expected, leaving room under the NDA targets to finance the higher-than-expected deficit.9

The program would not have allowed as much absorption as suggested by the NFA floors, however. Program ceilings on NDA would have kept absorption below these levels. Base money demand grew much faster than expected under the program. Had NFA increased only along the lines of the program floors, and had NDA ceilings been respected, money growth would have had to have been slower than was actually observed. Monetary policy would have been tighter and, given fiscal policy, would have encouraged less inflation and hence a more depreciated real exchange rate and less absorption.


Ethiopia experienced a surge in aid inflows beginning in 2001 and continuing through 2003. Overall during this period, only a small portion of the aid was absorbed into the economy, and little of the aid inflows was spent. The cautious fiscal and monetary response to the high levels of aid helped prevent a real exchange rate appreciation, and staved off any symptoms of Dutch disease. Noncoffee export growth remained strong, while the real effective exchange rate showed a slight depreciation.

The lack of Dutch disease symptoms in Ethiopia was due to a combination of factors, including a terms-of-trade shock from falling coffee prices, low inflation due to a food surplus, a fairly high import content to aid, and most importantly, the fact that, overall, very little of the aid was actually absorbed into the economy. Most of the aid inflows were used for reserve accumulation.

However, this overall picture masks two distinct policy episodes. Coming out of the war period in 2001, the authorities saved and did not absorb the aid, bringing the level of reserves up from very low levels. Because the deficit before grants did not expand (the aid was not spent), the reserve accumulation did not put pressure on domestic monetary policy. Indeed, at the same time, in order to stabilize the exchange rate and stabilize the level of domestic debt, the authorities undertook a fiscal contraction.

Aid increased again sharply in 2002–03 by a further 8 percent of GDP. In response, the authorities spent most of the additional aid, increasing the deficit before grants by a cumulative 4.7 percent of GDP. Absorption was only partial, with the non-aid current account deficit increasing by 3.1 percent of GDP.

This second episode illustrates some of the complexities in the relationship between spending and absorption. The monetary authorities took no obvious actions to reduce absorption, such as sterilizing the associated monetary expansion. The monetary authorities’ sales of foreign exchange of the aid inflows were dictated by the exchange rate peg to the U.S. dollar. It would have taken further action by the authorities to absorb more of the aid.

Without the freedom to sell foreign exchange to regulate aid absorption, the authorities were left with two choices, had they wanted to increase absorption. The first would have been to undertake a greater fiscal expansion, because higher fiscal deficits will tend to increase absorption if the associated monetary expansion is not sterilized. Alternatively, the authorities could have allowed looser monetary policy. This might have increased demand through the availability of credit to the private sector, and also increased inflation and hence appreciated the real exchange rate.

Greater absorption of the aid inflows might have been beneficial for Ethiopia. The country had accumulated a comfortable level of reserves by 2002. Ethiopia experienced a retraction in output in 2002 and 2003 due to a severe drought. Consumer demand was stagnating, and poverty levels remained high and persistent. While real annual GDP growth averaged 6.4 percent in 1999 through 2001, the growth rate fell to 1.6 percent in 2002 and to –3.9 percent in 2003. Greater spending or looser monetary policy might have helped to lessen the impact of this recession. It is likely that the resulting possibility of a higher real exchange rate through inflation and a higher national current account deficit could have been accommodated, given the higher aid inflows.

There were also reasons for caution in increasing absorption of the aid inflow. First, the wisdom of increasing aid absorption depends critically on the predictability and durability of the aid inflow. Both higher government spending and an appreciated real exchange rate create risks. If the aid flows recede after a few years, the country would then be in the position of facing difficult adjustment costs. The government would have to make painful spending cuts, and it would be costly for the export sector to recover from the period of appreciated exchange rates.

Second, to suggest that Ethiopia could have absorbed more of the aid through a greater fiscal deficit is to suggest that the country had worthy projects on which to spend the aid, as well as the capacity to implement or scale up those projects. It is possible that a more conservative approach to using the aid inflows in Ethiopia might have been appropriate if such projects did not exist or the capacity for project implementation was low.

Finally, the option of increasing absorption through looser monetary policy must also be carefully considered. The monetary authorities could have increased the money supply through purchasing bonds in open market transactions, in an attempt to stimulate demand. However, increasing the money supply to stimulate demand may be like “pushing on a string,” particularly in a country such as Ethiopia, with an undeveloped financial sector. The banking system already contained excess liquidity and there was little private sector demand for credit.

On balance, while there probably was some scope in Ethiopia to increase absorption through looser monetary policy or higher fiscal deficits, the cautious and gradual approach of the authorities was appropriate.


Ethiopia’s fiscal year begins on July 7. Throughout the chapter, the fiscal year is referred to by the last year. So for example, fiscal year 2000/01 is simply referred to as 2001.


The IMF’s ex-post assessment of its long-term engagement in Ethiopia found that fluctuations in rainfall have a substantial impact on real growth, with a change of 1 percent of rainfall leading to a change in real GDP of 0.3 percent in the next year (IMF, 2004).


The drought during 2002–03 was the most severe the country had seen since 1985, with cereal production declining by 6 percent in 2002 and by a further 26 percent in 2003.


The country reached the HIPC decision point in November 2001, and reached the enhanced HIPC completion point in April 2004.


Note the distinction between net budgetary aid, as captured by the fiscal data, and net public aid, as captured by the balance of payments data. We use data on net public aid when examining the balance of payments, and net budget aid for assessing fiscal policy. Differences exist between the two figures because not all aid to a country is necessarily channeled though the government’s budget, as a comparison of the aid data in Tables 3.4 and 3.5 shows. The IMF’s ex-post assessment of its long-term engagement in Ethiopia notes that aid flows have not been well captured by fiscal data, partly because of the need for better reporting of activities by donors (IMF, 2004).


Moreover, the finding that the Ethiopian government did not, on average, increase spending over the period needs to be qualified by noting that defense expenditure fell sharply during the aid-surge period, but total expenditure did not, thus allowing social spending to increase substantially.


Even as nonfood inflation rose, headline inflation rates actually fell to negative levels during this period. However, negative headline inflation was due to food surpluses from a bumper harvest and high amounts of food aid.


This ignores the possibility of tilting the pattern of aid-related expenditures further toward imported goods.


There was no target on base money, per se. This episode is an example of the fact that interpretation of NDA and domestic financing of the government is not straightforward. In particular, if nonabsorption of aid increases NFA, the resulting fall in NDA and domestic financing may not have the typically expected macroeconomic implications (see Chapter II).

    Other Resources Citing This Publication