## Appendix I. Technical Annex

## Source details

**Canonical URL:** [Appendix I. Technical Annex](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2011/_cr1133.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2011/_cr1133.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2011/_cr1133.pdf.json)

---

### Overview and main conclusions
- Eurobond financing is justified in the near term for Macedonia but the medium-term strategy should develop domestic long-maturity denar instruments as a complement.
- Recent spreads are higher than predicted by fundamentals, implying borrowing costs may decline as external conditions normalize.
- Market access is unpredictable and heavily influenced by external/global factors.

### Medium-term fiscal financing outlook and Eurobond trade-offs
- Market financing is currently a small portion of total debt; debt is dominated by official (multilateral and bilateral) loans and structural bonds at low interest rates.
- Market debt composition: T-bills (virtually all short-term) and two Eurobonds (original maturities of 10 and 3.5 years).
- Consequence of replacing non-market with market debt: higher interest rates requiring higher primary surpluses to stabilize debt ratios.
  - Example: meeting projected fiscal deficits through 2015 solely with Eurobonds at the present secondary market yield of 6 percent would raise interest payments by 0.3 percent of GDP, implying an equal increase in the debt-stabilizing primary balance.
- Benefits of Eurobond financing:
  - Helps build international reserves (if at least partially sterilized by NBRM instruments).
  - Lengthens debt maturities, reducing rollover risks.
  - May reduce crowding out by boosting reserves and allowing a more relaxed monetary stance.
  - Establishes a presence in international capital markets and broadens investor base, benefiting corporate access.
- Risks of Eurobond financing:
  - Sovereign debt markets are volatile for external reasons (e.g., spike after late 2008 global crisis and during the Greek crisis).
  - External volatility can close markets or demand high issuance premia, especially when existing Eurobonds mature, creating spikes in gross financing needs.
  - External debt exposes the country to foreign exchange risk, making exchange rate adjustment more expensive.
- Policy implication: combine Eurobond issuance with development of long-maturity domestic denar instruments to support private long-term finance and financial deepening (including demand from second pillar pension fund assets).

### Cross-country stylized facts on Eurobond debt
- Macedonia outstanding Eurobond stock: some 4½ percent of GDP (described as moderate internationally).
- If Macedonia finances deficits in 2011–15 fully with Eurobonds:
  - Eurobond stock would reach 11.6 percent of GDP.
  - Gross issuance over 2011–2015 would be 13.2 percent of GDP.
  - Net issuance over 2011–2015 would be 9.7 percent of GDP.
- Smaller countries and those with exchange rate pegs tend to have larger Eurobond stocks as a share of GDP.

### Empirical determinants of Eurobond costs — key findings
- Two modeling approaches:
  - Cross-country annual panel model of spreads at issuance (41 EMs, 1990–2010).
  - Single-country monthly secondary market spread model (higher frequency).
- Observed issuance spreads and model implications:
  - Macedonia issued at a 620 basis point spread in 2009, about 300 basis points above model-predicted levels for that issuance.
  - The annual model predicts Macedonia could issue at present at a spread of 250–270 basis points over German Bunds—more than 100 basis points tighter than present secondary market yields.
  - The inaugural 2005 10-year Eurobond was issued at a spread of 120 basis points, 90 basis points inside the model prediction for that time.
  - The second (higher-frequency) model shows global credit shocks have a much stronger short-run impact on Macedonia spreads than implied by the annual model.
- Interpretation:
  - Macedonia’s high 2009 spread likely reflects unsettled global financial conditions and a small issuance size (liquidity premium), not fundamentals.
  - Both models suggest spreads may decline as global conditions normalize, but market access remains largely determined by external factors.

### Quantitative model results and coefficients (cross-country annual panel model)
- Data: 41 emerging markets, 1990–2010; EMBI included at month of issuance.
- Estimated associations:
  - A higher government debt to GDP ratio of 1 percentage point is associated with higher spreads of 2–3 basis points.
  - A higher Eurobond debt stock of 1 percentage point of GDP is associated with another 5 basis points higher spreads (specification (3) in Table 1).
  - A one percent of GDP higher fiscal deficit is associated with higher spreads of 4–7 basis points.
  - Lower reserves by 1 month of imports corresponds to higher spreads of 20 basis points.
  - A higher current account deficit by 1 percentage point of GDP corresponds to higher spreads of 5 basis points.
  - Lower real output growth by 1 percentage point is associated with higher spreads of 9 basis points.
  - Higher inflation by 1 percentage point is associated with higher spreads of 5 basis points.
  - A higher EMBI spread of 1 percent is associated with higher spreads of 16–23 basis points.
  - The first issuance is associated with lower spreads of 70 basis points (specification (2) in Table 1).
  - Issuance size/liquidity premium: doubling issue size is associated with a reduction in spreads of 30 basis points (specification (5) in Table 1).
  - Maturity: maturity of the issuance is not found associated with spreads.
  - Non-linearities: determinants have stronger effects when EMBI spread is high (i.e., above 4 percent); non-linearities are particularly strong for reserves coverage and growth.
- Model performance and robustness:
  - The annual model would predict a spread of just 300 basis points for Macedonia’s 2009 issuance, leaving an unexplained residual of about 300 basis points.
  - Extensions (issuance size, non-linear effects) explain an additional 0.3 percent and 0.1 percent respectively, but leave around 300 basis points unexplained.
  - Robustness checks: similar findings from models of secondary market spreads, restricted samples, random-effects models, and country-and-year fixed effects specifications.

### Selected comparative observed spreads (Jul-Sep 2010 illustrative cases)
- Predicted vs actual spreads and residuals:
  - Poland: Predicted 273, Actual 122, Residual -151
  - Czech Republic: Predicted 231, Actual 125, Residual -106
  - Mexico: Predicted 251, Actual 166, Residual -85
  - Lithuania: Predicted 270, Actual 233, Residual -37
  - Turkey: Predicted 251, Actual 257, Residual 6
  - Chile: Predicted 698, Actual 213, Residual -485
  - Brazil: Predicted 102, Actual 233, Residual 131
  - Croatia: Predicted 190, Actual 357, Residual 167
  - Ukraine: Predicted 212, Actual 452, Residual 240
  - Belarus: Predicted 289, Actual 586, Residual 298

### Medium-term Eurobond spread projections and risks
- Predicted medium-term range for Macedonian Eurobond spreads: 2.5–2.7 percent.
- Assumption: Macedonian fundamentals improve and global conditions normalize.
- Exception: higher Eurobond debt stock by 10 percentage points of GDP could add 50 basis points to spreads if Eurobonds were used to finance projected deficits in 2011–15.
- Predicted yield implication: yields of around 5 percent, based on current German rates of 2.4 percent.
- Short-term shock risk: a 1 standard deviation shock to EMBI spreads (2.4 percentage points, which has a probability 5 percent of occurring over 3 months) would increase the MKD spread by 40 basis points.
- Additional risks:
  - German rates could rise.
  - Significant idiosyncratic shocks not captured by the model (example: large unexplained residual for the 2009 issuance).

### Single-country monthly spreads model — contagion and transmission evidence
- Model specification: monthly data in differences for each country over the period 2006 to present.
- Key findings:
  - Changes in the EMBI index translate almost one-to-one into Macedonian spreads (much higher than coefficients around 0.2 found in annual regressions).
  - Macedonian transmission of global shocks is very high relative to neighboring countries and other studies; Macedonia’s spreads shot up most and are slowest to come down.
  - Greek bond spreads show no statistically significant association with Macedonian spreads; Macedonian spreads respond to the common EM trend rather than idiosyncratic Greek movements.
  - The index of stock market volatility in advanced economies is negatively correlated with Macedonian spreads, indicating markets differentiate EMs and partially substitute towards them when advanced economies are in worse shape.
- Empirical table notes (selected rows preserved):
  - Table 2 model predictions (selected):
    - Baseline model, controlling for first issuance: Predicted spread 1.6-0.4 ; Unexplained residual 2.8 ; Predicted spread 3.4
    - Controlling for first issuance and issuance amount: Predicted spread 2.0-0.8 ; Unexplained residual 3.1 ; Predicted spread 3.1
    - Controlling for first issuance, issuance amount and non-linear effects of EMBI: Predicted spread 2.1-0.9 ; Unexplained residual 3.2 ; Predicted spread 3.0
  - Monthly-determinants regression summary (selected coefficients):
    - EMBI spread coefficients (significance): Romanian spread: 1.18***; Bulgarian spread: 0.87***; Croatian spread: 0.29**; Turkish spread: 0.08; Greek spread: 0.01; Macedonian spread column shows 0.04*** and 0.44**.
    - European VIX examples: Romanian spread: -0.05***; Bulgarian spread: -0.04**; Croatian spread: -0.02*; Macedonian-related columns include -0.02 and -0.00**.
    - Observations and R-squared (selected): Observations: 575, 327, 1750, 570, 575, 7; R-squared: 0.421, 0.543, 0.229, 0.207, 0.034, 0.126, 0.090.
  - Note: All regressions are estimated in first-differences. Data frequency is monthly. Significance notation: *** p<0.01, ** p<0.05, * p<0.1.

### Electricity sector in Macedonia — introduction and reform summary
- Transition: from centrally planned and managed to a decentralized regulated free market.
- Demand-side challenge: improve efficiency of energy usage, constrained by below-market pricing for all but the largest users.
- Supply-side constraints:
  - Lack of oil, gas or high quality coal reserves.
  - Ageing generation facilities.
  - Shortage of domestic generation capacity and relatively dirty generation facilities.
- Recent reform outcomes:
  - Signed the Energy Community Treaty (ECT) committing to implement relevant parts of the EU acquis communautaire.
  - New Energy Law drafted; expected to bring country into compliance once fully implemented.
  - Implicit electricity pricing subsidies decreased from an estimated 3.8 percent of GDP in 2008 to 0.7 percent of GDP in 2010.
  - Drivers of subsidy reduction: lower regional energy prices, increases in domestic price for subsidized end-users, full liberalization of the market for the 10 largest users.
  - Privatization of distribution in 2006 and subsequent investment reduced distribution losses and collection shortfalls.
- Policy goal: implement Energy Law to achieve full electricity market liberalization by 2015 to attract long-term private investment in generation capacity.
- Complementary measures: market-based pricing, promotion of energy awareness, and gasification to improve efficiency and contain demand increases.
- Macroeconomic impact: liberalized prices and private investment to help ensure sector does not become fiscal burden or source of balance of payments pressures.

### Structure of supply and demand, and medium- to long-term plans
- Consumption and supply trends (2000–2010):
  - Consumption rose 34 percent between 2000 and 2008, then decreased by 12 percent in 2009 due to collapse of electricity-intensive exports from heavy industry.
  - Demand growth averaged 1.9 percent per year over 2000–09, below average real GDP growth of 2.7 percent.
  - Domestic supply decreased; shortfalls met by imports costing an average of €95 million in 2003–09, peaking at €235 million (3.6 percent of GDP) in 2008.
- Demand composition (2005–08):
  - 10 largest industrial customers accounted for some 25 percent of total electricity consumption.
  - Households consumed around 50 percent of total.
- Supply composition:
  - Main domestic generation: lignite-fired thermal power plants (state-owned ELEM) providing some 80 percent of domestic supply.
  - Hydroelectric generation: about 20 percent of domestic supply.
- Generation capacity and environmental compliance:
  - Domestic generation capacity insufficient to meet peak demand; imports averaged 20 percent of demand between 2000 and 2009, peaking at 32 percent in 2008.
  - Lignite-fired plant emissions (total CO2-equivalent) are 60 percent higher per unit of GDP than the OECD average (IEA, 2009).
  - Measured at market exchange rates, Macedonian emissions are 5 times as high as the OECD average (note from source).
- Strategy for Energy Development (MANU, 2010) projection:
  - Renewables could contribute up to 4,600 GWh by 2030, 90 percent from hydro.
  - Some hydro projects economically viable; others depend on higher carbon prices or lower discount rates to be competitive with gas or lignite plants.
  - Less competitive projects could be justified to meet Macedonia’s renewable energy target of 21 percent of total consumption by 2020.
- Transmission and interconnection:
  - Government plans to build international interconnection capacity to alleviate high-voltage bottlenecks and reduce costs of imported power.

### Electricity subsidies — estimates, drivers, and phase-out plan
- Total implicit subsidies are estimated at about 2.3 percent of GDP in 2009 and 0.7 percent of GDP in 2010.
- Estimates based on a regional market price of €63 per MWh in 2009 and €46 per MWh in 2010.
- ELEM sells supply for tariff customers at €30 per MWh in 2009 and €33 per MWh in 2010.
- Out of the total subsidy in 2009, providing subsidized imports to customers costs some 0.2 percent of GDP, equivalent to €12 million in 2009.
- IMF (2009) total pricing subsidies were estimated at 3.8 percent of GDP in 2008.
- Drivers of subsidy decline from 2008 to 2010:
  - Obligation since September 2008 for the ten largest electricity customers to secure their own supplies on the regional market.
  - Regional market price fell from an average of €82 per MWh in 2008 to €46 per MWh in 2010.
  - Tariff prices were raised 13 percent in late 2008 and 10 percent at the start of 2010.
  - Favorable hydrological conditions in 2010 enabled ELEM to generate significantly more hydro electricity and thus cut imports.
- Other losses and distribution of costs after privatization:
  - After privatization and sale of distribution to EVN of Austria, unrecognized distribution losses and collection shortfalls no longer accrue to the state; they accrue to EVN.
  - Unrecognized distribution and collection losses estimated at 0.8 percent of GDP in 2009, declining to 0.6 percent of GDP in 2010.
  - Implicit subsidy calculation assumes average transmission price of €4 per MWh is included in the import price but excluded when calculating opportunity costs.
  - Distribution loss assumptions: 24 percent in 2008, 20 percent in 2009, 18 percent in 2010, with 11 percent distribution losses as the international standard.
  - EVN collection-rate assumptions: 75 percent in 2008, 85 percent in 2009, 90 percent in 2010.
- Phase-out and policy recommendations:
  - Full price liberalization by 2015 (as called for under the Energy Law) would eliminate pricing subsidies, contingent on strong political will.
  - Government’s new targeted lump-sum social benefit for heating expenses intended to cushion electricity price increases for poor households.
  - Scenario (Tieman, 2011): with further regulated price increases in line with 2009 and 2010 and reversion of the regional market price to its longer-term average (adjusted for inflation), implicit subsidies would remain at 2010 levels as a share of GDP.
- Longer-term outlook and recommended actions:
  - Without further action, a mismatch between supply and demand is projected to create a squeeze in the regional electricity market over the medium- to long-term.
  - Projected regional demand growth: 3.1 percent a year on average, or over 100 percent over 2005 levels by 2030.
  - Supply-side policy options:
    - Attract large private sector energy firms to invest in generation capacity.
    - Provide a clear multiyear time frame for market liberalization and a stable legal and regulatory environment (full adoption of the EU acquis recommended).
    - Improve collection and lower theft; regional cooperation to realize scale efficiencies.
  - Demand-side policy options:
    - Promote energy efficiency; promote better insulation of businesses and homes; pursue gasification to enable efficient gas-fired cogeneration.
  - Environmental/pricing instruments:
    - Establishing a carbon price in line with the European emission trading scheme is suggested.

### Key numeric facts and assumptions (preserved exactly)
- Macedonia outstanding Eurobond stock: some 4½ percent of GDP.
- If financing deficits in 2011–15 fully with Eurobonds:
  - Eurobond stock would reach 11.6 percent of GDP.
  - Gross issuance over 2011–2015 would be 13.2 percent of GDP.
  - Net issuance over 2011–2015 would be 9.7 percent of GDP.
- Example fiscal cost: present secondary market yield of 6 percent would raise interest payments by 0.3 percent of GDP.
- 2009 Macedonian issuance spread: 620 basis points (issued in 2009); model-predicted around 300 basis points for that issuance.
- Annual-model current predicted spread: 250–270 basis points over German Bunds.
- Inaugural 2005 10-year Eurobond issued at a spread of 120 basis points.
- Predicted medium-term Eurobond spreads: 2.5–2.7 percent.
- Predicted yields around 5 percent, based on current German rates of 2.4 percent.
- A 1 standard deviation shock to EMBI spreads: 2.4 percentage points (probability 5 percent of occurring over 3 months) would increase MKD spread by 40 basis points.
- Subsidies and electricity figures:
  - Total implicit subsidies: 2.3 percent of GDP in 2009; 0.7 percent of GDP in 2010.
  - Regional market price: €63 per MWh in 2009; €46 per MWh in 2010.
  - ELEM domestic sale price: €30 per MWh in 2009; €33 per MWh in 2010.
  - Subsidized imports cost: 0.2 percent of GDP or €12 million in 2009.
  - Total pricing subsidies in 2008: 3.8 percent of GDP.
  - Regional market price in 2008: €82 per MWh (average).
  - Tariff increases: 13 percent in late 2008; 10 percent at the start of 2010.
  - Transmission price for calculations: €4 per MWh.
  - Distribution loss assumptions: 24 percent (2008), 20 percent (2009), 18 percent (2010); international standard 11 percent.
  - EVN collection-rate assumptions: 75 percent (2008), 85 percent (2009), 90 percent (2010).
  - Consumption rose 34 percent between 2000 and 2008, then decreased by 12 percent in 2009.
  - Demand growth averaged 1.9 percent per year over 2000–09; average real GDP growth 2.7 percent.
  - Imports averaged 20 percent of demand between 2000 and 2009, peaking at 32 percent in 2008.
  - Renewables projection: up to 4,600 GWh by 2030 (90 percent hydro).
  - Renewable target: 21 percent of total consumption by 2020.
  - Projected regional demand growth: 3.1 percent a year on average; over 100 percent over 2005 levels by 2030.

*Source: Appendix I. Technical Annex, _cr1133 - Appendix I. Technical Annex*

### Appendix I. Technical Annex ............................................................................................

### Appendix I. Technical Annex

### Overview and main conclusions
- Eurobond financing is justified in the near term for Macedonia but the medium-term strategy should develop domestic long-maturity denar instruments as a complement.
- Recent spreads are higher than predicted by fundamentals, implying borrowing costs may decline as external conditions normalize.
- Market access is unpredictable and heavily influenced by external/global factors.

### Medium-term fiscal financing outlook
- Market financing is currently a small portion of total debt; debt is dominated by official (multilateral and bilateral) loans and structural bonds at low interest rates.
- Market debt composition: T-bills (virtually all short-term) and two Eurobonds (original maturities of 10 and 3.5 years).
- Consequence of replacing non-market with market debt: higher interest rates requiring higher primary surpluses to stabilize debt ratios.
  - Example: meeting projected fiscal deficits through 2015 solely with Eurobonds at the present secondary market yield of 6 percent would raise interest payments by 0.3 percent of GDP, implying an equal increase in the debt-stabilizing primary balance.

### Eurobond-financing trade-offs (benefits and risks)
- Benefits of Eurobond financing:
  - Helps build international reserves (if at least partially sterilized by NBRM instruments).
  - Lengthens debt maturities, reducing rollover risks.
  - May reduce crowding out by boosting reserves and allowing a more relaxed monetary stance.
  - Establishes a presence in international capital markets and broadens investor base, benefiting corporate access.
- Risks of Eurobond financing:
  - Sovereign debt markets are volatile for external reasons (e.g., spike after late 2008 global crisis and during the Greek crisis).
  - External volatility can close markets or demand high issuance premia, especially when existing Eurobonds mature, creating spikes in gross financing needs.
  - External debt exposes the country to foreign exchange risk, making exchange rate adjustment more expensive.
- Policy implication: combine Eurobond issuance with development of long-maturity domestic denar instruments to support private long-term finance and financial deepening (including demand from second pillar pension fund assets).

### Cross-country stylized facts on Eurobond debt
- Macedonia outstanding Eurobond stock: some 4½ percent of GDP (described as moderate internationally).
- If Macedonia finances deficits in 2011–15 fully with Eurobonds:
  - Eurobond stock would reach 11.6 percent of GDP.
  - Gross issuance over 2011–2015 would be 13.2 percent of GDP.
  - Net issuance over 2011–2015 would be 9.7 percent of GDP.
  - These levels would be high relative to comparator countries.
- Smaller countries and those with exchange rate pegs tend to have larger Eurobond stocks as a share of GDP.

### Empirical determinants of Eurobond costs — key findings
- Two modeling approaches:
  - Cross-country annual panel model of spreads at issuance (41 EMs, 1990–2010).
  - Single-country monthly secondary market spread model (higher frequency).
- Observed issuance spreads and model implications:
  - Macedonia issued at a 620 basis point spread in 2009, about 300 basis points above model-predicted levels for that issuance.
  - The annual model predicts Macedonia could issue at present at a spread of 250–270 basis points over German Bunds—more than 100 basis points tighter than present secondary market yields.
  - The inaugural 2005 10-year Eurobond was issued at a spread of 120 basis points, 90 basis points inside the model prediction for that time.
  - The second (higher-frequency) model shows global credit shocks have a much stronger short-run impact on Macedonia spreads than implied by the annual model.
- Interpretation:
  - Macedonia’s high 2009 spread likely reflects unsettled global financial conditions and a small issuance size (liquidity premium), not fundamentals.
  - Both models suggest spreads may decline as global conditions normalize, but market access remains largely determined by external factors.

### Quantitative model results and coefficients (cross-country annual panel model)
- Data: 41 emerging markets, 1990–2010; EMBI included at month of issuance.
- Estimated associations (effects reported as in source text):
  - A higher government debt to GDP ratio of 1 percentage point is associated with higher spreads of 2–3 basis points.
  - A higher Eurobond debt stock of 1 percentage point of GDP is associated with another 5 basis points higher spreads (specification (3) in Table 1).
  - A one percent of GDP higher fiscal deficit is associated with higher spreads of 4–7 basis points.
  - Reserves and current account:
    - Lower reserves by 1 month of imports corresponds to higher spreads of 20 basis points.
    - A higher current account deficit by 1 percentage point of GDP corresponds to higher spreads of 5 basis points.
  - Real output growth: lower growth by 1 percentage point is associated with higher spreads of 9 basis points.
  - Inflation: higher inflation by 1 percentage point is associated with higher spreads of 5 basis points.
  - EMBI spread: a higher EMBI spread of 1 percent is associated with higher spreads of 16–23 basis points.
  - First issuance effect: the first issuance is associated with lower spreads of 70 basis points (specification (2) in Table 1).
  - Issuance size/liquidity premium: lower issuance amounts significantly increase spreads; doubling issue size is associated with a reduction in spreads of 30 basis points (specification (5) in Table 1).
  - Maturity: maturity of the issuance is not found associated with spreads.
  - Non-linearities: determinants have stronger effects when EMBI spread is high (i.e., above 4 percent); non-linearities are particularly strong for reserves coverage and growth.
- Model performance and robustness:
  - The annual model would predict a spread of just 300 basis points for Macedonia’s 2009 issuance, leaving an unexplained residual of about 300 basis points.
  - Extensions (issuance size, non-linear effects) explain an additional 0.3 percent and 0.1 percent respectively, but leave around 300 basis points unexplained.
  - Robustness checks: similar findings from models of secondary market spreads, restricted samples, random-effects models, and country-and-year fixed effects specifications.

### Selected comparative observed spreads (illustrative cases from Jul-Sep 2010)
- Examples of predicted vs actual spreads and residuals (as reported):
  - Poland: Predicted 273, Actual 122, Residual -151
  - Czech Republic: Predicted 231, Actual 125, Residual -106
  - Mexico: Predicted 251, Actual 166, Residual -85
  - Lithuania: Predicted 270, Actual 233, Residual -37
  - Turkey: Predicted 251, Actual 257, Residual 6
  - Chile: Predicted 698, Actual 213, Residual -485
  - Brazil: Predicted 102, Actual 233, Residual 131
  - Croatia: Predicted 190, Actual 357, Residual 167
  - Ukraine: Predicted 212, Actual 452, Residual 240
  - Belarus: Predicted 289, Actual 586, Residual 298

### Technical annex notes
- The 2009 Macedonian issuance spread jumped to 6.3 percent relative to German yields from 1.2 percent in 2005.
- The 2009 issuance was one of the highest among EM issuances in 2008–2010; Lithuania’s summer 2009 issuance was the highest in that period.
- Potential drivers of Macedonia’s high 2009 cost highlighted: small issuance size (liquidity premium) and timing during turbulent global credit conditions.

*Source: Appendix I. Technical Annex, _cr1133 - Appendix I. Technical Annex*

### 5.      These results suggest that in the medium-term, as global conditions normalize

### _cr1133 - 5.      These results suggest that in the medium-term, as global conditions normalize

### Medium-term Eurobond spread projections and risks
- Predicted medium-term range for Macedonian Eurobond spreads: 2.5–2.7 percent.
- Assumption: Macedonian fundamentals improve and global conditions normalize.
- Exception: higher Eurobond debt stock by 10 percentage points of GDP could add 50 basis points to spreads if Eurobonds were used to finance projected deficits in 2011–15.
- Predicted yield implication: yields of around 5 percent, based on current German rates of 2.4 percent.
- Short-term shock risk: a 1 standard deviation shock to EMBI spreads (2.4 percentage points, which has a probability 5 percent of occurring over 3 months) would increase the MKD spread by 40 basis points.
- Additional risks noted:
  - German rates could rise.
  - Significant idiosyncratic shocks not captured by the model (example: large unexplained residual for the 2009 issuance).

### Single-Country Monthly Spreads Model — evidence on contagion and transmission
- Model specification: monthly data in differences for each country over the period 2006 to present.
- Key findings:
  - Changes in the EMBI index translate almost one-to-one into Macedonian spreads (much higher than coefficients around 0.2 found in annual regressions).
  - Macedonian transmission of global shocks is very high relative to neighboring countries and other studies; Macedonia’s spreads shot up most and are slowest to come down.
  - Greek bond spreads show no statistically significant association with Macedonian spreads, explained by:
    - Macedonian spreads responding to the common EM trend (which embeds Greek effects) rather than idiosyncratic Greek movements.
    - Short history and uniqueness of the recent Greek spreads shock.
  - The index of stock market volatility in advanced economies is negatively correlated with Macedonian spreads, indicating markets differentiate EMs and partially substitute towards them when advanced economies are in worse shape.

- Empirical table notes (as presented):
  - Table 2: Model predictions for Macedonian issuances (selected rows preserved as in source):
    - Baseline model, controlling for first issuance: Predicted spread 1.6-0.4 ; Unexplained residual 2.8 ; Predicted spread 3.4
    - Controlling for first issuance and issuance amount: Predicted spread 2.0-0.8 ; Unexplained residual 3.1 ; Predicted spread 3.1
    - Controlling for first issuance, issuance amount and non-linear effects of EMBI: Predicted spread 2.1-0.9 ; Unexplained residual 3.2 ; Predicted spread 3.0
  - Source: IMF Staff estimates.

- Monthly-determinants regression summary (selected coefficients preserved as in source Table 3):
  - EMBI spread coefficients (statistical significance noted):
    - Romanian spread: 1.18***
    - Bulgarian spread: 0.87***
    - Croatian spread: 0.29**
    - Turkish spread: 0.08
    - Greek spread: 0.01
    - Macedonian spread column shows 0.04*** and 0.44**
  - European VIX coefficient examples:
    - Romanian spread: -0.05***
    - Bulgarian spread: -0.04**
    - Croatian spread: -0.02*
    - Macedonian-related columns include -0.02 and -0.00** (as shown)
  - Observations and R-squared (selected):
    - Observations: 575, 327, 1750, 570, 575, 7 (as presented across columns)
    - R-squared: 0.421, 0.543, 0.229, 0.207, 0.034, 0.126, 0.090
  - Note: All regressions are estimated in first-differences. Data frequency is monthly.
  - Significance notation: *** p<0.01, ** p<0.05, * p<0.1.
  - Source: IMF Staff estimates.

### Data sources and country sample (methodological notes)
- Primary market Eurobond information: Dealogic database (only USD or EUR/German Mark issuances considered; local currency issuances excluded).
- Secondary market Eurobond yields: Bloomberg.
- Indicators of international credit conditions (EMBI index, US and German bond yields, VIX): Datastream.
- Country sample: all emerging, non-oil exporting economies that have issued Eurobonds.
- Comparator group selection criteria: country size, sovereign credit rating (Standard and Poor’s), and inclusion of relevant neighboring countries to Macedonia.
- “Other” group: larger emerging markets with deeper access to international markets for comparison.

### Electricity sector in Macedonia — introduction and summary
- Transition: from centrally planned and managed to a decentralized regulated free market.
- Demand-side challenge: improve efficiency of energy usage, constrained by below-market pricing for all but the largest users.
- Supply-side constraints:
  - Lack of oil, gas or high quality coal reserves.
  - Ageing generation facilities.
  - Shortage of domestic generation capacity and relatively dirty generation facilities.
- Recent reform outcomes:
  - Signed the Energy Community Treaty (ECT) committing to implement relevant parts of the EU acquis communautaire.
  - New Energy Law drafted; expected to bring country into compliance once fully implemented.
  - Implicit electricity pricing subsidies decreased from an estimated 3.8 percent of GDP in 2008 to 0.7 percent of GDP in 2010.
  - Drivers of subsidy reduction: lower regional energy prices, increases in domestic price for subsidized end-users, full liberalization of the market for the 10 largest users.
  - Privatization of distribution in 2006 and subsequent investment reduced distribution losses and collection shortfalls.
- Policy goal: implement Energy Law to achieve full electricity market liberalization by 2015 to attract long-term private investment in generation capacity.
- Complementary measures: market-based pricing, promotion of energy awareness, and gasification to improve efficiency and contain demand increases.
- Macroeconomic impact: liberalized prices and private investment to help ensure sector does not become fiscal burden or source of balance of payments pressures (example cited: 2008 spike in import prices and drop in domestic output causing imports to surge).

### Structure of electricity supply and demand
- Consumption and supply trends (2000–2010):
  - Consumption rose 34 percent between 2000 and 2008, then decreased by 12 percent in 2009 due to collapse of electricity-intensive exports from heavy industry.
  - Demand growth averaged 1.9 percent per year over 2000–09, below average real GDP growth of 2.7 percent, indicating decreased energy intensity of output.
  - Domestic supply decreased over the same period; shortfalls met by imports costing an average of €95 million in 2003–09, peaking at €235 million (3.6 percent of GDP) in 2008.
- Demand composition (2005–08):
  - 10 largest industrial customers (metals and mining) accounted for some 25 percent of total electricity consumption.
  - Households consumed around 50 percent of total.
  - Remaining consumption: small industry and SMEs, transmission and technical losses, and electricity theft.
- Supply composition:
  - Main domestic generation: lignite-fired thermal power plants (state-owned ELEM) providing some 80 percent of domestic supply.
  - Hydroelectric generation: about 20 percent of domestic supply, mostly owned and run by ELEM.
- Generation capacity and environmental compliance:
  - Domestic generation capacity insufficient to meet peak demand; imports averaged 20 percent of demand between 2000 and 2009, peaking at 32 percent in 2008.
  - Lignite-fired plant emissions (total CO2-equivalent) are 60 percent higher per unit of GDP than the OECD average (IEA, 2009).
  - Measured at market exchange rates, Macedonian emissions are 5 times as high as the OECD average (note from source).

### Medium- to long-term supply plans and regional integration
- Strategy for Energy Development (MANU, 2010) projection:
  - Renewables could contribute up to 4,600 GWh by 2030, 90 percent from hydro.
  - Some hydro projects economically viable; others depend on higher carbon prices or lower discount rates to be competitive with gas or lignite plants.
  - Less competitive projects could be justified to meet Macedonia’s renewable energy target of 21 percent of total consumption by 2020 (in line with EU commitments).
- Transmission and interconnection:
  - Government plans to build international interconnection capacity to alleviate high-voltage bottlenecks and reduce costs of imported power.

### European integration driving electricity reform
- Commitment: gradual adoption of EU energy policy through Energy Community Treaty (ECT).
- ECT goals: secure reliable and affordable energy supply in Southeastern Europe while safeguarding the environment; establish integrated market in natural gas and electricity; attract investment in transmission networks and generation capacity; foster competition and interconnectivity.
- Fixed adoption time frame: 2015 for the acquis.
- Progress since signing ECT:
  - Energy legislation adopted in 2006 and amended in 2008.
  - Liberalized market created for the ten largest industrial electricity users; legal framework for market regulator established.
  - Remaining shortcomings: market structure (ELEM monopoly), cost reflectivity, and implementation deadlines.
  - Ministry of Economy prepared new Energy Law expected to address monopoly, clarify provisions on accepted losses, and set clear liberalization deadlines.
  - Energy law includes social protection measures: lump-sum heating subsidies for vulnerable households in a liberalized market.

### Electricity subsidies — levels, inefficiencies, and phase-out plan
- Nature of implicit subsidies:
  - Tariff customers (all except the 10 largest users) face regulated prices below regional market price; price above short-run marginal cost but below market implies opportunity cost used to calculate implicit subsidies.
  - When tariff customers’ demand > supply, ELEM imports power and sells domestically at regulated price normally below import price.
- Inefficiencies and distributional concerns:
  - Subsidies distort incentives and stimulate inefficient usage.
  - As social policy, subsidies are wasteful because most electricity is used by richer households.
  - Below-market tariffs discourage domestic and foreign investment in the sector.
- Phase-out commitment:
  - Through ECT, government committed to fully liberalize market by 2015, implying gradual phasing out of subsidies and rising end-user prices.
  - At current tariff and market prices, required tariff increase: almost 40 percent over the next five years.

*Source: IMF Staff estimates and IMF document content.*

### 13.      The total amounts involved in these implicit subsidies are estimated at about

### 13.      The total amounts involved in these implicit subsidies are estimated at about

### Summary of subsidy estimates (2008–2010)
- Total implicit subsidies are estimated at about 2.3 percent of GDP in 2009 and 0.7 percent of GDP in 2010.
- These estimates are based on a regional market price of €63 per MWh in 2009 and €46 per MWh in 2010.
- The price at which ELEM sells its supply for tariff customers was €30 per MWh in 2009 and €33 per MWh in 2010.
- Out of the total subsidy in 2009, providing subsidized imports to customers costs some 0.2 percent of GDP, equivalent to €12 million in 2009.
- In IMF (2009) total pricing subsidies were estimated at 3.8 percent of GDP in 2008 (not distinguishing direct vs indirect subsidies).

### Drivers of the decline in subsidies from 2008 to 2010
- Obligation (since September 2008) for the ten largest electricity customers to secure their own supplies on the regional market at the regional market price (plus a transmission and service fee), eliminating subsidies to these large users.
- Regional market price fell from an average of €82 per MWh in 2008 to €46 per MWh in 2010.
- Tariff prices were raised 13 percent in late 2008 and 10 percent at the start of 2010.
- Favorable hydrological conditions in 2010 enabled ELEM to generate significantly more hydro electricity and thus cut imports; reversal to historic rainfall averages could increase subsidies again unless offset by price liberalization.

### Other losses and distribution of costs after privatization
- After privatization and sale of the distribution network to EVN of Austria, unrecognized distribution losses and collection shortfalls no longer accrue (indirectly) to the state; they accrue to EVN.
- Unrecognized distribution and collection losses amounted to an estimated 0.8 percent of GDP in 2009, declining to 0.6 percent of GDP in 2010, driven by considerable investment in the network by EVN.
- Calculation notes and assumptions:
  - The implicit subsidy calculation takes account of the average transmission price of €4 per MWh, which is included in the import price but excluded when calculating opportunity costs.
  - Assumes the regional market price includes the full cost of generation (including mining and depreciation), so cost estimates reflect the full implicit subsidy.
  - Distribution loss assumptions: 24 percent in 2008, 20 percent in 2009, 18 percent in 2010, with 11 percent distribution losses as the international standard (per Regulatory Commission formula).
  - EVN collection rate assumed to improve from 75 percent in 2008 to 85 percent in 2009 and 90 percent in 2010.

### Policy recommendations and scenarios
- Full price liberalization by 2015 (as called for under the Energy Law) would eliminate pricing subsidies, contingent on strong political will—especially if international prices rise, which would require higher domestic price increases.
- The government’s new targeted lump-sum social benefit for heating expenses is intended to cushion electricity price increases for poor households and increase social acceptance of price increases.
- A scenario noted (Tieman, 2011) indicates that with further regulated price increases in line with 2009 and 2010 and reversion of the regional market price to its longer-term average (adjusted for inflation), implicit subsidies would remain at 2010 levels as a share of GDP. Higher domestic price increases or a lower regional market price would lower estimated subsidies.

### Longer-term outlook and recommended actions
- Without further policy action, a mismatch between supply and demand is projected to create a squeeze in the regional electricity market over the medium- to long-term.
- Projected regional demand growth: 3.1 percent a year on average, or over 100 percent over 2005 levels by 2030.
- Contributing supply-side issues: many electricity generation and district heating facilities are ageing and will have to shut down or will operate less efficiently.
- Supply-side policy options:
  - Attract large private sector energy firms to invest in generation capacity.
  - Provide an enabling investment environment to reduce uncertainty: i) clear multiyear time frame for market liberalization (as envisaged in the new Energy Law); ii) stable and predictable legal and regulatory environment—full adoption of the EU acquis recommended; iii) improve collection and lower theft; iv) regional cooperation to realize scale efficiencies and strengthen security of supply.
- Demand-side policy options:
  - Promote energy efficiency (as envisaged in the government’s Energy Strategy and Renewable Energy Strategy).
  - Short- to medium-term measures: promote better insulation of businesses and homes; pursue gasification to enable natural gas for heating and facilitate efficient gas-fired cogeneration heat and power plants.
- Environmental/pricing instruments:
  - Establishing a carbon price in line with the European emission trading scheme is suggested as a market-friendly incentive for environmentally friendly solutions.

### Key numeric facts and assumptions (preserved exactly)
- 2.3 percent of GDP in 2009 (total implicit subsidies).
- 0.7 percent of GDP in 2010 (total implicit subsidies).
- Regional market price: €63 per MWh in 2009; €46 per MWh in 2010.
- ELEM domestic sale price: €30 per MWh in 2009; €33 per MWh in 2010.
- Subsidized imports cost: 0.2 percent of GDP or €12 million in 2009.
- Total pricing subsidies in 2008: 3.8 percent of GDP (IMF, 2009).
- Regional market price in 2008: €82 per MWh (average).
- Tariff increases: 13 percent in late 2008; 10 percent at the start of 2010.
- Transmission price for calculations: €4 per MWh.
- Distribution loss assumptions: 24 percent (2008), 20 percent (2009), 18 percent (2010); international standard 11 percent.
- EVN collection-rate assumptions: 75 percent (2008), 85 percent (2009), 90 percent (2010).
- Projected demand growth: 3.1 percent a year on average; over 100 percent over 2005 levels by 2030.
- Target for full price liberalization by 2015 (as called for under the Energy Law).

*Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2011/_cr1133.pdf*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2011/_cr1133.pdf_