## _cr08259

## Source details

**Canonical URL:** [_cr08259](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2008/_cr08259.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2008/_cr08259.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2008/_cr08259.pdf.json)

---

### Inflation in Ethiopia — recent developments and drivers
- Inflation reached about 40 percent in May 2008, far exceeding the authorities’ policy target of single-digit inflation.
- Inflation trend and characteristics:
  - Historical peak after a drought-related surge in 2003; receded to single digits; turned up in 2004 and gradually increased to about 40 percent (12-month basis) in May 2008.
  - Coincided with high economic growth; earlier inflation episodes were largely supply-shock driven (conflicts, droughts).
  - Initiated by sharp increases in nonfood prices in mid-2006 (retail petroleum adjustments and a construction boom in Addis Ababa); nonfood inflation subsided through 2007 while food inflation escalated subsequently.
  - Nonfood inflation rising again reflecting hikes in world nonfood commodity prices, especially oil.
- Contributing factors (studies cited: EDRI, 2007; FAO, 2008):
  - Increasing money supply.
  - Rising world commodity prices.
  - Continued good economic performance.
  - Housing shortages in urban areas.
  - Changes in farmers’ behavior (more uniform supply over the year due to improved credit, storage, marketing).
  - Increased local purchases by governmental food security institutions, agricultural cooperatives, and relief agencies.
- Role of inflation expectations: once inflation reaches 20–30 percent, past inflation and inflation expectations warrant special attention.

### Empirical evidence and structural change (model results)
- Two empirical approaches with limitations due to data scarcity and structural breaks:
  - Netsere (2007): SVAR (real output and CPI) using annual data 1965–98, 1965–2002, 1965–2005 — demand factors become important when recent periods included.
  - Ayalew (2007): general equilibrium model using annual data 1970–2005 with many structural/policy dummies — found supply shocks and consumer prices of major trading partners among most important determinants.
- Basic inflation model estimated for periods 1970–2001, 1970–2004, and 1970–2007 (structural-change evidence):
  - Δln Currency (t-1): 0.31 (t-value 2.9***); 0.33 (t-value 2.6***); 0.33 (t-value 2.5***)
  - Δln Broad money: 0.48 (t-value 5.9***); 0.49 (t-value 5.0***); 0.54 (t-value 5.5***)
  - Δln GDP: -0.89 (t-value -5.8***); -0.90 (t-value -4.8***); -0.72 (t-value -4.1***)
  - Δln Import price: 0.11 (t-value 2.2**); 0.12 (t-value 1.9*); 0.09 (t-value 1.4)
  - Observations: 32; 35; 38
  - Adjusted R2: 0.69; 0.51; 0.443
- Interpretation:
  - For 1970–2007 money has a more important role; GDP and import price coefficients decline.
  - Coefficient for lagged CPI generally stable — past inflation likely drives inflation expectations and gains impact when inflation is rising.
  - Estimated inflation based on earlier samples (1970–2001 and 1970–2004) for post-2002/2005 periods are far lower than actual inflation, indicating structural changes after 2004.

### External influences on prices
- Exchange rate:
  - Long remained stable vis-à-vis the U.S. dollar, moving by less than 5 percent in the last five years (prior to the report).
  - Since mid-2007 nominal effective exchange rate depreciated somewhat faster (about 10 percent annually) than in other countries in the region.
- International commodity prices and petroleum pricing:
  - Correlation between nonfood CPI and world industrial commodity prices = 70 percent (EDRI, 2007).
  - Retail petroleum prices are administered; adjustments in May and August 2006, January 2007, and January 2008. Price controls mitigated world oil-price hikes but increased implicit subsidies.
  - Nonfuel commodities represent only about 5 percent of Ethiopia’s imports (customs data).
  - From late 2007 indications that local food prices may be converging to high world prices as exports for some agricultural products increase.
- Imports:
  - Rapid import growth in recent years doubled the import share of GDP in the last ten years.
  - Food imports were significantly reduced in 2006 (after a 160 percent increase in 2005) and stayed low relative to the five-year average, dampening the import-led mitigation of food inflation.

### Monetary developments, velocity, and expectations
- Broad money growth accelerated since 2005, mainly driven by domestic credit to the public sector.
  - Significant NBE advances to government in 2005–2006 increased currency in circulation and enabled high public credit growth without crowding out private credit.
- Banking reserves and reserve requirements:
  - In late 2004 bank reserves exceeded required level by 35 percent.
  - Excess reserves reduced through greater use of treasury bills (fiscal financing), reserve requirement increases (July 2007 and April 2008), and increases in commercial bank credit to the private sector.
  - Excess reserves fell below 10 percent, close to the minimum by historical standards.
- Rising money velocity suggests inflation expectations are becoming ingrained:
  - Inflation eroded credibility of money as a store of value; real interest rates on deposits became increasingly negative, prompting shifts to consumption or alternative investments and adding price pressures.
- Overall assessment: external factors relatively limited to date; money appears to play a larger role following money growth since 2005; monetary factor alone cannot explain recent food inflation—structural factors also important.

---

### Box 3 — Public Sector Borrowing: estimation, evidence, and policy implications
- Estimation and correlation:
  - Domestic public sector borrowing estimated from annual changes in credit outstanding to the public sector from the NBE and commercial banks.
  - Domestic public sector borrowing is highly correlated with inflation.
- Historical magnitudes and disinflation episodes:
  - During previous disinflation periods (late 1980s and mid 1990s), inflation was lowered by 8–18 percentage points following reduction of public sector borrowing by 1½–4½ percent of GDP.
  - Cross-country result (Fischer et al., 2002): for high inflation countries, a 1 percentage point improvement (deterioration) in fiscal balance-to-GDP typically leads to a 4.2 percent decline (rise) in inflation, all else constant.
- Nature of spending matters:
  - Credit used for capital spending with high import content pressures international reserves more than domestic prices.
  - Credit used for recurrent outlays (including nontradables) may heighten inflation pressures.
- Monetary tightening and growth (cross-country evidence):
  - Countries that lowered money growth achieved low inflation; those with higher money growth faced higher inflation within two years.
  - Decisive disinflation tends to lead to a quick rebound in growth for countries experiencing low growth and high inflation.
  - Exchange rate: for most countries the nominal exchange rate was allowed to depreciate significantly; pace of depreciation slowed for most countries two years after inflation reached 20–40 percent.
- Selected conclusions (paragraphs 14–18):
  - 14. Ethiopia’s recent rise in inflation is unusual: it coincides with high growth rather than a negative supply shock.
  - 15. Both domestic and external factors explain recent inflation; Ethiopia-specific factors also important given inflation rates exceed those of trading partners and region.
  - 16. High money and credit growth in recent years likely played a key role; inflation expectations increasingly important. Imported goods prices have played a limited role for several reasons (small share of imported commodities, lagged domestic fuel price adjustments).
  - 17. Prices of exportable agricultural products have helped drive up domestic prices of some food items as they converge to higher world prices; this raises rural incomes but requires monitoring if domestic prices exceed international levels.
  - 18. International experience suggests significant macroeconomic tightening is essential once inflationary expectations emerge, with a proper policy mix between fiscal and monetary policies to avoid crowding out and harming long-term growth.

---

### Broad- and narrow-money velocity: cross-country evidence and implications for Ethiopia
- Sample and methodology:
  - Velocity curves constructed for 1995–2007 across IMF member countries/areas classified by 2006 per capita GNI into LICs, LMICs, HMICs, HICs.
  - Excluded hyperinflation countries (12-month inflation > 50 percent).
- Broad money (BM) velocity (1995–2007):
  - BM velocity generally declines over time across income groups.
  - Cumulative declines range from 17 to 35 percent.
  - Only 18 of 137 countries had higher BM velocity at end of the 13-year period.
  - BM velocity declines more slowly as countries become more developed.
- Narrow money (M1) velocity:
  - M1 velocity slopes downward but not strictly correlated with income.
  - Average M1 velocity higher for HMICs than LICs and LMICs; HICs have the lowest average M1 velocity.
- Key cross-sectional numbers (Broad Money Velocity in 2007, Table 2):
  - Ethiopia...3.01
  - LIC 37: 3.4 average; 6.3 maximum; 0.8 minimum
  - LMIC 43: 2.5 average; 9.3 maximum; 0.7 minimum
  - HMIC 30: 2.2 average; 12.3 maximum; 0.4 minimum
  - HIC 27: 1.4 average; 3.2 maximum; 0.3 minimum
- Drivers of velocity in LICs:
  - Opportunity cost of holding money (inflation):
    - Adaptive behavior: positive correlation between velocity and previous-period inflation.
    - Simple regression: a 1 percentage point increase in inflation increases velocity by 0.46 percent one year later.
    - Implication: on average, when inflation exceeds about 14 percent, velocity is more likely to rise than fall.
    - Probability of rising velocity, by inflation cutoff:
      - For Inflation > Cutoff: 5 → 0.37; 10 → 0.47; 15 → 0.52; 20 → 0.58; 30 → 0.70
      - For Inflation < Cutoff: 5 → 0.30; 10 → 0.28; 15 → 0.31; 20 → 0.32; 30 → 0.33
  - Return on alternative investments (T-bills):
    - Velocity increases with T-bill rates (lower money demand).
    - Ethiopia has the lowest nominal T-bill rate among all 18 LICs with data; combined with double-digit inflation, Ethiopia’s real T-bill return is almost 6.5 percentage points below the next lowest LIC.
  - Return on holding deposits:
    - Higher real deposit returns cause velocity to decline.
    - High inflation in Ethiopia significantly reduced real returns on saving deposits; only Haiti and Guinea had lower real returns among LICs.
- Ethiopia-specific dynamics:
  - BM velocity in Ethiopia relatively low compared with LIC average; before recent rise it was more in line with LMIC average.
  - Rising BM velocity in Ethiopia since 2003 contrasts with declining trends elsewhere; velocity bottomed out in 2003 and has sloped upward since.
  - M1 currency-in-circulation share increased from 34 percent in 2002 to 36 percent in 2006.
  - Ratio of saving and time deposits to nominal GDP has been declining since 2002.
  - Table 5 (04-07 averages, percent unless noted):
    - Velocity (04-07, cumulative in percentage points): Ethiopia 5.74; LICs -1.96
    - Inflation (2003-06): Ethiopia 10.50; LICs 7.89
    - T-bill (2003-06): Ethiopia 13.97; LICs -0.74
    - Saving deposit (2003-06): Ethiopia -11.90; LICs -0.17
    - Real return (2003-06): Ethiopia -9.02; LICs -1.06
    - Impact on velocity for Ethiopia: up (velocity), up (inflation), down (T-bill), up (saving deposit)
- Policy implications and recommendations:
  - Conservative money growth targets will be needed to offset the impact of rising velocity caused by Ethiopia’s high inflation.
  - A monetary program assuming declining or neutral velocity is likely to lead to excess money supply and higher inflation than targeted.
  - To achieve single-digit inflation, a much more conservative money growth taking into account rising velocity is called for.
  - Move real interest rates to positive territory to restore credibility of money as a store of value.
  - Promote healthy growth of bank deposits to aid financial intermediation and allow meaningful interest rates to allocate resources effectively.

---

### Real effective exchange rate (REER), competitiveness, and external sustainability
- Key overview and concerns:
  - Real exchange rate appreciated by 25 percent since 2004.
  - Import coverage of reserves declined from 5 months to under 2 months.
  - Appreciation largely reflects much higher inflation in Ethiopia compared with partner countries; trade-weighted nominal effective exchange rate tended to depreciate since 2002.
  - Real appreciation coincided with strong export growth but even faster import growth, widening the trade deficit.
  - Empirical results tentative due to data limitations, but suggest potential overvaluation.
- Balance of payments and reserves:
  - Deficit on goods and services running at about 20 percent of GDP.
  - Ratio of imports to GDP rose by 10 percentage points of GDP since 2000/01 to 32 percent in 2006/07.
  - Private and official transfers rose from 10 percent of GDP in 2000/01 to about 15 percent of GDP in 2006/07.
  - Since 2001/02 the capital account deteriorated mainly due to decline in net borrowing; cumulative disparity between capital and current account since 2003/04 reached 3 percent of GDP.
  - In 2006/07 overall balance of payments almost zero; reserve coverage stabilized at 2 months of imports.
- Export performance and sector specifics:
  - Coffee market share: about 0.6 percent of world coffee exports in 2003-04 rising to 0.9 percent in 2005-06.
  - Oilseeds and flowers largest export market gains: oilseeds change 2005-06 vs 2003-04 = 0.45 (percent of world exports); flowers = 0.55.
  - Flower sector: export earnings about US$65 million in 2006/07 and projected to double over the next few years.
    - Farm laborer wage reported at US$20 per month in Ethiopia versus about $60 per month in India.
    - Price of leased land about US$13 per hectare.
    - Government provision of 750 hectares of free land for investors in fruits and vegetables.
  - Wages: manual workers in Ethiopia earning about $60 a month (FDI magazine figure); ILO estimate average manufacturing worker wage in China about $190 per month.
- Infrastructure and logistics:
  - Air freight (million tonnes per km) rose by 46 percent between 2002 and 2005; air passengers rose by almost 9 percent between 2002 and 2005.
  - Road network rose by almost 5 percent per year between 2000 and 2006; World Bank road investment extended through 2012 to add about 11,000 km of federal roads and 5,500 km of regional roads (combined increase about 40 percent).
  - Telecommunications: about 20 per 1000 people currently own a cell phone; target coverage 60 per 1000 people by 2009/10.
- Agriculture cultivated area and yields (selected figures, percent changes or levels as presented):
  - Cereals cultivated area: 2004-05 = 9.3; 2005-06 = 5.7; 2006-07 = 4.7
  - Cereals yield (quintal per hectare): 2004-05 = 13.1; 2005-06 = 14.4; 2006-07 = 15.2
  - Pulses cultivated area: 2004-05 = 22.7; 2005-06 = -4.2; 2006-07 = 6.4
  - Oil seeds cultivated area: 2004-05 = 44.6; 2005-06 = -3.5; 2006-07 = -7.2
  - Coffee cultivated area: 2004-05 = 0.4; 2005-06 = 13.0; 2006-07 = 6.0
  - Coffee yield: 2004-05 = 6.6; 2005-06 = 8.2
- Imports surge and determinants (Box 2 regression, 1991/92–2006/07, all variables in logarithms):
  - Coefficient on Real GDP: 1.37
  - Coefficient on Real exchange rate: 0.24  *
  - No. of observations: 16
  - R squared: 0.94
  - Phillips–Perron statistic: -2.84  *
  - * indicates significance at the 10 percent level
- Counterfactual import-volume growth (2003/04–2006/07):
  - Actual import volumes grew by 14.1 percent per annum over the 3-year period.
  - If output fixed at its 10-year historical growth rate of 7 percent → import volume growth = 11 percent per annum.
  - If real exchange rate also fixed at 2003/04 level → import volume growth declines an additional 1.3 per annum to 9.7 percent.
- Competitiveness and equilibrium REER assessments:
  - Single-equation (“reduced form”) models find Ethiopia’s REER above equilibrium predicted by fundamentals.
  - Based on fitted values from Gilmour and Chudik’s long-run cointegrating relationship, REER about 25 percent above its level based on fundamentals in 2007.

---

### Model sensitivity, net foreign assets (NFA), and alternative RER estimates
- Sensitivity and omission of NFA:
  - Model results sensitive to assumption changes.
  - Shortcoming: NFA position not included in panel specification because of insignificant effect across countries as a group.
  - Economic mechanism: stronger NFA allows a higher REER (more room to borrow to finance a larger current account deficit).
  - Ethiopia’s NFA improved considerably after HIPC and MDRI initiatives.
- Including NFA in single-equation analysis:
  - When NFA/GDP included, estimated overvaluation largely disappears; amended estimates suggest:
    - REER close to equilibrium through 2007.
    - Subsequent real appreciation through April 2008 made REER somewhat overvalued.
- Box 3 — revised coefficient estimates (including NFA):
  - Openness: -1.32
  - Government consumption: -0.14
  - Growth differential: 0.86
  - Terms of trade (index): 0.39
  - Net foreign assets: 0.44
  - Model statistics (Revised specification):
    - No. of observations: 1/282616
    - R squared: 0.89
    - Phillips-Perron statistic: -3.21**
    - * indicates significance at the 5 percent level
    - 1/ For Chudik and Mongardini, number of observations are per country; there are 18 countries
- Flow-analysis scenarios for medium-term reserve cover target (mid-point of authorities’ target of 10-12 weeks of imports = 2 ¾ months by 2012/13):
  - Output-gap alternatives: high (5 percent of GDP), medium (3 percent of GDP), low (zero).
  - External financing scenarios: baseline IMF official financing; alternatives ±1 percent of GDP per annum over next five years.
  - Trade elasticities estimated from income and price elasticities.
  - Results range: required REER movement varies from a depreciation of 25 percent to an appreciation of 8 percent.
  - Mass of scenario observations lies above zero — most likely equilibrium outcome is a depreciation of the REER.
- Stock-analysis (debt sustainability) and implications:
  - Joint IMF–World Bank DSA baseline: ratio of debt to exports stays within prudent limits but room for maneuver slight; additional financing should be concessional.
  - Alternative financing scenarios (peak NPV of debt-to-exports ratios):
    - Market-term additional financing: ratio rises to 142 percent at peak (vs 119 percent baseline).
    - IDA-term financing: ratio reaches 127 percent at peak.
- Conclusions and policy implications:
  - Mixed evidence, but on balance likelihood that recent real appreciation has pushed REER above fundamentals-consistent level.
  - Positive offsets: rising export market share, low wages, rising productivity, improving infrastructure attract investment.
  - Risks: rapid import growth tied to infrastructure has widened trade deficit and reduced reserve cover; part of weakening balance of payments may be cyclical and could be eased by moderating growth.
  - Additional financing could alleviate external constraints but nonconcessional borrowing needs careful consideration.

*Italic line: IMF staff summary from the source content._cr08259_*

### 1.      Inflation in Ethiopia has been

### 1.      Inflation in Ethiopia has been

### Recent inflation developments
- Inflation reached about 40 percent in May 2008, far exceeding the authorities’ policy target of single-digit inflation.
- Inflation has been on the rise for nearly three years and, at these levels, risks escalating further as inflation expectations may become ingrained unless appropriate adjustment policies are implemented.
- Recent inflation:
  - Reached a historical peak after a drought-related surge in 2003, receded to single digits, then turned up in 2004 and gradually increased to about 40 percent (12-month basis) in May 2008.
  - Coincided with high economic growth; previously inflation was largely associated with supply shocks (conflicts, droughts).
  - Was initiated by sharp increases in nonfood prices in mid-2006 (partly reflecting upward adjustments in retail petroleum products and a construction boom in Addis Ababa); nonfood inflation subsided through 2007 while food inflation escalated subsequently.
  - Shows nonfood inflation on the rise, reflecting hikes in world nonfood commodity prices, especially oil.

### Factors attributed to recent inflation
- Studies and reports (EDRI, 2007; FAO, 2008) identify multiple contributing factors:
  - Increasing money supply.
  - Rising world commodity prices.
  - Continued good economic performance.
  - Housing shortages in urban areas.
  - Changes in farmers’ behavior to supply products more uniformly over the year (improvements in access to credit, storage facilities, marketing information, etc.).
  - Increased local purchases by governmental food security institutions, agricultural cooperatives, and relief agencies.
- Once inflation reaches 20–30 percent, the role of past inflation—and thus inflation expectations—warrants attention.

### Empirical findings on relative contributions
- Two empirical approaches were cited, each with limitations because of data scarcity and structural breaks:
  - Netsere (2007): structural vector autoregression (real output and CPI) using annual data for periods 1965–98, 1965–2002, and 1965–2005. When the sample included the most recent period, demand factors became important.
  - Ayalew (2007): general equilibrium model using annual data 1970–2005 with many dummy variables for structural and policy shifts. Found supply shocks and consumer prices of major trading partners among the most important determinants.
- EDRI (2007) cautions that past trends cannot fully explain current developments due to ongoing structural changes (e.g., rising import-to-GDP ratio since 1993, devaluation in 1992, tariff reductions, liberalization of price regulations).

### Structural-change evidence (Box 1 results)
- A basic inflation model estimated for periods 1970–2001, 1970–2004, and 1970–2007 indicates structural changes after 2004:
  - Δln Currency (t-1): coefficients 0.31 (t-value 2.9***), 0.33 (t-value 2.6***), 0.33 (t-value 2.5***)
  - Δln Broad money: coefficients 0.48 (t-value 5.9***), 0.49 (t-value 5.0***), 0.54 (t-value 5.5***)
  - Δln GDP: coefficients -0.89 (t-value -5.8***), -0.90 (t-value -4.8***), -0.72 (t-value -4.1***)
  - Δln Import price: coefficients 0.11 (t-value 2.2**), 0.12 (t-value 1.9*), 0.09 (t-value 1.4)
  - Observations: 32, 35, 38; Adjusted R2: 0.69, 0.51, 0.443
- For 1970–2007, money has a more important role for inflation, while GDP and import price coefficients decline. The coefficient for lagged CPI is generally stable—past inflation likely drives inflation expectations and gains impact when inflation is rising.
- Estimated inflation based on earlier samples (1970–2001 and 1970–2004) for post-2002/2005 periods are far lower than actual inflation, indicating structural changes.

### External developments affecting prices
- Exchange rate:
  - Long remained stable vis-à-vis the U.S. dollar, moving by less than 5 percent in the last five years (prior to the report).
  - Since mid-2007, nominal effective exchange rate depreciated somewhat faster (about 10 percent annually) than in other countries in the region, potentially adding upward pressure on imported-item prices since mid-2007.
- International commodity prices:
  - EDRI (2007) suggests recent increases in international commodity prices contribute to nonfood inflation; correlation between nonfood CPI and world industrial commodity prices = 70 percent.
  - Retail petroleum prices are administered; adjustments were made in May and August 2006, January 2007, and January 2008. Price controls have mitigated world oil-price hikes but increased implicit subsidies.
  - Nonfuel commodities represent only about 5 percent of Ethiopia’s imports (customs data); food prices were not strongly correlated with world food prices until recently because domestic food prices historically were well below world market prices.
  - From late 2007 there are indications local food prices may be converging to high world prices as exports for some agricultural products increase.
- Changes in imports:
  - Rapid growth in imports in recent years doubled the import share of GDP in the last ten years, which should expand supply and alleviate inflation pressures. However, food imports were significantly reduced in 2006 (after a 160 percent increase in 2005) and stayed low relative to the five-year average, dampening that mitigating effect.

### Monetary developments and money growth
- Broad money growth accelerated since 2005, mainly driven by domestic credit to the public sector.
  - Significant increases in National Bank of Ethiopia (NBE) advances to government in 2005–2006 pushed up currency in circulation and enabled high growth in public credit without crowding out private sector credit.
  - Relatively loose monetary conditions evidenced by prolonged excess reserves in the banking system.
    - In late 2004, bank reserves exceeded required level by 35 percent. Since then, excess reserves were gradually reduced through:
      - more use of treasury bills to finance the fiscal deficit;
      - increases in reserve requirements (July 2007 and April 2008);
      - increases in commercial bank credit to the private sector.
    - As a result, excess reserves fell below 10 percent, close to the minimum by historical standards.
- Rising money velocity suggests inflation expectations are becoming ingrained:
  - Inflation has eroded the credibility of money as a store of value; real interest rates on deposits became increasingly negative, prompting shifts to consumption or alternative investments and adding price pressures.
- Overall assessment: role of external factors is relatively limited to date (though future retail fuel price adjustments would add pressure); role of money appears larger following the rise in money growth since 2005; monetary factor alone cannot explain recent food inflation—structural factors (including convergence of some export agricultural prices to international prices) also play a role.

### Cross-country evidence and policy responses
- Among 152 developing countries, about 50 experienced high inflation (above 20 percent) in the last fifteen years (sample excludes some CIS countries that had high early-1990s inflation but were already trending down).
- Fischer et al. (2002) evidence: as inflation reaches high levels, probability rises that inflation continues upward as inflationary expectations build.
- Countries’ responses when inflation reached 20–40 percent (examined over two years after reaching that threshold):
  - Most countries tightened fiscal stance for two years after reaching inflation rates of 20–40 percent, reducing fiscal deficits by 0–5 percent of GDP. Most of these countries achieved lower inflation.
  - Some countries experienced an increase in fiscal imbalances and subsequently much higher inflation.

### Structure of paper (as presented)
- Section II: overview of recent inflation developments.
- Section III: exploration of contributing factors (studies, monetary and external review).
- Section IV: cross-country analysis with countries experiencing high inflation.
- Section V: main conclusions.

*Main contributor to this chapter: Jiro Honda.*

### Box 3. Public Sector Borrowing

### Box 3. Public Sector Borrowing

### Estimation and Correlation
- Domestic public sector borrowing is estimated based on annual changes in credit outstanding to the public sector from the NBE and commercial banks.
- According to this estimate, domestic public sector borrowing is highly correlated with inflation, likely reflecting the dominant role of the sector in the economy.
- While this fact alone would not rule out other factors causing inflation, it does suggest that public sector imbalances have played a significant role in Ethiopia’s price developments.

### Historical Evidence and Magnitudes
- For much of the past few decades, public enterprise borrowing from commercial banks was low, possibly because they were financed through other sources, including the government. In recent years, however, public enterprises have borrowed substantially from commercial banks.
- During previous disinflation periods (late 1980s and mid 1990s), inflation was lowered by 8–18 percentage points, following the reduction of the public sector borrowing by 1½–4½ percent of GDP.
- This is broadly equivalent with empirical results from cross-country analysis: Fischer et al. (2002), using fixed effects in panel of 94 developing countries, estimate that, for high inflation countries, a 1 percentage point improvement (deterioration) in the ratio of the fiscal balance-to-GDP typically leads to a 4.2 percent decline (rise) in inflation, all else constant.

### Nature of Spending and Price Effects
- The nature of spending by government and public enterprises would need to be examined further to assess its impact on prices:
  - If credit is used for capital spending with a high import content, higher borrowing should put more pressures on international reserves, rather than domestic prices.
  - If, however, more resources are spent on recurrent outlays (including nontradables), inflation pressures may heighten.

### Monetary Policy, Disinflation, and Growth — Cross-country Evidence
- There is a clear correlation between monetary tightening and lower inflation. Those countries which lowered money growth achieved low inflation. Some countries experienced significantly higher money growth, and all of them faced even higher inflation within the following two years.
  - Note 1: Monetary tightening is measured by the difference in broad money growth per annum between time t (when inflation reached 20-40 percent) and time t+2 (after two years).
  - Note 2: Changes in inflation are measured by the difference in a 12-month inflation between the two periods; t and t+2.
- Decisive disinflation tends to lead to a quick rebound in growth. Countries that were experiencing low growth and high inflation tended to experience higher growth when they reduced inflation. For other cases, growth was either unaffected or reduced somewhat, but there is no indication that the impact of anti-inflation policies on growth was strong. However, countries which failed to reduce inflation tended to see lower growth after some time.
  - Note 1: Changes in real growth are measured by the difference in a GDP growth between time t (when inflation reached 20-40 percent) and time t+2 (two years after).
  - Note 2: Changes in inflation is measured by the difference in a 12-month inflation between the two periods; t and t+2.
- For most countries, the exchange rate was not explicitly used to anchor domestic prices. The nominal exchange rate was allowed to depreciate significantly, and may even allow depreciation in real terms. However, the pace of depreciation slowed down significantly for most countries two years after inflation reached 20-40 percent, which often reinforce the move to lower inflation.
  - Note 1: Changes in nominal depreciation are measured by the difference in annual changes (in percent) in NEER between time t (when inflation reached 20-40 percent) and time t+2 (two years after).
  - Note 2: Changes in inflation is measured by the difference in a 12-month inflation (in percent) between the two periods; t and t+2.

### Selected Conclusions (paragraphs 14–18)
- 14. Ethiopia’s recent rise of inflation is unusual because it does not coincide, as in the past, with a negative supply shock but, instead, high growth. While Ethiopia has often been affected by domestic supply shocks, it has rarely experienced high inflation driven by demand factors.
- 15. Both domestic and external factors explain recent inflation. While estimating the precise impact of each factor is technically difficult, given data constraints and continuing structural changes, inflation rates in excess of those experienced by trading partners and in the region point to some Ethiopia-specific factors.
- 16. High money and credit growth in recent years has likely played a key role in increasing demand and prices. More recently, inflation expectations seem to play an increasing role as prices are steadily accelerating. By contrast, the prices of imported goods have played a limited role as (i) the bulk of Ethiopia’s imports are manufactured and intermediate goods, whose prices in world markets have increased only modestly; (ii) the share of imported commodities (for which world prices accelerated in recent years) is small; and (iii) the impact of rising oil prices has been mitigated through lags in adjusting domestic retail prices.
- 17. The prices of exportable agricultural products have played a significant role in driving up domestic prices of some food items as they have been converging to higher (and rising) world prices. This process is in principle desirable as it raises incomes of rural households through trade of their outputs, and price rises stemming from it should come to an end when international prices are reached. However, policy-makers need to pay careful attention if prices for these products begin to exceed their international level as this would indicate that domestic factors push prices beyond their competitive levels.
- 18. International experience suggests that, as soon as inflationary expectations begin to emerge, a significant tightening of macroeconomic policies is essential. In doing this, a proper policy mix between fiscal and monetary policies is key to avoid crowding out the private sector and affecting long-term growth potential.

*Source: IMF staff summary from Box 3, “Public Sector Borrowing,” _cr08259 - Box 3. Public Sector Borrowing_.*

### 8.      Bordo and Jonung further argue that the velocity turning point for narrow

### 8.      Bordo and Jonung further argue that the velocity turning point for narrow 

### Summary of empirical approach and sample
- Velocity curves by income groups constructed using data from 1995 to 2007.
- IMF member countries/areas with complete GDP and money aggregates series sorted into four groups based on 2006 per capital GNI: low-income countries (LICs); low-middle-income countries (LMICs); high-middle-income countries (HMICs); and high-income countries (HICs).
- Excluded from the sample are countries with hyperinflation, defined as 12-month inflation that exceeds 50 percent at any point.
- Results are broadly consistent with Bordo and Jonung’s findings for earlier data (1952–1982).

### Broad money (BM) velocity: cross-country patterns (1995–2007)
- BM velocity generally declines over time: during 1995–2007 average BM velocity for every income group slopes downward.
- Cumulative declines range from 17 to 35 percent.
- Decline is broad based; only 18 of 137 countries in the sample had higher velocity at the end of the 13-year period.
- BM velocity declines more slowly as countries become more developed (consistent with financial innovation and nonbank financial intermediation counteracting financial deepening).
- Levels of BM velocity vary widely across countries even within income groups.

### Narrow money (M1) velocity: patterns and interpretation
- M1 velocity also slopes downward but is not strictly correlated with income.
- Average M1 velocity higher for HMICs than for LICs and LMICs, yet HICs have the lowest average M1 velocity of the four income groups.
- Possible reasons: (1) shift from checking to saving deposits may be smaller than expected because more banks now pay interest on both; (2) M1 data are not available for Euro countries since the introduction of the Euro, and Euro countries—being at the high end of development—are more likely to see flat or rising M1 velocity.

### Key cross-sectional numbers (Broad Money Velocity in 2007, Table 2)
- ObservationsAverageMaximumMinimum
- Ethiopia...3.01......
- LIC37 3.4 6.3 0.8
- LMIC43 2.5 9.3 0.7
- HMIC30 2.2 12.3 0.4
- HIC27 1.4 3.2 0.3

### Drivers of velocity in peer countries (focus on LICs)
- Opportunity cost of holding money (inflation)
  - Expectation: high inflation reduces money demand and raises velocity.
  - Evidence for LICs supports adaptive behavior: positive correlation between velocity and inflation inferred from previous-period experience.
  - Simple regression: a 1 percentage point increase in inflation increases velocity by 0.46 percent one year later.
  - Implication: on average, when inflation exceeds about 14 percent, velocity is more likely to rise than fall.
  - Probability of rising velocity, by inflation cutoff (Table 3):
    - For Inflation > Cutoff: 5 → 0.37; 10 → 0.47; 15 → 0.52; 20 → 0.58; 30 → 0.70
    - For Inflation < Cutoff: 5 → 0.30; 10 → 0.28; 15 → 0.31; 20 → 0.32; 30 → 0.33
- Return on alternative investments (T-bills)
  - Velocity increases with T-bill rates (lower money demand).
  - Positive relation holds for nominal T-bill rates and for real returns (nominal minus CPI inflation).
  - Ethiopia has the lowest nominal T-bill rate among all 18 LICs with data; combined with double-digit inflation, Ethiopia’s real T-bill return is almost 6.5 percentage points below the next lowest LIC.
- Return on holding deposits
  - Higher real return on deposits causes velocity to decline (money demand increases).
  - High inflation in Ethiopia significantly reduced real returns on saving deposits; only Haiti and Guinea had lower real returns among LICs.

### Ethiopia: evidence and dynamics
- BM velocity in Ethiopia is relatively low compared with the LIC average; before the recent rise it was more in line with the LMIC average.
- Rising BM velocity in Ethiopia since 2003 contrasts with the declining trend in most countries; velocity bottomed out in 2003 and has sloped upward since.
- Low M1 velocity in Ethiopia relative to other countries is striking and supports the idea that economic activities are significantly under-recorded.
- Table 5: Ethiopia vs. Other LICs (04-07 averages in percent unless otherwise noted)
  - Velocity (04-07, cumulative in percentage points): Ethiopia 5.74; LICs -1.96
  - Inflation (2003-06): Ethiopia 10.50; LICs 7.89
  - T-bill (2003-06): Ethiopia 13.97; LICs -0.74
  - Saving deposit (2003-06): Ethiopia -11.90; LICs -0.17
  - Real return (2003-06): Ethiopia -9.02; LICs -1.06
  - Impact on velocity for Ethiopia: up (velocity), up (inflation), down (T-bill), up (saving deposit)
- Inflation and velocity dynamics
  - Before 2003 Ethiopia’s velocity followed the LIC downward trend as inflation was broadly below the LIC average.
  - In 2003 inflation in Ethiopia exceeded the LIC average and velocity began to rise, diverging from the LIC trend; divergence continued after 2003.
- Monetary and financial intermediation effects
  - Within M1, share of currency in circulation (CIS) increased from 34 percent in 2002 to 36 percent in 2006 (Ethiopia was already heavily cash-based).
  - High inflation eroded credibility of money as a store of value; negative real interest rates on deposit accounts led to switching into consumption or alternative investments.
  - Disappearance of saving deposits led to a noticeable rise of CIS as a share of BM; ratio of saving and time deposits to nominal GDP has been declining since 2002.
  - Decline in the money multiplier accompanied rising velocity and reduced the effectiveness of monetary policy: a given intervention by monetary authorities has less impact on broad money growth and inflation.

### Cross-country and historical corroboration
- Persistent rising velocity for multiple consecutive years is rare among LICs: for 1995–2007 only five of 37 LICs experienced rising velocity for four consecutive years, and seven for three consecutive years; most had double-digit inflation during the period of rising velocity.
- Box 1 examples:
  - China: five spikes of inflation over three decades were accompanied by rising broad money velocity starting in the second year of the inflation surge.
  - Vietnam: double-digit inflation in the mid-1990s accompanied rising broad money velocity; velocity later turned downward as inflation fell.
  - Tanzania: double-digit inflation and strong growth in the late 1990s saw rising velocity; successful disinflation around 2000 put velocity back on a downward path while growth remained strong.

### Conclusions and policy recommendations
- Experience for Ethiopia and a broad range of countries points to high inflation being associated with rising velocity.
- Monetary policy implications:
  - Conservative money growth targets will be needed to offset the impact of rising velocity caused by Ethiopia’s high inflation.
  - A monetary program based on a declining velocity or a neutral assumption on velocity (no change) is likely to continue to lead to excess money supply and higher inflation than targeted.
  - To achieve the goal of single-digit inflation, a much more conservative money growth taking into account the recent experience of rising velocity is called for.
- Other policy measures:
  - Move real interest rates to positive territory to restore the credibility of money as a store of value and create an environment for more effective monetary operations.
  - Promote healthy growth of bank deposits to aid financial intermediation and allow meaningful interest rates to allocate the economy’s resources more effectively.

*IMF staff analysis as presented in the source content.*

### 1.      The sharp rise in the real effective exchange rate in recent years merits close

### 1.      The sharp rise in the real effective exchange rate in recent years merits close

### Overview and key concerns
- The real exchange rate has appreciated by 25 percent since 2004.
- Import coverage of reserves has declined from 5 months to under 2 months.
- The appreciation largely reflects much higher inflation in Ethiopia compared with partner countries; the trade-weighted nominal effective exchange rate has tended to depreciate since 2002.
- Real appreciation has coincided with strong export growth but even faster import growth, producing a considerable widening of the trade deficit.
- Given data limitations (cyclical state of the economy, sustainability of external financing), empirical results are highly tentative but suggest potential overvaluation of the real exchange rate at present.

### Balance of payments, reserves, and recent trends
- Deficit on goods and services is currently running at about 20 percent of GDP.
- Ratio of imports to GDP rose by 10 percentage points of GDP since 2000/01 to 32 percent in 2006/07.
- Exports have grown relative to output but slower than imports.
- Private and official transfers rose from 10 percent of GDP in 2000/01 to about 15 percent of GDP in 2006/07, helping maintain the current account deficit at about 5 percent of GDP on average over this period.
- Since 2001/02 the capital account has deteriorated, mainly due to a fairly persistent decline in net borrowing.
- Since 2003/04 the cumulative disparity between the capital and current account has reached 3 percent of GDP.
- In 2006/07 the overall balance of payments was almost zero, allowing reserve coverage to stabilize at 2 months of imports.
- Reserve cover is cyclical: drought years 2001 and 2002 compressed imports and built reserves; subsequent rapid expansion eroded reserves to low levels posing medium-term risk.
- Decline in reserve cover reflects failure of external financing (including transfers and errors and omissions) to keep pace with the rapid widening of the goods and services deficit.

### Export competitiveness and structural factors
- Ethiopia’s export market shares:
  - Coffee: about 0.6 percent of world coffee exports in 2003-04 rising to an average of 0.9 percent in 2005-06.
  - Oilseeds and flowers showed the largest export market gain of almost ½ percent of the world market (change 2005-06 relative to 2003-04: oilseeds 0.45, flowers 0.55 in percent of world exports).
- Flower sector:
  - Export earnings about US$65 million in 2006/07 and projected to double over the next few years.
  - Farm laborer wage reported at only US$20 per month in Ethiopia versus about $60 per month in India.
  - Price of leased land about US$13 per hectare.
  - A number of Indian entrepreneurs are relocating to Ethiopia to develop the flower industry.
  - Kenya’s horticultural industry will be subject to a new 5 percent tariff in 2008 under Generalized System of Preferences.
  - Government announced provision of 750 hectares of free land for investors in fruits and vegetables.
- Wages and cost comparisons:
  - Manual workers in Ethiopia earning about $60 a month (FDI magazine figure cited).
  - ILO estimate: average wage for a manufacturing worker in China about $190 per month.
- Doing business and logistics:
  - It takes 16 days to set up a new business in Ethiopia; about 133 days to obtain a new license (time to start a business and investor protection indicators cited).
  - Ethiopia ranks in the top decile of countries that are the least competitive on the global competitiveness index, but has improved position over the past two years.
  - Logistics performance index percentile: Ethiopia at the 69th percentile (higher than its global competitiveness index ranking).
- Infrastructure trends:
  - Air freight (million tonnes per km) rose by 46 percent between 2002 and 2005; air passengers rose by almost 9 percent between 2002 and 2005.
  - Road network rose by almost 5 percent per year between 2000 and 2006; World Bank road investment extended through 2012 to add about 11,000 km of federal roads and 5,500 km of regional roads (combined increase of about 40 percent).
  - Telecommunications: only about 20 per 1000 people currently own a cell phone; target coverage of 60 per 1000 people by 2009/10.

### Agricultural production, wages, and productivity
- Trends in cultivated area and yields (selected figures):
  - Cereals cultivated area: 2004-05 = 9.3; 2005-06 = 5.7; 2006-07 = 4.7 (percent change).
  - Cereals yield (quintal per hectare): 2004-05 = 13.1; 2005-06 = 14.4; 2006-07 = 15.2.
  - Pulses cultivated area: 2004-05 = 22.7; 2005-06 = -4.2; 2006-07 = 6.4.
  - Oil seeds cultivated area: 2004-05 = 44.6; 2005-06 = -3.5; 2006-07 = -7.2.
  - Coffee cultivated area and yields: cultivated area 2004-05 = 0.4; 2005-06 = 13.0; 2006-07 = 6.0; coffee yield 2004-05 = 6.6; 2005-06 = 8.2.
- Low wages can promote agricultural export growth only if wages remain low and there are sufficient opportunities to increase production via acreage expansion and higher yields; improved use of fertilizer and better seeds supports continued productivity gains unless a sharp drought reoccurs.

### Import surge and determinants (Box 2)
- Rapid import growth partly reflects development focus on improving infrastructure, which has high import content—fastest growing component is capital goods (road and air infrastructure).
- Regression of import volume on real GDP and real exchange rate (1991/92–2006/07; all variables in logarithms):
  - Coefficient on Real GDP: 1.37
  - Coefficient on Real exchange rate: 0.24  *
  - No. of observations: 16
  - R squared: 0.94
  - Phillips–Perron statistic: -2.84  *
  - Notes: * indicates significance at the 10 percent level
- Interpretation: real exchange rate coefficient significantly greater than zero, supporting view that real appreciation stimulated import demand; elasticities within typical international ranges but subject to uncertainty due to structural changes.
- Counterfactual analysis of import volume growth (2003/04–2006/07):
  - Actual import volumes grew by 14.1 percent per annum over the 3-year period.
  - Scenario with output fixed at its 10-year historical growth rate of 7 percent yields average import volume growth of 11 percent per annum.
  - If real exchange rate also fixed at its 2003/04 level, import volume growth declines an additional 1.3 per annum to 9.7 percent.
- Composition of imports (in percent of GDP) shows rising shares across raw materials and semi-finished goods, fuel, consumer goods, and capital goods (chart referenced).

### Assessments of competitiveness and equilibrium real exchange rate
- IMF methodological approaches applied to assess whether the real exchange rate is at an equilibrium level consistent with fundamentals or a sustainable medium-term current account balance.
- Single equation (“reduced form”) models find Ethiopia’s real exchange rate is above the equilibrium level predicted by fundamentals.
- Gilmour and Chudik (2006) single-country analysis and Chudik and Mongardini (2007) panel analysis find real effective exchange rate positively related to:
  - Terms of trade,
  - Productivity in Ethiopia relative to major trading partners,
  - Government consumption in relation to GDP.
- Real exchange rate negatively related to openness (ratio of imports and exports to GDP).
- Based on fitted values from Gilmour and Chudik’s long-run cointegrating relationship, the real effective exchange rate was about 25 percent above its level based on fundamentals in 2007.

*Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2008/_cr08259.pdf*

### 18.      However, model results are sensitive to changes in assumptions. A shortcoming of

### _cr08259 - 18.      However, model results are sensitive to changes in assumptions. A shortcoming of

### Model sensitivity and role of net foreign assets (NFA)
- Model results are sensitive to changes in assumptions.
- Shortcoming: the net foreign asset (NFA) position is not included in the equilibrium exchange rate panel specification because of its insignificant effect when countries are treated as a group.
- Economic mechanism: A stronger NFA would, other things equal, allow a country to have a higher real exchange rate because it would provide room to borrow more to finance a larger current account deficit.
- Ethiopia’s NFA position has improved considerably in recent years following the HIPC and MDRI initiatives.
- When the ratio of NFA to GDP is included in the single equation analysis the estimated overvaluation of the real exchange rate largely disappears (Box 3).
- Amended estimates suggest:
  - the real exchange rate was close to equilibrium through 2007,
  - subsequent real appreciation through April 2008 has made the real exchange rate somewhat overvalued.

### Box 3 — Alternative estimates of the equilibrium real exchange rate (key results)
- When net foreign assets are included in a regression of Ethiopia’s real exchange rate:
  - NFA has significant explanatory power and the error term remains stationary.
  - The effect on the real exchange rate of openness and terms of trade is considerably stronger.
  - Government consumption has the wrong sign.
- Reported coefficient estimates (Revised specification, including NFA):
  - Openness: -1.32
  - Government consumption: -0.14
  - Growth differential: 0.86
  - Terms of trade (index): 0.39
  - Net foreign assets: 0.44
- Model statistics (Revised specification):
  - No. of observations: 1/282616
  - R squared: 0.89
  - Phillips-Perron statistic: -3.21**
- Notes:
  - * indicates significance at the 5 percent level
  - 1/ For Chudik and Mongardini, number of observations are per country; there are 18 countries
  - Determinants of the Real Effective Exchange Rate (In percent of GDP, unless otherwise stated)

### Flow analysis — external sustainability and equilibrium RER estimates
- Flow equilibrium method estimates how much the real exchange rate would need to change to bring the current account into line with a sustainable level of medium-term external financing assuming full capacity utilization.
- Method anchor: assumes a reserve cover target rather than an equilibrium current account estimated on fundamentals, justified because reserve cover is currently judged to be too thin to provide adequate cushion against external shocks.
- Key difficult-to-quantify assumptions required:
  - What is a reasonable/sustainable level of grant and debt financing flows?
  - What should be a satisfactory level of reserve cover in the medium term?
  - How overheated is the domestic economy at present?
  - What is the responsiveness of exports and imports to changes in the real exchange rate?
- For the medium-term reserve cover target the analysis assumes the mid-point of the authorities’ target of 10-12 weeks of imports.
- The method requires estimating:
  - the underlying current account (adjusting for output gap/overheating),
  - the trade elasticities to determine how much the real exchange rate must change to eliminate any gap between underlying and equilibrium current account.
- Scenarios and assumptions used:
  - Output gap alternatives: high (5 percent of GDP), medium (3 percent of GDP), low (zero).
  - External financing scenarios: one based on current IMF official financing estimates; two alternatives assuming additional/reduced 1 percent of GDP per annum borrowing over the next five years.
  - Target: 2 ¾ months of imports by 2012/13.
  - Trade elasticities: based on estimated income and price elasticities for imports (and a similar sized, oppositely signed, elasticity for exports).
- Results range:
  - Required real exchange rate movement varies from a depreciation of 25 percent to an appreciation of 8 percent.
  - Scenarios implying little need for depreciation (or possible appreciation) assume a large output gap and external financing above current expectations.
  - Scenarios implying large depreciation needs assume a smaller output gap and external financing below baseline; up to a 25 percent depreciation would be needed.
  - The mass of scenario observations lies above zero, indicating the most likely equilibrium outcome is a depreciation of the real exchange rate.

### Stock analysis — debt sustainability and implications for equilibrium RER
- Stock approach assesses whether the ratio of net foreign assets to GDP is on a stable path using long-term debt projection.
- Debt trajectory based on assumptions about exports, imports, and grant and loan financing and assessed against international debt sustainability benchmarks.
- If the debt trajectory is within reasonable bounds and component trajectories of the balance of payments are credible, the current real exchange rate is assumed to be in equilibrium.
- Joint IMF and World Bank Debt Sustainability Analysis (DSA) baseline:
  - Shows the ratio of debt to exports stays within prudent limits.
  - Room for maneuver is slight; any additional financing to shore up reserves should be concessional.
- Alternative financing scenarios and peak NPV of debt-to-exports ratios:
  - If additional financing obtained on market terms: ratio rises to 142 percent at its peak, compared with 119 percent under the baseline scenario.
  - If financing obtained on IDA terms: ratio reaches 127 percent at its peak.

### Conclusions and policy implications
- Competitiveness and real exchange rate analysis is mixed, but on balance there is a likelihood that real appreciation has resulted in a real exchange rate above the level associated with fundamentals and macroeconomic stability.
- Positive factors:
  - Ethiopia has succeeded in raising its global export market share.
  - Low wages, rising productivity, and improving infrastructure should help attract foreign investment and expand the export base.
- Pressures and risks:
  - Rapid growth in imports, largely associated with infrastructure development, has led to a wide trade deficit and low reserve cover.
  - Part of the weakening balance of payments is likely cyclical and could be eased by moderating growth to eliminate overheating pressures.
  - There is scope for more financing to alleviate the external constraint, but nonconcessional borrowing would need careful consideration.
- Overall assessment:
  - Analysis of a sustainable external position points to a degree of overvaluation.
  - Updated single equation estimates (including NFA) also suggest the real exchange rate may be moving above a level determined by fundamentals.

*Source: _cr08259 - 18.      However, model results are sensitive to changes in assumptions. A shortcoming of*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2008/_cr08259.pdf_
