## _wp0554

## Source details

**Canonical URL:** [_wp0554](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2005/_wp0554.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2005/_wp0554.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2005/_wp0554.pdf.json)

---

### Introduction — themes and objectives
- Over the last few decades, countries moved toward greater decentralization of fiscal revenue and spending responsibilities, making subnational policies central to macroeconomic stability.
- The paper assesses the impact on subnational budget deficits of different approaches to controlling subnational government borrowing and attempts to identify determinants of the effectiveness of each institutional arrangement.
- Structure:
  - Section II: sources of subnational fiscal indiscipline, classification of borrowing controls, literature review.
  - Section III: empirical approach and results for long-run and short-run dynamics.
  - Section IV: conclusions.

### A. Sources of subnational fiscal indiscipline
- Common pool problem
  - Separation of costs and benefits: jurisdictions benefit while paying only a small fraction of costs, leading to excessive spending and competition for federal transfers.
  - Regions may signal high need via large deficits or unsustainable debts to attract central grants.
- Soft budget constraints
  - Transfers based on ex post needs and perceived bailouts create soft budget constraints and weaken lenders’ incentives to police regional governments.
  - Default by one region can raise borrowing costs for others and create neighbor-bailout incentives.
- Interregional competition
  - Competition for mobile capital and labor can encourage inefficiently low taxes or excessive borrowing-financed infrastructure.
  - Under shared tax bases and multi-tier tax-setting, excessive taxation can result (Keen, 1998; Keen and Kotsogiannis, 2002).
- Unfunded federal mandates
  - Mandates without adequate federal funding create expectations of supplementary transfers and reduce subnational expenditure-management incentives.
- Short electoral cycles
  - Short horizons induce pre-election spending increases or tax cuts, with stronger effects at subnational level where local performance matters for reelection (Besley and Case (1995)).

### B. Role of vertical fiscal imbalances (VFI)
- Definition/measurement:
  - VFI measured as the ratio of transfers to total subnational government revenue.
- Effects:
  - High VFI exacerbates moral hazard, common pool problems, and bailout expectations.
  - High proportions of regional spending financed from common pool reduce local accountability and facilitate arguments blaming the central government for deficits.
  - Examples of tax-competition distortions under high VFI: China, Russia, United States.

### C. Classification of controls for subnational borrowing
- Four broad approaches (following Ter-Minassian (1997)):
  - Market discipline
    - No central limits; local governments free to borrow; examples: Canada, Switzerland, United States.
    - Preconditions: markets free/open; adequate borrower information; no perceived bailout chance; institutions responsive to market signals (Lane (1993)).
  - Administrative constraints
    - Central government directly controls borrowing (examples: Lithuania since 2001; Mexico; India; Bolivia; Latvia; Indonesia).
    - Advantages: strict control, preserves central fiscal flexibility, coordinates external borrowing.
    - Drawbacks: implicit/explicit guarantees can lead to bailouts; central authorities may lack local information.
  - Rule-based controls
    - Standing constitutional/legal rules: deficit limits (Austria, Spain), operating deficit restrictions (Norway), debt-servicing indicators (Spain, Japan, Brazil, Korea), accumulated debt limits (Hungary), spending restrictions (Belgium, Germany), "golden rules" for investment-only borrowing (Germany).
    - Advantages: clarity, transparency, monitoring ease.
    - Disadvantages: trade-off compliance vs flexibility; evasion risks (reclassification, off-budget accumulation, sale-and-lease-back, hidden debts).
  - Cooperative arrangements
    - Negotiated borrowing controls with subnational co-responsibility; example: Australian loan council (established 1929).
    - Strengths: dialogue, coordination, flexibility; weaknesses: hybrid flaws can soften constraints and hamper coordination if poorly implemented.

### D. Empirical literature — summary
- Case studies suggest:
  - Ter-Minassian and Craig (1997): sole reliance on market discipline unlikely; rule-based preferable to administrative for transparency; cooperative arrangements promising.
  - Ter-Minassian, Albino-War, and Singh (2004): self-imposed fiscal rules associated with better outcomes where bailout experience is limited.
  - Rodden and Eskeland (2003): containment requires strong hierarchical oversight or strong market mechanisms.
  - Rattsø (2002): decentralized governments can achieve stability through diverse means.
- Econometric evidence mixed:
  - von Hagen and Eichengreen (1996): introduction of borrowing constraints associated with increased subnational indebtedness (causality unclear).
  - Fornasari, Webb, and Zou (2000); Jin and Zou (2002): constraining borrowing did not show consistent effects.
  - Rodden (2002): largest deficits where reliance on federal transfers is high and borrowing is free.
- Methodological note:
  - Early studies used presence/absence dummies; Rodden used quasi-continuous index (1 to 5) to capture degree of autonomy but not specific control types.

### III. Empirical analysis — approach and data
- Objectives:
  - Cross-sectional analysis: impact of institutional features on average long-run subnational and consolidated fiscal balances.
  - Panel (dynamic) analysis: short-run responses and central government influence in a given year.
- Data:
  - Newly collected dataset on subnational borrowing regimes and fiscal outcomes for 44 countries, 1982–2000; primary source IMF Government Financial Statistics (GFS).
  - Countries classified by predominant borrowing-control approach; regime changes treated as separate cross-sectional units where applicable.
  - A country is coded as having bailout history if significant past bailouts occurred and expectations of future bailouts were unaltered by reform.
- Summary counts (Table 1: Subnational Borrowing Controls: Summary — Number of countries)
  - Unrestricted: 13 total; Emerging 5; Industrial 8; With Bailout History 6; Without Bailout History 7
  - Self-imposed rules: 3 total; Emerging 1; Industrial 2; With Bailout History 1; Without Bailout History 2
  - Centrally imposed rules: 12 total; Emerging 6; Industrial 6; With Bailout History 4; Without Bailout History 8
  - Administrative: 15 total; Emerging 13; Industrial 2; With Bailout History 7; Without Bailout History 8
  - Cooperative: 9 total; Emerging 2; Industrial 7; With Bailout History 4; Without Bailout History 5
  - Total: 52 total; Emerging 27; Industrial 25; With Bailout History 22; Without Bailout History 30

### Control variables and hypothesized effects
- Degree of decentralization (share of subnational expenditure in general government expenditure): higher decentralization may harm macro coordination or reflect shifted central burdens.
- Common standards of public expenditure management (PEM): dummy for presence expected to facilitate monitoring and enforcement.
- Governance features: corruption index, central bank independence (Cukierman (1992)).
- Cost of debt service: real interest rates, inflation, debt service ratio.
- Macroeconomic and political variables: unemployment rate, parliamentary democracy dummy, regional representation in upper chamber, election dummies, Herfindahl political fractionalization, elected subnational authorities dummy, ethno-linguistic fractionalization.
- Demographics and development: population density, dependency ratio, GDP per capita, emerging vs industrial dummy.
- Size and federal structure: government expenditure-to-GDP, area, federation dummy, average size of subnational jurisdictions, horizontal imbalances (not estimated due to data limits).

### Long-run cross-sectional specification
- Marginal effect of restricting subnational borrowing modeled as dependent on VFI:
  - αij = α0j + α1j * vfi_i
- Basic linear specification for subnational fiscal balances:
  - SS_it = β0 + α01 self_i + α02 coop_i + α03 central_i + α04 admin_i + β1 vfi_it + α11 self_i * vfi_it + α12 coop_i * vfi_it + α13 central_i * vfi_it + α14 admin_i * vfi_it + β2 CS_it + γ bailout_i + δ controls_it + ε_it
  - Definitions preserved: SS = subnational surplus-to-total revenue; CS = central government surplus-to-total central government revenue; regime dummies as defined.

### Key cross-sectional empirical findings (selected)
- Bailout history associated with statistically significant weaker subnational fiscal performance:
  - "Consistent with our priors, the presence of a bailout history is associated with a statistically significant weaker subnational fiscal performance (5.4 percentage points higher subnational"
- Selected coefficients from Table 2 (Dependent variable: Subnational Surplus to Revenue (SS); standard errors in parentheses; significance: * 10%, ** 5%, *** 1%):
  - CS (Column 1): 0.396 (0.101)***
  - Centrally imposed rule (Column 1): -0.089 (0.047)*
  - Cooperative controls (Column 1): -0.147 (0.066)**
  - Interaction VFI * central (Column 1): 0.219 (0.101)**
  - Interaction VFI * coop (Column 1): 0.201 (0.112)*
  - Bailout (Column 1): -0.054 (0.020)***
  - Observations/Groups/R2 (between): Column 1: 584 obs; 52 groups; R2 = 0.57
  - IV two-stage and subgroup columns broadly confirm qualitative patterns with some differences noted in text.

### Marginal effects of regimes and VFI (cross-section)
- Self-imposed fiscal rules outperform centrally imposed rules only at low VFI.
- Centrally imposed rules and cooperative arrangements’ marginal effects increase rapidly as VFI widens (positive and relatively large interaction coefficients).
- Administrative approach interaction positive but small and statistically insignificant in cross-section.
- Thresholds from marginal-effect analysis:
  - Self-imposed rules appear optimal at VFI below 16 percent (area includes only five countries).
  - Centrally imposed fiscal rules perform best at VFI above 47 percent (covers 15 countries).
  - Administrative controls seem superior between 16 percent and 47 percent.
- Cooperative arrangements and unrestricted borrowing never outperform alternatives in cross-section evidence.
- All marginal effects measured in percentage points of total subnational government revenue.
- Common PEM standards:
  - Success of cooperative arrangements depends on presence of common standards; PEM * coop interaction large and statistically significant (Column 3).

### Differences: industrial vs emerging economies
- Average subnational fiscal balance:
  - Industrial countries = -2.3 percent
  - Emerging economies = -4.2 percent
- Emerging economies more often use administrative rules; industrial countries predominantly use cooperative arrangements.
- For centrally imposed rules (6 emerging, 6 industrial in sample), centrally imposed rules start outperforming other approaches at substantially lower VFIs in emerging economies and their marginal effect grows faster as VFI widens.
- In presence of bailouts, cooperative arrangements’ performance deteriorates significantly.
- Estimate: a bailout history is estimated to increase the deficit by only 1.4 percent of total revenue at the mean level of VFI (statistically highly insignificant).

### Endogeneity concerns and two-stage estimation
- Potential endogeneity: regime choice may be influenced by average subnational fiscal balances.
- Two-stage estimator: first-stage multinomial logit (included GDP, federation dummy, VFI, constant; N = 52; Pseudo R2 = 0.29); predicted probabilities used in second stage.
- Two-stage results (Table 2, Column 2) largely confirm qualitative results.
- Distinctions: unrestricted borrowing associated with inferior outcomes in two-stage; positive marginal impact of cooperative arrangements decreases when VFI widens.

### Short-run dynamic panel specification and identification
- Panel model:
  - SS_it = α_i + φ SS_it-1 + β dec_it + δ vfi_it + λ CS_it + μ Controls_it + ε_it
  - Fixed effects α_i absorb time-invariant country characteristics.
- Estimation: Arellano and Bond (1991) GMM — first-differencing and instrumenting with higher-order lags to handle dynamic bias and endogenous regressors.
- Interaction specifications identify differential short-run responses by bailout history and by borrowing regime.

### Short-run empirical findings (selected from Table 3; standard errors in parentheses; significance: * 10%, ** 5%, *** 1%)
- Model A (Bailouts):
  - 1st lag SS: 0.266 (0.077)***
  - VFI * Bailouts: 0.113 (0.070)
  - VFI * No Bailouts: 0.154 (0.037)***
  - CS * Bailouts: 0.031 (0.048)
  - CS * No Bailouts: 0.093 (0.032)***
  - Interpretation: Without bailouts, federal transfers and central government fiscal balances have highly significant positive impacts on subnational fiscal outcomes; estimated gap-closing effect of federal transfers is 0.154, indicating most transfer money used to inflate expenditure rather than close deficits. With bailout history, responsiveness to transfers and CS drops and is statistically insignificant.
- Model B (Borrowing Regimes):
  - 1st lag SS: 0.255 (0.072)***
  - VFI * No Rules: -0.069 (0.084)
  - VFI * Coop: 0.116 (0.169)
  - VFI * Central: 0.106 (0.052)**
  - VFI * Admin: 0.180 (0.041)***
  - CS * No Rules: 0.136 (0.044)***
  - CS * Coop: 0.134 (0.049)**
  - CS * Central: 0.100 (0.049)**
  - CS * Admin: -0.018 (0.025)
  - Decentralization (dec): -0.291 (0.088)*** in Model A; -0.304 (0.105)*** in Model B
  - GDP (log, PPP): -0.122 (0.049)** in Model A; -0.155 (0.050)*** in Model B
  - Population density: 0.0006 (0.0003)* in Model A; 0.0008 (0.0002)*** in Model B
  - Number of observations: 480; Number of groups: 49
  - Tests of no 2nd order autocorrelation: Model A Z = -0.95 (p = 0.341); Model B Z = -1.14 (p = 0.255)
  - Interpretation: Centrally imposed rules achieve highest short-run fiscal coordination — positive, large, and statistically significant interaction terms with VFI and CS. Administrative rules deliver large positive VFI interaction but weak coordination with CS.

### Interpretation and policy-relevant implications from empirical analysis
- Bailouts undermine short-run fiscal discipline: presence of bailouts erodes the effectiveness of central instruments (transfers, CS) on subnational fiscal outcomes.
- Centrally imposed borrowing rules support stronger short-run fiscal coordination: both federal transfers (VFI) and central government surpluses (CS) positively affect subnational fiscal balances when centrally imposed rules are in place.
- Administrative controls can yield positive marginal effects of transfers (VFI * Admin = 0.180 (0.041)***), but coordination with central fiscal balances is weak (CS * Admin = -0.018 (0.025), not significant).
- Decentralization associated with more negative subnational surpluses to revenue in short run (dec coefficients -0.291 and -0.304, both significant at 1 percent).
- Larger GDP (log, PPP) associated with more negative subnational surpluses in these specifications.
- Population density positively associated with subnational surpluses at conventional significance levels.
- Policy takeaway: to enhance subnational fiscal discipline and make central instruments effective in the short run, avoid bailout expectations and consider centrally imposed borrowing rules where appropriate; ensure common PEM standards to improve cooperative arrangements.

### Broader conclusions and policy recommendations
- No single institutional arrangement is superior in all circumstances; optimal design depends on:
  - degree of vertical fiscal imbalances,
  - existence of bailout precedents,
  - quality of fiscal reporting.
- Specific recommendations:
  - Unconstrained borrowing authority for subnational governments is unlikely optimal.
  - At low VFI, self-imposed fiscal rules by subnational governments appear to lead to better outcomes.
  - As VFI widens, the positive effect of self-imposed rules declines; centrally imposed fiscal rules become preferable in short and long runs, especially in emerging economies.
  - For high VFI, administrative procedures may provide tighter short-run control but implicit guarantees tied to these controls undermine long-run discipline.
  - Adoption of common standards of financial reporting is crucial for success of cooperative arrangements and may increase effectiveness of centrally imposed rules.
  - Central governments should avoid bailouts whenever possible; when considering bailouts, weigh short-term benefits against long-term negative consequences. In presence of bailout experiences, centrally imposed fiscal rules seem most effective.

### Institutional design considerations and limits
- Fiscal rules vary in form, coverage, and enforcement; enforcement mechanisms include market discipline, cooperative sanctions, higher-level punishments, and bankruptcy procedures for subnational governments.
- Transfer system design affects incentives; identical VFI levels may capture different incentives across countries.
- Data limitations prevented analysis of many institutional interactions; further research needed.

### Data sources and appendices (selected)
- Central and subnational budgets: IMF Government Finance Statistics, 2002.
- Macroeconomic indicators: World Bank, World Development Indicators (2004).
- Subnational jurisdiction structure: World Bank (2000), Table A.1.
- Corruption indices: Transparency International.
- Political indicators: World Bank Database of Political Indicators (2000).
- Federal/unitary classification: Treisman (2002).
- Characterization of fiscal institutions, borrowing constraints, bailout history: Ter-Minassian (1997); Inter-American Development Bank (1997); Bird and Vaillancourt (1999); OECD (2000); Dabla-Norris and Wade (2002), Table 9; Dafflon (2002); OECD (2003); Rodden, Eskelund, and Litvak (2003); IMF staff estimates.
- Additional empirical appendices: Table 4 (Subnational Borrowing Controls Across the World), Table 5 (Statistically Insignificant Control Variables), Table 6 (Performance of Different Regimes in the Presence of Bailout History), Appendix: Data Sources.

*Source: _wp0554*

### References..............................................................................................................

### _wp0554 - References..............................................................................................................

### Tables
- 1. Subnational Borrowing Controls: Summary........................................................................11
- 2. Determinants of Subnational Fiscal Balances: Cross-Sectional Analysis. ..........................15
- 3. Dynamic Panel Analysis. .....................................................................................................22

### Figures
- 1. Marginal Effects of Different Subnational Borrowing Controls on Average                  
  Subnational Fiscal Balance ..............................................................................................17
- 2. Marginal Effect of Centrally Imposed Rules on Average Subnational Fiscal             
  Balance in Industrial and Emerging Countries ................................................................18
- 3. Marginal Effects of Different Subnational Borrowing Controls on                              
  Average Subnational Fiscal Balance ...............................................................................19

### Appendix
- Data Sources ............................................................................................................................30

### Appendix Tables
- 4. Subnational Borrowing Controls Across the World ............................................................31
- 5. Statistically Insignificant Control Variables: Impact on Subnational Fiscal Balance .........32
- 6. Performance of Different Regimes in the Presence of Bailout History...............................33

### Introduction — themes and objectives
- Over the last few decades, countries around the world have gradually moved towards the greater decentralization of fiscal revenue and spending responsibilities.
- As a result, subnational economic policies have taken an increasingly important role in ensuring macroeconomic stability.
- National governments have adopted different institutional responses to the difficulties of decentralized decision making, especially addressing the need to improve policy coordination across levels of government and contain subnational borrowing.
- Various case studies have identified different approaches to controlling subnational borrowing, but none has indicated which institutional arrangement is the most successful.
- Only a few papers have conducted cross-country empirical analysis and even these few focused on the presence of borrowing constraints rather than on their specific design.
- This paper attempts to bridge the gap with a cross-country analysis on a large sample of countries.
- The paper focuses on the impact on subnational budget deficits of different approaches to controlling subnational government borrowing, and attempts to identify the determinants of the effectiveness of each institutional arrangement.

*Source: _wp0554 - References..............................................................................................................*

### Section II briefly describes the main sources of fiscal indiscipline at the subnational level,

### _wp0554 - Section II briefly describes the main sources of fiscal indiscipline at the subnational level,

### Institutional framework — main themes
- Paper structure:
  - Section II: describes main sources of fiscal indiscipline at the subnational level, presents a classification of approaches to controlling subnational borrowing, and reviews existing empirical literature.
  - Section III: discusses empirical analysis issues and presents results for long-run and short-run dynamics.
  - Section IV: concludes.

### A. Sources of subnational fiscal indiscipline
- Mechanisms increasing subnational fiscal indiscipline:
  - Common pool problem
    - Separation of costs and benefits of public spending: jurisdictions benefit from projects while paying only a small fraction of costs, resulting in excessive spending and competition for federal transfers.
    - Regions may compete by proposing high-need signals such as large budget deficits or unsustainable debts to attract central grants.
  - Soft budget constraints
    - When transfers are based on ex post financial needs rather than ex ante characteristics, central government bailouts create "soft" budget constraints for subnational governments.
    - Lenders lose incentive to police regional governments if bailouts are perceived; central governments find no-bailout commitments hard to sustain in the short run.
    - A default by one region can raise borrowing costs for other regions, creating incentives for neighbor bailouts.
  - Interregional competition
    - Competition for mobile capital and labor can encourage inefficiently low taxes or excessive borrowing-financed infrastructure spending.
    - If regional governments collect but do not administer taxes, incentives exist to under-enforce tax collection, reducing de facto tax burdens (Cai and Treisman (2004)).
    - Shared tax bases with tax-setting powers at multiple tiers can produce excessive taxation because neither tier internalizes the full negative effect on the tax base (Keen, 1998; Keen and Kotsogiannis, 2002).
  - Unfunded federal mandates
    - Constitutional or legal mandates without adequate federal funding (example cited: 1992 German Federal Constitutional Court ruling for Saarland and Bremen) create expectations of supplementary transfers and weaken subnational incentives for sound expenditure management.
    - Voters may blame the central government for mismatches between revenue assignments and mandated expenditures, reducing electoral accountability of subnational officials.
  - Short electoral cycles
    - Politicians with short planning horizons may increase spending or cut taxes before elections, failing to internalize future borrowing costs (Besley and Case (1995)).
    - Short political cycles can have greater effects at subnational level where local economic performance strongly influences reelection.

### B. The role of vertical fiscal imbalances
- Definition and measurement:
  - Vertical fiscal imbalance: gap between subnational governments’ own revenue (e.g., property taxes) and their expenditure responsibilities; measured as the ratio of transfers to total subnational government revenue.
- Effects:
  - High vertical fiscal imbalances exacerbate moral hazard and adverse incentives created by the common pool problem and soft budget constraints.
  - High proportions of regional spending financed from the common pool and discretionary transfers increase bailout expectations and weaken subnational fiscal discipline.
  - High vertical fiscal imbalances facilitate arguments that central government is to blame for subnational deficits, reducing local accountability.
  - Tax competition can take more distortionary forms under high vertical fiscal imbalances (examples: China, Russia, United States).

### C. Classification of controls for subnational borrowing
- Four broad approaches (following Ter-Minassian (1997)):
  - Market discipline
    - Central government sets no limits; local governments free to borrow.
    - Self-imposed fiscal rules exist in Canada, Switzerland, and the United States.
    - Preconditions for effective market discipline (Lane (1993)):
      - Markets free and open with no privileged-borrower regulations.
      - Adequate borrower information on outstanding debt and repayment capacity.
      - No perceived chance of central government bailout.
      - Institutions ensuring policy responsiveness to market signals.
  - Administrative constraints
    - Central government directly controls subnational borrowing (examples):
      - Setting annual overall debt limits (Lithuania since 2001).
      - Special treatment or prohibition of external borrowing (Mexico).
      - Review and authorization of individual borrowing operations (India, Bolivia).
      - Centralization of all government borrowing with on-lending (Latvia, Indonesia).
    - Advantages: strict control, preserves flexible fiscal policy, coordinates external borrowing with macro policy.
    - Drawbacks: introduces implicit/explicit guarantees that can lead to bailouts; central authorities may lack information on local needs leading to average-quality project selection; local governments may over-submit projects expecting coverage by common pool.
  - Rule-based controls
    - Standing constitutional or legal rules to control subnational borrowing (examples by type and country):
      - Restrictions on overall budget deficits (Austria, Spain).
      - Operating budget deficits restrictions (Norway).
      - Indicators of debt-servicing capacity (Spain, Japan, Brazil, Korea).
      - Level of accumulated subnational debt (Hungary).
      - Level of spending restrictions (Belgium, Germany).
      - "Golden rules" limit borrowing to investment purposes (Germany).
    - Advantages: clarity, transparency, ease of monitoring, potential credibility.
    - Disadvantages:
      - Trade-off between compliance and flexibility.
      - Strict rules reduce maneuver room during downturns; flexible rules with escape clauses may be weak.
      - Evasion risks: reclassification of current to capital spending, off-budget accumulation, sale-and-lease-back and lease arrangements, hidden off-budget debts (examples: Australia 1982–1984, Denmark, United States, China).
  - Cooperative arrangements
    - Negotiation between federal and lower levels to design borrowing controls; subnational governments co-responsible for macro objectives and fiscal parameters.
    - Example: Australian loan council (established 1929) comprising state treasurers/heads and Commonwealth treasurer; analyzes and approves financing requirements and monitors execution.
    - Strengths: combines advantages of other approaches, fosters dialogue, coordination, flexibility, and subnational awareness of macro implications without automatic guarantees.
    - Weaknesses: hybrid nature can replicate flaws of other approaches if poorly implemented—weakens central leadership, softens budget constraints, promotes bargaining, and hampers coordination.

### D. Empirical evidence (literature review)
- Case-study and cross-country findings:
  - Ter-Minassian and Craig (1997):
    - Sole reliance on market discipline unlikely to succeed.
    - Rule-based approach preferable to administrative controls for transparency and certainty.
    - Cooperative arrangements promising for involving subnational governments in medium-term adjustments.
  - Ter-Minassian, Albino-War, and Singh (2004):
    - Self-imposed fiscal rules associated with better fiscal outcomes where bailout experience is limited.
    - Caveat: simple means analysis omits controls for factors like vertical fiscal imbalance.
  - Rodden and Eskeland (2003):
    - Effective containment requires either strong hierarchical oversight or strong market mechanisms.
  - Rattsø (2002):
    - European experience suggests decentralized governments can achieve fiscal stability in different ways; countries reviewed avoided serious fiscal imbalances.
- Econometric evidence — mixed results:
  - von Hagen and Eichengreen (1996), sample of 30 countries:
    - Observed that introduction of borrowing constraints increases subnational indebtedness; causality unclear and regressions did not control beyond GDP.
  - Fornasari, Webb, and Zou (2000), panel of 31 countries:
    - Constraining subnational borrowing did not show consistent effect on subnational fiscal deficits.
  - Jin and Zou (2002), 32 countries:
    - Similar findings for size of subnational governments.
  - Rodden (2002), panel of 33 countries:
    - Largest deficits found where subnational governments rely heavily on federal transfers and are free to borrow.
- Methodological evolution:
  - Early studies used a dummy for presence/absence of controls.
  - Rodden used a quasi-continuous index (IDB (1997) inspired) from 1 to 5 to reflect degree of borrowing autonomy based on institutional features (e.g., requirement for explicit upper-tier authorization, existence of formal limits, subnational ownership of banks/public enterprises).
    - Advantage: compresses borrowing-autonomy information into a single index preserving degrees of freedom.
    - Disadvantage: focuses on degree of borrowing autonomy and cannot identify which types of constraints perform better.
  - Overall implication: evidence supports necessity to control subnational borrowing, especially with high vertical fiscal imbalances, but optimal design remains unclear from econometric evidence alone.

### III. Empirical analysis — approach and data
- Short-run vs. long-run analysis
  - Goal: assess whether design of subnational borrowing controls leads to particular fiscal outcomes measured as aggregate subnational fiscal balance-to-revenue ratio.
  - Distinction:
    - Cross-sectional analysis: impact of institutional features and federal arrangements on average long-run subnational and consolidated fiscal balances.
    - Panel data analysis: how central government can affect subnational fiscal balance in a given year and short-run control effectiveness.
  - Cautions:
    - Short-run panel estimations may produce misleading long-run inferences.
    - Long-run analysis should assess institutional changes such as reducing vertical fiscal imbalances from 60 to 15 percent (moving to local fiscal autonomy); such dramatic changes are rare over short periods.
    - Panel data can infer long-run impacts only from local variations (e.g., around 15 percent and around 60 percent).
- Data
  - Dataset: newly collected on subnational borrowing regimes and fiscal outcomes of 44 countries (industrial and emerging, federal and unitary), 1982–2000.
  - Primary source: IMF Government Financial Statistics (GFS), unless otherwise indicated.
    - Note: first level of subnational government provides data for each country unless stated otherwise.
    - GFS transfer data may not be consistent across countries; shared revenues may be classified differently, potentially affecting vertical fiscal imbalance measures. Comparisons with IDB (1997), OECD (1999 and 2000) revealed differences for only a few countries.
  - Classification:
    - Countries classified according to design of subnational borrowing controls and regime changes identified; if a country changed regime (e.g., Germany, Hungary), it was included as two cross-sectional units (pre- and post-change).
    - For countries using multiple institutional features simultaneously, classification emphasized the predominant approach.
    - Panel is unbalanced; some specifications represented by fewer countries.
  - Bailout history:
    - A country considered to have a bailout history if significant bailouts of subnational governments occurred in the past and no institutional reform had altered expectations of future bailouts.
- Summary table (Table 1. Subnational Borrowing Controls: Summary — Number of countries)
  - Regime totals:
    - Unrestricted: 13 total; Emerging 5; Industrial 8; With Bailout History 6; Without Bailout History 7
    - Self-imposed rules: 3 total; Emerging 1; Industrial 2; With Bailout History 1; Without Bailout History 2
    - Centrally imposed rules: 12 total; Emerging 6; Industrial 6; With Bailout History 4; Without Bailout History 8
    - Administrative: 15 total; Emerging 13; Industrial 2; With Bailout History 7; Without Bailout History 8
    - Cooperative: 9 total; Emerging 2; Industrial 7; With Bailout History 4; Without Bailout History 5
    - Total: 52 total; Emerging 27; Industrial 25; With Bailout History 22; Without Bailout History 30
  - Sources for the table: Appendix I (as noted in the text).
- Control variables
  - Numerous factors indirectly related to design of borrowing constraints can affect subnational fiscal balances and are required as controls in the empirical work.

*Italic source attribution: _wp0554 - Section II briefly describes the main sources of fiscal indiscipline at the subnational level,*

### 1.     Degree of decentralization: A higher degree of decentralization (proxied by the share of

### Degree of decentralization: A higher degree of decentralization (proxied by the share of subnational government expenditure in general government expenditure)

### Determinants and hypothesized effects (variables examined)
- Degree of decentralization:
  - A higher degree of decentralization can lead to inferior fiscal outcomes in the long run (harms macroeconomic coordination) or in the short run (may reflect central government shifting fiscal burden onto subnational governments).
- Common standards of public expenditure management (PEM):
  - Common standards for subnational budgeting and financial reporting facilitate monitoring budget execution and make enforcement of rules more effective.
  - This study uses a dummy variable for the presence of common standards.
- Governance features:
  - Corruption (survey-based perception index) may have an (indirect) negative effect on subnational fiscal discipline via weak fiscal institutions.
  - Central bank independence (Cukierman’s (1992) index) may impose a harder budget constraint on all levels of government and improve fiscal outcomes.
- Cost of debt service:
  - Lower real interest rates, higher inflation, and lower debt service spending (ratio of debt servicing expenditure to gross national income) make borrowing cheaper and may encourage higher budget deficits.
- Macroeconomic situation:
  - Fixed-year effects capture the current phase of the global business cycle.
  - The unemployment rate for localized business cycles is chosen as a highly procyclical variable and its effect on subnational government surplus may be negative.
- Political variables:
  - Dummy for parliamentary democracies: Persson and Tabellini (2003) evidence of differences in fiscal behavior between presidential and parliamentary democracies.
  - Dummy for regional representation in the upper chamber: Rodden (2002) — upper chamber comprised of regional representatives increases regions’ veto power, negatively affecting fiscal discipline.
  - Dummy for national elections: central government transfers may be distributed more generously in election years, reducing subnational budget deficits; presidential elections used (or parliamentary elections for parliamentary democracies).
    - Note: If elections are held in the first quarter of the year, the previous year was coded as an election year instead.
  - Herfindahl index of political fractionalization: Rodden (2002) — political cohesion improves subnational fiscal outcomes.
  - Dummy for elected subnational authorities: elected authorities may have less fiscal discipline due to short-term electoral objectives.
  - Index of ethno-linguistic fractionalization: higher fractionalization may increase autonomy/bargaining power of ethnic regions, undermining central fiscal discipline.
- Demographic variables:
  - Population density: lower density raises provision costs for public goods, pressuring subnational finances.
  - Dependency ratio (share of population below age 15 or above age 64): higher dependency ratio raises fiscal pressure on subnational governments.
- Development variables:
  - GDP per capita and a dummy distinguishing emerging markets from industrial economies reflect better fiscal positions and institutions of developed countries.
- Size variables:
  - Size of general government (ratio of government expenditure to GDP): larger governments may find it harder to balance budgets.
  - Area: Rodden (2002) reported larger countries (by area) run higher fiscal surpluses on average.
- Federal structure:
  - Federation dummy: federal structure may increase bargaining power of regions, undermining fiscal discipline.
  - Average size of subnational jurisdictions (average subnational government expenditure–to-GDP ratio): large jurisdictions may find it easier to borrow and undertake large-scale investments requiring outside financing.
  - Horizontal imbalances (interregional disparities in per capita income and budget expenditure): equalization programs may raise reliance on federal transfers.
    - Note: Because of data constraints, the impact of horizontal fiscal imbalances on fiscal outcomes could not be estimated.

### Empirical analysis of the long run: model specification
- Marginal effect of restricting subnational borrowing is modeled to depend on vertical fiscal imbalance (vfi):
  - αij = α0j + α1j * vfi_i
  - vfi_i is the vertical fiscal imbalance in country i; α0j is the component not depending on vfi for approach j.
- Basic linear specification for subnational fiscal balances (similar to Rodden (2002)), including regime dummies, interaction terms with vfi, central government fiscal balance (CS), bailout history dummy, and controls:
  - SS_it = β0 + α01 self_i + α02 coop_i + α03 central_i + α04 admin_i + β1 vfi_it + α11 self_i * vfi_it + α12 coop_i * vfi_it + α13 central_i * vfi_it + α14 admin_i * vfi_it + β2 CS_it + γ bailout_i + δ controls_it + ε_it
  - Definitions:
    - SS: subnational government fiscal balance measured as ratio of budget surplus to total revenue.
    - self, coop, central, admin: dummy variables for self-imposed rules, cooperative arrangements, centrally imposed rules, administrative controls.
    - CS: central government surplus to total central government revenue.
    - controls: vector of control variables.
- Estimation approach:
  - Focus on average subnational outcomes; equation (2) estimated using a panel estimator that exploits cross-sectional differences between time averages of variables for each country.

### Key empirical finding reported in text
- "Consistent with our priors, the presence of a bailout history is associated with a statistically significant weaker subnational fiscal performance (5.4 percentage points higher subnational"

### Table 2 — Determinants of Subnational Fiscal Balances: Cross-Sectional Analysis (selected estimates and statistics)
- Dependent variable: Subnational Surplus to Revenue (SS)
- Central government surplus (CS):
  - Column 1 (Linear, Between): 0.396 (0.101)***
  - Column 2 (IV, Two-Stage): 0.346 (0.121)***
  - Column 3 (Common Standards, Between): 0.437 (0.120)***
  - Column 4 (Emerging Countries, Between): 0.398 (0.098)***
- Decentralization (dec):
  - Column 1: 0.006 (0.078)
  - Column 2: -0.107 (0.095)
  - Column 3: -0.005 (0.091)
  - Column 4: 0.011 (0.078)
- Self-imposed rule (self):
  - Column 1: 0.085 (0.151)
  - Column 2: 0.115 (0.441)
  - Column 3: 0.096 (0.147)
- Centrally imposed rule (central):
  - Column 1: -0.089 (0.047)*
  - Column 2: 0.046 (0.573)
  - Column 3: -0.084 (0.046)*
  - Column 4: -0.109 (0.047)**
- Cooperative controls (coop):
  - Column 1: -0.147 (0.066)**
  - Column 2: 0.097 (0.209)
  - Column 3: -0.150 (0.055)**
  - Column 4: -0.154 (0.064)**
- Administrative controls (admin):
  - Column 1: 0.008 (0.044)
  - Column 2: 0.105 (0.356)
  - Column 3: -0.025 (0.047)
  - Column 4: 0.016 (0.042)
- Interaction terms VFI * self:
  - Column 1: -0.453 (0.644)
  - Column 2: -0.132 (2.178)
  - Column 3: -0.511 (0.625)
- Interaction terms VFI * central:
  - Column 1: 0.219 (0.101)**
  - Column 2: 0.141 (0.501)
  - Column 3: 0.210 (0.091)**
  - Column 4: 0.203 (0.098)**
- Interaction terms VFI * coop:
  - Column 1: 0.201 (0.112)*
  - Column 2: -0.031 (0.395)
  - Column 3: 0.007 (0.136)
  - Column 4: 0.212 (0.109)*
- Interaction terms VFI * admin:
  - Column 1: 0.015 (0.077)
  - Column 2: -0.004 (0.173)
  - Column 3: 0.056 (0.083)
  - Column 4: 0.017 (0.075)
- Common standards of PEM * coop:
  - Column 3: 0.156 (0.065)**
- Common standards of PEM * central:
  - Column 3: 0.013 (0.043)
- VFI * centrally imposed rules * emerge:
  - Column 4: 0.170 (0.092)*
- GDP (log, in PPP terms):
  - Column 1: 0.017 (0.017)
  - Column 2: 0.023 (0.045)
  - Column 3: 0.005 (0.018)
  - Column 4: 0.021 (0.016)
- Size of government:
  - Column 1: -0.075 (0.122)
  - Column 2: -0.079 (0.186)
  - Column 3: -0.049 (0.119)
- Bailout history (bailout):
  - Column 1: -0.054 (0.020)***
  - Column 2: -0.045 (0.025)*
  - Column 3: -0.021 (0.019)
  - Column 4: -0.053 (0.019)***
- Constant:
  - Column 1: -0.091 (0.134)
  - Column 2: -0.200 (0.580)
  - Column 3: -0.021 (0.178)
  - Column 4: -0.144 (0.133)
- Sample and fit:
  - Number of observations: Column 1: 584; Column 2: 584; Column 3: 427; Column 4: 584
  - Number of groups: Column 1: 52; Column 2: 52; Column 3: 36; Column 4: 52
  - R2 (between): Column 1: 0.57; Column 2: 0.48; Column 3: 0.78; Column 4: 0.61

Notes from table:
- Standard errors in parentheses. Values significant at 10% level are marked with *, at 5% with **, at 1% with ***.
- First stage multinomial logit regression included GDP, federation dummy, vertical fiscal imbalance, and constant. N = 52; Pseudo R2 = 0.29.

*Source: _wp0554 - 1.     Degree of decentralization: A higher degree of decentralization (proxied by the share of subnational government expenditure in general government expenditure).*

### 3. VFI is vertical fiscal imbalance. PEM is public expenditure management.

### _wp0554 - 3. VFI is vertical fiscal imbalance. PEM is public expenditure management.

### Marginal effects of subnational borrowing regimes and vertical fiscal imbalance (VFI)
- Self-imposed fiscal rules perform better than centrally imposed rules, but only with low vertical fiscal imbalances.
- The marginal effect of centrally imposed rules and of cooperative arrangements increases rapidly as vertical fiscal imbalances widen (positive coefficients on interaction terms are relatively large and statistically significant).
- The coefficient on the interaction term for the administrative approach is positive, but relatively small and statistically insignificant.
- The area where self-imposed rules appear to be optimal is at vertical fiscal imbalances below 16 percent (this area includes only five countries).
- The area where centrally imposed fiscal rules perform best of all is at vertical fiscal imbalances above 47 percent (this area covers 15 countries).
- Administrative controls seem superior between the 16 percent and 47 percent thresholds.
- Cooperative arrangements and unrestricted borrowing never outperform the alternative institutional frameworks in the cross-section evidence.
- All marginal effects are measured in percentage points of total subnational government revenue.
- Caveat: “zero vertical fiscal imbalances” do not exist; the horizontal axis in figures corresponds to unrestricted subnational borrowing.

### Interaction with common standards of public expenditure management (PEM)
- The success of cooperative arrangements depends on the presence of common standards in financial reporting: the positive coefficient on the interaction term is large and statistically significant.
- Lack of sufficient observations did not allow inclusion of self-imposed fiscal rules in this specific interaction analysis.

### Differences between industrial and emerging economies
- Average subnational fiscal balance: industrial countries = -2.3 percent; emerging economies = -4.2 percent.
- Emerging economies more often adopt administrative rules; industrial countries almost exclusively utilize cooperative arrangements.
- For centrally imposed fiscal rules (available for six emerging economies and six industrial countries in the sample), centrally imposed rules start outperforming other approaches at substantially lower levels of VFI in emerging economies and their marginal effect grows faster as VFI widens.
- In the presence of bailouts, the performance of cooperative arrangements deteriorates significantly.
- A bailout history is estimated to increase the deficit by only 1.4 percent of total revenue at the mean level of the vertical fiscal imbalance (this estimate is highly statistically insignificant).

### Endogeneity concerns and two-stage estimation
- Potential endogeneity: subnational borrowing regimes may be endogenous because average subnational fiscal balances could influence regime choice.
- A two-stage estimator (multinomial logit first stage for mutually exclusive dummies; predicted probabilities used in second stage) was implemented to address endogeneity.
- Two-stage results (Column 2 of Table 2) largely confirm main qualitative results.
- Distinctions from two-stage estimates:
  - Unrestricted subnational borrowing is now associated with inferior subnational fiscal outcomes relative to any other approach (the horizontal axis plunges).
  - The positive marginal impact of cooperative arrangements decreases when VFI widens.
- Interpretation: countries without limits or with cooperative approaches have intrinsically more disciplined subnational governments; countries with less disciplined subnational governments opt for fiscal rules or administrative controls.

### Short-run (dynamic panel) analysis — specification and identification
- Panel specification (equation (3)):
  SS_it = α_i + φ SS_it-1 + β dec_it + δ vfi_it + λ CS_it + μ Controls_it + ε_it.
- Fixed effects α_i absorb time-invariant country characteristics.
- Arellano and Bond (1991) GMM approach used to handle dynamic panel bias and endogenous regressors by first-differencing and instrumenting with appropriate lags.
- Instruments: higher-order lags of endogenous variables (e.g., ∆CS_it-2, ∆CS_it-3, ...) used for ∆CS_it and similarly for other endogenous variables.
- Interaction specifications used to identify differential short-run responses by bailout history and by borrowing regimes.

### Short-run empirical findings (Table 3 summaries)
- Model A (Bailouts):
  - 1st lag of Subnational Surplus to Revenue (SS): 0.266 (0.077)***.
  - VFI * Bailouts: 0.113 (0.070).
  - VFI * No Bailouts: 0.154 (0.037)***.
  - CS * Bailouts: 0.031 (0.048).
  - CS * No Bailouts: 0.093 (0.032)***.
  - Interpretation: In the absence of bailouts, both federal transfers and central government fiscal balances have a highly significant positive impact on subnational fiscal outcomes. The estimate of the gap-closing effect of federal transfers is 0.154, indicating most transfer money will be used to inflate expenditure rather than close the fiscal deficit.
  - In the presence of a bailout history, the responsiveness of subnational fiscal balances to transfers and central government fiscal balances drops and becomes statistically insignificant.

- Model B (Borrowing Regimes):
  - 1st lag of SS: 0.255 (0.072)***.
  - VFI * No Rules: -0.069 (0.084).
  - VFI * Coop: 0.116 (0.169).
  - VFI * Central: 0.106 (0.052)**.
  - VFI * Admin: 0.180 (0.041)***.
  - CS * No Rules: 0.136 (0.044)***.
  - CS * Coop: 0.134 (0.049)**.
  - CS * Central: 0.100 (0.049)**.
  - CS * Admin: -0.018 (0.025).
  - Decentralization (dec): -0.291 (0.088)*** in Model A; -0.304 (0.105)*** in Model B.
  - GDP (log, in PPP terms): -0.122 (0.049)** in Model A; -0.155 (0.050)*** in Model B.
  - Population density: 0.0006 (0.0003)* in Model A; 0.0008 (0.0002)*** in Model B.
  - Size of government: 0.001 (0.130) in Model A; -0.041 (0.097) in Model B.
  - Constant: 0.003 (0.001)** in Model A; 0.003 (0.001)*** in Model B.
  - Number of observations: 480; Number of groups: 49.
  - Test of no 2nd order autocorrelation in residuals: -0.95 (p = 0.341) for Model A; -1.14 (p = 0.255) for Model B.
  - Interpretation: Centrally imposed rules achieve the highest level of fiscal policy coordination in the short run — coefficients on interaction terms with vfi and CS are positive, large, and statistically significant. Administrative rules show a large positive VFI interaction (0.180 (0.041)***), but CS * Admin is negative and not significant.

### Interpretation and policy-relevant implications from the empirical analysis
- Bailouts undermine short-run fiscal discipline: when bailouts are present, central government instruments (transfers and fiscal balance) lose short-run effectiveness in shaping subnational fiscal outcomes.
- Centrally imposed borrowing rules support stronger short-run fiscal coordination: both federal transfers (VFI) and central government fiscal surpluses (CS) positively affect subnational fiscal balances when centrally imposed rules are in place.
- Administrative controls can deliver positive marginal effects of transfers (VFI * Admin = 0.180 (0.041)***), but their coordination with central government fiscal balances is weak (CS * Admin = -0.018 (0.025), not significant).
- Decentralization is associated with more negative subnational surpluses to revenue in the short run (dec coefficients are -0.291 and -0.304, both significant at 1 percent).
- Larger GDP (log, PPP) is associated with more negative subnational surpluses to revenue in these specifications.
- Population density is positively associated with subnational surpluses to revenue at conventional significance levels.
- Policy takeaway: to enhance subnational fiscal discipline and make central instruments effective in the short run, avoid bailout expectations and consider centrally imposed borrowing rules where appropriate; ensure common PEM standards to improve cooperative arrangements where they are used.

*Italic: Source: _wp0554 - 3. VFI is vertical fiscal imbalance. PEM is public expenditure management.*

### 3. CG balance is central government surplus-to-revenue ratio

### 3. CG balance is central government surplus-to-revenue ratio

### Empirical findings on transfers, vertical fiscal imbalances (vfi), and subnational fiscal responses
- At low levels of vertical fiscal imbalances, transfers play a modest role; the responsiveness of subnational fiscal balances to the central government balance is captured by interaction terms CS*regime.
- The highest responsiveness of subnational fiscal balances to central government balance occurs when:
  - subnational governments are free to borrow (unrestricted borrowing), or
  - cooperative arrangements are in place.
- Under unrestricted borrowing and cooperative arrangements, policy coordination via transfers appears more difficult:
  - the coefficients on interaction terms with vfi are insignificant and even negative in the case of unrestricted borrowing.
- As vertical fiscal imbalances widen and transfers become the main policy instrument:
  - administrative controls increasingly deliver better responses of subnational fiscal balances to federal transfers;
  - the corresponding coefficient on the interaction term with vfi is the largest of all, and statistically significant at the one percent level.
- Interpretation: central government involvement in microlevel decisions can directly influence subnational fiscal outcomes in the short run via transfers, but may produce soft budget constraints and economic inefficiencies in the long run.
- Comparison with Rodden (2002):
  - This paper finds subnational fiscal balances depend significantly and positively on their lagged values; Rodden (2002) finds a significantly negative dependence (mean-reverting behavior).
  - This paper finds a significantly negative coefficient on the degree of decentralization (dec), consistent with greater decentralization leading in the short run to deterioration of subnational fiscal outcomes; Rodden (2002) reports a significantly positive effect.
  - Rodden’s estimates suggest no comovement of central and local fiscal balances (CS coefficient small and insignificant); Table 3 in this paper shows a positive and highly significant coefficient on CS, indicating comovement.
  - Positive marginal effects of vfi (from mostly positive coefficients on interaction terms involving vfi) are interpreted as reflecting higher transfers from center to regions/municipalities in a given year, which should have a positive impact on subnational public finance that year, even if vfi as an institutional feature is associated with inferior long-run fiscal outcomes.

### Conclusions and policy recommendations
- Broad conclusion: no single institutional arrangement is superior under all circumstances; appropriateness of borrowing constraints depends on:
  - degree of vertical fiscal imbalances,
  - existence of bailout precedents,
  - quality of fiscal reporting.
- Specific recommendations:
  - Giving unconstrained borrowing authority to subnational governments is unlikely to be optimal.
  - At low levels of vertical fiscal imbalances, fiscal rules adopted by subnational governments themselves seem to lead to better fiscal outcomes.
  - As vertical fiscal imbalances widen, the positive effect of self-imposed rules declines rapidly; centrally imposed fiscal rules seem to become the best option, both in the short and long runs, especially in emerging economies.
  - For high vertical imbalances, administrative procedures may provide the central government with tighter short-run control over subnational fiscal outcomes compared to fiscal rules and cooperative arrangements; however, the implicit guarantee of subnational debt related to these controls seems to undermine fiscal discipline in the long run.
  - Adoption of common standards of financial reporting is crucial for the success of cooperative arrangements and may increase the effectiveness of centrally imposed fiscal rules.
  - Central governments should avoid bailing out subnational governments whenever possible, as bailouts significantly erode the effectiveness of borrowing controls.
  - When considering a bailout, central governments should weigh short-term benefits against long-term negative consequences; in the presence of bailout experiences, centrally imposed fiscal rules seem to be the most effective.

### Institutional design considerations and limits of the analysis
- Fiscal rules vary in form (debt ceiling, target fiscal deficit, expenditure caps), coverage, and enforcement.
- Enforcement mechanisms include:
  - financial market discipline,
  - cooperative administrative and financial sanctions among subnational governments,
  - higher-level government punishment rights,
  - bankruptcy procedures for subnational governments to avoid bailouts and enable transparent resolution.
- Transfer systems vary; a given level of vertical fiscal imbalance might capture different incentives depending on transfer design.
- Data limitations prevented analysis of interactions among many institutional features; these are left for future research.

### Data sources and empirical appendices (selected)
- Central and subnational government budgets: IMF Government Finance Statistics, 2002.
- Macroeconomic indicators: World Bank, World Development Indicators (2004).
- Subnational jurisdictions structure: World Bank (2000), Table A.1.
- Corruption perception indices: Transparency International.
- Political indicators, election years: World Bank Database of Political Indicators (2000).
- Federal / unitary classification: Treisman (2002).
- Characterization of fiscal institutions, borrowing constraints, bailout history: Ter-Minassian (1997); Inter-American Development Bank (1997); Bird and Vaillancourt (1999); OECD (2000); Dabla-Norris and Wade (2002), Table 9; Dafflon (2002); OECD (2003); Rodden, Eskelund, and Litvak (2003); and IMF staff estimates.
- Index of horizontal imbalances: Rodden and Wibbels (2002), Table 2.
- Standards of budgeting and financial reporting: OECD (2003).
- Central Bank Independence index: Cukierman (1992).
- Ethnolinguistic Fractionalization: Roeder (2001) — measures probability two randomly selected individuals belong to different ethnolinguistic groups.
- Bailout history: OECD (2003); Rodden, Eskelund, and Litvak (2003); Ter-Minassian (1997); United Nations, 2002 (for Chile, Colombia, France); von Hagen and others (2000); and IMF staff estimates.
- Country-specific data: Russia — Statistical Yearbook of Russia (2003), Goskomstat, Moscow (Rossiyskiy Statisticheskiy Ezhegodnik 2003); India — Purfield (2004).

### Selected tabular evidence (high-level)
- Table 4: Subnational Borrowing Controls Across the World — catalogues country/regime/bailout history/years for 1982–2000 (with specific country-regime-year entries, and note on missing data for Brazil 1995–96 and an Australian bailout history reset).
- Table 5: Statistically Insignificant Control Variables — lists variables with insignificantly signed coefficients across cross-section and dynamic panel specifications and notes on consistency with prior expectations (symbols: +, -, ++, --; + (-) denotes positive (negative) but insignificant coefficient; ++ (--) denotes statistical significance approached 10% level in some specifications).

*Source: IMF Working Paper chapter titled "3. CG balance is central government surplus-to-revenue ratio" (PDF content provided).*

### 2. N/A appears in the following cases: (1) variables that do not have meaningful time series variation were used

### _wp0554 - 2. N/A appears in the following cases: (1) variables that do not have meaningful time series variation were used

### Notes on "N/A" usage
- "N/A" appears in the following cases: (1) variables that do not have meaningful time series variation were used only in cross-section; (2) the election dummy has no meaningful cross-sectional averages; and (3) there is no prior hypothesis about the effect of area.

### Table 6 — Performance of Different Regimes in the Presence of Bailout History (Dependent Variable: Subnational Surplus to Revenue (SS))
- Central government surplus (CS): 0.383 (0.101)***
- Decentralization (dec): 0.004 (0.086)
- Self-imposed rule (self): 0.007 (0.159)
- Centrally-imposed rule (central): -0.061 (0.065)
- Cooperative controls (coop): -0.109 (0.039)***
- Administrative controls (admin): 0.017 (0.040)
- Bailout history (bailout): -0.014 (0.031)
- VFI * self: -0.122 (0.687)
- VFI * central * Bailout: 0.105 (0.105)
  - ∆ = -0.035
  - F = 0.07
  - (p =0.79)
- VFI * central * nobailout: 0.140 (0.168)
- VFI * coop * bailout: -0.017 (0.044)
  - ∆ = -0.241
  - F = 4.48
  - (p = 0.04)**
- VFI * coop * no bailout: 0.224 (0.089)**
- VFI * admin * bailout: -0.064 (0.057)
  - ∆ = -0.186
  - F = 1.38
  - (p = 0.25)
- VFI * admin * nobailout: 0.122 (0.112)
- GDP (log, in PPP terms): 0.022 (0.017)
- Size of government: -0.140 (0.129)
- Cconstant: -0.124 (0.136)

### Key statistics and sample
- Number of observations: 584
- Number of groups: 52
- R2 (between): 0.59

### Notes on significance and terminology
- Standard errors in parentheses. Values significant at 10 percent level are marked with *, at 5 percent with **, at 1 percent with ***.
- VFI denotes vertical fiscal imbalance.

*Source: _wp0554 - 2. N/A appears in the following cases: (1) variables that do not have meaningful time series variation were used*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2005/_wp0554.pdf_