## _wp15177 - Variable Definition and Sources

## Source details

**Canonical URL:** [_wp15177 - Variable Definition and Sources](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2015/_wp15177.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2015/_wp15177.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2015/_wp15177.pdf.json)

---

### Major themes and objectives
- Investigate whether Regional Trade Agreements (RTAs) affect growth volatility and whether growth volatility influences a country's decision to join an RTA.
- Sample: 170 developed and developing countries over 1978-2012, divided into five-year sub-periods (up to 7 data points per country).
- Two main contributions:
  - Estimating the impact of RTAs on growth volatility controlling for trade openness and other determinants.
  - Examining whether countries vulnerable to growth shocks are more likely to join RTAs.

### Measurement and empirical framework
- Dependent variable: growth volatility (Vgrowth).
  - Long-term component of ln(real GDP) modeled as an AR(1) with trend: ln(y_{i,t}) = α_i + β_i ln(y_{i,t−1}) + γ_i t + ε_{i,t}.
  - Cyclical component ε̂_{i,t} obtained from country-specific AR(1) fit.
  - For each five-year sub-period: Vgrowth = sqrt( Σ_j=1^5 (ε̂_{i,t} − mean(ε̂_{i,t}))^2 / 4 ).
  - Robustness: standard deviation of real GDP growth rate over five-year period.
- Main RTA indicators:
  - RTA membership dummy (1 if member of at least one RTA; 0 otherwise).
  - Share of exports to RTA members = (exports to RTA members / total exports).
  - Share of imports from RTA members = (imports from RTA members / total imports).
  - Regional trade openness = (exports to + imports from RTA members) / (total exports + total imports).
- Control variables:
  - GDP per capita (log).
  - Trade openness: (exports + imports) / GDP (log).
  - Volatility of terms of trade (log).
  - Volatility of inflation (log).
  - Volatility of private credit ratio to GDP (log) and volatility of private credit growth (log).
  - Additional: inflation (log), Rule of Law (log) in some specifications.
- Baseline linear specification:
  - Vgrowth_{i,t} = α + β RTA_{i,t} + A X_{i,t} + u_i + ε_{i,t}.

### Descriptive patterns (figures and sample)
- Number of countries in at least one RTA rose from a little above 50 in the late 1970s to close to 200 by 2012 (driven by middle-income countries; low-income countries lagged).
- Average number of regional trade partners per country rose from 2 in 1978 to 24 in 2012.
- Cross-sectional patterns:
  - Weak and at best slightly negative correlation between growth volatility and trade openness after controlling for income.
  - Countries in an RTA tend to experience smaller growth volatility than non-RTA countries in all income groups except the lowest.
  - Countries that trade more with regional partners appear to enjoy smoother growth; high-income countries concentrated in the right tail (stable growth and higher share of trade with regional partners); low-income countries at opposite end.

### Main empirical findings — Fixed-effects estimates (selected)
- RTA membership dummy (Table 1, columns (2)-(4)):
  - Coefficients: -0.236, -0.256, -0.230; standard errors [0.079]***, [0.078]***, [0.078]*** (significant at the 1 percent level).
  - Interpretation: a country in a RTA would experience on average about 25 percent less volatility than a similar non-member.
- Share of imports from regional trade partners (Table 1, columns (5)-(7)):
  - Coefficients: -0.004, -0.005, -0.005; standard errors [0.002]**, [0.002]**, [0.002]**.
- Share of exports to regional trade partners (Table 2):
  - Coefficients: -0.004, -0.004, -0.004; standard errors [0.002]*, [0.002]**, [0.002]**.
- Regional trade openness (Table 2):
  - Coefficients: -0.004, -0.004, -0.004; standard errors [0.002]*, [0.002]**, [0.002]**.
- Control variables (fixed effects):
  - Volatility of inflation (log): positive and significant at the 1 percent level (example: 0.237, [0.034]***).
  - Volatility of terms of trade (log): positive association; sometimes significant (example: 0.052, [0.030]*).
  - Volatility of private credit ratio (log) and volatility of private credit growth (log): positive and significant (examples: 0.159 [0.043]*** and 0.163 [0.047]***).

- Illustrative quantitative example:
  - Burkina Faso: if regional trade openness rose from 5 percent to 70 percent, growth volatility would be reduced by 26 percent using the Table 2 coefficient ((70-5)*(-0.004)*100).

### Main empirical findings — System GMM estimates (selected)
- Results confirm fixed-effects findings: regional trade integration lowers growth volatility in almost all specifications.
- RTA membership (Table 3, columns (1)-(4)):
  - Coefficients: -0.406, -0.374, -0.433, -0.391; standard errors [0.093]***, [0.096]***, [0.090]***, [0.095]***.
- Share of imports from regional trade partners (Table 3):
  - Coefficients: -0.005, -0.005, -0.005, -0.005; standard errors [0.002]*, [0.003]*, [0.002]**, [0.002]**.
- Share of exports to regional trade partners and regional trade openness (Table 4):
  - Share of exports: -0.005, -0.006, -0.005, -0.005; standard errors [0.002]**, [0.002]***, [0.002]**, [0.002]**.
  - Regional trade openness: -0.005, -0.005, -0.005, -0.004; standard errors [0.002]*, [0.002]**, [0.002]**, [0.002]*.
- Control variables (System GMM):
  - Volatility of terms of trade (log): positive and significant in several specifications (example: 0.085 [0.051]*).
  - Volatility of inflation (log): positive and significant (example: 0.210 [0.057]***).
  - Inflation (log): positive and significant where included (example: 0.300 [0.132]**).
  - Volatility of private credit ratio and private credit growth: positive and significant (examples: 0.271 [0.077]***, 0.205 [0.075]***).
- Instrument validity and diagnostics (examples from Table 3):
  - Hansen test prob.: 0.17, 0.46, 0.46, 0.68 — null not rejected.
  - AR2 test prob.: 0.39, 0.58, 0.50, 0.91 — no evidence of second-order serial correlation.

### Robustness checks
- Alternative volatility measure (standard deviation of real GDP growth rate) — System GMM (Table 5):
  - RTA membership: -0.458 [0.094]***, -0.386 [0.087]*** (columns (1)-(2)); -0.215 [0.098]** (column (4)).
  - Share of imports from regional trade partners: -0.007 [0.002]*** (column (3)).
  - Share of exports to regional trade partners: -0.008 [0.002]*** (column (4)).
  - Regional trade openness: -0.007 [0.002]*** and -0.004 [0.002]* (columns (5)-(6)).
- Inclusion of a step dummy for the 2008 global financial crisis did not change results in a significant way.

### Additional analysis — Do countries join RTAs in response to growth volatility?
- RTA entry decision modeled as: RTA_{i,t} = φ + Σ_{j=2}^3 δ_j Vgrowth_{i,t−j} + B Z_{i,t} + u_i + ε_{i,t}.
  - Lagged Vgrowth (Lags 2 and 3) included to reflect that RTA decisions are taken years before formal entry and to maximize AIC.
- Motivation: countries vulnerable to shocks may join RTAs for diversification, insurance against shocks, and to reduce protectionist responses of partners.
- Estimation approach: panel logit fixed-effect estimator (lagged RDGP, DRGDP, DKL, DROWKL to address endogeneity); predicted probability of RTA from the logit is used in growth-volatility regressions estimated with System GMM.

- Table 6 (Explaining the Probability of a RTA) — selected coefficients:
  - Growth Volatility (Lag 3): 0.485, 0.416, 1.008, 1.078, 1.072 (significant in reported columns).
  - Growth Volatility (Lag 2): -0.248, -0.280, -0.311, -0.285, -0.266.
  - Standard Deviation of Real GDP Growth Rate (Lag 3): 0.847 and 0.853 (both ***).
  - Standard Deviation of Real GDP Growth Rate (Lag 2): -0.317 and -0.330.
  - Ratio of Average Growth Volatility in RTA to that of the ROW: -2.293 [1.089]** and -2.402 [1.060]**.
  - RGDP (Lag 1): 0.264, 1.226, 1.500, 1.844, 1.482, 1.845 (reported ***) across columns.
  - DRGDP (Lag 1): -0.289 [0.043]*** (one column).
  - DKL (Lag 1): 3.692, 5.709, 6.678, 5.663, 6.681 (reported ***).
  - DROWKL (Lag 1): -1.206, -1.492, -1.208, -1.505 (reported ***).
  - Observations by column: 511, 511, 466, 466, 466, 473, 473.
  - Number of countries by column: 80, 80, 72, 72, 72, 73, 73.
  - Pseudo R2: 0.53, 0.51, 0.64, 0.66, 0.67, 0.66, 0.67.
  - Predictive power: 67 percent correctly predicted (column 5).

- Table 7 (Predicted Probability of a RTA and Growth Volatility — System GMM) — selected:
  - GDP per capita (log): 0.114 [0.051]** and 0.111 [0.050]**.
  - Predicted Probability of a RTA: -0.240 [0.100]** and -0.310 [0.099]***.
  - Trade openness (log): 0.021 [0.181] and -0.021 [0.181] (not significant).
  - Volatility of Terms of Trade (log): 0.172 [0.053]***.
  - Volatility of Inflation (log): 0.106 [0.060]* and 0.099 [0.061].
  - Volatility of Private Credit Ratio (log): 0.290 [0.091]*** and 0.299 [0.094]***.
  - Observations: 622 and 622.
  - Number of countries: 151 and 151.
  - Hansen test prob.: 0.71 and 0.60.
  - AR2 test prob.: 0.71 and 0.86.

### Summary statistics (Appendix Table 1)
- Growth Volatility (log): Observations 737; Mean -3.8; Std. Dev. 0.7; Min -6.1; Max -1.0.
- GDP per Capita (log): 737; Mean 7.7; Std. Dev. 1.6; Min 4.7; Max 11.6.
- RTA Membership: 737; Mean 0.7; Std. Dev. 0.4; Min 0; Max 1.
- Share of Imports from Regional Trade Partners: 640; Mean 22.8; Std. Dev. 27.8; Min 0.0; Max 92.9.
- Share of Exports to Regional Trade Partners: 640; Mean 21.6; Std. Dev. 28.0; Min 0.0; Max 97.2.
- Regional Trade Openness: 640; Mean 22.2; Std. Dev. 27.5; Min 0.0; Max 93.5.
- Trade openness (log): 737; Mean 4.3; Std. Dev. 0.6; Min 2.6; Max 6.0.
- Volatility of Terms of Trade (log): 737; Mean -3.2; Std. Dev. 1.2; Min -10.5; Max -0.8.
- Inflation (log): 737; Mean 0.1; Std. Dev. 0.3; Min 0.0; Max 4.0.
- Volatility of Inflation (log): 737; Mean -3.5; Std. Dev. 1.5; Min -33.1; Max 0.7.
- Volatility of Private Credit Ratio (log): 698; Mean -2.6; Std. Dev. 0.8; Min -5.1; Max -0.1.
- Volatility of Private Credit Growth (log): 726; Mean -2.2; Std. Dev. 0.7; Min -5.0; Max -0.2.
- RGDP: 729; Mean 19.6; Std. Dev. 12.2; Min 0.0; Max 31.2.
- DRGDP: 729; Mean 8.2; Std. Dev. 9.4; Min 0.0; Max 29.4.
- DKL: 688; Mean 4.2; Std. Dev. 3.1; Min 0.1; Max 11.6.
- DROWKL: 688; Mean 3.2; Std. Dev. 3.0; Min 0.0; Max 10.2.
- Ratio of Average Growth Volatility in RTA to that of the ROW: 729; Mean 0.8; Std. Dev. 0.5; Min 0.0; Max 1.4.

### Interpretation and policy implications
- Core empirical result: RTAs are consistently associated with lower growth volatility across descriptive statistics and econometric estimations for 1978-2012.
- Robustness: findings hold across estimators (fixed-effect and System GMM), different volatility measures, and different RTA measures (dummy; export/import shares; regional trade openness).
- Determinants of RTA formation:
  - Past large growth shocks increase likelihood of participating in an RTA (positive Growth Volatility (Lag 3) effect).
  - Probability to form an RTA declines with higher growth volatility of regional trade partners relative to ROW (negative Ratio of Average Growth Volatility in RTA to that of the ROW).
  - RGDP, DRGDP, DKL, DROWKL exhibit expected signs and statistical significance.
- Policy recommendations:
  - Countries vulnerable to growth shocks, particularly low-income countries, would benefit from joining a RTA or deepening trade with existing regional partners to cope with growth shocks.
  - RTAs can strengthen policy credibility and coordination, reduce likelihood of conflicts, and reduce risks of growth volatility.
  - Low-income countries would benefit from RTAs involving advanced economies, as richer partners tend to have more stable growth and can minimize shock transmission and increase resilience through access to larger markets.

*Source: IMF staff paper content unit "_wp15177 - 3. Variable Definition and Sources ....................................................................................."*

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

### _wp15177 - References

### Tables
- 1. Impact of RTA Membership and Import Share of Regional Trade Partners on Growth Volatility: Fixed-Effect Estimator ...................................................................................... 17
- 2. Impact of Export Share of Regional Trade Partners and their Total Trade Share on Growth Volatility: Fixed-Effect Estimator ...................................................................................... 18
- 3. Impact of RTA Membership and Import Share of Regional Trade Partners on Growth Volatility: System GMM Estimator ................................................................................... 21
- 4. Impact of Export Share of Regional Trade Partners and their Total Trade Share on Growth Volatility: System GMM Estimator ................................................................................... 22
- 5. Robustness Analysis with Growth Volatility Measured by the Standard Deviation of Growth Rate: System GMM Estimator .............................................................................. 23
- 6. Explaining the Probability of a RTA: the Role of Growth Volatility: Panel Logit Estimator ............................................................................................................................................ 28
- 7. Predicted Probability of a RTA and Growth Volatility: System GMM Estimator ............. 29

### Figures
- 1. Number of Countries Member of at Least one RTA, 1978-2012 ......................................... 4
- 2. Average Number of Regional Trading Partners per Country, 1978-2012 ............................ 5
- 3. Growth Volatility and Trade Openness .............................................................................. 13
- 4. Growth Volatility and Membership of Regional Trade Agreement ................................... 14
- 5. Growth Volatility and Trade Intensity with Regional Trade Partners ................................ 15
- 6. Growth Volatility and GDP per Capita ............................................................................... 19

### Appendix Tables
- 1. Summary Statistics.............................................................................................................. 35
- 2. Correlation Matrix .............................................................................................................. 36

*Source: _wp15177 - References*

### 3. Variable Definition and Sources .....................................................................................

### _wp15177 - 3. Variable Definition and Sources .....................................................................................

### Major themes and objectives
- Investigate whether Regional Trade Agreements (RTAs) affect growth volatility and whether growth volatility influences a country's decision to join an RTA.
- Use a worldwide sample of 170 developed and developing countries over the period 1978-2012, divided into five-year sub-periods (up to 7 data points per country).
- Focus on two main contributions:
  - Estimating the impact of RTAs on growth volatility controlling for trade openness and other determinants.
  - Examining whether countries vulnerable to growth shocks are more likely to join RTAs.

### Measurement and empirical framework
- Dependent variable: growth volatility (Vgrowth).
  - Long-term component of ln(real GDP) is modeled as an AR(1) with trend: ln(y_{i,t}) = α_i + β_i ln(y_{i,t−1}) + γ_i t + ε_{i,t}.
  - The cyclical component ε̂_{i,t} is obtained from the country-specific fit of the AR(1) model.
  - For each five-year sub-period, Vgrowth is calculated as the standard error of ε̂_{i,t}: Vgrowth = sqrt( Σ_j=1^5 (ε̂_{i,t} − mean(ε̂_{i,t}))^2 / 4 ).
  - Robustness checks also use the standard deviation of real GDP growth rate.
- Main variable of interest: Regional Trade Agreement (RTA).
  - Four indicators used:
    - RTA membership dummy (1 if country is member of at least one RTA; 0 otherwise).
    - Share of exports to RTA members (exports to RTA members / total exports).
    - Share of imports from RTA members (imports from RTA members / total imports).
    - Regional trade openness = (exports to + imports from RTA members) / (total exports + total imports).
- Control variables:
  - Level of economic development: GDP per capita (log).
  - Trade openness: (exports + imports) / GDP (log).
  - Volatility of terms of trade (log) — measured by change in ratio of export prices to import prices.
  - Volatility of inflation (log) — volatility in consumer price index.
  - Financial instability: volatility of private credit ratio to GDP (log) and volatility of private credit growth (log).
  - Additional controls in some specifications: inflation (log), Rule of Law (log).

- Baseline linear specification:
  - Vgrowth_{i,t} = α + β RTA_{i,t} + A X_{i,t} + u_i + ε_{i,t}.

### Descriptive patterns (from figures and sample)
- Number of countries in at least one RTA rose from a little above 50 in the late 1970s to close to 200 by 2012 (driven by middle-income countries; low-income countries lagged).
- Average number of regional trade partners per country rose from 2 in 1978 to 24 in 2012.
- Cross-sectional patterns:
  - Weak and at best slightly negative correlation between growth volatility and trade openness after controlling for income.
  - Countries in an RTA tend to experience smaller growth volatility than non-RTA countries in all income groups except the lowest.
  - Countries that trade more with their regional partners appear to enjoy smoother growth; high-income countries concentrated in right tail (stable growth and higher share of trade with regional partners); low-income countries at opposite end.

### Main empirical findings — Fixed-effects estimates (selected results)
- RTA membership dummy:
  - Coefficient estimates (fixed effects) for RTA membership: -0.236, -0.256, -0.230 (reported in Table 1 columns (2)-(4)); all are significant at the 1 percent level ([0.079]***, [0.078]***, [0.078]***).
  - Interpretation reported: a country in a RTA would experience on average about 25 percent less volatility than a similar non-member.
- Share of imports from regional trade partners (fixed effects):
  - Coefficients: -0.004, -0.005, -0.005 (Table 1 columns (5)-(7)); standard errors [0.002]** / [0.002]** / [0.002]**.
- Share of exports to regional trade partners (Table 2):
  - Coefficients: -0.004, -0.004, -0.004; standard errors [0.002]* / [0.002]** / [0.002]**.
- Regional trade openness (Table 2):
  - Coefficients: -0.004, -0.004, -0.004; standard errors [0.002]* / [0.002]** / [0.002]**.
- Control variables (fixed effects):
  - Volatility of inflation (log): positive and significant at the 1 percent level across specifications (e.g., 0.237, [0.034]*** in Table 1 column (1)).
  - Volatility of terms of trade (log): positive association with growth volatility but only significant in a few specifications (e.g., 0.052, [0.030]* in Table 1 column (1)).
  - Volatility of private credit ratio (log) and volatility of private credit growth (log): positive and significant (e.g., 0.159 [0.043]*** and 0.163 [0.047]*** respectively in Table 1).

- Illustrative quantitative statement from the paper:
  - Example using Burkina Faso: If Burkina Faso’s regional trade openness rose from 5 percent to 70 percent (as Italy’s with EU), its growth volatility would be reduced by 26 percent using coefficient from Table 2 ((70-5)*(-0.004)*100).

### Main empirical findings — System GMM estimates (selected results)
- Results confirm fixed-effects findings: regional trade integration lowers growth volatility in almost all specifications.
- RTA membership (System GMM):
  - Coefficients: -0.406, -0.374, -0.433, -0.391 (Table 3 columns (1)-(4)); standard errors [0.093]***, [0.096]***, [0.090]***, [0.095]***.
- Share of imports from regional trade partners (System GMM):
  - Coefficients: -0.005, -0.005, -0.005, -0.005 (Table 3); standard errors [0.002]* / [0.003]* / [0.002]** / [0.002]**.
- Share of exports to regional trade partners and regional trade openness (Table 4):
  - Share of exports: -0.005, -0.006, -0.005, -0.005 (various columns) with standard errors [0.002]** / [0.002]*** / [0.002]** / [0.002]**.
  - Regional trade openness: -0.005, -0.005, -0.005, -0.004 with standard errors [0.002]* / [0.002]** / [0.002]** / [0.002]*.
- Control variables (System GMM):
  - Volatility of terms of trade (log): positive and significant in several specifications (e.g., 0.085 [0.051]*, Table 3 column (1)).
  - Volatility of inflation (log): positive and significant (e.g., 0.210 [0.057]***, Table 3 column (1)).
  - Inflation (log): positive and significant in specifications where included (e.g., 0.300 [0.132]**, Table 3).
  - Volatility of private credit ratio and of private credit growth: positive and significant (e.g., 0.271 [0.077]***, 0.205 [0.075]*** respectively in Table 3).
- Instrument validity and diagnostics (System GMM):
  - Hansen test probability values reported across tables (examples): 0.17, 0.46, 0.46, 0.68 (Table 3 columns) — null not rejected.
  - AR2 test probability values reported (examples): 0.39, 0.58, 0.50, 0.91 (Table 3 columns) — no evidence of second-order serial correlation.

### Robustness checks
- Growth volatility measured alternatively by the standard deviation of real GDP growth rate — System GMM results (Table 5) confirm that:
  - RTA membership: -0.458 [0.094]*** and -0.386 [0.087]*** in columns (1)-(2); -0.215 [0.098]** in column (4).
  - Share of imports from regional trade partners: -0.007 [0.002]*** (column (3)).
  - Share of exports to regional trade partners: -0.008 [0.002]*** (column (4)).
  - Regional trade openness: -0.007 [0.002]*** and -0.004 [0.002]* (columns (5)-(6)).
- Inclusion of a step dummy for the 2008 global financial crisis did not change results in a significant way.

### Additional analysis — Do countries join RTAs in response to growth volatility?
- Framework for RTA entry decision follows Baier and Bergstrand (2004) extended to include lagged growth volatility:
  - Estimated model: RTA_{i,t} = φ + Σ_{j=2}^3 δ_j Vgrowth_{i,t−j} + B Z_{i,t} + u_i + ε_{i,t}.
  - Vgrowth lags (second and third) included to reflect that RTA decisions are taken years before formal entry and to maximize AIC.
- The paper motivates inclusion of growth volatility as a determinant because:
  - Countries vulnerable to shocks may join RTAs to gain product and market diversification, insurance against shocks, and to reduce the likelihood of protectionist measures from partners.
- The paper builds on Baier and Bergstrand (2004) determinants (geography, economic size, factor endowments) and Whalley (1998) political/strategic motives, while explicitly adding growth volatility to the set of economic determinants.

### Synthesis of substantive conclusions
- Empirical evidence (fixed effects and System GMM) consistently indicates that RTA membership and deeper trade intensity with regional partners are associated with lower growth volatility.
- Estimated magnitudes:
  - RTA membership dummy: fixed-effects ~ -0.236; System GMM ~ -0.406 (both significant at 1 percent).
  - Trade-intensity measures (shares and regional trade openness): coefficients around -0.004 to -0.007 across specifications (significant at 1–10 percent depending on specification).
- Key drivers of growth volatility in the models: volatility of inflation, volatility of private credit measures, and volatility of terms of trade (in some specifications) — all positively associated with growth volatility.
- Trade openness (overall) is not robustly significant, consistent with theoretical ambiguity and the possibility that some channels (e.g., terms of trade shocks) are accounted for separately.
- Robustness checks, alternative volatility measures, and dynamic-panel instruments support the main conclusion that RTAs tend to reduce growth volatility.

*Source: IMF staff paper content unit "_wp15177 - 3. Variable Definition and Sources ....................................................................................."*

### 1. RDGP, the sum of the logs of real GDPs of countries in the RTA and DRGDP is

### _wp15177 - 1. RDGP, the sum of the logs of real GDPs of countries in the RTA and DRGDP is

### Variable definitions and construction
- RDGP: Sum of the logs of real GDPs of countries in the RTA (RGDP in tables).
- DRGDP: Absolute value of the difference between the log of real GDP of the country and the average of the other countries in the RTA.
- DKL: Absolute value of the difference between the logs of the capital–labor ratios of the country and the average of the other countries in the RTA.
- DROWKL: Difference between the average capital–labor ratio of the RTA member countries and that of the ROW.
- Growth Volatility (main measure): The standard error of the residual of the log of real GDP regressed on its lags value and a time trend (assuming an AR(1) process with a trend), calculated over a five-year period. Alternative measure: The standard error of annual real GDP growth rate over a five-year period.
- Ratio of Average Growth Volatility in RTA to that of the ROW: Average growth volatility of the country's regional partners in an RTA divided by the average growth volatility of non-RTA members.
- RTA Membership: dummy variable equal to 1 when a country has signed at least one regional trade agreement, and zero otherwise.

### Econometric specification and estimation approach
- Panel logit model (logit fixed effect estimator) used to explain probability of joining an RTA; lagged explanatory variables RDGP, DRGDP, DKL and DROWKL are used to address potential endogeneity.
- Predicted probability of an RTA derived from the panel logit (Equation 5) is then used to replace the RTA dummy in the growth-volatility regressions (Equation 4).
- Growth-volatility regressions estimated with System GMM to address endogeneity; fixed-effect estimator used for robustness and to control for country-specific effects.
- Lag structure: Growth Volatility (Lag 2) and Growth Volatility (Lag 3) included; alternative specification uses Standard Deviation of Real GDP Growth Rate (Lag 2) and (Lag 3).

### Main empirical findings (Table 6: Explaining the Probability of a RTA)
- Growth Volatility (Lag 3): positive and significant in multiple specifications
  - Coefficients reported: 0.485, 0.416, 1.008, 1.078, 1.072 (standard errors in brackets; significance up to *** in columns reported).
- Growth Volatility (Lag 2): coefficients reported: -0.248, -0.280, -0.311, -0.285, -0.266 (noted with reported standard errors).
- Standard Deviation of Real GDP Growth Rate (Lag 3): 0.847 and 0.853 (both reported as ***).
- Standard Deviation of Real GDP Growth Rate (Lag 2): -0.317 and -0.330 (reported in two columns).
- Ratio of Average Growth Volatility in RTA to that of the ROW: coefficients -2.293 and -2.402 (both [1.089]** and [1.060]** respectively) — negative sign implies higher probability to join an RTA when regional partners are relatively more stable compared to ROW.
- RGDP (Lag 1): coefficients reported across columns: 0.264, 1.226, 1.500, 1.844, 1.482, 1.845 (all with [..]***).
- DRGDP (Lag 1): -0.289 [0.043]*** in one reported column.
- DKL (Lag 1): coefficients 3.692, 5.709, 6.678, 5.663, 6.681 (all [..]***).
- DROWKL (Lag 1): coefficients -1.206, -1.492, -1.208, -1.505 (all [..]***).
- Model fit and sample:
  - Observations: 511, 511, 466, 466, 466, 473, 473 (by column).
  - Number of countries: 80, 80, 72, 72, 72, 73, 73.
  - Pseudo R2: 0.53, 0.51, 0.64, 0.66, 0.67, 0.66, 0.67.
- Predictive power: 67 percent of values of the dependent variable correctly predicted (column 5).

### Main empirical findings (Table 7: Predicted Probability of a RTA and Growth Volatility — System GMM)
- Dependent variable: Growth Volatility (two measures: "Our measure of Growth Volatility" and "Standard Deviation of Real GDP Growth Rate").
- GDP per capita (log): 0.114 [0.051]** and 0.111 [0.050]**.
- Predicted Probability of a RTA: -0.240 [0.100]** and -0.310 [0.099]*** — negative and statistically significant, indicating predicted RTA membership associated with lower growth volatility.
- Trade openness (log): 0.021 [0.181] and -0.021 [0.181] (not significant).
- Volatility of Terms of Trade (log): 0.172 [0.053]*** (positive and significant).
- Volatility of Inflation (log): 0.106 [0.060]* and 0.099 [0.061] (weakly significant).
- Volatility of Private Credit Ratio (log): 0.290 [0.091]*** and 0.299 [0.094]*** (positive and significant).
- Constant: -3.161 [0.641]*** and 1.722 [0.658]***.
- Sample and tests:
  - Observations: 622 and 622.
  - Number of countries: 151 and 151.
  - Hansen test prob.: 0.71 and 0.60.
  - AR2 test prob.: 0.71 and 0.86.

### Summary statistics (Appendix Table 1)
- Growth Volatility (log): Observations 737; Mean -3.8; Std. Dev. 0.7; Min -6.1; Max -1.0.
- GDP per Capita (log): 737; Mean 7.7; Std. Dev. 1.6; Min 4.7; Max 11.6.
- RTA Membership: 737; Mean 0.7; Std. Dev. 0.4; Min 0; Max 1.
- Share of Imports from Regional Trade Partners: 640; Mean 22.8; Std. Dev. 27.8; Min 0.0; Max 92.9.
- Share of Exports to Regional Trade Partners: 640; Mean 21.6; Std. Dev. 28.0; Min 0.0; Max 97.2.
- Regional Trade Openness: 640; Mean 22.2; Std. Dev. 27.5; Min 0.0; Max 93.5.
- Trade openness (log): 737; Mean 4.3; Std. Dev. 0.6; Min 2.6; Max 6.0.
- Volatility of Terms of Trade (log): 737; Mean -3.2; Std. Dev. 1.2; Min -10.5; Max -0.8.
- Inflation (log): 737; Mean 0.1; Std. Dev. 0.3; Min 0.0; Max 4.0.
- Volatility of Inflation (log): 737; Mean -3.5; Std. Dev. 1.5; Min -33.1; Max 0.7.
- Volatility of Private Credit Ratio (log): 698; Mean -2.6; Std. Dev. 0.8; Min -5.1; Max -0.1.
- Volatility of Private Credit Growth (log): 726; Mean -2.2; Std. Dev. 0.7; Min -5.0; Max -0.2.
- RGDP: 729; Mean 19.6; Std. Dev. 12.2; Min 0.0; Max 31.2.
- DRGDP: 729; Mean 8.2; Std. Dev. 9.4; Min 0.0; Max 29.4.
- DKL: 688; Mean 4.2; Std. Dev. 3.1; Min 0.1; Max 11.6.
- DROWKL: 688; Mean 3.2; Std. Dev. 3.0; Min 0.0; Max 10.2.
- Ratio of Average Growth Volatility in RTA to that of the ROW: 729; Mean 0.8; Std. Dev. 0.5; Min 0.0; Max 1.4.

### Interpretation and policy implications (Conclusion)
- Core result: RTAs are consistently associated with lower growth volatility across descriptive statistics and econometric estimations for a sample of 170 countries covering 1978-2012.
- Robustness: Results hold across estimators (fixed-effect and System GMM), different measures of growth volatility, and different measures of RTAs (RTA dummy, ratio of exports to RTA members to total exports, ratio of imports from RTA members to total imports, and sum of exports to and imports from RTA members divided by sum of total exports and imports).
- Determinants of RTA formation:
  - Countries experiencing large past growth shocks are more likely to participate in an RTA (positive Growth Volatility (Lag 3) effect).
  - The probability to form an RTA declines with higher growth volatility of regional trade partners relative to ROW (negative Ratio of Average Growth Volatility in RTA to that of the ROW).
  - Other variables (RGDP, DRGDP, DKL, DROWKL) show expected signs and are statistically significant.
- Policy recommendations:
  - Countries vulnerable to growth shocks, particularly low-income countries, would benefit from joining a RTA or deepening trade with existing regional partners to cope with growth shocks.
  - RTAs can strengthen policy credibility and coordination, reduce likelihood of conflicts, and reduce risks of growth volatility.
  - Low-income countries would benefit from RTAs involving advanced economies, as richer partners tend to have more stable growth and can minimize shock transmission and increase resilience through access to larger markets.

*Source: _wp15177 - 1. RDGP, the sum of the logs of real GDPs of countries in the RTA and DRGDP is*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2015/_wp15177.pdf_