## _cr0855

## Source details

**Canonical URL:** [_cr0855](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2008/_cr0855.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2008/_cr0855.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2008/_cr0855.pdf.json)

---

### Overview of vulnerabilities
- Objective: To provide an integrated assessment of Serbia’s vulnerabilities.
- Main findings:
  - Despite improvements in public sector finances, external and financial vulnerabilities increased in 2006–07, reflected in a rapidly rising current account deficit, external debt, and highly euroized private credit.
  - High official reserves may be only partly available to mitigate risks due to large short-term obligations of the central bank.
- Policy implications:
  - A comprehensive approach, underpinned by structural reforms and strong fiscal and financial sector policies, is critical to reduce the growing imbalances.

### Public sector vulnerabilities
- Key developments and indicators (2002–2007 Proj.):
  - Fiscal balance: -4.2, -3.0, 0.0, 0.7, -1.5, -1.0
  - primary balance: -3.7, -2.0, 1.2, 2.2, -0.1, -0.1
  - Consolidated public sector debt 1/: 81.8, 79.6, 68.0, 59.1, 48.7, 47.2
  - General government debt: 77.4, 74.7, 63.1, 54.1, 39.6, 37.6
  - of which foreign currency denominated: 53.3, 49.0, 54.1, 49.2, 35.4, 32.9
  - Gross official reserves 2/: 14.4, 15.4, 17.3, 22.3, 37.4, 34.2
  - In months of prospective imports: 3.3, 3.6, 4.3, 4.7, 7.1, 7.5
  - In billions of U.S. dollars: 2.3, 3.1, 4.2, 5.8, 11.9, 13.9
- Memorandum items:
  - Stock of NBS securities (in RSD bn.) 1/: 1.5, 2.2, 1.8, 16.8, 148.8, 200.4
  - in percent of GDP: 0.2, 0.2, 0.1, 1.0, 7.0, 8.3
  - Privatization and license sales receipts: 1.8, 4.2, 0.5, 2.4, 7.2, 2.7
- External public debt, Aug. 2007 (USD bn., Share (percent)):
  - IFIs 4.25 2
  - Paris club 2.22 8
  - Other bilateral 0.45
  - Private - London Club 1.11 4
  - Other private 0.12
  - Total 8.1 100
- Interpretation:
  - Over 80 percent of external public debt is owed to non-private creditors; maturities are predominantly medium- or long-term.
- Recent risks and trends:
  - Fiscal slippages in 2006–07, large public-sector pay raises, tighter monetary policy, central bank quasi-fiscal losses (over one percent of GDP in 2006), accelerated spending of privatization receipts, and growing central bank securities stock (~8 percent of GDP).

### External sector vulnerabilities
- Structural constraints and external debt:
  - Slow transformation of productive sectors led import demand and capital inflows to outpace domestic supply; private external debt more than doubled in the past three years.
  - Over three quarters—or 100 percent—of exports is owed by the domestic non-bank sector which is not fully hedged.
- Export performance:
  - Volume of exports: 27 percent of GDP.
  - Exports increased by over 8 percentage points of GDP since 2002 but remain below Emerging Europe average.
  - Metals and food items represent over 40 percent of exports.
- Comparative vulnerability:
  - In a sample of 17 emerging European market economies, only Bulgaria, Estonia and Latvia have larger current account deficits than Serbia; those are EU members with export sectors 2–3 times larger.

### Reserve coverage and central bank contingent liabilities
- Gross official reserves:
  - Tripled since 2004, reaching USD 14 billion or 7.5 months of imports in 2007.
- Effective reserve coverage adjustments:
  - Commercial banks’ forex deposits at the central bank ≈ USD 5 billion; after adjusting:
    - Import cover falls to 4.6 months of imports.
    - After subtracting commercial banks’ deposits and stock of NBS bills, reserves ≈ 3 months of imports.
- Reserve coverage in selected SEE economies (2006):
  - Serbia: Reserves 1/ = 7.1; Share of com. banks' deposits (percent) 2/ = 36; Reserves net of com. banks' deposits 1/ = 4.6
  - Bosnia: 5.1, 0, 5.1
  - Romania 3/: 4.8, 193.9
  - Bulgaria: 4.3, 123.8
  - Ukraine: 4.2, 4.2
  - Albania: 4.0, 103.6
  - Croatia: 4.0, 262.9
  - Slovenia: 3.1, 13.1
- Risks from NBS securities:
  - Over USD 3 billion invested by commercial banks in two-week repo securities (NBS bills) driven by high dinar/euro interest differentials; reversal of carry trades could unwind reserves.

### Financial sector vulnerabilities
- Key drivers:
  - Rapid household credit growth, high euroization, surging offshore borrowing by enterprises.
  - Credit euroization: share of forex-denominated and forex-indexed domestic credit exceeds 70 percent.
- Bank-related considerations:
  - High commercial bank forex deposits at the central bank reduce usable reserves in stress scenarios.
  - Potential under-pricing of risk by foreign banks pursuing aggressive strategies.
- Policy implications:
  - Continue tight prudential and supervision policies to prevent excessive household credit risk buildup.
  - Strengthen structural policies (privatization, domestic capital markets) to reduce euroization-related vulnerabilities.

### Banking sector transformation and foreign presence
- Foreign presence: over 70 percent of the market.
- Bank assets: more than quadrupled since 2002; composition more diverse.

### Household credit: scale, composition, and risks
- Growth and scale (as of October 2007):
  - Household credit: 22 percent of bank assets; 12 percent of GDP.
  - Household credit in GDP remains among the lowest in the region.
- Composition (September 2007):
  - Consumer loans: 58 percent of total household credit.
  - Mortgage: 27 percent.
  - Other: 16 percent.
- Euroization and risks:
  - Forex-denominated/indexed domestic credit > 70 percent; elsewhere noted as over 80 percent in parts of the chapter.
  - Most household loans carry variable interest rates.
- Loan quality and NPLs:
  - Over 20 percent of bank loans classified risky (C, D, E categories).
  - NPL ratio (nine largest banks, June 2007): overall 10.4 percent; household NPL ratio 4.4 percent.
  - Rise in household NPLs by 43 percent in first half of 2007.
- NBS prudential responses (2006–Aug 2007):
  - Raised reserve requirements on forex liabilities.
  - Limited retail lending to 150 percent of banks’ capital (August 2007).
  - Shortened maturity of cash loans from ten to two years.
  - Effects: disintermediation, slowed domestic lending to enterprises, rise in direct off-shore corporate borrowing.

### Household-sector stress and policy implications
- Stress-test example:
  - Dinar depreciation of 25 percent could lower banking-sector capital-to-asset ratio from 19 percent to about 14½ percent on an FX/indexed household portfolio.
  - Stress-test would not push risk-weighted capital adequacy below 12 percent but could constrain new credit.
- Policy recommendations:
  - Strengthen prudential regulation further; accompany with supportive macroeconomic policies.
  - Continue financial literacy, credit registry use, and supervision improvements (Basel II strategy, expanded regulatory scope).

### Corporate sector: balance-sheet indicators and vulnerabilities
- Aggregate balance-sheet (2002–2006, RSD bn.):
  - Assets: 4,018; 4,355; 4,443; 5,465; 6,360
  - Fixed assets: 2,182; 2,376; 2,802; 3,462; 3,911
  - Current assets: 1,049; 1,205; 1,491; 1,810; 2,219
  - Liabilities: 1,694; 1,850; 2,218; 2,707; 3,121
  - Equity: 2,324; 2,505; 2,225; 2,757; 3,239
  - GDP, RSD bn.: 1,020; 1,172; 1,431; 1,750; 2,126
  - Ratio of GDP to total assets, percent: 25.4; 26.9; 32.2; 32.0; 33.4
- Leverage and maturity (2006 comparison):
  - Serbia: Leverage 49, Maturity 35
  - Median developed: Leverage 52, Maturity 36
  - Median emerging: Leverage 49, Maturity 36
- Corporate debt and external borrowing:
  - Corporate debt rose from 13 percent of GDP in 2002 to over 37 percent in August 2007.
  - Over 50 percent of outstanding external debt (11 percent of 2007 GDP) was to be repaid or refinanced before 2010 (as of June 2007).
  - Arrears on cross-border loans rose to 7.7 percent in 2003–06, declined to 6 percent in first eight months of 2007.
- Liability euroization:
  - Over 70 percent of domestic credit indexed/denominated in euros.
  - External credit 54 percent of corporate debt.
  - Effective euroization of corporate sector financial liabilities exceeded 85 percent in August 2007.

### Corporate sector stress-test results (aggregate balance-sheet scenarios)
- Assumptions:
  - Liability euroization of 80 percent (alternative 50 percent for liquid assets).
  - Dinar denomination of all fixed assets, and tax-related assets and liabilities.
- Aggregate balance sheet (pre-shock 2006, RSD bn., sample entries):
  - Assets pre-shock: 6,360 (80% scenario)
  - Fixed assets (dinar): 3,911
  - Short-term assets pre-shock: 2,219
  - Liabilities pre-shock: 3,121
  - Equity pre-shock: 3,239
- Shock impacts (percent changes):
  - Change in liabilities (80% euro-indexed liquid assets): 10% shock 7.9; 20% shock 15.7; 50% shock 39.3
  - Change in equity (80% euro-indexed liquid assets): 10% shock -4.0; 20% shock -8.0; 50% shock -19.9
  - Change in equity (50% euro-indexed liquid assets): 10% shock -5.3; 20% shock -10.7; 50% shock -26.7
  - Short-term assets to short-term liabilities (percent): pre-shock 100; shocks fall to 98, 95, 90 (80% scenario) and 96, 92, 83 (50% scenario)
  - Liabilities to total assets (percent): pre-shock 49; shocks rise to 52, 55, 63 (80% scenario) and 52, 56, 65 (50% scenario)
- Interpretation:
  - A 10 percent nominal depreciation raises corporate liabilities by roughly 8 percent and squeezes net worth by 4 percent under baseline assumptions.
  - In the worst scenario—50 percent depreciation and only half of liquid assets in euros—liabilities rise by about 40 percent while net worth shrinks by over a quarter.
  - Results capture first-round balance-sheet effects; second-round effects (higher lending rates, defaults, output declines) could amplify losses.

### Policy implications for the corporate sector
- Priorities:
  - Productivity-enhancing structural reforms and development of domestic currency capital markets.
  - Prudential and structural measures to contain exchange-rate exposures, maturity and rollover risks, and strengthen capacity to absorb foreign credit safely.

### REER assessment and external sustainability
- CGER-type REER assessment:
  - REER found overvalued in the range of 5–16 percent (results uncertain).
  - Overvaluation reflects unbalanced policy mix.
- Macroeconomic balance approach:
  - Estimated current account norm: -3.8 (percent of GDP)
  - Underlying current account: -7.9 (percent of GDP)
  - Difference: 4.0 (percent of GDP)
  - Elasticity of CA to REER: -0.25
  - REER adjustment needed: -16.4 percent
- External sustainability scenarios (sensitivity to NFL and potential growth):
  - Stabilize NFL at 66% of GDP, potential growth 5.5%:
    - Required CA: -4.8; Underlying CA: -7.9; Difference: 3.1; REER adjustment: -12.4%
  - Stabilize NFL at 83% of GDP, potential growth 5.5%:
    - Required CA: -6.1; Underlying CA: -7.9; Difference: 1.8; REER adjustment: -7.3%
  - Stabilize NFL at 66% of GDP, potential growth 6.2%:
    - Required CA: -5.2; Underlying CA: -7.9; Difference: 2.7; REER adjustment: -10.8%
  - Stabilize NFL at 83% of GDP, potential growth 6.2%:
    - Required CA: -6.6; Underlying CA: -7.9; Difference: 1.3; REER adjustment: -5.3%
- Policy implications:
  - Accelerate structural reforms and tighten fiscal policy to allow less restrictive monetary policy and reduce current account deficit.

### Twin deficits and fiscal–current account link (empirical results)
- Objective: Quantify relationship between fiscal and current account balances in Serbia.
- Main results:
  - Fiscal balance elasticity to current account between 0.5 and 1.1.
  - Baseline estimate: increase in fiscal deficit by 1 percent of GDP raises current account deficit by 1 percent of GDP over two quarters.
  - Baseline specification: +1 percent of GDP fiscal deficit widens current account by 0.6 percent of GDP same quarter and 0.5 percent next quarter.
- Single-equation OLS (2000Q1–2007Q3) selected coefficients:
  - Fiscal balance (excl. grants): 0.57** (Spec (1)); 0.50** (Spec (2))
  - Government expenditure (contemporaneous): -0.38***; lagged effects also significant.
  - Credit to nongovernment: contemporaneous -0.32**; lagged values up to -0.75***.
  - Change in trend GDP: large contemporaneous negative effects in some specs (e.g., -3.18***).
  - REER: -0.11*** (contemporaneous).
- Diagnostics:
  - Number of observations: 29; Adjusted R-squared: 0.76 (Spec (1)), 0.75 (Spec (2)), 0.72 (Spec (3)).
- Policy implication:
  - Fiscal tightening can be expected to reduce the current account deficit in the short run.

### Model structure, calibration, baseline, and risk simulations
- Model: Four-equation small-open-economy (Serbia home, euro area foreign): IS (output gap), Phillips (headline and core inflation), uncovered interest parity (real exchange rate), monetary policy rule.
- Key calibrated parameter values and steady states:
  - β_lag = 0.5; β_ld = 0.10
  - α_{πld} = 0.3; α_{ygap} = 0.25; α_z = 0.3
  - δ = 0.4 (real exchange rate expectations)
  - γ_{ygap} = 0.5; γ_π = 2.0; γ_{RSlag} = 0.5
  - Inflation target = 4 percent; long-term real interest = 4 percent; long-run potential growth = 5½ percent
- Data and forecast horizons:
  - Historical: 1999q1–2007q4; Baseline forecasts: 2008q1–2012q4.
- Baseline (selected annual averages, IMF staff projections):
  - Short-term interest rate: 2007 = 10.5; 2008 = 9.2; 2009 = 8.6
  - RPI inflation: 2007 = 6.5; 2008 = 9.3; 2009 = 6.4
  - Core RPI inflation: 2007 = 3.8; 2008 = 6.1; 2009 = 3.5
  - Exchange rate (€/Din.): 2007 = 0.0125
  - Real GDP growth: 2007 = 7.0; 2008 = 5.8; 2009 = 5.6
  - Output gap: 2007 = 0.7; 2008 = 0.5; 2009 = 0.2
- Baseline judgment:
  - Assumes unwinding in 2008–09 of tight 2006–07 monetary stance, real depreciation, and modest growth slowing toward potential.
- Risk simulations (selected shocks and qualitative outcomes):
  - 10 percent one-quarter real appreciation: lowers inflation and rates, creates persistent negative output gap, short-term growth reduction.
  - 10 percent one-quarter real depreciation: triggers inflation, monetary tightening, temporary growth boost.
  - 3 percentage points temporary nominal rate hike (negative interest shock): appreciation, lower inflation, slower growth; recovery of growth takes about two years.
  - 50 percent oil price rise for 2 quarters: headline inflation spike, central bank raises rates for 6 quarters, output contraction over about two years.
  - Temporary expansionary fiscal/domestic demand shock: output gap +2 percentage points first quarter; later real appreciation and growth slowdown.
  - Euro area +2 percentage points interest rate shock: through exchange channel, raises Serbian inflation and lowers growth.
  - Permanent reduction in inflation target: rapid disinflation, nominal/real appreciation, drop in growth with multi-year recovery.
  - Increase in reserve requirements: slows growth; partial compensation via exchange rate depreciation.
- Usage notes and limitations:
  - Model useful for consistency checks and shock analysis; subject to data limitations, simplified supply side, and sensitivity to steady-state choices.

### Fiscal impact of privatization (sample of 30 firms privatized in 2005)
- Total estimated fiscal contribution (RSD million):
  - 2003: 778.6
  - 2004: 675.0
  - 2005: 1,504.3
  - 2006: 2,549.7
  - 2007 H1: 1,368.1
- In percent of GDP:
  - 2003: 0.066%
  - 2004: 0.047%
  - 2005: 0.086%
  - 2006: 0.120%
  - 2007 H1: 0.113%
- Component highlights (RSD million):
  - CIT: 2003 3.5; 2004 1.8; 2005 3.4; 2006 88.9; 2007 H1 123.9
  - Wage Tax: 2003 234.8; 2004 191.1; 2005 290.7; 2006 456.5; 2007 H1 245.5
  - Contributions: 2003 551.0; 2004 479.0; 2005 773.4; 2006 1,131.0; 2007 H1 643.1
  - Budget financing costs (impact): 2005 -439.3; 2006 -840.1; 2007 H1 -335.4
- Privatization sales and cash receipts:
  - Sample sold for €89.1 million.
  - Cash collections: 2005 €34.2 million; 2006 €34.9 million; 2007 H1 ~ €0.2 million.
  - Receipts reduced domestic financing costs by €5.3 million in 2005, €10.0 million in 2006, €4.2 million in 2007 H1.
- Arrears toward EPS (RSD million, end-year stock):
  - 2003: 32.3; 2004: 27.0; 2005: 36.3; 2006: 22.6; 2007 H1: 14.7
- Employment and SSC observations:
  - Employed sample fell from 8,569 (2003) to 6,639 (2006) while SSC per employee increased; SSC sample per capita (000 RSD): 2003 64.3; 2004 58.7; 2005 103.4; 2006 170.3
- Extrapolation and caveats:
  - Extrapolating sample to ~200 companies privatized in 2005 implies potential fiscal contribution ≈ ¾ of one percent of GDP in 2006.
  - Caveats: sample representativeness, underestimation of subsidies, exclusion of some quasi-fiscal items and secondary effects.

*Prepared by Tokhir Mirzoev (EUR); content drawn from the IMF staff analytical chapter on vulnerabilities, 2006–07.*

### 2. External Public Debt, Aug. 2007.............................................................................. 7

### 2. External Public Debt, Aug. 2007.............................................................................. 7

### Overview of vulnerabilities
- Objective: To provide an integrated assessment of Serbia’s vulnerabilities.
- Main findings:
  - Despite improvements in public sector finances, external and financial vulnerabilities increased in 2006–07, reflected in a rapidly rising current account deficit, external debt, and highly euroized private credit.
  - High official reserves may be only partly available to mitigate risks due to large short-term obligations of the central bank.
- Policy implications: A comprehensive approach, underpinned by structural reforms and strong fiscal and financial sector policies, is critical to reduce the growing imbalances.

### Public sector vulnerabilities
- Key developments and indicators:
  - General government debt more than halved since 2002 due to Paris and London Club debt restructuring, strong fiscal performance in 2004–05 and large privatization-related receipts since 2005.
  - Fiscal balances and related indicators (2002–2007 Proj.):
    - Fiscal balance: -4.2, -3.0, 0.0, 0.7, -1.5, -1.0
    - primary balance: -3.7, -2.0, 1.2, 2.2, -0.1, -0.1
    - Consolidated public sector debt 1/: 81.8, 79.6, 68.0, 59.1, 48.7, 47.2
    - General government debt: 77.4, 74.7, 63.1, 54.1, 39.6, 37.6
    - of which foreign currency denominated: 53.3, 49.0, 54.1, 49.2, 35.4, 32.9
    - Gross official reserves 2/: 14.4, 15.4, 17.3, 22.3, 37.4, 34.2
    - In months of prospective imports: 3.3, 3.6, 4.3, 4.7, 7.1, 7.5
    - In billions of U.S. dollars: 2.3, 3.1, 4.2, 5.8, 11.9, 13.9
  - Memorandum items:
    - Stock of NBS securities (in RSD bn.) 1/: 1.5, 2.2, 1.8, 16.8, 148.8, 200.4
    - in percent of GDP: 0.2, 0.2, 0.1, 1.0, 7.0, 8.3
    - Privatization and license sales receipts: 1.8, 4.2, 0.5, 2.4, 7.2, 2.7
  - Notes:
    - 1/ Includes central bank obligations. The 2007 projection is based on the stock of NBS bills as of October 2007.
    - 2/ As of end-October 2007.
- Table 2 summary (external public debt, Aug. 2007):
  - Creditor composition (in USD bn., Share (percent)):
    - IFIs 4.25 2
    - Paris club 2.22 8
    - Other bilateral 0.45
    - Private - London Club 1.11 4
    - Other private 0.12
    - Total 8.1 100
  - Interpretation:
    - Over 80 percent of external public debt is owed to non-private creditors, implying limited rollover risks.
    - Maturities are predominantly medium- or long-term.

- Risks and recent trends:
  - Fiscal slippages in 2006–07 including generous public sector pay raises reversed prior improvements, forcing tighter monetary policy.
  - Tight monetary policy contributed to a rise in the stock of central bank securities (about 8 percent of GDP) and an appreciation, resulting in quasi-fiscal losses of the central bank (over one percent of GDP in 2006).
  - Accelerated spending of one-time privatization receipts and persistent imbalances risk reviving public sector vulnerabilities.

### External sector vulnerabilities
- Structural constraints:
  - Slow transformation of productive sectors left import demand and capital inflows (half debt-creating) outpacing domestic supply, translating into sharply rising current account deficits.
  - Private external debt more than doubled in the past three years; over three quarters—or 100 percent—of exports is owed by the domestic non-bank sector which is not fully hedged against exchange rate risks.
- Export performance and composition:
  - Volume of exports stands at 27 percent of GDP.
  - Exports increased by over 8 percentage points of GDP since 2002, but remain below the Emerging Europe average.
  - Exports are concentrated: metals and food items represent over 40 percent of Serbia’s exports, increasing vulnerability to terms-of-trade shocks.
  - High import content of exports lowers effective foreign exchange receipts.
- Comparative vulnerability:
  - In a sample of 17 emerging European market economies, only Bulgaria, Estonia and Latvia have current account deficits larger than Serbia’s.
  - Unlike Serbia, those countries are EU members and have export sectors 2–3 times larger than Serbia’s.

### Reserve coverage and central bank contingent liabilities
- Gross official reserves:
  - Gross official reserves have more than tripled since 2004, reaching USD 14 billion or 7.5 months of imports in 2007.
- Effective reserve coverage adjustments:
  - Commercial banks’ foreign currency deposits at the central bank are about USD 5 billion, reflecting increased reserve requirements and prudential tightening in 2006; these deposits obscure usable reserves in distress.
  - After adjusting for commercial banks’ deposits, Serbia’s import cover drops to 4.6 months of imports.
  - After subtracting both commercial banks’ deposits and the stock of NBS bills, official reserves drop to just about 3 months of imports.
- Table 3 (Reserve Coverage in Selected SEE Economies, 2006):
  - Reserves 1/ and Share of com. banks' deposits (percent) 2/ and Reserves net of com. banks' deposits 1/:
    - Serbia: 7.1 36 4.6
    - Bosnia: 5.1 0 5.1
    - Romania 3/: 4.8 193.9
    - Bulgaria: 4.3 123.8
    - Ukraine: 4.2 4.2
    - Albania: 4.0 103.6
    - Croatia: 4.0 262.9
    - Slovenia: 3.1 13.1
  - Notes:
    - 1/ Gross reserves in months of imports.
    - 2/ Share of forex deposits of commercial banks at the central bank in gross reserves.
    - 3/ Com. bank deposits were estimated based on forex liabilities of com. banks and the 40% reserve requirement.
- Risks from NBS securities:
  - Tight monetary policy in 2006–07 and high dinar/euro interest rate differentials attracted over USD 3 billion in commercial banks’ investments in two-week repo securities (NBS bills).
  - These investments were a major driver of reserve accumulation but could unwind via carry trade reversals, creating pressure on reserves.
  - The large stock of short-term central bank securities presents an additional risk to official reserves.

### Financial sector vulnerabilities
- Drivers of financial vulnerability:
  - Rapid growth of household credit, high euroization, and surging offshore borrowing by enterprises have increased financial sector vulnerabilities.
  - Credit euroization is high; household and corporate unhedged foreign currency exposures are significant risks.
- Bank-related considerations:
  - High commercial bank foreign currency deposits at the central bank reduce effective reserve usability in a stress scenario.
  - Potential under-pricing of risk by foreign banks pursuing aggressive strategies could lead to pressures on their balance sheets.
- Policy implications:
  - Preventing excessive risk buildup in household credits requires continued tight prudential and supervision policies.
  - Strengthening structural policies, including privatization and development of domestic capital markets, would help reduce euroization-related vulnerabilities and corporate sector risks.

*Prepared by Tokhir Mirzoev (EUR); content drawn from the IMF staff analytical chapter on vulnerabilities, 2006–07.*

### 14.      The banking sector underwent a

### _cr0855 - 14.      The banking sector underwent a

### Banking sector transformation and foreign presence
- Increased foreign presence (over 70 percent of the market) brought technological innovation, new credit products, and access to credit lines from abroad.
- Bank assets more than quadrupled since 2002 and their composition is more diverse.

### Household credit: scale, composition, and risks
- Household credit expanded almost tenfold since being virtually non-existent before 2002, reaching:
  - 22 percent of bank assets (as of October 2007)
  - 12 percent of GDP (as of October 2007)
- Serbia’s household credit in GDP is still among the lowest in the region.
- Credit euroization and household debt:
  - The share of forex-denominated and forex-indexed domestic credit exceeds 70 percent.
  - High credit euroization exposes borrowers to larger currency risks.
- Risks and constraints:
  - Rapid household credit growth may reflect banks’ ambitious market share targets and can raise concerns over asset quality due to the rapid growth (pace) rather than the level of household credit.
  - External vulnerabilities (low exports, rapidly growing euroized liabilities in the corporate sector, and other external vulnerabilities) may constrain households’ effective debt-carrying capacity even when debt-to-GDP ratios are low.
  - On balance, the report judges that the current rapid growth of household credit is making the country more vulnerable.

### Household loan quality and central bank prudential response
- Over 20 percent of bank loans are classified as risky (share of loans in C, D, and E risk categories).
- The overall NPL ratio of the nine largest banks surveyed by the NBS in June 2007 stood at 10.4 percent.
- The household NPL ratio of these nine banks stood at 4.4 percent.
- A survey of the nine largest banks revealed a rise in household NPLs by 43 percent during the first half of 2007.
- NBS regulatory actions:
  - In 2006, the NBS raised reserve requirements on commercial banks’ forex liabilities.
  - In August 2007, the NBS introduced additional regulations:
    - limiting retail lending to 150 percent of banks’ capital
    - shortening the maturity of cash loans from ten to two years
- Effects of prudential measures:
  - Disintermediation occurred: domestic lending to enterprises slowed and its share in banks’ assets dropped by ten percentage points since 2005.
  - Direct off-shore borrowing by enterprises rose sharply over the same period (Table 5).
  - Disintermediation lowered some domestic banking-sector credit risks and lowered interest rates, but increased corporate sector exchange rate risks and moral hazard (more risky borrowers less able to borrow off-shore).
  - An initial slowdown of household credit growth in September 2007 was followed by acceleration in October and November, indicating circumvention.

### Macro and policy implications (concluding remarks for household/financial stability)
- Recommended threefold comprehensive approach:
  - Strengthen aggregate supply (strong structural reforms: continued restructuring, privatization, improvements in the business environment).
  - Contain growth of domestic demand (constrain public spending to create space for rising private demand, reduce the external deficit, and alleviate monetary policy pressures).
  - Minimize vulnerabilities from rising mismatches in the private sector (financial sector policies to limit exposures to exchange rate risks).
- Greater absorptive capacity and higher domestic savings would help sustainably absorb foreign inflows and lower pressures on the public sector.

### Corporate sector: balance-sheet indicators and vulnerabilities (overview)
- Positive trends:
  - Number of socially-owned companies more than halved since 2002, while total number of companies rose.
  - Assets under public and social ownership declined from over 40 percent in 2003 to 25 percent in 2006; assets in private companies grew strongly.
  - Economy-wide return on assets increased from 25 percent in 2002 to 33 percent in 2006.
  - Corporate sector posted aggregate profit of over 6 percent of GDP in 2006.
- Leverage and maturity (Table 1 comparison, 2000 data source):
  - Serbia (2006): Leverage 49, Maturity 35
  - Median for developed countries: Leverage 52, Maturity 36
  - Median for emerging countries: Leverage 49, Maturity 36
- Caveats on indicators:
  - Balance sheet is aggregated but not consolidated (inter- and intra-sectoral obligations included).
  - Maturity structure often based on origination definitions, not remaining repayment schedules.
  - Data does not distinguish currency denomination of assets and liabilities, masking exchange rate exposure.

### Corporate credit growth, external borrowing, and term structure
- Corporate debt rose from 13 percent of GDP in 2002 to over 37 percent in August 2007.
- The bulk of corporate debt expansion was due to sharp growth in external borrowing.
- Since 2004:
  - Share of loans maturing in 1–3 years nearly doubled while the share of over-5-year loans dropped markedly (Figure 3).
  - Almost all bank loans and over three quarters of loans from non-banks have adjustable interest rates (Table 2), exposing firms to interest rate risks.
- Repayment and rollover risks:
  - As of June 2007, over 50 percent of outstanding external debt—11 percent of the 2007 GDP—was to be repaid or refinanced before 2010.
- Arrears:
  - Arrears on cross-border loans rose in 2003–06, nearly doubling to 7.7 percent, then declined to 6 percent in the first eight months of 2007.
  - Many cross-border loans are backed by guarantees from resident banks, increasing domestic financial sector vulnerabilities.

### Liability euroization and exchange-rate exposure
- Corporate sector liability euroization:
  - Over 70 percent of domestic credit either indexed or denominated in euros.
  - External credit reached 54 percent of the corporate debt.
  - Effective euroization rate of corporate sector financial liabilities exceeded 85 percent in August 2007.
  - By this measure Serbia was one of the most vulnerable countries in the region as of mid-2007.
- Economic costs of currency mismatch:
  - A depreciation can reduce firms’ net worth, raise risk premia, increase borrowing costs, expose maturity mismatches, aggravate rollover risks, and lead to defaults—potentially outweighing competitiveness gains from depreciation.

### Corporate sector stress-test results (aggregate balance-sheet scenarios)
- Assumptions:
  - Liability euroization of 80 percent (with an alternative of 50 percent for liquid assets).
  - Dinar denomination of all fixed assets, and tax-related assets and liabilities.
- Pre-shock (2006) aggregate balance sheet (in RSD billions):
  - Assets 6,360 (pre-shock), 6,476 (10% shock), 6,592 (20% shock), 6,940 (50% shock) under 80% euro-indexed liquid assets scenario; analogous series for 50% euro-indexed liquid assets scenario: 6,432; 6,505; 6,722.
  - Fixed assets (dinar-denominated) 3,911 across scenarios.
  - Short-term Assets: 2,219 (pre-shock) rising to 2,799 under a 50% depreciation with 80% euro-indexed liquid assets.
  - Liabilities: 3,121 (pre-shock) rising to 4,347 under a 50% depreciation with 80% euro-indexed liquid assets.
  - Equity: 3,239 (pre-shock) falling to 2,593 under a 50% depreciation with 80% euro-indexed liquid assets; under 50% euro-indexed liquid assets and 50% depreciation equity falls to 2,376.
- Impact summary (percent changes):
  - Change in liabilities (80% euro-indexed liquid assets): 10% shock 7.9; 20% shock 15.7; 50% shock 39.3.
  - Change in equity (80% euro-indexed liquid assets): 10% shock -4.0; 20% shock -8.0; 50% shock -19.9.
  - Change in liabilities (50% euro-indexed liquid assets): 10% shock 7.9; 20% shock 15.7; 50% shock 39.3 (same liability percent changes shown).
  - Change in equity (50% euro-indexed liquid assets): 10% shock -5.3; 20% shock -10.7; 50% shock -26.7.
  - Short-term assets to short-term liabilities (percent): pre-shock 100; under shocks fall to 98, 95, 90 (80% scenario) and 96, 92, 83 (50% scenario).
  - Liabilities to total assets (percent): pre-shock 49; under shocks rise to 52, 55, 63 (80% scenario) and 52, 56, 65 (50% scenario).
- Interpretation:
  - A nominal depreciation of 10 percent would raise corporate sector financial liabilities by roughly 8 percent and squeeze net worth by 4 percent under baseline assumptions.
  - In the worst scenario—50 percent depreciation and only half of liquid assets in euros—the liabilities rise by 40 percent while net worth would shrink by over a quarter.
  - These results capture first-round balance-sheet effects; second-round effects could involve lenders raising interest rates, further increasing liabilities and pushing firms toward default and/or forcing output declines.

### Policy implications for the corporate sector
- Productivity-enhancing structural reforms and development of domestic currency capital markets are instrumental to reduce risks stemming from rising euroized liabilities.
- Prudential and structural measures are needed to contain exchange-rate exposures, maturity and rollover risks, and to strengthen the corporate sector’s ability to absorb foreign credit safely.

*Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2008/_cr0855.pdf*

### 33.      The considerations above suggest that the high growth of credit in recent years

### _cr0855 - 33.      The considerations above suggest that the high growth of credit in recent years

### Corporate-sector vulnerability (summary points from opening paragraphs and Table 4)
- High growth of credit in recent years has made Serbian companies more vulnerable; still low leverage may insulate the sector in the short run but present trends could dissolve this advantage.
- Calls for continuous monitoring and analysis of “early warnings” based on underlying trends.
- Developing local capital markets would help companies hedge against exchange rate risks and stimulate dinar borrowing, relieving euroization-related vulnerabilities.
- Table 4 (Aggregate Corporate Sector Balance Sheet, In billions of RSD / In percent of total assets) highlights expanding assets and liabilities across 2002–2006 (number of reporting companies: 67,248; 74,510; 76,372; 75,878; 76,638) and shows:
  - Assets: 4,018; 4,355; 4,443; 5,465; 6,360 (2002–2006)
  - Fixed assets: 2,182; 2,376; 2,802; 3,462; 3,911 (2002–2006)
  - Current Assets: 1,049; 1,205; 1,491; 1,810; 2,219 (2002–2006)
  - Liabilities: 1,694; 1,850; 2,218; 2,707; 3,121 (2002–2006)
  - Equity: 2,324; 2,505; 2,225; 2,757; 3,239 (2002–2006)
  - GDP, RSD bn.: 1,020; 1,172; 1,431; 1,750; 2,126 (2002–2006)
  - Ratio of GDP to total assets, percent: 25.4; 26.9; 32.2; 32.0; 33.4 (2002–2006)

### Household credit — objectives, main findings, policy implications
- Objective: Assess macro-financial vulnerabilities related to household credit growth.
- Main findings:
  - Household balance sheet risks originate from currency mismatches due to credit euroization, which is an indirect credit risk for banks.
  - Overall level of household credit is still low and buffers have been installed through prudential measures.
- Policy implications: Financial vulnerabilities are on the rise and amplifying macro imbalances. Ensuring financial stability will entail further strengthening of prudential regulation accompanied by supportive macroeconomic policies.

### Background and recent dynamics (key statistics and facts)
- Serbia: overall private sector credit about 10 percent of GDP as of end-2006.
- Household credit rose from a marginal share in 2000 to 11.7 percent of GDP as of September 2007.
- Household share of total private sector credit increased to 41 percent (over the same period up to September 2007).
- Composition of household credit (September 2007):
  - Consumer loans: 58 percent of total household credit.
  - Mortgage: 27 percent (text) — noted as growing rapidly.
  - Other: 16 percent.
- Most household loans carry variable interest rates and are denominated in or indexed to foreign exchange.
- Cross-border capital flows and foreign-bank market-share strategies have significantly augmented the availability of funds for household credit.

### Key issues and vulnerabilities
- Rapid household credit growth driven by convergence expectations, financial liberalization, lower long-term interest rates, decline in inflation, and aggressive foreign-bank lending into consumer finance.
- Macro and prudential concerns include:
  - Balance sheet mismatches from credit euroization (over 80 percent of household loans linked to foreign currencies).
  - Interest rate exposure—most household loans carry variable rates.
  - Potential asset-market imbalances (real estate), with mortgage lending growing rapidly; exemption from mandated reserve requirements for mortgage loans insured by National Mortgage Insurance Company contributed to sharp increase in housing loans.
  - Large public-sector wage increases of 20–30 percent and fiscal relaxation during electoral cycles contributed to higher household leverage and widening current account deficits.
- Relationship to external balances:
  - Growth of aggregate demand outpaced supply, leading to a widening of the current account deficit; chapter V finds credit important in explaining the rising current account deficit.
- Household net worth and debt-servicing:
  - Despite rising indebtedness, Serbian households appear to have positive net worth (tentative).
  - Debt-servicing burden relatively low due to prudential limit that monthly installments on household debt cannot exceed 30 percent of net monthly income.
- Non-performing loans and vulnerability metrics:
  - A dinar depreciation of 25 percent could lower banking-sector capital-to-asset ratio from 19 percent to about 14½ percent (example stress test on FX/indexed household portfolio).
  - Stress-test result would not push risk-weighted capital adequacy below minimum requirement of 12 percent but could constrain new credit extension.
  - Survey of nine largest banks showed a 43 percent increase in non-performing household loans during the first half of 2007; household NPL ratio was 4.4 percent in June 2007.
  - Overall NPL ratio of these nine banks reached 10.4 percent in June 2007, and was 3.5 percent after accounting for provisions.
  - Household assets (foreign-currency deposits, remittances) could mitigate exchange-rate shock impacts, but borrower–saver overlap may be limited.

### Policy developments, prudential measures, and international lessons
- NBS regulatory and supervisory actions implemented and expanded:
  - Tightened foreign exchange exposure limits.
  - Increased capital requirements.
  - Broadened reserve requirements on foreign liabilities.
  - Extended regulatory coverage to leasing companies.
  - Banking supervision improved after new banking law in October, 2006.
  - Central Credit Registry established mid-2002; public information/financial literacy campaigns launched.
  - Credit bureau improved availability of credit information.
  - Monitoring of credit risk induced by foreign exchange risk to be improved; new rules for unhedged borrowers to be implemented.
  - Capital regulation for market-risk management issued.
  - Formal MOU with national supervisors in Greece; informal contact maintained with other home supervisors.
  - Bank Supervision Department working on strategy for Basel II implementation.
- Box 1 — Recent Prudential Measures (exact measures):
  - HH debt to income ratio: monthly installment cannot exceed 30% of net monthly income. This can go up to 50 percent including mortgages.
  - Maximum maturity of cash loans to households was lowered to 2 years.
  - Limit the minimum monthly repayment on revolving credit cards (5 percent) and maximum maturity of certain consumer loans (5 to 7 years).
  - A risk weight of 125 percent to be used in the calculation of risk-weighted assets that are FX or FX indexed loans if the borrower can’t rely on inflow in the same currency. Currently this is only for exposures exceeding dinars 10 million. In 2008, the minimum threshold will be lifted.
  - HH lending cap to be reduced from 200 percent of Tier I capital to 150 percent (from 2008). All mortgages are included in this calculation.
- International parallels and measures observed in emerging Europe:
  - Higher/differentiated risk-weights on unhedged foreign-currency loans (Croatia, Poland, Slovakia).
  - Higher liquidity requirements (Croatia, Romania).
  - More frequent stress tests (Czech Republic, Hungary, Poland).
  - Some countries have resorted to administrative measures like credit ceilings as last resort.
  - Enhanced supervision of FX-lending banks and public-awareness campaigns (e.g., Hungary).

*Source: _cr0855 - 33.*

### 47.      These measures, however, have shown mixed results in curbing credit growth,

### _cr0855 - 47.      These measures, however, have shown mixed results in curbing credit growth,

### Mixed results of credit measures and recent country experiences
- Measures (administrative or prudential) have generally strengthened the health of the financial system by building buffers but have shown mixed results in curbing credit growth.
- Credit growth, especially to households, remains strong in many emerging European countries, aided by foreign banks’ ability to obtain funds through rapid deposit growth and borrowing from abroad (including from parent banks).
- Circumvention of measures by borrowers through direct foreign borrowing or from less supervised financial institutions (including through leasing and credit cards) has been common.
- Country examples:
  - Croatia: direct credit controls led to circumvention and disintermediation.
  - Estonia and Romania: increased financing from abroad continued to support strong credit growth.
  - Bulgaria: administrative limits on bank credit had some impact but broader credit growth remained unabated.
  - Ukraine and Bosnia: saw some slowdown following tightening measures.
  - Serbia: it remains to be seen whether the effectiveness of macro-prudential and supervisory policy measures will be durable.

### Elements of a way forward — risks and rationale for continued policy action
- Key vulnerabilities and mitigating factors:
  - Rapid growth of consumer credit raises vulnerabilities despite low levels of consumer credit.
  - Concerns are mitigated by a relatively low level of consumer indebtedness, the tight prudential stance of the NBS, and a perception that the majority of Serbian banks (in foreign ownership) would likely receive support from their parent offices in times of distress.
  - Potential under-pricing of risk by foreign banks pursuing aggressive market share targets could lead to pressures on their balance sheet.
  - Further risks from a sudden stop in capital flows, especially if tight global credit conditions persist and/or domestic political issues remain unresolved.
  - Rising external vulnerabilities may pose substantial risks to households’ balance sheets.
- Policy recommendation:
  - Authorities are well advised to continue to take measures to prevent excessive build-up of risks in the household and financial sectors through an appropriate mix of macro and prudential policies.

### Designing policies to safeguard macroeconomic and financial stability
- Policy design principles:
  - Safeguard macroeconomic and financial sector stability by managing macroeconomic risks, reducing the probability of financial distress, and increasing resilience to adverse shocks by building buffers.
  - Prudential policies are effective when accompanied by macroeconomic policies including structural reforms.
  - Ensure sound lending practices with high origination standards (e.g., through conservative loan-to-value, debt-service to income and loan-to-income ratios).
  - Administrative controls can generally be circumvented beyond the short term and should be used judiciously (Hilbers and others, 2005).
  - In developing prudential measures, consider potential disintermediation and weak enforcement of creditors’ rights due to legal/judicial limitations and a weak credit culture.

### Strengthening financial sector surveillance and information
- Recommendations to improve surveillance:
  - Widen the information base to better assess debtor characteristics.
  - Use micro-level data from household surveys to analyze household balance sheets and assess household debt sustainability (example cited: Hungary).
  - The central bank could augment the information collected by the Credit Bureau with asset side data (e.g., on real estate prices from the National Mortgage Insurance Corporation) to broaden overall data analysis.

### Monitoring banks’ risk management and supervisory capacity
- Key supervisory priorities:
  - Close monitoring of banks’ risk management and resilience to shocks.
  - Continue to build supervisory capacity in a fast growing industry.
  - The NBS’ new financial stability unit conducting stress tests would help Serbia better prepare for systemic risks.
  - The NBS, along with the Bankers’ Association, should continue to raise borrowers’ risk awareness.
  - Strengthen dialogue between home and foreign supervisors to facilitate cross-border supervisory arrangements and crisis management in an environment dominated by foreign banks.
  - Continuously monitor banks’ liquidity in foreign exchange, noting that the NBS cannot act as a lender of last resort in foreign exchange.
  - In the medium term, development of a well-developed capital market could help diversify the financial system and improve the efficacy of monetary policy transmission in a highly euroized economy.

### CGER-type assessments of the dinar real effective exchange rate (REER) — objective and main results
- Objective:
  - Explaining the assessment of the level of the dinar real effective exchange rate carried out using CGER-type methodologies.
- Main results:
  - The REER is found to be overvalued in the range of 5–16 percent. The results are to be taken with caution given the significant uncertainties in the calculations.
- Policy implications:
  - The overvaluation reflects an unbalanced policy mix. It calls for an acceleration of structural reforms and a tightening of fiscal policy that would make room for a less restrictive monetary policy.

### Data limitations and key assumptions in the REER assessment
- Significant data limitations in Serbia due to the break-up of former Yugoslavia and separation of Montenegro in 2006.
- Problematic series:
  - Remittances: declined from 13 percent of GDP to about 6 percent of GDP in 2007; data quality is inadequate with frequent and ample revisions.
  - Net External Asset Position (NEAP): no available estimate; stocks available for some components only.
- Staff assumptions and estimates:
  - Following discussions with the Central Bank, staff assumed remittances would gradually return over the medium-term closer to, but below, their historical average of about 11 percent.
  - Foreign debt assets were broadly estimated at US$ 2 billion in consultation with the NBS.
  - At end-2007, the NEAP is projected at -66 percent of GDP.
- Underlying current account estimation:
  - Using staff’s medium-term baseline projections, stripped of projected future changes in the REER and terms of trade.
  - Assuming remittances return to about 9 percent of GDP, the current account deficit after grants is projected at 13.8 percent of GDP in 2012.
  - Projected REER appreciation of about 15 percent would explain 3.6 percentage points of this deficit.
  - Projected deterioration in the terms of trade by about 5 percent would explain another 2.3 percentage points.
  - Underlying current account deficit after removing these two factors: 7.9 percent of GDP.

### Macroeconomic balance approach — findings and calibration
- Method:
  - Assesses the underlying current account against a norm consistent with fundamentals estimated from a panel dataset of 54 advanced and emerging market economies over 1973–2004.
- Results and parameters:
  - Estimated current account norm: -3.8 (percent of GDP)
  - Underlying current account: -7.9 (percent of GDP)
  - Difference between norm and underlying CA: 4.0 (percent of GDP)
  - Elasticity of the CA to the REER: -0.25
  - REER adjustment needed to bring the underlying CA to the level of the norm (percent): -16.4
  - The change in the REER was calculated assuming an elasticity of the current account to the REER estimated at 0.25.
  - Fundamentals were expressed relative to Serbia’s four main trading partners: the European Union, Russia, Bosnia and Herzegovina, and China (together about 70 percent of Serbia’s trade).

### External sustainability approach — scenarios, results, and sensitivity
- Approach:
  - Assesses the underlying current account against the current account that would stabilize the NEAP at benchmark values given potential growth, inflation, and rates of return.
- Calibration challenges:
  - Defining benchmark NEAP is tricky; two scenarios were run:
    - Stabilize NFL at projected end-2007 value (66 percent of GDP).
    - Allow further build-up of liabilities; target mid-point between projected 2007 value and regional maximum observed (100 percent of GDP in Estonia).
- Sensitivity to potential growth assumptions:
  - Two potential growth assumptions:
    - Staff’s baseline: 5.5 percent.
    - Higher plausible growth with clearer EU prospects: 6.2 percent (2003–07 average for 15 Eastern European countries).
- Summary results (by scenario):
  - Stabilizing NFL at 66 percent of GDP, with 5.5 percent potential growth:
    - Required current account balance (percent of GDP): -4.8
    - Underlying current account (percent of GDP): -7.9
    - Difference between required and underlying CA (percent of GDP): 3.1
    - Elasticity of the CA to the REER: -0.25
    - REER adjustment needed (percent): -12.4
  - Stabilizing NFL at 83 percent of GDP, with 5.5 percent potential growth:
    - Required current account balance (percent of GDP): -6.1
    - Underlying current account (percent of GDP): -7.9
    - Difference between required and underlying CA (percent of GDP): 1.8
    - Elasticity of the CA to the REER: -0.25
    - REER adjustment needed (percent): -7.3
  - Stabilizing NFL at 66 percent of GDP, with 6.2 percent potential growth:
    - Required current account balance (percent of GDP): -5.2
    - Underlying current account (percent of GDP): -7.9
    - Difference between required and underlying CA (percent of GDP): 2.7
    - Elasticity of the CA to the REER: -0.25
    - REER adjustment needed (percent): -10.8
  - Stabilizing NFL at 83 percent of GDP, with 6.2 percent potential growth:
    - Required current account balance (percent of GDP): -6.6
    - Underlying current account (percent of GDP): -7.9
    - Difference between required and underlying CA (percent of GDP): 1.3
    - Elasticity of the CA to the REER: -0.25
    - REER adjustment needed (percent): -5.3

### Policy implications from REER assessment
- Consistent message: the dinar is overvalued in real terms (range 5–16 percent across approaches and scenarios).
- Diagnosis:
  - Overvaluation reflects an unbalanced policy mix: insufficiently rapid structural reforms, loose fiscal policy and restrictive monetary policy have put the current account on an unsustainable path.
- Recommended policy response:
  - Address problem at its roots by accelerating structural reforms and tightening fiscal policy, thereby making room for a less restrictive monetary policy.
  - This rebalancing would contribute to reducing the current account deficit and bringing it more in line with fundamentals.

*Source: _cr0855 - 47.      These measures, however, have shown mixed results in curbing credit growth,*

### References

### References

### Objective and Main Results
- Objective: To quantify econometrically the relationship between the fiscal and the current account balances—the so-called “twin deficits” in Serbia.
- Main results:
  - The fiscal balance has a strong and significant impact on the current account in Serbia, with elasticity estimates between 0.5 and 1.1.
  - Baseline estimate: an increase in the fiscal deficit by 1 percent of GDP raises the current account deficit by 1 percent of GDP over two quarters.
  - Baseline specification: an increase in the fiscal deficit by 1 percent of GDP widens the current account deficit by 0.6 percent of GDP in the same quarter and by 0.5 percent of GDP in the following quarter.
  - For a panel of SEE countries, the twin deficits relationship is more elusive and is overwhelmed by the impact of investment on the current account.

### Policy Implications
- Because fiscal policy has a strong impact on external balances, fiscal tightening can be expected to reduce the current account deficit in the short run.

### Literature Overview (A)
- Accounting identity: current account = national savings − investment; for a given private sector savings-investment balance, an increase in government deficit equals an increase in the current account deficit (“twin deficits”).
- Ricardian equivalence channel:
  - Full Ricardian equivalence implies higher private savings offset fiscal expansion; more likely in advanced economies with developed financial markets and fewer liquidity constraints.
  - Full Ricardian offset more likely if fiscal policy change is perceived as permanent rather than temporary.
- Labor supply and productivity channel: households may increase labor supply in response to fiscal expansion, potentially boosting productivity and private investment and amplifying fiscal effects.
- Real exchange rate and interest rate channels:
  - Fiscal expansion can appreciate the real exchange rate, reducing net exports and widening the current account deficit.
  - If fiscal expansion raises interest rates, it can reduce private investment (and possibly increase savings); persistent real appreciation may instead boost investment.
- Output and cycle channel:
  - During output growth, current account may deteriorate due to higher imports from investment, while fiscal balance may improve due to higher tax receipts → “twin divergence.”

### Empirical Evidence Summary (B)
- Mixed results across methods and samples:
  - General equilibrium models: typically find a twin deficit relation but small in magnitude; results sensitive to model calibration and non-Ricardian households.
  - VAR techniques: varying results; identify fiscal shocks and trace impacts controlling for output-driven fiscal endogeneity.
  - Panel data analyses: generally find positive but small twin deficit relationship.
  - Time-series single-equation estimates: evidence exists but magnitude depends on method.
- Methodological limitations:
  - Single-equation and panel estimates face endogeneity and simultaneity; GMM and instrumental variables can partially address this.
  - Dynamic general equilibrium models reflect underlying assumptions about household behavior.
  - VAR analyses require long time series; short samples reduce reliability.
- Short-term determinants often excluded from medium- to long-term studies: capital account developments, financing constraints, large capital flow changes, financial crises, terms-of-trade shocks—important for emerging countries like Serbia.

### Data Constraints and Strategy for Serbia (B, C)
- Serbian time-series limitations:
  - Reliable statistics only available since around 2000, coinciding with a structural break (political transition and reintegration).
  - Annual single-equation estimates have too short a span; quarterly data provide more observations but still limited.
  - Panel data for Southeastern European countries can increase observations but results apply to the sample aggregate, not specifically Serbia.
- Empirical approach taken:
  - Single-equation OLS estimates on quarterly data for Serbia (2000Q1–2007Q3), following a Kanda (2006)-like setup, with variables (except terms of trade and REER) expressed in percent of nominal quarterly GDP.
  - Annual panel estimates for several Southeastern European countries as a complementary approach.

### Current Account and Savings-Investment Balances in Serbia (C)
- Casual observation:
  - Main driver of current account deficits is the nongovernment sector (private sector and public enterprises).
  - Correlation coefficients:
    - Level correlation between external current account and non-government savings-investment balance: 0.8.
    - Change correlation between external current account and non-government savings-investment balance: 0.5.
  - Changes in the fiscal balance are positively correlated with changes in the external current account with correlation coefficient of 0.35.
  - In 5 out of 7 years (2000–06), the current account responded in the expected direction to changes in the fiscal stance.
- Table 1 (selected series, in percent of GDP, 2000–2006):
  - External current account (underlying): 1/ −4.7 −7.4 −11.0 −9.3 −12.4 −10.9 −12.2
  - Change (external current account): −0.6 −2.7 −3.6 1.7 −3.1 1.5 −1.4
  - General government fiscal balance: −0.2 −1.0 −4.2 −3.0 0.0 0.7 −1.5 0.14
  - Change (fiscal): 1.4 −0.8 −3.1 1.2 3.0 0.8 −2.3 0.35
  - Non-government savings-investment balance: −1.5 −2.4 −3.9 −4.0 −10.2 −10.2 −10.0 0.83
  - Change (non-government): 4.4 −0.9 −1.6 −0.1 −6.1 −0.1 0.2 0.47
  - Note: External current account adjusted in 2004 and 2005 for the introduction of the VAT in 2005.
- Sources: Serbian authorities, and IMF Staff Reports.
- Footnote: State- and socially owned enterprises are part of the nongovernment sector in national accounts.

### Single-Equation Estimates for Serbia (D)
- Estimation setup:
  - Dependent variable: current account (or trade balance).
  - Explanatory variables include fiscal balance, fiscal components (government expenditure and revenue), domestic credit to nongovernment, change in trend GDP, terms of trade, real effective exchange rate (REER), and seasonal dummies.
  - Estimation method: OLS on quarterly data 2000Q1–2007Q3; variables in percent of quarterly GDP except REER and terms of trade.
  - Notes: Unit root tests (ADF) suggest absence of unit roots for all variables. Specification selection by lag inclusion and significance. Specification (2) includes amortization of FFCDs as expenditure and nonbank external borrowing as credit.
- Key estimated elasticities and impacts (Table 2, selected):
  - Fiscal balance (excl. grants): 0.57** (Specification (1)); 0.50** (Specification (2)).
  - Fiscal balance (lagged 1 quarter): 0.50** (Specification (1)).
  - Interpretation: Short-term elasticity of current account to fiscal balance ≈ 0.6 same quarter and 0.5 following quarter → about 1.1 percent of GDP widening after two quarters for a 1 percentage point of GDP fiscal deficit increase (Equation 1).
  - Government expenditure coefficients (Specification (3)):
    - Gov. expenditure: −0.38*** (contemporaneous)
    - Gov. expenditure (lagged 1 quarter): −0.26**
    - Gov. expenditure (lagged 2 quarters): −0.38*
    - Interpretation: government expenditure elasticity large (about −1) acting over three quarters.
  - Government revenue (lagged 2 quarters): 0.53**.
  - Credit to non-government:
    - Contemporaneous: −0.32**
    - 1Q lag: −0.75*** (Spec (1)), −0.60** (Spec (2))
    - 2Q lag: −0.58*** (Spec (1)), −0.50*** (Spec (2)), −0.67** (Spec (3))
    - Interpretation: credit expansion has a negative effect on the current account with elasticities between −0.8 to −1.3 over several quarters in other descriptions.
  - Change in trend GDP:
    - Contemporaneous: −3.18*** (Spec (1)), −1.56** (Spec (2)), −3.15*** (Spec (3))
    - 1Q lag: −1.18** (Spec (3))
    - 2Q lag: 2.28*** (Spec (3))
    - Interpretation: economic growth initially worsens the current account, later improves it.
  - Terms of trade: 0.54** (contemporaneous in one spec); Terms of trade (1Q lag): −0.78***; Terms of trade (2Q lag): 0.21**.
  - REER (real effective exchange rate): −0.11*** (contemporaneous).
- Goodness of fit and diagnostics:
  - Number of observations: 29 for each specification.
  - Adjusted R-squared: 0.76 (Spec (1)), 0.75 (Spec (2)), 0.72 (Spec (3)).
  - Significance markers: ***, **, * denote significance at 1, 5, and 10 percent, respectively.
  - Sample: 2000Q1–2007Q3. Current account excluding grants, corrected for VAT introduction in January 2005.
- Additional note:
  - Using Specification (2) (including FFCD amortization and nonbank external borrowing) reduces estimated elasticity to 0.5 and makes the effect contemporaneous; smaller fiscal impact may reflect smaller demand effect of FFCD amortization compared with other public expenditures.

### Other Determinants and Interpretation
- Domestic credit growth, economic activity (trend GDP changes), REER, and terms of trade significantly affect the current account alongside fiscal variables.
- Credit to nongovernment negatively affects the current account over several quarters.
- Economic growth first reduces the current account (higher demand, lower savings) but later increases it (higher net exports and savings).
- Real appreciation is associated with a deterioration in the current account.
- Terms-of-trade effects vary by specification: in one specification improvement benefits current account after a lag; in another, immediate improvement is followed by deterioration, possibly due to increased demand.

*Source: IMF Staff calculations and IMF Staff Reports.*

### 81.      The current account in Eastern European countries experiencing economic

### _cr0855 - 81.      The current account in Eastern European countries experiencing economic

### Background and motivation
- Transition and convergence in Eastern European economies were associated with large and persistent current account deficits, led in part by high investment and consumption.
- Fiscal stance may be insignificant in driving the current account in these economies, although it is expected to have some impact at the margin.
- Empirical test focuses on the presence of a "twin deficit" relationship (fiscal deficit ↔ current account deficit) in Eastern and Southeastern Europe (ESEE) during transition, reproducing Bartolini and Lahiri (2006) with annual data for 15 countries over 1995–2006.

### Methodology
- Two-step panel analysis:
  - Step 1: Estimate response of private consumption to changes in the general government fiscal balance, controlling for public consumption, public debt, real GDP growth, and population growth.
  - Step 2: Estimate response of the current account to changes in the fiscal balance using the same specification to identify whether investment or private consumption changes transmit fiscal effects to the current account.
- Panel fixed effects estimation. Sample 1995-2006. Variables in percent of GDP or in percent. Significance symbols: ***, **, * denote 1, 5, and 10 percent.

### Key empirical findings for ESEE panel (selected coefficients and interpretations)
- Private consumption regressions (Table 3, Columns 1–4) show differing signs depending on specification:
  - Gen. Gov. Fiscal balance: 0.29*** (Col 1); -0.18* (Col 2); -0.52*** (Col 3)
  - Gen. Gov. Fiscal balance (lagged): 0.44** (Col 3); 0.20* (Col 4)
  - Public consumption: -0.82*** (Col 1); -0.44** (Col 2); -0.63*** (Col 3); -0.50*** (Col 4)
  - Real GDP growth: -0.14* (Col 1); 0.32*** (Col 2); 0.32*** (Col 3); 0.33*** (Col 4)
  - Population growth: -0.83** (Col 1); -0.49 (Col 2); -0.46 (Col 3); -0.32 (Col 4)
  - Gen. Gov. Debt: -0.01 (Col 1); 0.05*** (Col 2); 0.06*** (Col 3); 0.03** (Col 4)
  - Gross capital formation: -0.54*** (Col 4)
- Memorandum items (responses under policy actions):
  - Fiscal balance-increasing policies:
    - Tax increase: 0.29 (Col 1); -0.18 (Col 2); -0.08 (Col 3); 0.20 (Col 4)
    - Expenditure cuts: 1.11 (Col 1); 0.26 (Col 2); 0.55 (Col 3); 0.70 (Col 4)
  - Fiscal balance-neutral policy:
    - Expenditure cuts with tax cuts: 0.82 (Col 1); 0.44 (Col 2); 0.63 (Col 3); 0.50 (Col 4)
- Interpretation:
  - Baseline fixed-effects specification (Column 2): an increase in the deficit (through tax cut, without change in public consumption) reduces private consumption and improves the current account by 0.18, suggesting a “twin divergence” consistent with Ricardian behavior.
  - An increase in public consumption (financed by higher taxes) reduces private consumption and worsens the current account by 0.44; thus a deficit-financed increase in public spending would worsen the current account by only 0.26 (0.44 worsening from public consumption offset by 0.18 improvement from deficit via tax cut).
  - Evidence of strong Ricardian effects in some specifications: a tax cut may actually reduce private consumption (Column 1).
  - Public and private consumption appear to be close substitutes in several specifications.

### Role of investment and specification sensitivity
- Including the lag of the fiscal balance restores a twin deficit relationship in the second year (Column 3).
- Introducing gross capital formation (investment) suggests:
  - A lagged twin deficit relation.
  - Investment is a strong and significant determinant of the current account deficit (Column 4: Gross capital formation coefficient -0.54*** in the private consumption equation and prominence in current account regressions).
- Controlling for other short-term shocks to the current account (e.g., terms-of-trade shocks) does not significantly affect results.

### Serbia-specific results and elasticities (section F: Conclusions)
- For Serbia, controlling for bank credit, growth, the real exchange rate, and terms of trade:
  - Fiscal balance has a strong and significant impact on the current account.
  - Elasticity estimates between 0.5 and 1.1.
  - Baseline specification: an increase in the fiscal deficit by 1 percent of GDP widens the current account deficit by 0.6 percent of GDP in the same quarter and by 0.5 percent of GDP in the following quarter.
  - Decomposed timing: government revenue acts with a two-quarter lag; government expenditure acts both immediately and with lags.
  - Bank credit has a significant negative impact on the current account.

### Overall conclusions
- For the ESEE panel, evidence of a twin deficits relationship is elusive overall; strong Ricardian effects and dominant investment dynamics during transition may explain weak panel-wide twin deficits.
- For Serbia, fiscal policy appears to exert a sizable and timely influence on the current account, with elasticities materially different from the panel average.
- Country-specific studies are useful to assess circumstances explaining weak or strong twin deficit relationships in the region.

### Limitations and cautions
- Serbia data series are relatively short, reducing robustness of econometric estimates.
- Different specifications of key variables yield results that can differ substantially in magnitude (though often not in sign).
- Potential multicollinearity: exogenous variables explaining trade or current account balances may not be independent (e.g., fiscal policy may affect private credit demand).
- Results should be interpreted with caution for forecasting and policymaking.

### Appendix: summary of related empirical literature (select highlights)
- General equilibrium model findings range from current account elasticities of 0.05 to 0.5 for different fiscal compositions and model features; Kumhof and Laxton (2007) find elasticity about 0.6–1 after about 5 years in a model with non-Ricardian features.
- VAR and panel studies report elasticities varying widely:
  - Bartolini and Lahiri (2006): elasticity 0.38 (26-country sample) and 0.30 (18 OECD sample).
  - Bussière, Fratzscher, and Müller (2004): elasticity between 0.06 and 0.25.
  - Chinn and Prasad (2003): long-run elasticity for developing countries about 0.4; insignificant for advanced economies.
  - Time-series country studies report mixed results with average long-run elasticities around 0.4 where twin deficits exist.

*Source: IMF Staff calculations and text from _cr0855 - 81. The current account in Eastern European countries experiencing economic (PDF chapter/section).*

### 92.      The model consists essentially of four equations: (i) an aggregate demand or output

### _cr0855 - 92.      The model consists essentially of four equations: (i) an aggregate demand or output

### Model structure and purpose
- Four core equations:
  - Aggregate demand / output gap (IS curve): relates real activity to expected and past real activity, the real interest rate, the real exchange rate, and foreign activity.
  - Price-setting (Phillips) curve for headline inflation: relates inflation to past and expected inflation, the output gap, the real exchange rate, and the price of oil.
  - Uncovered interest rate parity condition: determines the real exchange rate.
  - Monetary policy rule: sets the policy (nominal) interest rate as a function of real activity and expected inflation.
- Augmentation:
  - Second price-setting equation for core inflation (does not depend directly on oil prices, but includes lagged effect of headline inflation on core inflation).
- Model scope:
  - Two-country model: home country = Serbia; foreign country = euro area. Home is small and open; foreign is large and treated as exogenous to the home country.

### Key equations (summary)
- Output gap equation (ygap):
  - ygap_t = β_ld ygap_{t+1} + β_lag ygap_{t–1} – β_RRgap RRgap_{t–1} + β_zgap zgap_{t–1} + β_EA ygap_{EA,t} – β_RRgapEA RRgap^{EA}_{t-1} + ε^{ygap}_t
  - RRgap^{EA} = RR^{EA}(1 + ResReq_{Serbia}) – RR^{Serbia*}
  - Interpretation: current output gap depends on expected and lagged gaps, lagged real interest rate gap (negative), lagged real depreciation (positive), foreign (euro area) output gap, lagged effective foreign real interest rate (augmented by reserve requirement cost, negative), and a disturbance.
- Headline inflation (π):
  - π_t = α_{πld} π4_{t+4} + (1–α_{πld}) π4_{t–1} + α_{ygap} ygap_{t–1} + α_z (z_t – z_{t-1}) + α_0 π_{rpoil,t} + α_1 π_{rpoil,t–1} + ε^{π}_t
  - Interpretation: inflation is a function of expected inflation (four-quarter ahead y-o-y), lagged inflation, lagged output gap, real depreciation, oil price changes, and a disturbance.
- Core inflation (π_c):
  - π_{c,t} = α_{c,πld} π4_{c,t+4} + (1–α_{c,πld}) π4_{c,t–1} + α_{c,ygap} ygap_{t–1} + α_{c,z} (z_t – z_{t-1}) + α_{c,3}(π4_{t-1} – π4_{c,t-1}) + ε^{πc}_t
  - Last term captures lagged difference headline–core (pass-through from oil into core).
- Real exchange rate (z):
  - z_t = z^e_{t+1} – (RR_t – RR^{EA}_t – ρ^*_t) / 4 + ε^z_t
  - Expected real exchange rate:
    - z^e_{t+1} = δ_z z^*_{t+1} + (1 – δ_z) z_{t–1}
  - Interpretation: uncovered interest parity linking real exchange rate to expected future real exchange rate and real interest differential adjusted for currency risk premium; expectations mix forward- and backward-looking components.
- Monetary policy rule (RS nominal rate):
  - RS_t = γ_{RSlag} RS_{t–1} + (1 – γ_{RSlag}) [RR^*_t + π4_t + γ_π (0.9 π4_{c,t+4} + 0.1 π4_{t+4} – π^*_{t+4}) + γ_{ygap} ygap_t] + ε^{RS}_t
  - Central bank targets core inflation 90 percent and headline 10 percent in the forward inflation gap.

### Implementation and calibration
- Data and forecast horizons:
  - Historical quarterly data: 1999q1 to 2007q4.
  - Baseline forecasts: 2008q1 to 2012q4.
- Output gap construction:
  - Calculated as difference between trend-cycle component of seasonally adjusted GDP and Hodrick-Prescott filtered GDP series, to minimize excessive fluctuations and produce a generally positive output gap in recent quarters and short-term projection period.
- Serbia-specific modeling considerations:
  - High euroization of deposits and loans → strong dependence of domestic activity on euro area real interest rates; introduces RRgap^{EA} term, augmented by reserve requirement cost ResReq_{Serbia}.
  - Parameters calibrated to capture high exchange rate pass-through and the National Bank of Serbia’s focus on core inflation.
- Parameter selection process:
  - Based on experience with other country models; iterative adjustment to minimize model residuals for recent years while acknowledging nonzero residuals due to model simplicity and regime change in mid-2006 (transition to inflation targeting).

### Key parameter choices and numerical values
- Output gap equation:
  - β_lag = 0.5
  - β_ld chosen at mid-point of 0.05–0.15 range → 0.10
  - β_zgap chosen high (reflecting exchange rate channel)
  - β_EA_ygap chosen low (Serbia’s modest openness)
  - β_RRgap set relatively low; β_RRgapEA set relatively high (reflecting weak dinar channel and predominance of foreign interest rates)
- Inflation equations:
  - α_{πld} = 0.3 (forward component for headline inflation)
  - α_{ygap} = 0.25
  - α_z = 0.3 (short-term exchange rate pass-through)
  - Oil price responsiveness: higher than many countries due to large CPI share of oil and oil-related products
  - Core inflation:
    - α_{c,πld} and α_{c,z} set slightly higher than headline counterparts
    - α_{c,3} = 0.25 (lagged headline–core linkage)
- Real exchange rate expectations:
  - δ = 0.4 (mix of forward- and backward-looking expectations; < 0.5 to avoid relying on fully forward-looking Dornbusch overshooting dynamics)
- Monetary policy rule:
  - γ_{ygap} = 0.5
  - γ_π = 2.0
  - γ_{RSlag} = 0.5
  - Central bank target weight: core 90 percent, headline 10 percent in forward inflation gap
- Steady-state (long-run) values used:
  - Inflation target = 4 percent
  - Long-term real interest rate = 4 percent
  - Long-term potential output growth = 5½ percent

### Supply side and potential output
- Potential output exogenous to model; current growth of potential output depends on long-run potential growth and changes in oil prices:
  - 400 (y^*_t – y^*_{t–1}) = g^*_t – υ rpoil π4_{rpoil,t} + ε^{y^*}_t
  - y^* = potential output; g^* = long-term growth of potential output; π4_{rpoil,t} = four-quarter change in relative price of oil; υ parameter captures oil impact.

### Model diagnostics and historical fit
- Historical robustness:
  - Model residuals for 2005q1 to 2007q4 are relatively small overall, suggesting broad ability to replicate recent data but not perfect.
  - Notable deviations where model struggled:
    - Jump in inflation in 2005q1 due to VAT introduction (exogenous shock).
    - Persistence of high inflation in 2006q2 despite strong real appreciation.
    - More rapid-than-predicted disinflation in late 2006 and 2007q1.
    - Spike in inflation in late 2007 despite real appreciation (partly supply shocks).

### Practical implications and usage notes
- Two-country setup: Serbia’s equations incorporate euro area activity and interest rates; euro area treated as exogenous (home country does not affect foreign output gap or inflation).
- Euroization and reserve requirements are explicitly modeled via RRgap^{EA} to capture transmission from foreign rates into Serbian activity.
- Monetary policy interpretation:
  - Central bank reacts aggressively to inflation deviations (γ_π = 2.0) while smoothing interest rates (γ_{RSlag} = 0.5) and paying attention to real activity (γ_{ygap} = 0.5).
- Limitations:
  - Simplified supply side and exogenous potential output.
  - Sensitivity to chosen steady-state values for long-horizon projections.
  - Residuals indicate model cannot capture all idiosyncratic shocks and structural regime changes.

*Source: IMF staff (model description, calibration, and implementation for Serbia and euro area; historical data 1999q1–2007q4; baseline forecasts 2008q1–2012q4).*

### 112.     In the first instance, we use the forecasts of the Serbian authorities or the IMF staff,

### _cr0855 - 112.     In the first instance, we use the forecasts of the Serbian authorities or the IMF staff,

### Baseline analysis and model forecast
- Approach:
  - Use forecasts of the Serbian authorities or the IMF staff (judgmental forecasts) rather than model-produced forecasts for the baseline.
  - The model is used to assess consistency of forecasts and to simulate plausible shocks to assess risks and analyze endogenous policy reaction.
- Judgment on forecasts:
  - Judgmental forecasts are expected to perform better in the short run than purely model-based forecasts for a simple model.
- Baseline narrative (judgmental):
  - Assumes an unwinding in 2008–09 of the tight monetary stance of 2006–07.
  - High interest rates in 2006–07 contributed to significant real exchange rate appreciation and brought inflation down from double-digit rates.
  - Lagged effect of real appreciation on growth will slow growth to or below potential, removing inflationary pressures from excess demand.
  - Gradually lower interest rates will generate a slight nominal and real depreciation, consistent with low inflation while sustaining medium-term growth.
- Selected baseline figures (Annual average in percent, unless otherwise indicated; Source: IMF staff projections; Table 3):
  - Short-term interest rate: 2007 = 10.5; 2008 = 9.2; 2009 = 8.6; deviation (Model vs Baseline) for 2008 = 4.2 p. points; for 2009 = 1.7 p. points
  - RPI inflation: 2007 = 6.5; 2008 = 9.3; 2009 = 6.4; deviation for 2008 = -0.4 p. points; for 2009 = -1.5 p. points
  - Core RPI inflation: 2007 = 3.8; 2008 = 6.1; 2009 = 3.5; deviation for 2008 = 0.5 p. points; for 2009 = 1.2 p. points
  - Exchange rate (€/Din.): 2007 = 0.0125; model deviation percent for 2008 = 3.3; for 2009 = 6.2 percent
  - Real GDP growth: 2007 = 7.0; 2008 = 5.8; 2009 = 5.6; deviation for 2008 = -1.4 p. points; for 2009 = -0.3 p. points
  - Output gap: 2007 = 0.7; 2008 = 0.5; 2009 = 0.2; deviation for 2008 = -1.3 p. points; for 2009 = -1.7 p. points
- Model vs baseline implication:
  - Purely model-based forecast suggests need for monetary tightening in face of rising inflation.
  - Model projects somewhat higher short-term interest rates to maintain a real interest rate differential with the euro area, supporting a more appreciated exchange rate to reduce headline and core inflation convergence—at cost of lower growth than baseline.

### Risk analysis — overview
- Purpose:
  - Simulate a series of plausible shocks to assess risks to the baseline and the impact of the endogenous policy reaction.
- Main shock categories analyzed:
  - Real exchange rate appreciation (dinar appreciation)
  - Real exchange rate depreciation (dinar depreciation)
  - Negative interest rate shock in Serbia (excessive tightening)
  - Oil price shock
  - Fiscal or domestic demand shock
  - Foreign demand shock
  - Reduction in the inflation target
  - Domestic inflation shock (one-off increase in energy or utility tariffs)
  - Euro area interest rate shock
  - Increase in reserve requirements
  - Productivity or positive supply shock

### Real exchange rate appreciation (Table 4)
- Shock characterization:
  - Modeled as temporary exogenous decline in currency risk premium leading to a one-quarter nominal exchange rate appreciation of 10 percent.
  - High pass-through reduces headline and core inflation significantly.
- Policy and macro effects:
  - Central bank lowers interest rates.
  - Real appreciation negatively impacts growth short run and creates a persistently negative output gap.
  - Because reaction function places low weight on output gap, model does not call for aggressive rate cuts to boost growth; real appreciation persists.
- Selected deviations from baseline (Table 4; Deviation from baseline, in percentage points or in percent; Source: IMF staff projections):
  - Short-term interest rate (2008, 2009): 9.2, 8.6; deviations 2008 = -1.6 p. points, 2009 = -1.0 p. points
  - RPI inflation (2008, 2009): 9.3, 6.4; deviations 2008 = -1.3 p. points, 2009 = -0.2 p. points
  - Core RPI inflation (2008, 2009): 6.1, 3.5; deviations 2008 = -0.9 p. points, 2009 = -0.3 p. points
  - Exchange rate (€/Din.) deviations: 2008 = 7.5 percent, 2009 = 2.6 percent (appreciation); modeled path shows persistent appreciation
  - Real GDP growth (2008, 2009): 5.8, 5.6; deviations 2008 = -0.7 p. points, 2009 = -0.2 p. points
  - Output gap (2008, 2009): 0.5, 0.2; deviations 2008 = -0.7 p. points, 2009 = -0.5 p. points

### Real exchange rate depreciation (Table 4)
- Shock characterization:
  - Exogenous increase in risk premium causing nominal exchange rate to depreciate by 10 percent for one quarter.
- Policy and macro effects:
  - Inflation rises in short run, triggering monetary response via higher interest rates.
  - Real depreciation provides short-term boost to growth and a lasting positive output gap due to one-time increase in output.
  - After a few quarters shock unwinds; monetary response (vigorous) brings nominal and then real exchange rates to appreciate, slowing inflation and growth back to baseline.

### Negative interest rate shock in Serbia (Table 4)
- Shock characterization:
  - Temporarily higher nominal interest rate than baseline (e.g., misjudged tightening): nominal interest rate increases by 3 percentage points in one quarter, then follows model reaction function.
- Policy and macro effects:
  - Exchange rate appreciates in nominal and real terms; inflation drops; growth slows.
  - Central bank reduces interest rates below baseline after three quarters; it takes two years for growth to return to baseline and for output gap to close.
  - Rational expectations moderate impact because agents anticipate central bank return to reaction function.

### Oil price shock (Table 5)
- Shock characterization:
  - Price of oil rises by 50 percent relative to baseline during 2 quarters, then returns to baseline over next 1.5 years.
  - Small pass-through to core inflation (by assumption).
- Policy and macro effects:
  - Headline inflation rises; central bank raises interest rates during 6 quarters.
  - Exchange rate depreciates on impact because inflation rises more than nominal rates; later appreciates above baseline after two quarters as inflation is contained and real rates rise.
  - Potential output growth drops due to direct oil price effect, reducing output for about two years.
  - Monetary policy response is unwound during the second year through significantly lower interest rates than baseline; subsequent drop in oil prices and lower rates boost output to close output gap.
- Selected deviations from baseline (Table 5; Annual average; Source: IMF staff projections):
  - Short-term interest rate (2008, 2009): 9.2, 8.6; deviations 2008 = 1.8 p. points, 2009 = -0.8 p. points
  - RPI inflation (2008, 2009): 9.3, 6.4; deviations 2008 = 2.7 p. points, 2009 = -0.2 p. points
  - Core RPI inflation (2008, 2009): 6.1, 3.5; deviations 2008 = 0.1 p. points, 2009 = 0.0 p. points
  - Exchange rate (€/Din.) deviations: 2008 = -0.2 percent, 2009 = 3.1 percent (initial depreciation then appreciation)
  - Real GDP growth (2008, 2009): 5.8, 5.6; deviations 2008 = -0.5 p. points, 2009 = -1.1 p. points
  - Output gap (2008, 2009): 0.5, 0.2; deviations 2008 = -0.1 p. points, 2009 = -1.0 p. points

### Fiscal or domestic demand shock (Table 5)
- Shock characterization:
  - Temporary increase in domestic demand (e.g., expansionary fiscal policy).
  - Output gap increases by 2 percentage points for one quarter, phased out gradually by 85 percent per quarter.
- Policy and macro effects:
  - Due to low parameters on output gap → inflation impact and low weight on output gap in reaction function, effect on inflation and nominal interest rates is relatively benign.
  - Growth increases temporarily; central bank raises nominal interest rates above baseline.
  - Exchange rate appreciates, keeping inflation down but causing sharp slowdown in growth to below potential after about a year—illustrates how surge in domestic demand can eventually choke off growth via sharp real appreciation.

### Foreign demand shock (Table 5)
- Shock characterization:
  - Euro area output gap increases by 1 percentage point for one quarter, phased out by 85 percent per quarter (e.g., stronger export demand from euro area).
- Policy and macro effects:
  - Temporary increase in foreign demand raises domestic growth mildly short term.
  - Without increased potential output, inflationary pressures rise and central bank raises interest rates.
  - Most impact runs through euro area response: ECB raises interest rates to dampen euro area demand shock, causing dinar depreciation due to drop in real interest rate differential; exchange rate channel transmits inflationary and demand pressures to Serbia prompting significant monetary tightening.

### Reduction in inflation target (Table 6)
- Shock characterization:
  - Permanent reduction in numerical value of targeted inflation.
- Policy and macro effects:
  - In forward-looking model, rapid disinflation occurs within a few quarters.
  - Rapid disinflation increases real interest rate and causes nominal and real exchange rate appreciation, resulting in a drop in growth.
  - Central bank reduces nominal rates only gradually; growth returns to baseline after three years, but cumulative output loss is significant.

### Domestic inflation shock (Table 6)
- Shock characterization:
  - Temporary increase in non-core prices (one-off increase in energy or utility tariffs).
  - Footnote: Headline inflation increases by 3 percentage points (year-on-year) above the baseline for one quarter.
- Policy and macro effects:
  - Central bank gradually raises interest rates because reaction function includes headline inflation and anticipates feedback to core inflation.
  - Real interest rates rise; real exchange rate appreciates; GDP growth slows.
  - Growth recovers to baseline after three years, leaving a significant and persistent negative output gap.
- Selected deviations from baseline (Table 6; Annual average; Source: IMF staff projections):
  - Short-term interest rate (2008, 2009): 9.2, 8.6; deviations 2008 = -0.2 p. points, 2009 = -1.8 p. points (for Euro area interest rate shock) and other scenarios listed in table
  - RPI inflation (2008, 2009): 9.3, 6.4; deviations for reduction in inflation target and domestic inflation shock reflected in table entries (e.g., domestic inflation shock RPI deviation 2008 = 3.9 p. points, 2009 = 1.7 p. points)
  - Core RPI inflation, exchange rate, real GDP growth, output gap deviations as reported in Table 6 for each scenario.

### Euro area interest rate shock (Table 6)
- Shock characterization:
  - Temporary increase in euro area interest rates by 2 percentage points for two quarters.
- Policy and macro effects:
  - Leads, through exchange rate depreciation caused by lower real interest rate differential, to slight rise in Serbian inflation requiring some monetary tightening.
  - Euroized domestic environment effects also lead to lower growth.

### Increase in reserve requirements (Table 6)
- Shock characterization:
  - Permanent increase in reserve requirements above baseline.
- Policy and macro effects:
  - Model suggests some slowdown in growth.
  - Central bank does not react significantly with interest rates.
  - Slowdown in real growth partially compensated by exchange rate depreciation.

### Productivity shock / positive supply shock
- Effect:
  - Increase in potential output translates into increase in growth; no other variables affected in this model due to absence of imbalances.

### Appendix I: Definition of variables (selected entries)
- ygap: output gap (y_t – y_t^*), percentage points
- y_t: log of real GDP
- y_t^*: log of potential real GDP
- g_t^*: growth rate of potential GDP, quarter/quarter at annual rate, percentage point
- π: CPI inflation, quarterly at annualized rate, percentage points
- π^*: target inflation rate, annualized rate, percentage points
- π_roil,t: change in the relative price of oil, quarterly at annualized rate, percentage points
- RS: nominal interest rate, in percentage points
- RR: real interest rate, in percentage points
- z: log of the real exchange (an increase implies a depreciation)
- zgap: real exchange rate gap (z – z^*), percentage points
- S: nominal exchange rate, value of foreign currency in local currency

### Appendix II: Data
- Historical quarterly data for Serbia and the euro area: 1999q1 to 2007q4.
- Forecasts: 2008q1 to 2012q4 (IMF staff projections in biannual World Economic Outlook).
- Serbia data sources and definitions:
  - GDP: constant prices of 2002, source Serbian Statistics Office (SSO), seasonally adjusted.
  - Ygap: Difference between trend GDP (trend-cycle term in seasonal adjustment) and potential GDP (HP filter with no priors).
  - CPI: Retail price index, SSO, seasonally adjusted.
  - CPI_X: Core retail price index; includes only 48 percent of full CPI; main excluded items energy, utilities, public transportation; calculated by IMF staff, seasonally adjusted.
  - Interest rate: Before 2005 weighted average of interest rate on NBS bills; from 2005 onward weighted average of interest rate on NBS repo operations; source NBS. Equilibrium interest rate is HP-filtered series.
  - Exchange rate: Dinar/euro, source NBS. Equilibrium exchange rate is HP-filtered.
  - Price of oil: IMF WEO price of oil (simple average of UK Brent, Dubai, WTI spot prices) in dollars converted to domestic currency; source IMF staff.
- Euro area data: IMF staff based on Eurostat and WEO forecasts. Potential GDP calculated using enhanced HP filter. Euro area nominal interest rate is ECB’s main refinancing rate.

### VII. The fiscal impact of privatization — objective, results, policy implications
- Objective:
  - Assess potential budgetary impact of privatization using sample of 30 socially owned enterprises privatized in Serbia in 2005.
- Main results:
  - Fiscal contribution of privatization—via increased taxes and social security contributions, reduced government subsidies and arrears to public utilities, and interest savings—can be sizeable.
- Policy implications:
  - Privatization can raise efficiency of companies and the economy and have significant positive impact on fiscal performance.

### Analysis framework for fiscal impact (Section A)
- Quantification approach:
  - Observe financial flows in 2003–07 between public sector and sample of 30 companies privatized in 2005.
  - Fiscal contribution flows (annual differences): Δ Fiscal contribution = Δ Tax revenue – Δ Subsidies – Δ Budget financing costs – Δ Quasi-fiscal deficits.
- Components:
  - Direct tax revenues: corporate income tax (CIT), wage tax, social security contributions (SSC), property tax.
  - Subsidies: mostly soft credits with interest rates below market rates.
  - Privatization revenues, if saved, reduce budget financing costs: sales proceeds transferred to budget can permanently lower stock of Treasury bills; savings measured by applying historic period T-bill rate on stock reduction (impact measured by interest saved).
  - Quasi-fiscal deficits: mainly arise from arrears towards public utility companies; measured by annual change in arrears of sample towards EPS (data on other utilities not available).

*Source: IMF staff projections and analysis as presented in the provided IMF chapter text.*

### 129.     The total fiscal contribution of the sample improved significantly during the

### _cr0855 - 129.     The total fiscal contribution of the sample improved significantly during the

### Fiscal contribution summary (Table 1)
- Total estimated fiscal contribution (RSD million):
  - 2003: 778.6
  - 2004: 675.0
  - 2005: 1,504.3
  - 2006: 2,549.7
  - 2007 H1: 1,368.1
- In percent of GDP:
  - 2003: 0.066%
  - 2004: 0.047%
  - 2005: 0.086%
  - 2006: 0.120%
  - 2007 H1: 0.113%
- Component breakdown (RSD million):
  - 1.1. CIT:
    - 2003: 3.5
    - 2004: 1.8
    - 2005: 3.4
    - 2006: 88.9
    - 2007 H1: 123.9
  - 1.2. Wage Tax:
    - 2003: 234.8
    - 2004: 191.1
    - 2005: 290.7
    - 2006: 456.5
    - 2007 H1: 245.5
  - 1.3. Contributions:
    - 2003: 551.0
    - 2004: 479.0
    - 2005: 773.4
    - 2006: 1,131.0
    - 2007 H1: 643.1
  - 1.4. Property Tax:
    - 2003: 7.8
    - 2004: 7.8
    - 2005: 21
    - 2006: 19.5
    - 2007 H1: 12.3
  - 2. Subsidies:
    - 2003: 18.5
    - 2004: 10
    - 2005: 11.5
    - 2006: 0
    - 2007 H1: 0
  - 3. Budget financing costs 1/:
    - 2003: 0
    - 2004: 0
    - 2005: -439.3
    - 2006: -840.1
    - 2007 H1: -335.4
  - 4. Quasi-fiscal deficit 2/ 3/ ...:
    - 2003: -5.3
    - 2004: 12.0
    - 2005: -13.7
    - 2006: -7.9
- Notes on calculations:
  - Discount rates used: 15.50%, 14.44%, 6.01%. Those are multiplied by the annual stock of cash accumulated from privatization of our sample during the period of observation.
  - Quasi-fiscal deficit: First 9 months in 2007.
  - Corrected in 2005 for debt-write offs of 2.7 million RSD.
- Data sources: Ministry of Finance, Tax Administration, Ministry of Economy and Regional Development, Privatization Agency, Share fund, Statistical Office, and staff calculations.

### Corporate income tax (CIT) performance (Table 2)
- Operating revenue (RSD million):
  - 2003: 11,298
  - 2004: 13,526
  - 2005: 16,468
  - 2006: 23,739
  - 2007 H1: ...
- Operating revenue (in percent of GDP):
  - 2003: 0.96%
  - 2004: 0.95%
  - 2005: 0.94%
  - 2006: 1.12%
  - 2007 H1: ...
- CIT Serbia (RSD million):
  - 2003: 5,933
  - 2004: 6,922
  - 2005: 10,307
  - 2006: 18,313
  - 2007 H1: 17,273
- CIT sample / CIT Serbia (percent):
  - 2003: 0.06%
  - 2004: 0.03%
  - 2005: 0.03%
  - 2006: 0.49%
  - 2007 H1: 0.72%
- Observations:
  - Corporate income taxes collected from the sample rose strongly after privatization.
  - Sample firms increased output and profitability with rationalization and measures to improve operating efficiency.
  - The CIT rate was lowered in 2004 from 14 to 10 percent.

### Wage taxes and social security contributions (SSC) (Table 3)
- SSC sample (per capita, 000 RSD):
  - 2003: 64.3
  - 2004: 58.7
  - 2005: 103.4
  - 2006: 170.3
- SSC Serbia (per capita, 000 RSD) 1/:
  - 2003: 27.8
  - 2004: 35.6
  - 2005: 42.2
  - 2006: 53.7
  - Note: Excluding the agricultural producers pension fund.
- Employment (annual average):
  - Employed sample:
    - 2003: 8,569
    - 2004: 8,158
    - 2005: 7,478
    - 2006: 6,639
  - Employed Serbia 2/ 3/:
    - 2003: 2,041,395
    - 2004: 2,050,854
    - 2005: 2,068,964
    - 2006: 2,025,627
    - Note: Excluding agriculture.
- Observations:
  - Total collections of wage taxes and SSC from the sample doubled in the observed period, although employment numbers fell by 23 percent.
  - Each employee of the sample contributed 3 times more to social security in 2006 than the average Serbian worker.

### Subsidies, privatization receipts, and budget financing costs
- Government subsidies to the sample ceased after privatization.
  - Before privatization, subsidies distributed to the sample were higher than the CIT collected from it.
- Privatization sales:
  - The sample companies were sold for a total of €89.1 million.
  - Cash collections (installments reduced cash received):
    - 2005: €34.2 million
    - 2006: €34.9 million
    - 2007 H1: around €0.2 million
- Impact on domestic financing costs:
  - Receipts reduced domestic financing costs by:
    - €5.3 million in 2005
    - €10.0 million in 2006
    - €4.2 million in the first half of 2007

### Arrears toward EPS (Table 4)
- Stock of arrears towards EPS (Sample Companies) (RSD million), end of year stock:
  - 2003: 32.3
  - 2004: 27.0
  - 2005: 36.3
  - 2006: 22.6
  - 2007 H1: 14.7
- Debt write-offs:
  - 2003: 0.0
  - 2004: 0.0
  - 2005: 2.7
  - 2006: 0.0
  - 2007 H1: 0.0
- Observations:
  - Privatization appears to have improved payment discipline, visible in the drop in the stock of arrears toward EPS, except during the year of privatization (2005), when arrears and debt write-offs spiked.

### Concluding remarks and extrapolation
- Main finding:
  - Privatization of the sampled firms is associated with sizeable fiscal savings and higher tax and contribution collections.
- Linear extrapolation:
  - Extrapolating the sample results for these 30 companies to the total population of enterprises privatized in 2005 of about 200 companies would result in a potential fiscal contribution of about ¾ of one percent of GDP in 2006.
- Caveats and limitations:
  - Representativeness: The calculation assumes the sample is representative of the total pool of privatized companies, which may not be the case, particularly regarding increased efficiency post-privatization. Expanding the sample and observation period would improve assessment.
  - Direct vs. secondary effects: The analysis captures only the direct impact of privatization on the tax base; there are indications of secondary effects (contracting out to private companies) that could further boost tax payments.
  - Subsidy measurement: The reduction in subsidies may be underestimated because the data only captures distributions through the Development Fund; subsidies from other government levels and institutions are not fully captured.
  - Quasi-fiscal activities: Comprehensive quantification was not possible due to lack of readily available data. The analysis excludes arrears to locally owned companies (mainly utilities) and to the state-owned oil company NIS.

*Sources: Ministry of Finance; Tax Administration; Ministry of Economy and Regional Development; Privatization Agency; Share fund; Statistical Office; Solvency Centre; EPS; staff calculations.*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2008/_cr0855.pdf_