## _wp06213

## Source details

**Canonical URL:** [_wp06213](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2006/_wp06213.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2006/_wp06213.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2006/_wp06213.pdf.json)

---

### Introduction
- IMF conditionality toolkit: prior actions, performance criteria, structural benchmarks.
- Prior actions: measures the member must implement up-front, prior to Board approval.
- Dual purposes:
  - Ensure up-front implementation of measures key to program objectives.
  - Signal authorities’ ownership and commitment where track record is weak.
- Prior actions are a screening device but implementation is not necessarily an unambiguous sign of commitment.

### Stylized facts (programs commencing 1992–2004; conditionality counted through end-2005)
- Data normalization and sample handling:
  - Data normalized by arrangement duration; permanently interrupted and ongoing programs assume end-date = last completed review (or approval date if no review) + nine months.
- Prior actions and total conditions (averages and evolution):
  - Average number of prior actions:
    - Rose from about 3½ (1992–97) to about 6 (1998–2000), then declined following streamlining initiative to about 3½–6 depending on program type.
  - GRA-supported (stand-by and EFF) prior actions:
    - About 3½ (1992–97) → about 6 (1998–2000) → about 6 per annum over 2001–04 for non-precautionary GRA-supported arrangements.
  - PRGF-supported programs:
    - Rose to four conditions during 2001–04, about one higher than average during 1992–97.
  - Total conditions:
    - GRA-supported programs spiked in 1998–2000 then declined but remained above pre-Asian crisis levels.
    - PRGF-supported programs had a substantial decline in 1998–2000 relative to before, with partial recovery thereafter.
- Implementation index and stoppage frequency:
  - Implementation index (excludes prior actions; varies between 0 and 2):
    - Improved from 1 in 1992–97 to 1.2 (1998–2000) and 1.5 (2001–04).
  - Frequency of stoppages declined for GRA-supported programs.
- Selected table excerpts (values normalized by duration):
  - GRA-supported programs:
    - 1992-97: No. of observations 43.5; No. of prior actions 9.9; Total no. of conditions 1.0; Implementation index 0.4
    - 1998-2000: No. of observations 305.9; No. of prior actions 19.1; Total no. of conditions 1.2; Implementation index 0.4
    - 2001-04: No. of observations 235.0; No. of prior actions 16.7; Total no. of conditions 1.5; Implementation index 0.2
  - Non-precautionary:
    - 1992-97: No. of observations 783.8; No. of prior actions 10.1; Total no. of conditions 0.9; Implementation index 0.5
    - 1998-2000: No. of observations 179.5; No. of prior actions 26.7; Total no. of conditions 1.3; Implementation index 0.4
    - 2001-04: No. of observations 116.3; No. of prior actions 21.7; Total no. of conditions 1.6; Implementation index 0.1
  - PRGF-supported programs:
    - 1992-97: No. of observations 492.8; No. of prior actions 20.7; Total no. of conditions 1.4; Implementation index 0.5
    - 1998-2000: No. of observations 264.3; No. of prior actions 14.7; Total no. of conditions 1.3; Implementation index 0.3
    - 2001-04: No. of observations 353.9; No. of prior actions 15.1; Total no. of conditions 1.4; Implementation index 0.3
- Regional variation (selected):
  - Transition (Former Soviet Union, excl. Baltics): No. of observations 365.8; No. of prior actions 18.5; Total no. of conditions 1.3; Implementation index 0.5
  - Baltics: No. of observations 182.7; No. of prior actions 7.6; Total no. of conditions 1.5; Implementation index <0.05
  - Central and Eastern Europe: No. of observations 337.5; No. of prior actions 19.7; Total no. of conditions 1.5; Implementation index 0.3
  - Africa: No. of observations 862.2; No. of prior actions 13.6; Total no. of conditions 1.4; Implementation index 0.3
  - Asia/Pacific: No. of observations 224.9; No. of prior actions 17.6; Total no. of conditions 1.4; Implementation index 0.5
  - Middle East: No. of observations 166.1; No. of prior actions 16.4; Total no. of conditions 1.4; Implementation index 0.4
  - Western Hemisphere: No. of observations 522.6; No. of prior actions 13.2; Total no. of conditions 1.3; Implementation index 0.4
- Track record association:
  - Prior actions used more in countries with relatively weak track records (implementation index < 1.42 median for 1995–2003).
  - Since 2001, average prior actions per program year:
    - Relatively strong track records: about 3
    - Relatively weak track records: about 5½
  - Country examples:
    - Bulgaria (1997): 13 prior actions implemented following weak performance under previous programs.
    - Romania (2001): about 24 prior actions per program year; five previous programs had gone off track; 2004 program continued to use many prior actions.

### Determinants of prior actions (empirical analysis 1992–2004)
- Dependent variable: ratio of prior actions to total number of structural conditions set at approval (period t).
- Key explanatory variables: lagged implementation index, initial macroeconomic conditions (budget balance % of GDP, current account balance % of GDP, reserves/imports ratio, inflation rate, external debt/GDP), institutional factors (political freedom/democracy, duration of government, executive variation, government stability), share of prior actions in previous program, regional dummies.
- Main regression highlights (Table 3):
  - Lagged implementation index in GRA programs: coefficient -0.08 * (significant at 10 percent) — weaker previous implementation ⇒ higher share of prior actions.
  - Lagged implementation index in ESAF/PRGF programs: coefficient -0.05 (not significant).
  - Lagged prior actions: coefficient 0.33 *** (p < 0.01) — persistence/hysteresis in use of prior actions.
  - Initial budget position: coefficient -2.44 ** (significant at 5 percent) — stronger initial budget ⇒ fewer prior actions.
  - Initial external current account position: coefficient 1.34 ** (significant at 5 percent) — stronger current account associated with higher share of prior actions (effect disappears when government balance is dropped).
  - Transition economies dummy: coefficient 0.21 ** (significant at 5 percent) — higher share of prior actions in European transition countries.
  - No. observations: 73; R-squared: 0.48.
  - Other institutional variables not statistically significant in this sample.

### Prior actions and program success

A. Prior actions and program implementation
- Research question: Does a higher share of prior actions at approval improve implementation of subsequent structural benchmarks and performance criteria?
- Method: Change in implementation rate of performance criteria and structural benchmarks between successive programs regressed on predicted share of prior actions at approval (from Table 3). Sample: programs 1992–2004.
- Key results (Table 4):
  - Dependent variable: Change in implementation of non-prior action conditionality.
  - Coefficients (predicted prior actions ratio interacted with program type):
    - Ratio of prior actions in GRA-supported programs in the CEE/CIS countries (excl. Baltics): 1.00 * (significant at 10 percent)
    - Ratio of prior actions, GRA-supported programs in non-CEE/CIS countries: 2.12 * (significant at 10 percent)
    - Ratio of prior actions, PRGF-supported programs: 1.73 ** (significant at 5 percent)
  - Interpretation:
    - Increase in share of prior actions by 10 percentage points:
      - Improves implementation between successive programs by 0.10 in CEE/CIS GRA programs (≈10 percent relative to sample mean ≈0.98).
      - Improves implementation by about 21 percent in non-transition GRA programs.
      - Improves implementation by about 17 percent in PRGF programs.
  - Sample: 71 observations; R-squared: 0.08.
- Caveats:
  - Implementation of prior actions is not an absolute sign of commitment; nominal implementation without real commitment may not achieve program goals.
  - Adding too many prior actions could dilute the positive effect.

B. Prior actions and macroeconomic policy
- Fiscal impact (middle-income countries):
  - Dependent variable: change in overall government balance normalized by lagged output.
  - Significant cyclicals: U.S. growth (positive), lagged fiscal balance (negative, mean reversion), nominal interest rate (significant).
  - Conditionality effects:
    - Number of structural performance criteria and benchmarks, and the difference between prior actions set at outset versus other times are positive and statistically significant.
    - Estimated effect: countries with prior actions set at outset achieve budgetary improvement of about 0.2 percent of GDP more than countries without initial prior actions, holding other factors constant.
- Inflation regression:
  - Significant determinants: lagged inflation, change in terms of trade, change in real exchange rate, lagged fiscal balance, conditionality variables.
  - Implementation index of previous program significant: full implementation associated with a 4 percentage point decline in inflation.
  - Number of structural performance criteria/benchmarks and difference between prior actions at outset vs rest not significant for inflation (signs negative).
- Conclusion: Prior actions at approval act as strong screening devices and assist achievement of macroeconomic targets for middle-income countries.

C. Prior actions and macroeconomic performance (growth)
- Hypothesized channels for prior actions to raise medium-term growth:
  1. Implementation rate of prior actions is 100 percent by definition.
  2. Screening effect: separates committed from less-committed reformers.
  3. Up-front measures may be more substantive.
- Growth regression findings:
  - Program-period results (Table 5):
    - Log normalized number of prior actions and normalized number of PCs and SBs (interacted with implementation index): not significant determinants of growth during program period.
    - Implementation index of previous program: positive and significant — full implementation would boost growth by about ¾ of a percentage point per year (0.004*2).
    - Cyclical variables significant: higher inflation and higher interest rates dampen growth; overvalued real exchange rate lowers growth; U.S. growth positive.
  - Screening versus quantity:
    - Ratio of prior actions at outset vs rest of program (time-profile) is significantly positive for growth; total number of prior actions is not significant when both included.
    - Interpretation: demonstrable screening effect — up-front prior actions associated with higher growth, independent of total number.
  - Medium-term effects (3 years after program start; Table 6):
    - Prior actions at program approval relative to rest (log) significant positive for both non-transition and transition countries:
      - Non-transition: coefficient 0.015 *** (significant at 1 percent)
      - Transition countries: coefficient 0.010 *** (significant at 1 percent) and 0.009 ** in alternate column (significant at 5 percent)
    - Normalized number of prior actions during program (log) not significant in main specification; becomes significant if time-profile variable excluded.
    - Quantitative illustration: Countries with twice the number of prior actions set at outset relative to other times grow by about ¾ percent of GDP per annum faster than countries without up-front prior actions, holding other factors constant.
    - Other significant long-run determinants: law and order (elasticity ~1.6), U.S. growth positive; investment rate not significant in these specifications.
  - Sample notes:
    - Middle-income countries analysis excluded PRGF-eligible and very small countries; self-selection correction applied via inverse Mills ratio.

### Conclusions and policy implications
- Empirical evidence supports prior actions as a useful signaling/screening device, especially for countries with weak track records and in GRA-supported programs.
- Key findings:
  - Weaker prior implementation (lower implementation index) associated with a higher share of prior actions at approval for GRA programs.
  - Share of prior actions persists across programs (hysteresis).
  - A higher share of prior actions at approval improves implementation rates of other conditionality over successive programs, though diminishing returns and dilution risk exist.
  - Prior actions at approval strengthen macroeconomic targets: fiscal balance improves during program and inflation declines (latter not always significant).
  - Screening effect on growth: the ratio of prior actions at the outset versus rest of program is positively associated with medium-term growth; total number of prior actions is not robustly significant once time-profile accounted for.
- Policy cautions:
  - Prior actions are an effective screening tool but not a panacea; nominal compliance without underlying commitment may not yield successful outcomes.
  - Excessive prior actions could reduce domestic support and dilute positive effects.
- Suggested future work:
  - Decompose prior actions by policy focus to ascertain differential impacts on growth.

*Source: IMF working paper section "1. Prior Actions and Track Record" (extracted from _wp06213).*

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

### _wp06213 - References..........................................................................................................

### References
- References................................................................................................................................22

### Appendix
- Appendix..................................................................................................................................24

### Tables
- 1. Characteristics of Programs by Type of Arrangement ..........................................................7
- 2. Characteristics of Programs by Region, 1992–2004..............................................................8
- 3. Determinants of Prior Action...............................................................................................11
- 4. Prior Actions and Program Implementation References .....................................................14
- 5. Prior Actions and Macroeconomic Policies.........................................................................18
- 6. Prior Actions and Growth References .................................................................................20

### Figure
- Figure

*Source: _wp06213 - References..........................................................................................................*

### 1. Prior Actions and Track Record ......................................................................................

### 1. Prior Actions and Track Record

### Introduction
- The IMF conditionality toolkit comprises prior actions, performance criteria, and structural benchmarks.
- Prior actions: measures the member must implement up-front, prior to Board approval of an IMF-supported program.
- Dual purposes of prior actions:
  - Ensure up-front implementation of measures key to program objectives.
  - Signal authorities’ ownership and commitment where track record is weak.
- Prior actions are a screening device but their implementation is not necessarily an unambiguous sign of commitment.

### Literature
- Signaling analogy to Spence (1973): prior actions represent a costly signal separating “committed” from “uncommitted” governments.
- Drazen (2001): IMF must account for political constraints; failure can result if domestic constraints are not understood during program formulation.
- Empirical parallels:
  - Dollar and Svensson (2000): program failure predicted by a small set of domestic political economy variables.
  - Ivanova and others (2001): program completion depends on ethnic/linguistic divisions, special interests, political cohesion; IMF staff hours not significant.
  - Dreher (2004): program interruptions more likely before elections, with high government consumption, high short-term debt, low output per capita.
  - Dreher and Vaubel (2004): number of conditions positively influenced by size of budget deficit and money growth, negatively by change in reserves.

### Stylized Facts of IMF-Supported Programs (sample: programs commencing 1992–2004; conditionality counted through end-2005)
- Data normalized by arrangement duration; permanently interrupted and ongoing programs assume end-date = last completed review (or approval date if no review) + nine months.
- Average number of prior actions:
  - Rose from about 3½ (1992–97) to about 6 (1998–2000), then declined following streamlining initiative to about 3½–6 depending on program type.
  - Stand-by and EFF (GRA-supported) prior actions: about 3½ (1992–97) → about 6 (1998–2000) → about 6 per annum over 2001–04 for non-precautionary GRA-supported arrangements (decline by three to average about six per annum over 2001–04 quoted).
  - PRGF-supported programs: rose to four conditions during 2001–04, about one higher than average during 1992–97.
- Total conditions in GRA-supported programs spiked in 1998–2000 then declined but remained above pre-Asian crisis levels; PRGF-supported programs had a substantial decline in 1998–2000 relative to before, with partial recovery thereafter.
- Implementation index and stoppage frequency:
  - Implementation index improved from 1 in 1992–97 to 1.2 (1998–2000) and 1.5 (2001–04) (index varies between 0 and 2; excludes prior actions).
  - Frequency of stoppages declined for GRA-supported programs.

- Table 1 excerpt (characteristics by type of arrangement; values normalized by duration):
  - GRA-supported programs:
    - 1992-97: No. of observations 43.5; No. of prior actions 9.9; Total no. of conditions 1.0; Implementation index 0.4
    - 1998-2000: No. of observations 305.9; No. of prior actions 19.1; Total no. of conditions 1.2; Implementation index 0.4
    - 2001-04: No. of observations 235.0; No. of prior actions 16.7; Total no. of conditions 1.5; Implementation index 0.2
  - Non-precautionary:
    - 1992-97: No. of observations 783.8; No. of prior actions 10.1; Total no. of conditions 0.9; Implementation index 0.5
    - 1998-2000: No. of observations 179.5; No. of prior actions 26.7; Total no. of conditions 1.3; Implementation index 0.4
    - 2001-04: No. of observations 116.3; No. of prior actions 21.7; Total no. of conditions 1.6; Implementation index 0.1
  - PRGF-supported programs:
    - 1992-97: No. of observations 492.8; No. of prior actions 20.7; Total no. of conditions 1.4; Implementation index 0.5
    - 1998-2000: No. of observations 264.3; No. of prior actions 14.7; Total no. of conditions 1.3; Implementation index 0.3
    - 2001-04: No. of observations 353.9; No. of prior actions 15.1; Total no. of conditions 1.4; Implementation index 0.3

- Regional variation (Table 2 excerpt; values normalized by duration):
  - Transition (Former Soviet Union, excl. Baltics): No. of observations 365.8; No. of prior actions 18.5; Total no. of conditions 1.3; Implementation index 0.5
  - Baltics: No. of observations 182.7; No. of prior actions 7.6; Total no. of conditions 1.5; Implementation index <0.05
  - Central and Eastern Europe: No. of observations 337.5; No. of prior actions 19.7; Total no. of conditions 1.5; Implementation index 0.3
  - Africa: No. of observations 862.2; No. of prior actions 13.6; Total no. of conditions 1.4; Implementation index 0.3
  - Asia/Pacific: No. of observations 224.9; No. of prior actions 17.6; Total no. of conditions 1.4; Implementation index 0.5
  - Middle East: No. of observations 166.1; No. of prior actions 16.4; Total no. of conditions 1.4; Implementation index 0.4
  - Western Hemisphere: No. of observations 522.6; No. of prior actions 13.2; Total no. of conditions 1.3; Implementation index 0.4

- Prior actions used more in countries with relatively weak track records (implementation index < 1.42 median for 1995–2003):
  - Since 2001, average prior actions per program year:
    - Relatively strong track records: about 3
    - Relatively weak track records: about 5½
  - Examples:
    - Bulgaria (1997): 13 prior actions implemented following weak performance under previous programs.
    - Romania (2001): about 24 prior actions per program year; five previous programs had gone off track; 2004 program continued to use many prior actions.

### Determinants of Prior Actions (empirical analysis 1992–2004)
- Dependent variable: ratio of prior actions to total number of structural conditions set at approval (period t).
- Key explanatory variables: lagged implementation index (previous program), initial macroeconomic conditions in t-1 (budget balance % of GDP, current account balance % of GDP, reserves/imports ratio, inflation rate, external debt/GDP), institutional factors (political freedom/democracy, duration of government, executive variation, government stability), share of prior actions in previous program, regional dummies.
- Main regression results (Table 3 highlights):
  - Lagged implementation index in GRA programs: coefficient -0.08 * (significant at 10 percent) — weaker previous implementation ⇒ higher share of prior actions.
  - Lagged implementation index in ESAF/PRGF programs: coefficient -0.05 (not significant).
  - Lagged prior actions: coefficient 0.33 *** (p < 0.01) — persistence/hysteresis in use of prior actions.
  - Initial budget position: coefficient -2.44 ** (significant at 5 percent) — stronger initial budget ⇒ fewer prior actions.
  - Initial external current account position: coefficient 1.34 ** (significant at 5 percent) — stronger current account associated with higher share of prior actions (note: this effect disappears when government balance is dropped).
  - Transition economies dummy: coefficient 0.21 ** (significant at 5 percent) — higher share of prior actions in European transition countries.
  - No. observations: 73; R-squared: 0.48.
  - Other institutional variables not statistically significant in this sample.

### Prior Actions and Program Success

A. Prior Actions and Program Implementation
- Research question: Does a higher share of prior actions at approval improve implementation of subsequent structural benchmarks and performance criteria?
- Method: Change in implementation rate of performance criteria and structural benchmarks between successive programs regressed on predicted share of prior actions at approval (from Table 3 regression). Sample: programs 1992–2004.
- Key results (Table 4):
  - Dependent variable: Change in implementation of non-prior action conditionality.
  - Coefficients (predicted prior actions ratio interacted with program type):
    - Ratio of prior actions in GRA-supported programs in the CEE/CIS countries (excl. Baltics): 1.00 * (significant at 10 percent)
    - Ratio of prior actions, GRA-supported programs in non-CEE/CIS countries: 2.12 * (significant at 10 percent)
    - Ratio of prior actions, PRGF-supported programs: 1.73 ** (significant at 5 percent)
  - Interpretation:
    - Increase in share of prior actions by 10 percentage points:
      - Improves implementation between successive programs by 0.10 in CEE/CIS GRA programs (~10 percent relative to sample mean ~0.98).
      - Improves implementation by about 21 percent in non-transition GRA programs.
      - Improves implementation by about 17 percent in PRGF programs.
  - Sample: 71 observations; R-squared: 0.08.
- Caveats:
  - Implementation of prior actions is not an absolute sign of commitment; nominal implementation without real commitment may not achieve program goals.
  - Adding too many prior actions could dilute the positive effect.

B. Prior Actions and Macroeconomic Policy
- Fiscal impact regression (middle-income countries sample):
  - Dependent variable: change in overall government balance normalized by lagged output.
  - Significant cyclicals:
    - U.S. growth: positive impact on fiscal position.
    - Lagged fiscal balance: negative (mean reversion).
    - Nominal interest rate: significant (to finance debt service).
  - Conditionality effects:
    - Number of structural performance criteria and benchmarks, and the difference between prior actions set at outset versus other times are positive and statistically significant.
    - Estimated effect: countries with prior actions set at outset achieve budgetary improvement of about 0.2 percent of GDP more than countries without initial prior actions, holding other factors constant.
- Inflation regression:
  - Significant determinants include lagged inflation, change in terms of trade, change in real exchange rate, lagged fiscal balance, conditionality variables.
  - Implementation index of previous program significant: full implementation associated with a 4 percentage point decline in inflation.
  - Number of structural performance criteria/benchmarks and difference between prior actions at outset vs rest not significant for inflation (signs negative).
- Conclusion: Prior actions at approval act as strong screening devices and assist achievement of macroeconomic targets for middle-income countries.

C. Prior Actions and Macroeconomic Performance (Growth)
- Hypotheses why more prior actions may raise medium-term growth:
  1. Implementation rate of prior actions is 100 percent by definition.
  2. Screening effect: separates committed from less-committed reformers.
  3. Up-front measures may be more substantive.
- Growth regressions (instrumenting endogenous condition counts; controls for macro policies, exogenous conditions, long-run factors):
  - Program-duration results (Table 5):
    - Log normalized number of prior actions and normalized number of PCs and SBs (interacted with implementation index): not significant determinants of growth during program period.
    - Implementation index of previous program: positive and significant — full implementation would boost growth by about ¾ of a percentage point per year (0.004*2).
    - Cyclical variables significant: higher inflation and higher interest rates dampen growth; overvalued real exchange rate lowers growth; U.S. growth positive.
    - Corruption: negative effect on inflation; corruption positively associated with fiscal balance over long run.
  - Screening vs quantity test:
    - The ratio of prior actions at outset vs rest of program (time-profile) is significantly positive for growth; total number of prior actions is not significant when both are included.
    - Interpretation: demonstrable screening effect — up-front prior actions associated with higher growth, independent of total number.
  - Medium-term effects (3 years after program start; Table 6):
    - Prior actions at program approval relative to rest (log) significant positive for both non-transition and transition countries:
      - Non-transition: coefficient 0.015 *** (significant at 1 percent)
      - Transition countries: coefficient 0.010 *** (significant at 1 percent) and 0.009 ** in alternate column (significant at 5 percent)
    - Normalized number of prior actions during program (log) not significant in main specification; becomes significant if the time-profile variable excluded.
    - Quantitative illustrative result: Countries with twice the number of prior actions set at outset relative to other times grow by about ¾ percent of GDP per annum faster than countries without up-front prior actions, holding other factors constant.
    - Other significant long-run determinants: law and order (elasticity ~1.6), U.S. growth positive; investment rate not significant in these specifications.
  - Sample notes:
    - Middle-income countries analysis excluded PRGF-eligible and very small countries; self-selection correction applied via inverse Mills ratio.

### Conclusions
- Empirical evidence supports prior actions as a useful signaling/screening device for countries with weak or nonexistent track records of program implementation (especially in GRA-supported programs).
- Key findings:
  - Weak prior implementation (lower implementation index) is associated with a higher share of prior actions at approval for GRA programs.
  - Share of prior actions persists across programs (hysteresis).
  - A higher share of prior actions at approval improves implementation rates of other conditionality over successive programs, though diminishing returns and dilution risk exist.
  - Prior actions at approval strengthen macroeconomic targets: fiscal balance improves during program and inflation declines (latter not always significant).
  - Screening effect on growth: the ratio of prior actions at the outset versus rest of program is positively associated with medium-term growth; the total number of prior actions is not robustly significant once time-profile accounted for.
- Policy implications and caveats:
  - Prior actions are an effective screening tool but not a panacea; nominal compliance without underlying commitment may not yield successful outcomes.
  - Excessive prior actions could reduce domestic support and dilute positive effects.
  - Future work suggested: decompose prior actions by policy focus to ascertain differential impacts on growth.

*Source: IMF working paper section "1. Prior Actions and Track Record" (extracted from _wp06213)._*

### References

### _wp06213 - References

### Major cited works
- Acemoglu, D., J. Robinson, and S. Johnson, 2001, “The Colonial Origins of Comparative Development: An Empirical Investigation,” American Economic Review, Vol. 91, pp. 1369–1401.
- Alcala, F., and A. Ciccone, 2004, “Trade and Productivity,” Quarterly Journal of Economics, Vol. 119, No. 2, pp. 613–46.
- Alesina, A., and A. Drazen, 1991, “Why are Stabilizations Delayed?” American Economic Review, Vol. 81, pp. 1170–88.
- Barro, R. J., 1996, “Determinants of Democracy,” Journal of Political Economy, Vol. 107 (6/2) Supplement.
- Boughton, J., 2001, Silent Revolution: the International Monetary Fund, 1979–1989 (Washington: International Monetary Fund).
- Caselli, F., G. Esquivel, and F. Lefort, 1996, “Reopening the Convergence Debate: A New Look at Cross-Country Growth Empirics,” Journal of Economic Growth, Vol. 1, No. 3, pp. 363–89.
- Conway, P., 1994, “IMF Lending Programs: Participation and Impact,” Journal of Development Economics, Vol. 45, pp. 365–91.
- Cukierman, A., and N. Liviatan, 1992, “Dynamics of Optimal Gradual Stabilization,” World Bank Economic Review, Vol. 6, pp. 439–58.
- Dollar, D., and J. Svensson, 2000, “What Explains The Success or Failure of Structural Adjustment Programs?” Economic Journal, 110, pp. 894–917.
- Dollar, D., and A. Kraay, 2003, “Institutions, Trade, and Growth,” Journal of Monetary Economics, Vol. 50, No. 1, pp. 133–62.
- Drazen, A., 2001, “Conditionality and Ownership in IMF Lending: A Political Economy Approach,” paper presented at the Second Annual IMF Research conference, Washington, November.
- Dreher, A., 2004, The Influence of IMF Programs on the Re-Election of Debtor Governments,” Economics and Politics, Vol. 16, No. 1, pp. 53–75.
- Dreher, A., and R. Vaubel, 2004, “Does the IMF Cause Moral Hazard and Political Business Cycles? Evidence from Panel Data, Open Economics Review, Vol. 15, No. 1, pp. 5–22.
- Easterly, W., and R. Levine, 2001, “It’s Not Factor Accumulation: Stylized Facts and Growth,” World Bank Economic Review, Vol. 15, No. 2.
- Glaesar, E., R. la Porta, F. Lopez-de-Silanes, and A. Shleifer, 2004, “Do Institutions Cause Growth? (unpublished; Cambridge, Massachusetts: Harvard University).
- Haggard, S., and S.B. Webb, 1994, Voting for Reform: Democracy, Political Liberalization, and Economic Adjustment (New York: Oxford University Press).
- Ivanova, A., W. Mayer, A. Mourmouras, and G. Anayiotos (2001), “What Determines the Success Or Failure of IMF-supported Programs?”, Paper presented at the Second Annual Research Conference, Washington, November.
- Joyce, J. J, 1992, “The Economic Characteristics of IMF Program Countries,” Economics Letters, Vol. 38, No. 2, 237–42.
- Kaufman, D., A. Kraay, and P. Zoido-Lobaton, 2002, “Governance Matters II,” World Bank Working Paper No. 2772 (Washington: World Bank).
- Kim, J., 2005, “Catalytic Finance and Fund Program Design: a Model,” (unpublished; Washington: International Monetary Fund).
- Knight, M., and J. Santaella, 1997, “Economic Determinants of IMF Programs,” Journal of Development Economics, Vol. 54, pp. 405–36.
- Laban, R., and F. Sturzenegger, 1994, “Distributional Conflict, Financial Adaptation, and Delayed Stabilization,” Economics and Politics, Vol. 6, pp. 257–76.
- Levine, R., and D. Renelt, 1992, “A Sensitivity Analysis of Cross-Country Growth Regressions,” American Economic Review, Vol. 82, No. 4, pp. 942–63.
- Rodrik, D., A. Subramanian, and F. Trebbi, 2004, “Institutions Rule: the Primacy of Institutions over Geography and Integration in Economic Development,” Journal of Economic Growth, Vol. 9, pp. 131–65.
- Sachs, J., and A. Warner, 1995, “Economic Reform and the Process of Global Integration,” Brookings Papers on Economic Activity, No. 1, pp. 1–118.
- Sala-i-Martin, X., 1997, “I Just Ran 2 Million Regressions,” American Economic Review, Vol. 87, No. 2, pp. 178–83.
- Spence, M., 1973, “Job Market Signaling,” Quarterly Journal of Economics, Vol. 87, pp. 355–79.
- Svensson, J., 1998, “Investment, Property Rights, and Political Instability: Theory and Evidence,” European Economic Review, Vol. 42, pp. 1317–41.
- Tabellini, G., and A. Alesina, 1990, “Voting on the Budget Deficit,” American Economic Review, Vol. 80, pp. 37–49.
- Tommasi, M., and A. Velasco, 1996, “Where Are We in the Political Economy of Reform?” Journal of Policy Reform, No. 1 (April), pp. 173–238.

### Appendix — Variables used in regressions on implementation
- Institutional/political economy explanatory variables include:
  - Quality of the bureaucracy (ICRG database; based on perceptions).
  - Government stability (ICRG database): has a 12-point scale and measures both the government’s ability to carry out its program and its ability to stay in office. To limit the variability in the estimate, this variable is considered in logarithmic terms.(Source: International Country Risk Guide—ICRG).
- Democracy and autocracy variables (University of Maryland, Polity IV project):
  - Democracy indicator ranges from 1 to 10.
  - Autocracy indicator ranges from -10 to -1.
  - Dummy variables constructed:
    - Democracy dummy: value of zero for polity values less than zero, and the actual value for polity values greater than zero.
    - Autocracy dummy: value of zero for polity values greater than zero, and minus the actual value for polity values less than zero.
- Change in government:
  - Dummy variable recorded as one if the election of the executive is within one year of the commencement of the IMF-supported program.
- Length of term in office:
  - Number of years that the current executive has been in office at the time of program approval.
  - Both change-in-government and length-of-term variables sourced from the World Bank’s database.

### Appendix — Variables used in growth regression
- Variable definitions and measurement conventions:
  - Variable changes are defined either as logarithmic changes in the underlying variables (CPI, U.S. output) or as nominal changes divided by GDP in U.S. dollars (budget balance).
  - Output per capita: output in U.S. dollars valued at the PPP exchange rate and divided by the population.
  - Overvaluation: defined as the logarithmic difference between the real exchange rate measured using the CPI and a 25-year HP filter trend line.
  - Real interest rate: measured as the 3-month t-bill rate minus the inflation rate.
  - Macroeconomic variables source: World Economic Outlook.
  - Law and order: assessment of the strength and impartiality of the legal system and of the popular observance of the law (ICRG database).

*Source: _wp06213 - References*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2006/_wp06213.pdf_