## _wp06123

## Source details

**Canonical URL:** [_wp06123](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2006/_wp06123.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2006/_wp06123.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2006/_wp06123.pdf.json)

---

### Introduction and research question
- Research question:
  - Do budgetary institutional structures matter for fiscal outcomes once a comprehensive set of economic and political conditioning factors is controlled for?
- Empirical focus and sample:
  - New and potential member states of the European Union between 1997 and 2003, in the run-up to accession on May 1, 2004.
- Methodological highlights:
  - Isolate role of budgetary institutional structure while controlling for economic and political factors.
  - Use approaches that evaluate time-invariant country characteristics in transmitting shocks to budgetary outcomes (Blanchard and Wolfers (2000) approach adapted).
  - Multidimensional characterization of historical and political determinants treated as groups.

### Conceptual framework and determinants of fiscal performance
- Framework components (as represented in source Figure 2):
  - Fiscal Institutions
  - Contemporary Politics
  - Long-Term Structural Determinants
  - Fiscal Outcome
- Structural (time-invariant) determinants:
  - Proportional electoral rules increase representativeness but tend to reduce fiscal discipline; majoritarian systems tend to encourage accountability.
  - Sample observations:
    - None of the sample countries has a majoritarian system.
    - Average district magnitude in the sample is 29.
    - District magnitude range in the sample: low of 8 in Bulgaria and Romania and high of 150 in the Slovak Republic.
- Contemporary (time-varying) political determinants:
  - Government fragmentation, coalition ideology, fiscal decentralization, and voter participation vary within countries over 1997–2003.
  - Theoretical expectation: greater fragmentation allows multiple constituencies to claim fiscal resources without fully bearing taxation costs; higher voter participation appears associated with looser budgetary constraints.

### Box 1 — Fragmentation and fiscal discipline (mechanisms)
- Common-pool problem:
  - Budget behaves like a common-pool resource; more claimants create pressure to overconsume.
  - Dynamic outcome: short-run spending pressures can draw down national wealth or accumulate debt (Velasco, 1999).
- Parties and coalitions:
  - Coalitions exacerbate common-pool pressures; aggregate effect: coalitional governments spend more than single-party governments.
- Ideology and participation:
  - Ideology effects ambiguous; left/right not unidimensional.
  - Voter participation: higher turnout associated with larger governments and slower growth (Mueller and Stratmann, 2003).
- Institutional remedies:
  - Two perspectives: centralized/delegation approach (concentrate budgetary power) and decentralized/contract approach (collective negotiation, MTBFs).
  - Government-type dependence:
    - One-party: delegation effective.
    - Coalition (especially minority): contract-based approach with monitoring/enforcement preferable.
- Empirical sample note:
  - MTBFs in early stages in sample; both delegation and effective delegation to finance ministers have helped contain fiscal pressures.

### Budget institutions index — construction and cross-country assessment
- Sample: ten countries — Estonia, Bulgaria, the Czech Republic, Hungary, Latvia, Lithuania, Poland, Romania, the Slovak Republic, and Slovenia.
- Index dimensions (three stages):
  - Preparation stage (variables 1–4)
  - Authorization stage (variables 5–8)
  - Implementation stage (variables 9–12)
- Notable institutional facts (preserved from source):
  - Permanent constraints on budgetary parameters present in only three countries—Estonia, Latvia, and Poland.
  - Multiannual targets under an MTBF: only Slovenia uses MTBF multiannual targets (variable 2).
  - Finance minister blocking authority for revenue shortfalls or overspending: Bulgaria, Estonia, Latvia, and Lithuania grant this authority.
- Aggregate assessment:
  - Estonia and Slovenia have done consistently well.
  - Poland has considerably improved over the period.
  - Hungary’s position was low and deteriorated.

### Empirical framework and key econometric issues
- Objective:
  - Explain short-term dynamics of the primary budget balance using pooled observations from the ten countries over 1997–2003.
- Main empirical specification (notation preserved conceptually):
  - y_it = α + β x_it + δ w_it + φ s_it + γ v_i + φ u_t + ε_it
    - y_it: fiscal primary balance in country i and year t
    - x_it: economic control variables
    - w_it: political control variables
    - s_it: fiscal institutions index
    - v_i: unchanging country-specific effects (country dummies)
    - u_t: period effects (time dummies)
- Three econometric problems highlighted:
  1. Reverse causality / endogeneity between budget outcomes and fiscal institutions; working assumption that institutions do not change quickly.
  2. Omitted variables; partial solution: within-country (fixed-effects) analysis to eliminate v_i.
  3. Loss of cross-country heterogeneity when eliminating v_i; addressed by allowing time-invariant characteristics to condition responses to shocks (Blanchard and Wolfers-style nonlinear specification).

### Data, sample construction, and descriptive statistics
- Intended panel: 70 observations (7 years × 10 countries); final working sample: 63 observations after dropping cases with missing ideology data and extreme inflation.
- Time span: 1997–2003.
- Selected country means (Table 1, preserved exactly):
  - Bulgaria: Ethnic fractionalization 0.40; Average district magnitude 7.7; Voter turnout 66.9
  - Czech Republic: Ethnic fractionalization 0.32; Average district magnitude 25.0; Voter turnout 69.4
  - Estonia: Ethnic fractionalization 0.51; Average district magnitude 9.2; Voter turnout 61.5
  - Hungary: Ethnic fractionalization 0.15; Average district magnitude 12.2; Voter turnout 66.4
  - Latvia: Ethnic fractionalization 0.59; Average district magnitude 20.0; Voter turnout 76.2
  - Lithuania: Ethnic fractionalization 0.32; Average district magnitude 35.3; Voter turnout 62.1
  - Poland: Ethnic fractionalization 0.12; Average district magnitude 16.7; Voter turnout 48.7
  - Romania: Ethnic fractionalization 0.31; Average district magnitude 7.8; Voter turnout 70.7
  - Slovak Republic: Ethnic fractionalization 0.25; Average district magnitude 150.0; Voter turnout 76.6
  - Slovenia: Ethnic fractionalization 0.22; Average district magnitude 11.0; Voter turnout 76.7
- Selected descriptive statistics (Table 3, preserved exactly where shown):
  - Unemployment rate overall 11.29; Std. Dev. 4.23; Min 4.30; Max 20.00
  - Inflation overall 9.17; Std. Dev. 11.12; Min -1.20; Max 59.10
  - Openness index overall 0.90; Std. Dev. 0.25; Min 0.46; Max 1.36
  - Fiscal institutions index overall 2.12; Std. Dev. 0.45; Min 1.37; Max 2.72
  - Government fragmentation overall 0.63; Std. Dev. 0.20; Min 0.27; Max 1.00
  - Left/right overall 11.11; Std. Dev. 3.53; Min 5.81; Max 17.35
  - Nationalism overall 11.44; Std. Dev. 3.48; Min 6.80; Max 17.57
  - Fiscal centralization overall 10.46; Std. Dev. 2.00; Min 7.09; Max 13.46
  - Ethnic fractionalization overall 0.31; Std. Dev. 0.14; Min 0.12; Max 0.59
  - District magnitude (logarithms) overall 2.83; Std. Dev. 0.81; Min 2.00; Max 5.00
  - Voter turnout overall 67.00; Std. Dev. 8.39; Min 48.70; Max 76.70
  - Debt-to-GDP (lagged) overall 33.99; Std. Dev. 23.34; Min 5.80; Max 107.50

### Key empirical insights and principal numeric findings
- Principal conclusions (authors’ stated conclusions):
  - The quality of institutions matters strongly for fiscal outcomes even when conditioning variables are considered.
  - Contemporary politics is important and often trumps economic conditioning variables; features associated with less fiscal conservatism include:
    - More fragmented government coalitions.
    - An ideological disposition toward the “left.”
    - Greater nationalism.
    - Greater fiscal decentralization.
    - Greater voter participation appears to loosen budgetary constraints.
  - Deeper historical/political influences (ethnic fractionalization, district magnitude) suggest more inclusiveness hurts budgetary outcomes but evidence is unstable.
  - Policy implication: improvements in fiscal institutions can materially help curb fiscal pressures.
- Fixed-effects within-country results (Table 6; selected numeric estimates preserved exactly where shown):
  - Fiscal institutions index coefficients: 7.52 (2.08)*** and 7.97 (2.12)*** in columns where included.
  - Unemployment rate examples: coefficient -0.34 with (0.16)** (sign negative; higher unemployment reduces primary surplus).
  - Openness index examples: coefficients reported include 1.58, 4.78, 7.89, 7.95 in various specifications (some significant).
  - Interpretation: a movement of the index from the 25th percentile to the median results in an improvement of about 2 percentage points of GDP (within-country estimate noted in text).
- Cross-country regressions (Table 5; selected numeric examples preserved exactly):
  - Debt-to-GDP ratio coefficients across specifications: 0.06, 0.05, 0.08, 0.06, 0.06.
  - Fiscal institutions index coefficients reported: 2.74 (1.21)* and 1.52 (1.18) in select specifications.
  - District magnitude coefficients: -1.5 and -1.03 (with standard errors noted).
- Alternative estimators and dynamics (Table 8; selected entries preserved exactly):
  - Debt-to-GDP ratio (lagged): 0.09 (0.02)*** in column 1 (random effects).
  - Fiscal institutions index examples: 5.09 (1.19)***, 6.20 (2.13)***, 9.03 (4.42)** across columns.
  - Arellano-Bond column: Sargan test Pr≥χ2 = 0.27; AR(2) test Pr≥z = 0.93.
  - Observations vary by estimator: 63, 43, 53.
- Nonlinear least squares allowing time-invariant conditioning (Tables 10–11; selected numeric estimates preserved exactly):
  - Fiscal institutions index point estimates reported: 5.47 (2.45)**, 5.38 (2.40)**, 5.86 (2.33)**, 6.28 (2.52)** (NLS specifications).
  - Effect-size interpretation: move from the 25th to the 50th percentile leads to improvement in the primary surplus of about 1½ percent of GDP (authors’ stated interpretation).
  - Government fragmentation coefficients large and significant in several NLS specifications: e.g., 8.12 (2.87)***, 7.69 (2.83)**.
  - Government ideology examples (NLS): Left/right 0.46 (0.18)***; Nationalism -0.77 (0.21)***; Fiscal decentralization 0.95 (0.30)***.
  - Time-invariant variables examples: District magnitude negative e.g., -1.44 (0.85)*; Voter turnout negative e.g., -0.13 (0.05)**.
- Expenditure vs revenue channels (Tables 13–14; selected numeric estimates preserved exactly):
  - Seemingly unrelated regressions show discipline operates mainly through expenditure reduction.
  - Lagged expenditure coefficient example: 0.38 (0.08)***.
  - Lagged revenue coefficient example: 0.26 (0.08)***.
  - Fiscal institutions index effects on expenditure and revenue (examples): -3.59 (1.64)** on expenditure; -1.02 (1.11) on revenue in one specification.
  - Interaction patterns (examples): Interacted shock × District Magnitude 0.47 (0.26)* on Exp-to-GDP in one column; voter turnout operating through higher expenditures when oil price shock occurs.

### Robustness, sensitivity, and methodological notes
- Fixed-effects (within) preferred benchmark given small T and N and concerns about omitted variables; Hausman test reportedly does not reject random-effects in some cases.
- Dynamic estimators:
  - Arellano-Bond used but cautioned given small sample; Kiviet bias adjustment applied for unbalanced small-N panels.
  - Lagged dependent variable generally not statistically significant in benchmark specifications.
- Outlier checks:
  - Excluding one country at a time indicates fiscal institutions index relevance not driven by any single country.
  - Fiscal institutions index significance levels when excluding countries:
    - Significant at the 7 percent level when either Hungary or Poland is dropped.
    - Significant at the 5 percent level when one of the other countries or both Poland and Romania are dropped.
- Time-invariant conditioning (Box 2) produces unstable coefficients on deeper determinants due to small sample (ten countries) and short time series; directions often plausible but significance variable.

### Policy implications and conclusions
- Contemporary politics substantially influences budgets through fragmentation, ideological predispositions, decentralization, and voter participation.
- Quality of fiscal institutions materially affects budgetary discipline, primarily by constraining expenditures.
- Breadth of evidence:
  - Institutional effects are robust across specifications, estimators, and sensitivity checks in this sample.
  - Evidence on mechanisms (delegation vs contract) is less definitive; both delegation- and contract-based indices are statistically important in the sample.
- Practical implication:
  - Despite historical and political constraints, policy initiatives that establish checks in the competition for fiscal resources (improved fiscal institutions, MTBFs, monitoring and enforcement) can materially help curb fiscal pressures.

### Appendix notes (indices, data sources)
- Budget institutions index construction:
  - Variables scored 0–4 across 12 items grouped into three stage indices (preparation, authorization, implementation); overall index = simple average of three components.
  - Delegation-approach index formed by summing items 2–7 and 9–12; contract-based index by summing items 1, 5–7, and 9–11.
- Data sources (selected, preserved as in source):
  - Dependent variables: fiscal notifications to the European Commission; IMF World Economic Outlook where notifications unavailable.
  - Time-varying economic variables: IMF World Economic Outlook.
  - Political variables: Database on Political Institutions (2000, updated 2004); Database on Party Policy in Modern Democracies by Benoit and Laver (2005); Parties and Elections in Europe; EIU reports.
  - Ethnic fractionalization: Alesina and others (2003).
  - District magnitude: Database on Political Institutions, 2000 (updated 2004).
  - Voter turnout: Institute for Democracy and Electoral Assistance.

*Source: _wp06123 - References (PDF content provided).*

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

### _wp06123 - References..............................................................................................................

### Introduction
- Research question and context:
  - Do budgetary institutional structures matter for fiscal outcomes once a comprehensive set of economic and political conditioning factors is controlled for?
  - Empirical focus: the new and potential member states of the European Union between 1997 and 2003, in the run-up to accession on May 1, 2004.
- Motivation for the sample:
  - Common commitment to budgetary discipline (including adoption of the euro) provides a historical experiment with variation in fiscal outcomes despite similar stated objectives.
  - Time-series variation in the quality of fiscal institutions across countries—some improved institutions as part of transition reforms; others experienced slippages—permits analysis beyond cross-sectional relationships.
- Methodological approach highlights:
  - Isolation of the role of budgetary institutional structure while controlling for a more comprehensive set of economic and political factors than in many earlier studies.
  - Use of approaches that evaluate time-invariant country characteristics in transmitting shocks to budgetary outcomes (Blanchard and Wolfers (2000) approach adapted by Persson (2002, 2004), Persson and Tabellini (2003, 2004), Milesi-Ferretti, Perotti, and Rostagno (2002)).
  - Multidimensional characterization of historical and political determinants, considered as groups rather than single variables, to capture direct and indirect impacts on policy formulation.

### Determinants of Fiscal Performance: Summary of Theory and Evidence
- Conceptual framework (Figure 2 representation):
  - Fiscal Institutions
  - Contemporary Politics
  - Long-Term Structural Determinants
  - Fiscal Outcome
- Structural determinants (those that change little over time):
  - Population diversity and electoral systems shape the trade-off between representativeness and fiscal accountability.
  - Proportional electoral rules increase representativeness but tend to reduce fiscal discipline; majoritarian systems tend to encourage accountability.
  - Empirical observations:
    - None of the sample countries has a majoritarian system.
    - Average district magnitude in the sample is 29.
    - District magnitude range in the sample: low of 8 in Bulgaria and Romania and high of 150 in the Slovak Republic.
    - Correlation between ethnic fractionalization and district magnitude is weak in the sample (see Table 2 referenced).
  - Literature cited: Persson and Tabellini (2003, 2004); Persson, Roland, and Tabellini (2005); Aghion, Alesina, and Trebbi (2004); Lijphart (1994); Hallerberg and Marier (2004).
- Contemporary politics (time-varying political factors):
  - Government fragmentation and government ideologies show significant within-country variation during the sample period (see Table 3 referenced).
  - Theory and international evidence suggest:
    - Greater fragmentation allows multiple constituencies to claim fiscal resources without fully bearing taxation costs (see Box 1 referenced).
    - Early empirical findings (Roubini and Sachs, 1989) and subsequent cross-country studies link more fragmented coalitions to larger budget deficits.
    - Alternative fragmentation measures (e.g., number of ministries) may be relevant (Perotti and Kontopoulos, 2002).
  - Voter participation: greater voter participation appears associated with looser budgetary constraints (Mueller and Stratmann, 2003).

### Key empirical insights and principal findings (authors’ stated conclusions)
- Three main conclusions emphasized by the authors:
  - The quality of institutions continues to matter strongly in determining fiscal outcomes, even when a variety of conditioning variables are considered.
  - Contemporary politics is important and appears to trump the economic conditioning variables. Specific features associated with less fiscal conservatism include:
    - More fragmented government coalitions.
    - An ideological disposition toward the “left.”
    - Greater nationalism.
    - Greater fiscal decentralization.
    - Greater voter participation appears to loosen budgetary constraints.
  - “Deeper” historical and political influences (e.g., ethnic fractionalization and district magnitude) support the idea that more inclusiveness hurts budgetary outcomes; however, evidence on these deeper determinants is unstable, possibly reflecting theoretical ambiguity and limited time variation in the small country sample.
- Policy implication highlighted:
  - Despite the strong influence of history and politics, policy initiatives that establish checks in the competition for fiscal resources (i.e., improvements in fiscal institutions) can materially help curb fiscal pressures.

### Data, scope, and notable sample characteristics
- Time span: 1997–2003 for the main empirical focus; accession date: May 1, 2004 (contextual milestone).
- Cross-country/time-series advantages:
  - Time-series variation in fiscal institutions within countries is a particular strength of the sample.
  - Small-country sample limits: potential absence of time variation in deep structural variables and risk of insufficient information to draw certain conclusions, but principal findings are reported as robust and consistent with related literature.
- Observed expenditure patterns (qualitative examples from the sample):
  - Estonia: managed public finances well, running a primary surplus in some years.
  - Poland: improved fiscal position in the late 1990s (surplus in 1999 and 2000), then drifted back to deficits.
  - Hungary: budget balance generally worsened over the period.
  - Cross-country variation in government size: Baltic nations have small governments; Hungary and Poland and, to a lesser extent, the Czech Republic have large government sizes (expenditure/GDP) that lie above the line showing the tendency for governments to be bigger as per capita incomes increase (Figure 1 referenced).

### Methodological notes and avenues for further analysis
- Authors’ methodological choices:
  - Control for a comprehensive set of economic and political conditioning factors while isolating budget institutions.
  - Grouping historical and political variables to capture multidimensional influences and interactions (e.g., between coalition fragmentation and ideology).
  - Examination of time-invariant country characteristics conditioning the transmission of shocks to fiscal outcomes using Blanchard and Wolfers (2000)-style approaches.
- Suggested further work:
  - Investigation of rule-based institutional measures across a broader set of countries to assess generality.
  - Exploration of nonlinearities and thresholds in the relationship between proportionality/representativeness and fiscal outcomes (as cautioned by Hallerberg and Marier, 2004).

*Source: _wp06123 - References (PDF content provided).*

### Box 1: Fragmentation and Fiscal Discipline

### Box 1: Fragmentation and Fiscal Discipline

### Common-pool problem and political fragmentation
- Analogy: the budget behaves like a common-pool resource—where many can claim benefits but pay only part of the cost, there is pressure to overconsume (Shepsle and Weingast, 1981; Weingast, Shepsle, and Johnson, 1981).
- Implication: tendency to increase public spending in favor of interest groups that bear only a portion of the taxes needed to finance those expenditures; the larger the number of interest groups, the greater the induced spending.
- Dynamic outcome (Velasco, 1999): short-run spending pressures lead to a drawdown of national wealth (or accumulation of debt); deficits continue until debt crosses a threshold and distortionary taxes must be raised when the “writing is on the wall.”

### Parties, coalitions, and strategic spending incentives
- Parties and coalitions aggravate common-pool pressures:
  - Hallerberg and von Hagen (1999): short-term discipline worsens when governing parties can shift taxes onto constituents supporting other parties.
  - Persson, Roland, and Tabellini (2005): competition within coalitions gives each coalition member an incentive to provide narrowly targeted public goods/subsidies to secure loyalties; costs are borne by coalition partners and general taxpayers.
  - Aggregate effect: coalitional governments spend more than single-party governments.

### Ideology, voter participation, and fiscal outcomes
- Ideological effects are ambiguous:
  - Tests of traditional left-right distinctions yield ambiguous results.
  - Cukierman and Tomassi (1998): leftist governments may credibly pursue budgetary conservatism in certain circumstances (“a Nixon to go to China” argument).
  - Ideology is multidimensional and does not map simply to left-right or preferences on fiscal decentralization/nationalism.
- Voter participation:
  - Mueller and Stratmann (2003) find higher voter participation associated with larger governments and slower growth.
  - Interpretation: greater participation increases pressure on governments to deliver, or may increase involvement of uninformed voters leading to worse policies.
  - Empirical treatment: voter participation is treated as an unchanging variable across the sample because variations in the sample are small.

### Fiscal institutions as remedies (C. Budget Institutions and Fiscal Performance)
- Fiscal institutions defined: rules and procedures of budget formation; viewed as a form of “self control” imposed by fiscal actors (Poterba, 1996).
- Objective: improve quality of fiscal decisions, not necessarily depoliticize them (Eichengreen, Hausmann, and von Hagen, 1999).
- Evidence: international and U.S. state evidence suggests fiscal institutions do matter (Alesina and Perotti, 1999).
- Two perspectives to overcome the common-pool problem (see Hallerberg and von Hagen, 1999):
  - Centralized/delegation approach: concentrate budgetary power in key policymakers (e.g., prime minister or finance minister) who internalize costs and benefits.
  - Decentralized/contract approach: cooperative decision making via collective negotiation and commitment to detailed multiannual fiscal targets (e.g., MTBFs).
- Complementary needs: transparent and efficient monitoring and enforcement structures/devices promote fiscal discipline.
- Government-type dependence:
  - One-party governments: strong finance minister can be effective as central delegate.
  - Coalition governments (especially minority coalitions): strong finance minister may be infeasible; contract-based approach with well-informed and transparent rules is preferable. Finance minister’s role shifts toward monitoring/enforcement of the contract rather than proactive budget formation.
- Empirical note for new and potential EU members in the sample: contract-based approach appears appropriate and important; MTBFs are in early stages, while effective delegation to finance ministers with discretionary authority has also helped contain fiscal pressures.

### Budget institutions index and cross-country assessment (III)
- Sample: ten countries—Estonia, Bulgaria, the Czech Republic, Hungary, Latvia, Lithuania, Poland, Romania, the Slovak Republic, and Slovenia.
- Index structure: institutional features grouped in three dimensions:
  - Preparation stage (variables include fiscal rules, quantitative budgetary targets, dominance of finance minister/prime minister).
  - Authorization stage (limits on scope of amendments, sequence of legislative decision making, executive vs parliament relative power, role of the president).
  - Implementation stage (firmness of execution and procedures for adjustments to revenue shortfalls or overspending).
- Key findings (preparation stage):
  - Permanent constraints on budgetary parameters (variable 1): present in only three countries—Estonia, Latvia, and Poland.
    - Estonia: restrictions on foreign borrowing by the state (total foreign outstanding debt and new borrowing) and on state guarantees for foreign loans.
    - Latvia: public borrowing permitted only for financing capital expenditure.
    - Poland: 1997 constitution stipulated public debt-to-GDP ratio must not exceed 60 percent; detailed regulations in the organic budget law of 1999.
  - Multiannual targets under an MTBF (variable 2): only Slovenia uses MTBF multiannual targets.
  - Sequence improvements: Bulgaria and the Czech Republic moved to setting targets before ministries develop budget requests; Lithuania, Poland, and Romania moved to setting targets after receipt of initial requests.
  - Centralization of negotiation structure (variable 3): in most countries the ministry of finance compiles the draft budget; finance minister conducts bilateral negotiations with spending ministries.
  - Concentration of power in reconciling disputes (variable 4): Latvia and Slovenia—prime minister can overrule cabinet; Lithuania (since 1999) and Poland—more decentralized reconciliation with whole cabinet involvement.
- Key findings (authorization stage):
  - Offsetting amendments for higher expenditure (variable 5): Estonia, Lithuania, and Slovenia require higher expenditure amendments to be offset by specific financing to keep overall budget target unchanged. Poland and Bulgaria introduced formal limits on legislative scope to amend the government budget in 1998 and 2003, respectively.
  - Parliamentary sequence (variable 6): only the Czech Republic has the budget committee initially review the draft budget and fix limits to total revenue, expenditure, and deficit that cannot be changed in subsequent readings; in most countries parliament votes on budget aggregates at the end.
  - Executive vs parliament power (variable 7): institutional arrangements favoring the executive in conflicts are considered more conducive to fiscal discipline; only Poland contemplates (i) calling for a vote of confidence in connection with the budget vote, (ii) using the draft budget as the provisional budget if parliament has not adopted a budget before the fiscal year start, and (iii) dissolving parliament if it fails to approve the budget in due time. Czech rules were diluted in 2001 for late approval; Estonia introduced the possibility of a government vote of confidence in 2003.
  - Presidential veto power (variable 8): except for Hungary, Slovenia, and more recently Poland, the president has had some form of power in the budget process. The index presumes presidential veto creates fragmentation and is less conducive to fiscal discipline.
- Key findings (implementation stage):
  - Variables 9, 10, and 11: degree of firmness in budget execution.
  - Variable 12: procedures for fiscal adjustment to unforeseen revenue shortfalls or unexpected overspending.
  - Hungary: position weakened and deteriorated in 2002 due to an amendment allowing considerable leeway to undertake additional spending without supplementary appropriations and parliamentary approval—undermining commitment function and parliamentary authorization.
  - Czech Republic: improved substantially with 2000 budget law requiring higher expenditure to be compensated by decreases elsewhere; position worsened in 2001 when unused funds could be carried into the following year.
  - Finance minister blocking authority in reaction to negative fiscal shocks: Bulgaria, Estonia, Latvia, and Lithuania grant the finance minister authority to block expenditures if unforeseen revenue shortfalls or overspending occur; in remaining countries the cabinet can block expenditure without parliamentary approval.
- Aggregate assessment:
  - Estonia and Slovenia have done consistently well.
  - Poland has considerably improved over the period.
  - Hungary’s position was low and deteriorated.
  - Table 4 shows rankings for 1997 and 2003 (table referenced in source).

### Empirical framework and econometric issues (IV. Benchmark results)
- Objective: explain short-term dynamics of the primary budget balance (balance most under national authorities’ control), using pooled observations from the ten countries over the seven years, 1997 through 2003.
- Empirical specification (equation (1) reproduced in source):
  - y_it = α + β x_it + δ w_it + φ s_it + γ v_i + φ u_t + ε_it  (notation as in source; the source presents the equation form and variable groupings)
  - y_it: fiscal primary balance in country i and year t
  - v_i: unchanging country-specific effects (proxied by country dummies)
  - u_t: effects common to all countries in period t (time dummies)
  - x_it: economic control variables
  - w_it: political control variables
  - s_it: fiscal institutions index
- Three econometric problems highlighted:
  1. Reverse causality / endogeneity: budget outcomes may influence evolution of fiscal institutions; identifying an instrument that affects fiscal institutions but not budget outcomes is difficult. Prior literature notes this unresolved issue (Alesina and Perotti, 1999; Stein, Talvi, and Grisanti, 1999; Knight and Levinson, 2000; Perotti and Kontopoulos, 2002). Working assumption: budget performance cannot quickly feed back into alteration of fiscal institutions since these are costly to change. Arellano-Bond procedure reported as addressing the problem but not conclusive.
  2. Omitted variables: omitted factors could be driving budget outcomes, leading to overstatement of fiscal institutions effect. Partial solution: focus on within-country variations over time to eliminate v_i; this benchmark within-country analysis is used where data permit.
  3. Loss of informative cross-country heterogeneity: sweeping away v_i via within-country analysis discards unchanging country effects of analytical interest. Blanchard and Wolfers (2000) propose allowing reactions to time-varying variables to be conditioned by unchanging country characteristics (see Box 2 in source).
- Data pooling period: 1997 through 2003.
- Figure 3 (Hungary, 1997-2003): source shows primary balance (in percent of GDP, LHS) moving from 1.3 through 1.7 across 1997–2003 and fiscal institutions quality index (RHS); data series listed in the source: Primary balance values—1.3, 1.3, 1.4, 1.4, 1.5, 1.5, 1.6, 1.6, 1.7 (plotted across 1997–2003 in figure).

*Italic: Box 1: Fragmentation and Fiscal Discipline, extracted from the provided IMF working paper content.*

### Appendix B provides details of the explanatory variables used and their data sources. Tables 2

### _wp06123 - Appendix B provides details of the explanatory variables used and their data sources. Tables 2

### Data scope and cross-sectional summaries
- Sample: ten countries.
- Cross-sectional regressions are briefly presented (in Table 5) to contrast with subsequent results.
- Appendix B lists explanatory variables and data sources; Tables 2 and 3 summarize means and correlations and show considerable heterogeneity even within this relatively homogenous group.
- Cross-section findings:
  - Higher public debt-to-GDP ratio is associated with a larger primary surplus (or smaller deficit) and is significant in some specifications, suggesting greater fiscal effort when debt payments increase.
  - Of four democratic-practice variables (fragmentation of the government coalition and three ideology variables), only the nationalism variable shows modest significance in the cross section: more nationalistic ideology tends to reduce the budget surplus.
  - Among longer-term determinants, district magnitude (logged) shows a larger district magnitude results in a smaller long-term surplus (log used to limit influence of Slovak Republic).
  - An IMF-supported program tends to be associated with a larger primary surplus.
  - Controlling for an IMF-supported program, there is some hint that better budgetary institutions help increase the primary surplus.
  - These cross-sectional results align with Gleich (2003) and Yläoutine (2004).

### Panel sample construction and within-country analysis
- Intended panel observations: seven observations for each of the ten countries, potential total of 70 observations.
- Lost observations:
  - Five observations lost because ideological orientation of some political parties could not be identified: 1997 for Lithuania; 1997 and 1998 for Latvia; 1997 and 1998 for the Slovak Republic.
  - Two observations dropped (Bulgaria and Romania in 1997) due to very high inflation.
- Final working sample: 63 observations.
- Estimation approach: include country fixed effects (explaining deviations from country mean) and year dummies to control for common shocks.

### A. Economic influences (within-country, Table 6)
- Debt and fiscal effort:
  - Higher debt level apparently induces greater fiscal effort, increasing the primary balance.
  - Sign on debt variable is always positive but only sometimes statistically significant at conventional levels.
- Unemployment:
  - Unemployment rate has a negative sign and is more often close to statistical significance: an increase in the unemployment rate reduces the primary surplus (increases the deficit).
- Inflation:
  - Higher inflation rate associated with a larger primary surplus, consistent with inflation reducing the real value of expenditures without compromising tax receipts (consistent with Perotti and Kontopoulos (2002), though their significance was higher).
- Trade openness:
  - Country openness to external trade typically becomes solidly significant when other conditioning variables are included; more open countries tend toward greater fiscal conservatism.
  - Significance of economic variables typically falls when pitted against political variables, especially in nonlinear regressions.
- Budgetary institutions index:
  - Stronger budgetary institutions are associated with a larger primary surplus (or smaller deficit).
  - Coefficient significant at the conventional 5 percent level of significance.
  - A movement of the index from the 25th percentile to the median results in an improvement of about 2 percentage points of GDP.
- External anchors:
  - EU dummy (timed to take value 1 from the year negotiations for EU entry were initiated): sign is negative, indicating the prospect of EU entry may raise the deficit, though effect not statistically significant.
  - IMF-supported program dummy: not significant in the within regression; the cross-section association of IMF programs with smaller deficits disappears once unobserved country factors are controlled for.

### B. Political influences (time-varying democratic-practice variables)
- Variables considered: coalition fragmentation, three ideology variables, voter participation (treated as time-invariant for later analysis).
- Single-variable tests: political variables show little action when considered one at a time (columns 1–4).
- Combined variables:
  - Putting the three ideological variables together increases significance (column 5), though not especially high.
  - Including coalitional fragmentation with the three ideology variables sharply increases statistical significance (column 6).
- Methodological implications:
  - Fragmentation and ideology need to be examined together; ideology is multifaceted rather than unidimensional.
  - A larger number on the fragmentation variable (Herfindahl index from coalition shares) indicates less fragmentation; positive coefficient implies a larger surplus with reduced fragmentation.
- Ideology effects:
  - Coalitions leaning to the right, not highly nationalist, and favoring centralization of public finances are likely to deliver a conservative budget.
  - In the sample, leftist coalitions have been less fragmented; some rightist coalitions have had nationalistic tendencies.
  - Only when dimensions are considered simultaneously do effects become clear.

### C. Economics versus politics (robustness and substantive conclusions)
- Estimation approaches and robustness:
  - Random-effects estimation (column 1) uses cross-country and within-country variation but requires stringent conditions.
  - Fixed-effects (within-country) models (columns 2 and 3) suggest omitted-variable correlation may not be serious; Hausman test does not reject random-effects.
  - Lagged debt coefficient changes significantly between random- and fixed-effects: across countries higher debt induces greater fiscal conservatism, but same effect not observed within a country over time (debt varies little within country relative to cross-country variation).
  - Serial correlation: Lagrange multiplier test following Baltagi (2005) suggests no serial correlation (test approximate for unbalanced panels).
  - Dynamic estimators:
    - Arellano-Bond estimator allows for persistence and reverse causality (Table 8, column 4) but performs better in large samples.
    - Kiviet bias adjustment works better for unbalanced panels with a small number of units (Table 8, column 5).
  - Given small sample size and sample drops when using lags, heavy reliance on dynamic estimators is cautioned against.
  - Lagged dependent variable not statistically significant.
  - Benchmark used: fixed-effects (within-country) estimates without lagged dependent and debt variables (column (5) of Table 8).
- Outlier sensitivity:
  - Excluding one country at a time (Milesi-Ferretti, Perotti, and Rostagno (2002) approach) indicates the fiscal institutions index relevance is not driven by any single country.
  - Specific significance levels for fiscal institutions index when excluding countries:
    - Significant at the 7 percent level when either Hungary or Poland is dropped.
    - Significant at the 5 percent level when one of the other countries or both Poland and Romania are dropped.
- Substantive comparison:
  - When economic conditioning variables are combined with political variables (Table 9), politics tends to dominate: economic variables decline in strength though signs remain.
  - Among political variables, left-right and nationalist ideology measures remain most clearly significant.
  - Inclusion of EU dummy, IMF dummy, and timing of elections does not change results.
  - Overall index of budgetary institutions remains highly significant when added to specifications including economic and political conditioning variables.
  - Principal finding: politics has a strong influence on the budget, but discipline is possible through checks and balances of budgetary processes and institutions.

### V. Role of deeper (time-invariant) determinants
- Concern: political determinants and budgetary institutions effects may reflect deeper underlying, largely time-invariant variables.
- Methodology:
  - Use interaction of shocks with time-invariant variables in a nonlinear regression (Box 2), following Blanchard and Wolfers (2000).
  - Two shock types: country-specific shocks (time-varying country variables) and common shocks (time dummies).
  - Method evaluates whether shocks persist longer when deeper determinants are more salient; longer persistence in short span implies higher average realization.
- Theoretical caution: expectations on direction of influence of time-invariant political institutions may be ambiguous and nonlinear; references to Aghion, Alesina, and Trebbi (2004) noted.

*Content derived from the provided document excerpt.*

### Box 2: Shocks and Institutions

### Box 2: Shocks and Institutions

### Methodology and Empirical Strategy
- Objective: disentangle influence of economic shocks (changes in oil prices and shifts in productivity trends) from that of sluggish labor market institutions on evolution of unemployment; apply approach to fiscal outcomes and institutions.
- Estimation approach:
  - Use country-fixed effects to eliminate unobserved, unchanging country influences (νi+ηzi).
  - Estimate a nonlinear specification that allows time-invariant variables (zi) to condition responses to common shocks (ut) and country-specific shocks (xit).
  - Crucial parameter λ captures how time-invariant institutions (zi) shape the common response βut, where zi is entered as deviation from sample mean so βut is the response when zi is at the mean.
  - Coefficient estimates obtained by nonlinear least squares; country-fixed effects control for country-specific averages.
- Practical considerations:
  - Specification implies single common mediating coefficient λ across shocks; allowing differential responses increases parameterization rapidly and may be difficult to identify.
  - Sample size constraints noted: sample of only ten countries increases instability of coefficients on time-invariant variables.

### Main Results (Steps and Robustness)
- Stepwise presentation of findings (referring to Tables 10–11):
  - Step 1 (columns 1–3, Table 10): allow time-invariant variables to mediate domestic shocks (inflation, employment, trade-to-GDP) and common shocks (time dummies).
    - Greater ethnic diversity appears to amplify adverse shocks; sign not significant in this specification.
    - Larger district magnitude and greater voter turnout also amplify adverse shocks; statistical significance varies between the 5 and 10 percent range.
  - Step 2 (columns 4–6, Table 10): add lagged debt-to-GDP ratio as a time-invariant variable.
    - Time variation in debt-to-GDP had limited bearing on movements in the primary surplus; much variation is across countries rather than within short periods.
    - Level of debt conditions responses: positive shocks to the budget are muted and negative shocks are amplified in countries with heavier debt ratios—contrary to simple mean reversion expectations.
    - Interpretation: in short run, countries with higher debt may be less conservative when budget improves; debt may need to cross a threshold before inducing tightening.
    - With debt ratio included, most results are strengthened.
    - Contemporary political variables (coalitional fragmentation and three ideological variables) become highly significant when time-invariant variables are included.
  - Effect size for fiscal institutions:
    - Budgetary institutions index remains strongly significant.
    - Point estimate relative to earlier estimates is now somewhat lower, at about 5½–6.
    - Suggested interpretation: a move in institutional quality from the 25th to the 50th percentile leads to an improvement in the primary surplus of about 1½ percent of GDP.
  - Steps 3–4 (Table 11): allow shocks to fiscal institutions to be conditioned by time-invariant variables (columns 1–3 without debt ratio; columns 4–6 with debt ratio).
    - Each country’s response to improvements in fiscal institutions varies with its time-invariant variables.
    - Stronger fiscal institutions help everywhere except possibly in the Slovak Republic, which combines a large district magnitude and high voter participation.
    - A linear conditioning response is imposed; nonlinear possibilities not pursued in depth due to paper’s scope.

- Notes on statistical stability:
  - Effects of time-invariant variables are sensitive to specifications; direction often plausible but strength and statistical significance are unstable.
  - Greater instability in coefficients on time-invariant variables is expected given small sample (ten countries) and short time series.

### Model Predictions and Examples
- Two illustrative country fits (Figure 4): Poland and Hungary (1997–2003).
  - Predictions match actual primary balance values reasonably well in magnitudes and directional changes.
  - Poland: early improvement partly reflects improvement in domestic fiscal institutions; subsequent deterioration reflects fall in inflation exposing underlying weaknesses (some apparent strength due to higher inflation).
  - Hungary: worsening predicted budget balance in last two years reflects worsening institutions.
- Estimation basis shown in figure: Nonlinear least squares estimates, in percent of GDP, covering 1997–2003.

### Components of the Budget Institutions Index
- Decomposition (Table 12) into three stages: preparation, authorization (bargaining/legal authorization), and implementation.
  - All three components have strong and independent influence, with the implementation stage most relevant.
  - Interpretation: rules and checks in preparation and authorization matter, but the greatest danger arises from weak budget implementation; checks are needed to ensure first two phases are not undone.
  - Political fragmentation and ideological variables remain strong influences across stages.
  - Voter participation:
    - Strong influence in preparation and authorization phases.
    - Smaller influence in implementation phase, where actions are presumably less visible to voters.
- Substitution test (column 4):
  - Created an index taking the lowest of the three phase values per country-year to test whether strong performance in some phases can compensate for weak links.
  - Results suggest phases are not substitutes; a weakest phase can undermine fiscal responsibility—policy attention required to all phases.
- Delegation vs contract-based approaches (columns 5–6):
  - Delegation-approach index formed by items 2–7 and 9–12; contract-based index by items 1, 5–7, and 9–11.
  - Both indices are statistically important in the sample.
  - Finding does not align cleanly with Hallerberg and von Hagen (1999) hypothesis that contract-based approach is most relevant for multiparty coalitions; delegation and contract indices may be correlated when institutions are “young.”
  - Possible future divergence in central and eastern Europe analogous to EU-15 evidence after 1998, but interpretation requires country-specific analysis.
  - Caution: empirical evidence that institutions matter is stronger than evidence on mechanisms by which they matter.

### Expenditures and Revenues (Box 3 and Tables 13–14)
- Joint estimation of expenditure and revenue equations using seemingly unrelated regressions to allow correlated errors and improve efficiency.
- Parametrization used in Box 3:
  - For each equation, country-fixed effects remove unchanging country features; observed variables proxy common shocks (ut).
  - Observable shocks used: GDP growth in the EU-15, oil price movements, and average of lagged primary balances in Germany and France.
- Key findings:
  - Tax-to-GDP ratio movements are difficult to explain; inertia in tax receipts (significant coefficient on lagged tax variable) explains part of the movement.
  - External shocks play a more salient role for tax receipts:
    - Rise in oil prices most consistently related to higher tax receipts (possibly via import and excise duties).
    - Large surplus in Germany and France associated with a lower tax-to-GDP ratio.
  - Discipline effect mainly operates through the expenditure side rather than revenues; externally induced gains on expenditures are partially relinquished on tax receipts.
- Interaction of external shocks with time-invariant variables:
  - Fiscal institutions remain strong and robust in affecting expenditures (Table 14).
  - Interaction patterns are mixed and shock-specific:
    - Larger district magnitude weakly associated with higher expenditures when EU growth increases.
    - Larger district magnitude associated with lower tax receipts when oil prices rise.
    - Larger voter turnout appears to operate through higher expenditures when there is an oil price shock.

### Conclusions and Policy Implications
- Politics substantially influences budget outcomes for the examined group of new and potential EU members over 1997-2003.
  - Politics operates through both long-term determinants and contemporary democratic practice; contemporary politics exerts independent influence on budgets.
  - Contemporary political influence best represented by a vector of attributes: government fragmentation, ideological predispositions (left-right, nationalism, emphasis on decentralization), and voter participation.
  - Government fragmentation’s importance often depends on ideological orientation.
  - Voter participation findings (consistent with Mueller and Stratmann 2003) suggest greater democratic participation is accommodated by increased budgetary indiscipline.
- Hope and role of institutions:
  - Quality of fiscal institutions materially affects budgetary discipline, primarily by constraining expenditures.
  - Determinants of good fiscal institutions are not addressed in depth; evidence suggests a window of opportunity during transition can allow reform, but progress is uneven and slippages occur.
  - Fiscal institutions correlate with contemporary politics and long-term structural features; their effectiveness may decline if political forces move away from fiscal conservatism.
  - Principal message: long-term determinants do not fully predetermine outcomes—institutions and contemporary politics matter, and reforms can improve discipline, but outcomes depend on political behavior and country-specific conditions.

*Source: Box 2: Shocks and Institutions, _wp06123*

### REFERENCES

### REFERENCES

### Bibliographic list (selected from source)
- Acemoglu, D., 2005, “Constitutions, Politics, and Economics,” Journal of Economic Literature, Vol. 43, No. 4, pp. 1025-1048.
- Aghion, P., A. Alesina, and F. Trebbi, 2004, “Endogenous Political Institutions,” Quarterly Journal of Economics, Vol. 119 (May), pp. 565–612.
- Alesina, A., R. Hausmann, R. Hommes, and E. Stein, 1999, “Budget Institutions and Fiscal Performance in Latin America,” Journal of Development Economics, Vol. 59, No. 2, pp. 253–273.
- Alesina, A., and R. Perotti, 1999, “Budget Deficits and Budget Institutions,” in Fiscal Institutions and Fiscal Performance, ed. by J. Poterba and J. von Hagen (Chicago: University of Chicago Press).
- Allen R., and D. Tommasi, 2001, Managing Public Expenditure: A Reference Book for Transition Countries (Paris: OECD).
- Alt, J., and R. Lowry, 1994, “Divided Government, Fiscal Institutions, and Budget Deficits: Evidence from the States,” American Political Science Review, Vol. 88, pp. 811–828.
- Annett, A., J. Decressin, and M. Deppler, 2005, “Reforming the Stability and Growth Pact, IMF Policy Discussion Paper, PDP/05/2 (Washington D.C.: International Monetary Fund).
- Baltagi, B. H., 2005, Econometric Analysis of Panel Data (Chichester; Hoboken, NJ: J. Wiley and Sons).
- Besley, T., and A. Case, 2003, “Political Institutions and Policy Choices: Evidence from the United States,” Journal of Economic Literature, Vol. 41, No. 1, pp. 7–73.
- Blanchard, O., 2005, “European Unemployment: The Evolution of Facts and Ideas,” Massachusetts Institute of Technology, Department of Economics, Working Paper 05/24 (Cambridge: Massachusetts).
- Blanchard, O., and J. Wolfers, 2000, “The Role of Shocks and Institutions in the Rise of European Unemployment: The aggregate Evidence,” 1999 Harry Johnson Lecture, Economic Journal, Vol. 100 (March), pp. C1-33.
- Benoit, K., and M. Laver, 2006, Party Policy in Modern Democracies (London: Routledge).
- Bohn, H., 1998, “The Behavior of U.S. Public Debt and Deficits,” Quarterly Journal of Economics, Vol. 113, pp. 949-963.
- Bruno G., 2005, “Estimation and Inference in Dynamics Unbalanced Panel Data Models with a Small Number of Individuals,” Stata Journal, StataCorp LP, Vol. 5, No. 4, pp. 473–500.
- Cukierman, A. and M. Tomassi, 1998, “When Does It Take a Nixon to Go to China?” American Economic Review, Vol. 88, No. 1, pp. 180–197.
- Eichengreen, B., R. Hausmann, J. von Hagen, 1999, “Reforming Budgetary Institutions in Latin America: The Case for a National Fiscal Council,” Open Economies Review, Vol. 10, pp. 415–442.
- Franzese, R., 1999, “Partially Independent Central Banks, Politically Responsive Governments, and Inflation,” American Journal of Political Science, Vol. 43, No. 3, pp. 681–706.
- Gleich, H., 2003, “Budget Institutions and Fiscal Performance in Central and Eastern European Countries,” European Central Bank, Working Paper No. 215 (Frankfurt, Germany: European Central Bank).
- Hahm, S., M. Kamlet, and D. Mowery, “The Political Economy of Deficit Spending in Nine Industrialized Parliamentary Democracies,” Comparative Political Studies, Vol. 29, No. 1, pp. 52–77.
- Hallerberg, M., and P. Marier, 2004, “Executive Authority, the Personal Vote, and Budget Discipline in Latin American and Caribbean Countries,” American Journal of Political Science, Vol. 48, No. 3, pp. 571–87.
- Hallerberg, M., R. Strauch, and J. von Hagen, 2004, “Budgeting in Europe After Maastricht: Patterns of Reforms and Their Effectiveness,” Hacienda Publica Espanola, No. 167, pp. 201–25.
- Hallerberg, M., and J. von Hagen, 1999, “Electoral Institutions, Cabinet Negotiations, and Budget Deficits in the European Union” in Fiscal Institutions and Fiscal Performance, ed. by J. Poterba and J. von Hagen (Chicago: University of Chicago Press).
- Knight, B. and A. Levinson, 2000, “Fiscal Institutions in US States,” in Institutions, Politics, and Fiscal Policy, ed. By R. Strauch and J. von Hagen (Boston: Kluwer Academic Publishers).
- Kontopoulos, Y., and R. Perotti, 1999, “Government Fragmentation and Fiscal Policy Outcomes: Evidence from OECD Countries,” in Fiscal Institutions and Fiscal Performance, ed. by J. Poterba and J. von Hagen (Chicago: University of Chicago Press).
- Lijphart, A, 1994, Electoral Systems and Party Systems (Oxford: Oxford University Press).
- Milesi-Ferretti, G.M., R. Perotti, and M. Rostagno, 2002, “Electoral Systems and Public Spending,” Quarterly Journal of Economics (May), pp. 609–657.
- Mueller Dennis, and Thomas Stratmann, 2003, “The economic effects of democratic participation,” Journal of Public Economics, Vol. 87, Nos. 9–10, pp. 2129–2155.
- Perotti, R. and Y. Kontopoulos, 2002, “Fragmented Fiscal Policy,” Journal of Public Economics, Vol. 86, pp. 191–222.
- Persson, T., 2002, “Do Political Institutions Shape Economic Policy?” Econometrica, Vol. 70 (May), pp. 883–905.
- Persson, T., 2004, “Consequences of Constitutions,” Journal of the European Economic Association, Vol. 2, Nos. 2–3, pp. 139–161.
- Persson, T., G. Roland, and G. Tabellini, 2005, “Electoral Rules and Government Spending in Parliamentary Democracies,” unpublished.
- Persson, T., and G. Tabellini, 2000, Political Economics: Explaining Economic Policy, (Cambridge, Massachusetts: MIT Press).
- Persson, T., and G. Tabellini, 2003, The Economic Effects of Constitutions, (Cambridge, Massachusetts: MIT Press).
- Persson, T., and G. Tabellini, 2004, “Constitutional Rules and Fiscal Policy Outcome,” American Economic Review, Vol. 94 (March), pp.25–45.
- Poterba, J., 1994, “State Responses to Fiscal Crises: The Effects of Budgetary Institutions and Politics,” Journal of Political Economy, Vol. 102, No. 4, pp. 799–821.
- Poterba, J., 1996, “Do Budget Rules Work?” National Bureau of Economic Research Working Paper 5550 (Cambridge, Massachusetts).
- Potter, B., and J. Diamond, 1999, Guidelines for Public Expenditure Management, (Washington D.C.: International Monetary Fund).
- Roubini, N. and J. Sachs, 1989, “Government Spending and Budget Deficits in the Industrial Democracies,” Economic Policy, Vol. 8, pp. 100–132.
- Shepsle, K. and B. Weingast, 1981, “Political Preferences for the Pork Barrel: A Generalization,” American Journal of Political Science, Vol. 25, No. 1, pp. 96–111.
- Stein, E., E. Talvi, and A. Grisanti, 1999, “Institutional Arrangements and Fiscal Performance: The Latin American Experience,” in Fiscal Institutions and Fiscal Performance, ed. by J. Poterba and J. von Hagen (Chicago: University of Chicago Press).
- Velasco, A., 1999, “A Model of Endogenous Fiscal Deficits and Delayed Fiscal Reforms,” in Fiscal Institutions and Fiscal Performance, ed. by J. Poterba and J. von Hagen (Chicago: University of Chicago Press).
- Velasco, A., 2000, “Debts and Deficits With Fragmented Fiscal Policymaking,” Journal of Public Economics, Vol. 76 (April), pp.105–25.
- von Hagen, J., and I. Harden, 1995, “Budget Processes and Commitment to Fiscal Discipline,” European Economic Review, Vol. 39 (April), pp.771–9.
- Weingast, B., K. Shepsle, and C. Johnson, 1981, “The Political Economy of Benefits and Costs: A Neoclassical Approach to Distributive Politics,” Journal of Political Economy, Vol. 89, pp. 642–664.
- Wyplosz, C., 2005, “Fiscal Policy: Institutions versus Rules,” National Institute Economic Review, Vol. 191 (January), pp. 64–78.
- Yläoutinen, S., 2004, “Fiscal Frameworks in the Central and Eastern European Countries,” Finnish Ministry of Finance Discussion Paper No. 72 (Ministry of Finance of Finland).

*Source: content unit _wp06123 - REFERENCES*

---

### Data description and descriptive statistics

### Table 1: Country means (selected)
- Bulgaria: Ethnic fractionalization 0.40; Average district magnitude 7.7; Voter turnout 66.9
- Czech Republic: Ethnic fractionalization 0.32; Average district magnitude 25.0; Voter turnout 69.4
- Estonia: Ethnic fractionalization 0.51; Average district magnitude 9.2; Voter turnout 61.5
- Hungary: Ethnic fractionalization 0.15; Average district magnitude 12.2; Voter turnout 66.4
- Latvia: Ethnic fractionalization 0.59; Average district magnitude 20.0; Voter turnout 76.2
- Lithuania: Ethnic fractionalization 0.32; Average district magnitude 35.3; Voter turnout 62.1
- Poland: Ethnic fractionalization 0.12; Average district magnitude 16.7; Voter turnout 48.7
- Romania: Ethnic fractionalization 0.31; Average district magnitude 7.8; Voter turnout 70.7
- Slovak Republic: Ethnic fractionalization 0.25; Average district magnitude 150.0; Voter turnout 76.6
- Slovenia: Ethnic fractionalization 0.22; Average district magnitude 11.0; Voter turnout 76.7

Source: Authors' calculations.

### Table 2: Correlations (selected pairwise entries)
- Government fragmentation with Government ideology (nationalism): -0.04
- Government fragmentation with Government ideology (decentralization): -0.07
- Government fragmentation with Government ideology (left/right): -0.37
- Fiscal institutions index with Government ideology (left/right): -0.32
- Ethnic Fractionalization with Government ideology (left/right): 0.53
- District magnitude with Voter turnout: 0.30

Source: Authors' calculations.

### Table 3: Descriptive statistics (selected variables)
- Unemployment rate overall 11.29; Std. Dev. 4.23; Min 4.30; Max 20.00
  - between 4.06; 6.16; 18.11
  - within 1.98; 6.05; 15.65
- Inflation overall 9.17; Std. Dev. 11.12; Min -1.20; Max 59.10
  - between 10.20; 1.22; 37.15
  - within 5.75; -12.68; 31.12
- Openness index overall 0.90; Std. Dev. 0.25; Min 0.46; Max 1.36
- Fiscal institutions index overall 2.12; Std. Dev. 0.45; Min 1.37; Max 2.72
- Government fragmentation overall 0.63; Std. Dev. 0.20; Min 0.27; Max 1.00
- Left/right overall 11.11; Std. Dev. 3.53; Min 5.81; Max 17.35
- Nationalism overall 11.44; Std. Dev. 3.48; Min 6.80; Max 17.57
- Fiscal centralization overall 10.46; Std. Dev. 2.00; Min 7.09; Max 13.46
- Ethnic fractionalization overall 0.31; Std. Dev. 0.14; Min 0.12; Max 0.59
- District magnitude (logarithms) overall 2.83; Std. Dev. 0.81; Min 2.00; Max 5.00
- Voter turnout overall 67.00; Std. Dev. 8.39; Min 48.70; Max 76.70
- Debt-to-GDP (lagged) overall 33.99; Std. Dev. 23.34; Min 5.80; Max 107.50

Source: Authors' calculations.

---

### Key empirical results and numeric estimates

### Cross-country regressions (Table 5)
- Debt-to-GDP ratio coefficients across specifications: 0.06, 0.05, 0.08, 0.06, 0.06 (standard errors reported in table).
- IMF program dummy coefficients: 2.79, 3.09, 1.909 (standard errors reported; some significant at 10% or 5%).
- Fiscal institutions index coefficients reported: 2.74 (1.21)* and 1.52 (1.18) in select specifications.
- District magnitude coefficients: -1.5 and -1.03 (standard errors including (0.5)** and (0.53)).

Note: Table 5 reports the primary balance-to-GDP regressions with country counts: Number of countries 10.

### Panel data regressions (Fixed Effects) (Table 6)
- Fiscal institutions index coefficients: 7.52 (2.08)*** and 7.97 (2.12)*** in columns where included.
- Unemployment rate coefficients: e.g., -0.34 with (0.16)** and significance levels reported across columns.
- Openness index coefficients: e.g., 1.58, 4.78, 7.89, 7.95 with appropriate standard errors; some significant at 10% or 5%.
- Observations: 63; Number of countries: 10.

### Government fragmentation and ideology (Table 7 and Table 9)
- Government fragmentation coefficients vary: e.g., -0.82 (1.81) in one column; 6.23 (2.74)** in another.
- Government ideology (Left/right) positive and often significant: examples include 0.07 (0.13), 0.16 (0.12), 0.49 (0.19)**.
- Government ideology (Nationalism) negative and often significant: examples include -0.24, -0.39, -0.66 with standard errors and significance levels indicated.
- Fixed effects panel (Table 9) coefficients summary:
  - Unemployment rate: -0.33 (0.17)*, -0.36 (0.17)**, -0.33 (0.17)*, -0.35 (0.18)*
  - Openness index: 8.97 (4.41)**, 8.88 (4.38)**, 9.19 (4.72)*, 9.02 (4.85)*
  - Government fragmentation: 4.60 (2.91), 4.19 (2.87), 4.71 (2.82), 4.26 (3.05)

### Alternative panel estimators and dynamics (Table 8)
- Random effects and fixed effects specifications reported; example coefficients:
  - Debt-to-GDP ratio (lagged): 0.09 (0.02)*** in column 1.
  - Fiscal institutions index: 5.09 (1.19)***, 6.20 (2.13)***, 9.03 (4.42)** across columns.
  - Government fragmentation: 1.78 (1.72), 6.78 (2.96)**, 4.39 (2.84) with various estimator columns.
- Column 4 uses Arellano-Bond; Sargan test reports Pr≥χ2 = 0.27 and Pr≥z = 0.93 for AR(2) test in that specification.
- Observations vary by column: 63, 43, 53 depending on estimator.

### Nonlinear least squares: deeper determinants (Tables 10 and 11)
- Fiscal institutions index positive and significant across NLS specifications: examples include 5.47 (2.45)**, 5.38 (2.40)**, 5.86 (2.33)**, up to 6.28 (2.52)**.
- Government fragmentation large positive coefficients: examples 8.12 (2.87)***, 7.69 (2.83)**, 6.10 (2.72)**, 9.74 (2.99)***.
- Government ideology:
  - Left/right: positive and significant (e.g., 0.46 (0.18)***).
  - Nationalism: negative and significant (e.g., -0.77 (0.21)***).
  - Fiscal decentralization: positive and significant (e.g., 0.95 (0.30)***).
- Time-invariant variables:
  - Ethnic fractionalization coefficients reported (e.g., -4.34, -3.59) with standard errors and significance in some columns.
  - District magnitude negative: e.g., -1.44 (0.85)*, -1.34 (0.70)*, or -1.18 (0.64)*.
  - Voter turnover negative: e.g., -0.13 (0.05)**, -0.12 (0.05)**, -0.09 (0.04)**.
- Adjusted R-squared reported: examples 0.63, 0.64 in different specifications.
- Table 11 replicates similar patterns with alternate specifications and reports Adjusted R-squared up to 0.67 in some columns.

### Budget process stages and delegation/contract indices (Table 12)
- Fiscal institutions index by stage (examples):
  - Budget preparation phase: Fiscal institutions index 2.9 (1.41)** in column 1.
  - Budget authorization phase: 1.33 (0.53)** in column 2.
  - Budget implementation phase: 3.79 (1.36)*** in column 3.
- Delegation approach and Contract-based approach index coefficients:
  - Delegation approach: 14.99 (11.28) constant reported; fiscal institutions index 4.49 (1.47)*** in column for delegation.
  - Contract-based approach: constant 6.02 (7.98); fiscal institutions index 2.95 (0.95)***.
- Government fragmentation and ideology effects remain statistically significant across these stage-specific specifications.
- Observations: 63; Adjusted R-squared examples: 0.61, 0.62, 0.64, 0.65.

### Expenditure and revenue decompositions (Tables 13 and 14)
- Seemingly unrelated regressions for Exp-to-GDP and Rev-to-GDP (multiple common shocks included):
  - Lagged expenditure coefficient: 0.38 (0.08)*** in several columns.
  - Lagged revenue coefficient: 0.26 (0.08)***.
  - Unemployment rate positive effect on expenditure: 0.41 (0.13)*** and smaller on revenue: 0.08 (0.08).
  - Openness index negative effect on expenditure in some specifications: e.g., -6.46 (3.33)*.
  - Fiscal institutions index negative effect on expenditure and revenue in these specifications: -3.59 (1.64)** and -1.02 (1.11).
  - Shock variables (EU15 growth, oil price, Germany/France primary balance) have varying impacts; examples:
    - Shock coefficient -0.56 (0.28)** on Exp-to-GDP in one specification; 0.96 (0.25)*** in another.
- Table 14 extends results with interacted terms:
  - Interacted shock × District Magnitude example: 0.47 (0.26)* on Exp-to-GDP in one column; -0.07 (0.05) in another with significance reported.
  - Fiscal institutions index remains negative on expenditure: -3.83 (1.62)** and on revenue: -0.92 (1.11) in one specification.
- Observations in these tables: 63; time and country dummies included.

---

### Appendix A — Construction of the Budget Institutions Index

- Numerical indices assigned values ranging from 0 to 4 to qualitative features of budget institutions following Gleich (2003), updated with Yläoutinen (2004), IMF fiscal ROSC reports, and direct contact with authorities.
- Budget process stages and variables:
  - Preparation stage variables (Variables 1–4):
    - Variable 1: strictness of permanent constraints on budgetary parameters (legal limits on deficits/borrowing).
    - Variable 2: control by the finance minister in setting fiscal targets and ceilings.
    - Variable 3: power of the finance minister in compiling and negotiating the draft budget.
    - Variable 4: how remaining disputes are reconciled in the executive (degree of centralization/delegation).
  - Authorization stage variables (Variables 5–8):
    - Variable 5: formal constraints on legislature’s scope to amend the government budget.
    - Variable 6: sequence of decision making during budget deliberation (whether major aggregates decided before details).
    - Variable 7: institutional devices summarizing executive strength vis-à-vis parliament during deliberation.
    - Variable 8: power of the president in the budget process.
  - Implementation stage variables (Variables 9–12):
    - Variable 9: requirement of parliamentary approval for supplementary budgets.
    - Variable 10: requirement of parliamentary approval for transfers between ministries.
    - Variable 11: inability to carry over unused funds to the next year (seen as discipline-enhancing).
    - Variable 12: finance minister’s ability to block expenditures (ranked by whether blocking requires minister, cabinet, parliament, or no approval).
- Indices construction formulas:
  - Budget preparation index = 1/4 ∑_{i=1}^{4} x_i
  - Budget authorization index = 1/4 ∑_{i=5}^{8} x_i
  - Budget implementation index = 1/4 ∑_{i=10}^{13} x_i
  - Overall index = simple average of the three component indices.
- Delegation and contract-based indices:
  - Delegation-approach index formed by summing items 2–7 and 9–12 (emphasizing centralized fiscal authority).
  - Contract-based index formed by summing items 1, 5–7, and 9–11 (emphasizing well-informed and transparent rules; minister of finance in monitoring/enforcement role).
- Appendix notes that Appendix Table 2 (in source) reports underlying sub-measures and year(s) of change for underlying items.

*Source: content unit _wp06123 - REFERENCES*

### 1. Existence of statutorily mandated fiscal rules0.25

### 1. Existence of statutorily mandated fiscal rules0.25

### A. Preparation Stage
- 1. Existence of statutorily mandated fiscal rules0.25
  - a. Balanced budget rule.4.00
  - b. Limits on public borrowing.2.00
  - c. No legal limits on borrowing.0.00

- 2. Sequence of budgetary decision making0.25
  - a.  MF sets forth aggregate and specific budget targets in initial budget circular4.00
  - b. MF proposes, cabinet decides on targets for budget aggregates and spending limits are assigned to each ministry before spending ministries develop3.00 budget requests.
  - c. MF proposes, cabinet decides on targets for budget aggregates before spending ministries develop budget requests.2.00
  - d. Budgetary targets are set on the basis of preliminary budget requests.1.00
  - e. No budget targets are determined.0.00

- 3. Compilation of the draft budget0.25
  - a. Finance ministry holds bilateral negotiations with each spending ministry.4.00
  - b. Finance ministry holds bilateral negotiations, other parties included.2.00
  - c. Finance ministry only collects budget requests and compiles summary for0.00 cabinet session.

- 4. Members of executive responsible for reconciling conflicts over budget bids0.25
  - a. MF or PM can veto or overrule cabinet decision.4.00
  - b. Senior cabinet committee, then whole council of ministers or cabinet.2.00
  - c. Executive collectively (e.g., council of ministers or cabinet).0.00

### B. Legislation
- 5. Constraints on the legislature to amend the government's draft budget0.25
  - a. Deficit provided in the draft budget cannot be exceeded, or4.00 individual amendments have to indicate offsetting changes.
  - b. No restrictions.0.00

- 6. Sequence of votes0.25
  - a. Initial vote on total budget revenues, expenditures, and the deficit.4.00
  - b. Final vote on budget aggregates.0.00

- 7. Relative power of the executive vis-à-vis the parliament0.25
  - a. Cabinet can combine a vote of confidence with a  vote on the budget.0.334.00
  - b. Draft budget is executed if parliament fails to adopt the budget before the start of the fiscal year.0.334.00
  - c. Parliament can be dissolved if it fails to adopt the budget in due time.0.334.00

- 8. Authority of the national president in the budget procedure0.25
  - a. No special authority.4.00
  - b. President has veto right (president elected by parliament).2.67
  - c. President has veto right (president directly elected by citizens).1.33
  - d. President has veto right (qualified majority required to override veto).0.00

### C. Implementation
- 9. Flexibility to change budget aggregates during execution.0.25
  - a. Any increase in total revenues, expenditures and the deficit needs to be approved by parliament in a supplementary budget.4.00
  - b. Revenue windfalls can be used to increase expenditure without the approval of parliament as long as the deficit is not increased.2.67
  - c. Simultaneous changes in revenue and expenditures allowed without approval of parliament if budget balance is not changed.1.33
  - d. At total or large discretion of government.0.00

- 10. Transfers of expenditures between chapters (i.e. ministries' budgets)0.25
  - a. Require approval of parliament.4.00
  - b. FM or cabinet can authorize transfers between chapters.2.67
  - c. Limited.1.33
  - d. Unrestricted.

- 11. Carryover of unused funds to next fiscal year0.25
  - a. Not permitted.4.00
  - b. Only if provided for in initial budget or with finance ministry approval.2.67
  - c. Limited.1.33
  - d. Unlimited.0.00

- 12. Procedure to react to a deterioration of the budget deficit (due to unforeseen revenue shortfalls or expenditure increase)0.25
  - a. MF can block expenditures .4.00
  - b. The cabinet can block expenditures.2.67
  - c. Approval of the parliament necessary to block expenditures.1.33
  - d. No action is taken.0.00

*Sources: IMF staff; and Gleich (2003).*

### Table A1 note
- Table A1 heading: Construction of the Index: Fiscal Institutions and Their Index Parameters
- Weighting factors are indicated in the source.

### Table A2 note (selected excerpts as presented)
- Table A2 heading: Construction of the Index: Scores
- Sections: A. Preparation Stage; B. Authorization Stage; C. Implementation Stage; Overall Index
- Country score excerpts (as presented in source; values preserved exactly where shown):
  - Bulgaria: 0 0 4 0 1.00 1.75 0 0 1.33 1.33 0.67 0.67 1.33 2.67 4 4 3.00 3.00 1.54 1.79 3 1/
  - Czech Republic: 0 0 4 0 1.00 1.75 0442.67 2.67 2.34 1.33 2.67 4 2.67 2.67 2.67 2.09 2.23 3 1/ 2.67 2/ 4 3/ 1.33 2/ (as presented)
  - Estonia: 2 3 4 0 2.25 2.25 4 0 1.33 2.67 2.00 2.34 442.674 3.67 3.67 2.61 2.72 2.67 4/ (as presented)
  - Hungary: 0 2 4 0 1.50 1.50 0 0 1.33 4 1.33 1.33 2.67 1.33 1.33 2.67 2.00 1.33 1.59 1.37 0 5/ (as presented)
  - Latvia: 2 2 2 4 2.50 2.50 0 0 0 2.67 0.67 0.67 4 4 2.67 4 3.67 3.67 2.26 2.26 (as presented)
  - Lithuania: 0 0 4 0 1.00 1.75 4 0 1.33 1.33 1.67 1.67 4 2.67 4 4 3.67 3.00 2.09 2.12 1 6/ 2 6/ 1.33 2/ (as presented)
  - Poland: 0042 1.50 2.25 0040 1.00 3.00 4 2.67 2.67 2.67 3.00 3.00 1.82 2.72 2 1/ 1 6/ 4 1/ 4 1/ (as presented)
  - Romania: 0 1 4 0 1.25 1.50 00 0 1.33 0.33 1.33 4 2.67 4 2.67 3.34 3.34 1.62 2.04 2 4/ 4 4/ (as presented)
  - Slovak Republic: 0 1 2 0 0.75 0.75 0 0 2.67 2.67 1.34 1.34 2.67 2.67 1.33 2.67 2.34 2.34 1.46 1.46 (as presented)
  - Slovenia: 0344 2.75 2.75 4 0 1.33 4 2.33 2.33 2.67 2.67 2.67 2.67 2.67 2.67 2.56 2.56 (as presented)

*Sources: Gleich (2003); Ylaoutinen (2004); fiscal ROSC reports; and information provided directly by the authorities.*

### Appendix B — Variables and Data Sources (selected)
- Dependent Variables
  - General government primary balance (fiscal notifications to the European Commission).
  - Total revenue and grants, total expenditure and net lending minus interest payments.
  - Where fiscal notifications unavailable, IMF World Economic Outlook (WEO) database used.

- Time-Varying Economic Conditioning Variables
  - Public debt as percent of GDP (lagged); unemployment rate; openness index (imports plus exports normalized by GDP); output gap (Hodrick-Prescott filter to GDP).
  - Data source: IMF World Economic Outlook.

- Dummies
  - Dummy for an IMF program: 1 if the country had a program with the Fund during that year, 0 otherwise.
  - Dummy for the preparation to EU accession: value of 1 from the year the country was invited to start the negotiations on aquis communitaire chapters.

- France/Germany primary balance
  - Average of the lagged primary balances in France and Germany.
  - Data source: EUROSTAT database.

- Time-Varying Political Variables
  - Government fragmentation: Herfindhal index (sum of squares of the shares of each party in the government coalition); ranges from 0 to 1.
    - Data sources: Database on Political Institutions, 2000 (updated in 2004); Parties and Elections in Europe; Economist Intelligence Unit reports.
  - Ideological orientation of governing coalitions: three dimensions — (a) left/right (larger = greater leaning to the right); (b) nationalism; (c) centralization.
    - Government coalition position estimated as sum of each party’s position weighted by party’s seats.
    - Data source: Database on Party Policy in Modern Democracies by Benoit and Laver (2005).

- Election dummy
  - Value of 1 for the year of elections, 0 otherwise.

- Fiscal institutions indices
  - Described in Appendix I.
  - Sources: Gleich (2003), Yläoutinen (2004), IMF fiscal ROSC reports, and direct information from the authorities.

- Other variables
  - Ethnic fractionalization: one minus the sum of squares of the shares of identified ethnic groups. Source: Alesina and others (2003).
  - District magnitude: number of representatives elected by a single district. Data source: Database on Political Institutions, 2000 (updated in 2004).
  - Voter turnout: percent of voting age population. Data source: Institute for Democracy and Electoral Assistance.

*End of content unit.*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2006/_wp06123.pdf_