## Do fiscal rules cause fiscal discipline over the electoral cycle? Evidence from developing countries

## Source details

**Canonical URL:** [Do fiscal rules cause fiscal discipline over the electoral cycle? Evidence from developing countries](https://www.imf.org/-/media/files/publications/wp/2019/wpiea2019291-print-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2019/wpiea2019291-print-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2019/wpiea2019291-print-pdf.pdf.json)

---

### Abstract and key finding
- Sample: 67 developing countries, 1985-2007.
- Main quantitative result: in election years with fiscal rules in place, public consumption is reduced by 1.65% point of GDP as compared to election years without these rules.
- The reduction is described as equivalent to a reduction by a third of the volatility of public consumption in the sample.
- Effectiveness of fiscal rules depends on:
  - type of rule;
  - institutional design (monitoring outside government, coverage);
  - duration since adoption;
  - degree of competitiveness of elections (stronger in democratic/competitive elections).

### Empirical strategy and identification
- Dependent variable: general government final consumption expenditure as a share of GDP.
- Baseline panel specification:
  - Y_it = β1 elec_it + β2 elec_it × FR_it + β3 FR_it + β4 X_it + α_i + trend_i + ε_it.
  - Hypotheses: β1 > 0 (political budget cycle increases public consumption around elections); β2 < 0 (fiscal rules dampen the political budget cycle).
- Identification threats and mitigation:
  - Omitted variable bias addressed with country fixed effects and country-specific linear trends.
  - Reverse causality addressed with an IV approach exploiting geographic diffusion.

### Instrumental variable approach
- Instrument: number of neighboring countries in the subregion with fiscal rules in place (Neighbors_it).
- First step: Probit predicting national fiscal rule adoption:
  - Pr(FR=1 | X_it, Neighbors_it) = Φ(δ0 + δ1 Neighbors_it + δ2 X_it) (δ1 > 0 under diffusion).
- Two-step IV (Wooldridge, 2010) for endogenous binary regressor:
  - Use predicted probability (ˆFR) from the Probit as instrument.
  - First-stage equations instrumenting FR and elec × FR:
    - FR = θ1 ˆFR_it + θ2 elec_it + θ3 elec_it × ˆFR_it + θ4 X_it + α_i + trend_i + η_it.
    - elec_it × FR_it = γ1 elec_it × ˆFR_it + γ2 elec_it + γ3 ˆFR_it + γ4 X_it + α_i + trend_i + ξ_it.
- Controls to mitigate regional shock concerns: economic growth, trade openness, public debt; specifications with sub-regional and global economic growth estimated.
- Estimation: within IV estimator with clustered standard errors at country level.

### Data and sample details
- Cross-country panel: 67 developing countries, 1985-2007.
- Sample selection: data availability; restricted to fiscal rules adopted before the financial crisis of 2008-2009.
- Countries with national fiscal rules over the period: 13 total, comprised of:
  - 11 Balanced Budget Rules (BBR)
  - 4 Expenditure Rules (ER)
  - 8 Debt Rules (DR)
  - 1 Revenue Rule (RR)
  - Note: majority have combinations of rules (totals do not sum to 13).
- Main variables and sources:
  - Dependent: general government final consumption expenditure (% of GDP) (WDI).
  - Fiscal rules: IMF Fiscal Rules Dataset, 2016 (Schaechter et al., 2012), year-by-year dummies; definition per Kopits and Symansky (1998).
  - Election timing: NELDA (Hyde and Marinov, 2011); election index per Franzese (2000).
  - Fiscal rule strength index (robustness): Schaechter et al. (2012) stringency index.
- Controls: real economic growth, trade openness, net public aid per capita, GDP per capita, lagged public debt (% of GDP), sub-regional economic growth, global economic growth; political/institutional controls including government fragmentation and party orientation.

### Baseline results and magnitudes
- OLS vs IV comparison for interaction effect (election years with fiscal rules vs without):
  - OLS: public consumption reduced by 0.45% point of GDP.
  - IV: public consumption reduced by 1.65% point of GDP.
- Democratic elections subsample (IV): public consumption reduced by 2.46% point of GDP in democratic election years with fiscal rules.
- Interpretation:
  - IV estimates larger than OLS suggest OLS may understate the disciplining effect once endogeneity is addressed.
  - Coefficient on the interaction term larger in absolute value than the election-term coefficient alone implies rules neutralize political budget cycles rather than merely dampen them.

### Heterogeneity: rule design, type, and experience
- Institutional design features with stronger disciplining effects:
  - Monitoring of compliance outside the government: IV estimate shows public consumption reduced by 1.88% point of GDP in election years with such rules.
  - Coverage at the general government level: IV estimate shows public consumption reduced by 2.07% point of GDP in election years with such rules.
- By rule type (OLS marginal effects in election years):
  - Balanced Budget Rules (BBR): public consumption reduced by 0.97% point of GDP.
  - Debt Rules (DR): public consumption reduced by 0.42% point of GDP.
  - Expenditure Rules (ER): public consumption reduced by 0.39% point of GDP.
- IV estimates by rule type: larger magnitudes; only Expenditure Rules display a statistically significant IV effect in reported specification:
  - IV (column (6)): public consumption reduced by 2.44% points of GDP during election years with expenditure rules.
  - Caution: small sample variability requires careful interpretation of type-specific IV results.
- Experience/entrenchment:
  - Rules become more effective over time.
  - Fiscal rules not statistically significant for the first election after adoption.
  - IV (column (4)): fiscal rules reduce public expenditure by 1.68% point of GDP in the second election after adoption.

### Robustness and validity checks
- Additional instrument: one-year lagged fiscal rule strength index (Schaechter et al., 2012) used to exploit persistence; combined with baseline instruments.
- Identification tests:
  - Angrist and Pischke (2009) test p-values allow rejection of weak identification at 5% level for baseline instruments.
  - With lagged index instrument: cannot reject validity of overidentifying restrictions at conventional levels; reject null of weak identification at conventional levels.
- Additional regressions include sub-regional and global economic growth; results remain robust (details in Appendix E).
- Conclusion from robustness: causal disciplining effect of fiscal rules on political budget cycles in public consumption is supported.

### Findings — determinants of fiscal rule adoption (selected)
- Regional diffusion: number of neighboring countries with fiscal rules positively and significantly affects likelihood of adoption.
- Macroeconomic correlates:
  - Economic growth positively correlated with likelihood of adopting Balanced Budget Rules and Expenditure Rules.
  - GDP per capita positively correlated with likelihood of adopting an Expenditure Rule (column-specific result).
  - High inflation negatively correlated with presence of fiscal rules.
  - Foreign aid negatively correlated with probability of having fiscal rules.
- Political correlates:
  - Right wing governments have lower probability of adopting Balanced Budget Rules and Debt Rules.

### Institutional features, trade-offs, and limitations
- Design matters: escape clauses should be well defined to avoid frequent amendment and loss of credibility.
- In decentralized contexts, rules targeting the general government may be more effective.
- Risk: overly rigid rules may constrain countercyclical fiscal policy and limit smoothing of business cycles.
- Note: the discipline effect found using the fiscal rules dummy is not robust to use of the strength index.

### Policy implications
- Well-designed fiscal rules can reduce macroeconomic volatility by curbing politically motivated public expenditures in developing countries.
- Features associated with discipline over the electoral cycle:
  - Expenditure targets.
  - Coverage of all levels of government (general government).
  - Independent external monitoring of compliance.
- Policymaker recommendations:
  - Define escape clauses clearly to preserve long-run credibility.
  - Consider applying fiscal rules at the general government level in decentralized settings.
  - Ensure sufficient flexibility in rule design to allow countercyclical fiscal responses when needed.

*Source: IMF Working Paper — "Do fiscal rules cause fiscal discipline over the electoral cycle? Evidence from developing countries," Kodjovi M. Eklou and Marcelin Joanis, wpiea2019291-print-pdf.*

### 1.6 percentage point of GDP as compared to election years without these rules. This impact is

### Do fiscal rules cause fiscal discipline over the electoral cycle? Evidence from developing countries

### Abstract and key finding
- This paper estimates the causal effect of fiscal rules on political budget cycles in a sample of 67 developing countries over the period 1985-2007.
- Main quantitative result: in election years with fiscal rules in place, public consumption is reduced by 1.65% point of GDP as compared to election years without these rules.
- The reduction is described as equivalent to a reduction by a third of the volatility of public consumption in the sample.
- The effectiveness of fiscal rules depends on:
  - their type;
  - their institutional design;
  - whether they have been in place for a long time;
  - the degree of competitiveness of elections.

### Empirical strategy and identification
- Dependent variable: general government final consumption expenditure as a share of GDP (a proxy for discretionary fiscal policy).
- Baseline panel specification:
  - Y_it = β1 elec_it + β2 elec_it × FR_it + β3 FR_it + β4 X_it + α_i + trend_i + ε_it,
  - where elec_it is the election index (Franzese, 2000), FR_it is a fiscal rule dummy, X_it are controls, α_i are country fixed effects, and trend_i are country-specific linear trends.
- Main hypotheses:
  - β1 > 0 (presence of a political budget cycle increases public consumption around elections);
  - β2 < 0 (fiscal rules dampen the political budget cycle).
- Two main identification threats addressed:
  1. Omitted variable bias (e.g., voter preferences for fiscal discipline): addressed by including country fixed effects and country-specific linear trends.
  2. Reverse causality (budgetary outcomes driving adoption of rules): addressed with an IV approach exploiting geographic diffusion.

### Instrumental variable approach
- Instrument: number of neighboring countries in the subregion with fiscal rules in place (Neighbors_it).
- First step: Probit model to predict the probability of national fiscal rule adoption:
  - Pr(FR=1 | X_it, Neighbors_it) = Φ(δ0 + δ1 Neighbors_it + δ2 X_it),
  - relying on diffusion argument (δ1 > 0).
- Two-step IV following Wooldridge (2010) for an endogenous binary regressor:
  - Use predicted probability (ˆFR) from the Probit as an instrument.
  - Two first-stage equations to instrument FR and the interaction elec × FR:
    - FR = θ1 ˆFR_it + θ2 elec_it + θ3 elec_it × ˆFR_it + θ4 X_it + α_i + trend_i + η_it.
    - elec_it × FR_it = γ1 elec_it × ˆFR_it + γ2 elec_it + γ3 ˆFR_it + γ4 X_it + α_i + trend_i + ξ_it.
- Controls included to mitigate regional shock concerns: economic growth, trade openness, public debt; specifications with sub-regional and global economic growth also estimated.
- Estimation: within IV estimator with clustered standard errors at country level.

### Results and heterogeneity
- Baseline causal finding: fiscal rules cause fiscal discipline over the electoral cycle, quantified as a reduction in public consumption of 1.65% point of GDP in election years with rules versus without rules.
- Heterogeneous effectiveness:
  - Expenditure rules are effective in dampening political budget cycles in government consumption.
  - Rules covering the general government (versus narrower scopes) are more effective.
  - Rules characterized by a monitoring body outside the government dampen political budget cycles.
  - Rules that have been in place for a long time show stronger effects.
  - The discipline effect is particularly strong during competitive elections (democracies).

### Relation to literature
- Extends theoretical arguments (Rogoff and Sibert, 1998; Rose, 2006) that binding rules eliminate political budget cycles by preventing deficit financing and making fiscal signaling less effective.
- Complements empirical work on developed countries (Rose, 2006; Krogstrup and Wälti, 2008; Grembi et al., 2016) by focusing on developing countries and addressing endogeneity via geographic diffusion instruments.
- Improves on prior developing-country studies by using a larger sample of countries with fiscal rules and an IV strategy that addresses reverse causality and omitted variables.

### Data and sample
- Sample: 67 developing countries, 1985-2007.
- Dependent variable: general government final consumption expenditure (% of GDP).
- Election timing captured by Franzese (2000) election index.
- Fiscal rule indicator: dummy for presence of a national fiscal rule in a given year.
- Regional focus for instrument construction: Latin America, Sub-Saharan Africa, and South Asia.

### Robustness and validity checks
- Arguments for exclusion restriction: popularity of rules in subregion affects country-level public consumption only through adoption of national fiscal rules; controls for subregional shocks included.
- Additional regressions include sub-regional economic growth and global economic growth to control for regional and global common shocks; results remain robust (details referenced to Appendix E).

*Source: IMF Working Paper — "Do fiscal rules cause fiscal discipline over the electoral cycle? Evidence from developing countries," Kodjovi M. Eklou and Marcelin Joanis, September, 2019.*

### 1.2    Data

### 1.2    Data

### Data and sample
- Cross-country panel dataset covering 67 developing countries over the period 1985-2007.
- Sample selection based on data availability and restricted to fiscal rules adopted before the financial crisis of 2008-2009.
- Sample includes 13 countries with national fiscal rules over the period, among which:
  - 11 have Balanced Budget Rules (BBR)
  - 4 have Expenditure Rules (ER)
  - 8 have Debt Rules (DR)
  - 1 has a Revenue Rule (RR)
- Majority of countries have combinations of rules (hence totals do not sum to 13).

### Main variables
- Dependent variable: general government final consumption expenditure as a share of GDP (World Development Indicators, WDI).
- Fiscal rules: IMF Fiscal Rules Dataset, 2016 (Schaechter et al., 2012), coded as year-by-year dummy variables; fiscal rule defined per Kopits and Symansky (1998).
- Elections: National Elections across Democracy and Autocracy (NELDA) database (Hyde and Marinov, 2011); election timing index computed per Franzese (2000): elec = (m−1) + d/D12, preelectoral index = 1 − elec.
- Fiscal rule strength index (used in robustness): Schaechter et al. (2012) stringency index (legal basis, coverage, enforcement, expenditure ceilings, fiscal responsibility laws, independent monitoring, etc.).

### Controls
- Macroeconomic: real economic growth (WDI), trade openness, net public aid per capita, GDP per capita, lagged public debt (share of GDP), sub-regional economic growth, global economic growth.
- Political/institutional: government fragmentation (probability that two randomly picked legislators are from different parties), party orientation (Right wingit dummy = 1 if conservative/Christian democratic/right-wing), partisan and common-pool considerations noted.

### Institutional detail notes
- Monitoring of compliance outside the government is coded when an independent council monitors rule compliance.
- Coverage distinguishes rules applying to general government versus central government only.
- Election classifications: all elections, “democratic” elections (positive Polity2), and constitutionally predetermined election dates.

### Empirical strategy
- Two-step IV approach:
  - First step: Probit estimation of determinants of fiscal rule adoption (equation (2)); spatial diffusion measured by number of neighboring countries with fiscal rules.
  - Second step: First-stage equations (3) and (4) for endogenous regressors (fiscal rule dummy and interaction term); instrument interaction between election and fiscal rules by interaction between election and predicted probability ĤFR.
  - Second-stage: equation (1) estimated by OLS and IV to measure effect of fiscal rules on political budget cycles in public consumption.
- Additional IV robustness: use one-year lagged fiscal rule strength index as instrument (persistence in legislative process), combined with baseline instruments to test overidentification and weak identification.

### Appendix and supplementary material
- Appendix B: descriptive statistics of variables used.
- Appendix C: detailed information about rules.
- Appendix D: evidence that political budget cycles in public consumption are present in the sample.

### Identification tests
- Angrist and Pischke (2009) test p-values allow rejection of weak identification at 5% level for baseline instruments.
- Robustness specifications with the lagged fiscal rule index: cannot reject null of validity of overidentifying restrictions at conventional levels; reject null of weak identification at conventional levels.

---

### Findings — Determinants of fiscal rule adoption
- Regional diffusion: number of neighboring countries with fiscal rules has a positive and statistically significant effect on the likelihood of adoption.
- Macroeconomic correlates:
  - Economic growth positively correlated with likelihood of adopting Balanced Budget Rules and Expenditure Rules.
  - GDP per capita positively correlated with likelihood of adopting an Expenditure Rule (column-specific result).
  - High inflation negatively correlated with presence of fiscal rules.
  - Foreign aid negatively correlated with probability of having fiscal rules.
- Political correlates:
  - Right wing governments have lower probability of adopting Balanced Budget Rules and Debt Rules.

### Baseline results — fiscal rules and political budget cycles
- General result: fiscal rules have a negative and statistically significant effect on public consumption during election years.
- Comparison OLS vs IV:
  - OLS (column (1)): in election years with fiscal rules, public consumption is reduced by 0.45% point of GDP relative to election years without such rules.
  - IV (column (2)): in election years with fiscal rules, public consumption is reduced by 1.65% point of GDP relative to election years without such rules.
  - Magnitude interpretation: IV impact is equivalent to a reduction by a third of the volatility of public consumption in the sample.
- Democratic elections subsample:
  - IV (column (4)): public consumption is reduced by 2.46% point of GDP in democratic election years with fiscal rules (larger effect than full-sample IV of 1.65% point).
- Interpretation:
  - IV estimates larger than OLS suggests OLS may overestimate effects due to omitted variable bias related to better macroeconomic conditions increasing likelihood of rule adoption.
  - Coefficient on interaction term larger in absolute terms than coefficient on election years alone suggests fiscal rules neutralize political budget cycles (discipline effect) rather than merely dampening them.

### Design of fiscal rules, experience with rules, and the cycle
- Institutional design features matter:
  - Monitoring of compliance outside the government and coverage at the general government level strengthen the disciplining effect.
  - IV results (columns (2) and (4)) indicate public consumption is reduced by:
    - 1.88% point of GDP in election years with fiscal rules that feature monitoring of compliance outside the government.
    - 2.07% point of GDP in election years with fiscal rules covering the general government.
- Type of rules (Table 5):
  - OLS marginal effects in election years:
    - Balanced Budget Rules (BBR): public consumption reduced by 0.97% point of GDP.
    - Debt Rules (DR): public consumption reduced by 0.42% point of GDP.
    - Expenditure Rules (ER): public consumption reduced by 0.39% point of GDP.
  - IV estimates larger in magnitude; only expenditure rules display a statistically significant IV effect in the reported specification:
    - IV (column (6)): public consumption reduced by 2.44% points of GDP during election years with expenditure rules in place.
  - Caution: small sample variability implies interpreting IV results on types of rules carefully.
- Experience and entrenchment (Table 6):
  - Effect stronger for the second election after rule adoption than for the first.
  - Fiscal rules not statistically significant for the first election after adoption.
  - IV (column (4)): fiscal rules reduce public expenditure by 1.68% point of GDP in the second election.
  - Interpretation: rules become more credible and effective as they become more entrenched.

### Robustness
- Additional instrument: one-year lagged fiscal rule strength index (Schaechter et al., 2012) used to exploit time variation and persistence in legislative process.
- Results with augmented instrument set are similar to baseline: fiscal rules dampen the political budget cycle.
- Overidentification and weak-identification tests:
  - Cannot reject validity of overidentifying restrictions at conventional levels.
  - Reject null of weak identification at all conventional levels.
- Conclusion: robustness checks confirm a causal disciplining effect of fiscal rules on political budget cycles in public consumption.

*Source: wpiea2019291-print-pdf - 1.2    Data*

### Conclusion

### Conclusion

### Main findings
- The paper exploits the geographical pattern in the adoption of fiscal rules to empirically investigate whether the constraint that fiscal rules may impose on discretionary fiscal policy is binding during election years.
- The empirical analysis provides a novel, robust identification strategy to estimate the causal effect of fiscal rules using a sample of developing countries.
- Fiscal rules matter for fiscal discipline over the electoral cycle.
- Specific features of fiscal rules that are particularly relevant to achieve fiscal discipline:
  - Expenditure rules.
  - Rules targeting all levels of government (general government).
  - Rules that are monitored by an independent body outside the government.
- Fiscal rules are more effective in constraining political budget cycles the longer they are in place.
- Note: the discipline effect found using the fiscal rules dummy is not robust to the use of the index.

### Institutional features and rule design
- The design of fiscal rules is crucial to their effectiveness.
- Escape clauses should be well defined to avoid changing or amending fiscal rules too often and thus undermining their credibility in the long run.
- In increasingly decentralized developing countries, fiscal rules might be more effective if they target the general government.
- Caution: if fiscal rules are not well-designed to be flexible enough to accommodate cyclical fluctuations, they may constrain the ability of fiscal policy to smooth business cycles.

### Policy implications
- Well-designed fiscal rules can reduce macroeconomic volatility by curbing politically motivated public expenditures in developing countries.
- Rules that combine expenditure targets, coverage of all levels of government, and independent external monitoring are associated with discipline over the electoral cycle.
- Policymakers should:
  - Define escape clauses clearly to preserve long-run credibility.
  - Consider applying rules at the general government level in decentralized settings.
  - Ensure sufficient flexibility in rule design to allow countercyclical fiscal responses when needed.

*Source: wpiea2019291-print-pdf - Conclusion; https://www.imf.org/-/media/files/publications/wp/2019/wpiea2019291-print-pdf.pdf*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2019/wpiea2019291-print-pdf.pdf_