Appendix D — Paper Template
D.1 Purpose
This appendix provides a complete template for the final region-specific gravity paper. Replace placeholders with the student’s own region, dataset, estimates, tables, and interpretation. Do not invent results before estimation is complete.
D.2 Title Page
Title: [Institutional Variable] and Bilateral Trade in [Region]: Gravity Evidence from [Years]
Author: [Student Name]
Course: NREC6006 - International Trade
Date: [Submission Date]
Replication base: Post-Soviet gravity workflow translated to Python
D.3 Abstract
Suggested length: 150-200 words.
Template:
This paper studies whether
[policy or institution]is associated with bilateral trade among[region]countries during[years]. Using a dyad-year gravity dataset built from[data sources], the paper estimates baseline OLS, fixed-effects OLS, PPML, and robustness specifications. The main dependent variable is[trade flow measure]. The core institutional variable is[policy variable]. Results show that[brief findings placeholder]. The analysis contributes to applied trade research by linking a reproducible Python workflow to a region-specific policy question.
D.4 Keywords
Use 4-6 keywords.
Suggested format:
gravity model; international trade; [region]; [institution]; PPML; Python replication
D.5 Introduction
Suggested length: 800-1,000 words.
The introduction should answer four questions:
- What policy or trade problem motivates the paper?
- What is the research question?
- What is the empirical contribution?
- What are the main findings?
Suggested structure:
- Paragraph 1: Regional or policy motivation.
- Paragraph 2: Why bilateral trade is the right outcome.
- Paragraph 3: Research question and institutional variable.
- Paragraph 4: Data, period, and estimators.
- Paragraph 5: Findings preview without overstating causality.
- Paragraph 6: Roadmap of the paper.
Common mistake: beginning with broad trade theory rather than the specific empirical problem.
D.6 Literature Review
Suggested length: 900-1,200 words.
Organize the review around the paper’s design rather than listing papers one by one.
Recommended subsections:
- Gravity foundations and structural gravity.
- Estimation issues: OLS, fixed effects, PPML, and zeros.
- Regional integration and trade agreements.
- Literature specific to
[region]or[policy variable].
Each paragraph should explain how the cited literature informs the current paper.
D.7 Data Section
Suggested length: 700-900 words.
Report:
- Data source.
- Countries included.
- Years covered.
- Unit of observation.
- Dependent variable.
- GDP and distance variables.
- Institutional variables.
- Sample restrictions.
- Missing-value handling.
- Zero-flow treatment.
Template:
The dataset is a directed dyad-year panel covering
[N]exporters,[N]importers, and[years]. The dependent variable is[flow variable]. Economic size is measured using[GDP variables], and bilateral trade costs are proxied by[distance and controls]. The main institutional variable is[policy variable], coded as[definition].
Required data table:
| Item | Value |
|---|---|
| Observations | [N] |
| Exporters | [N] |
| Importers | [N] |
| Years | [range] |
| Zero-flow observations | [N] |
| Positive-flow observations | [N] |
D.8 Methodology
Suggested length: 900-1,200 words.
Include the model sequence:
- Baseline OLS.
- Fixed-effects OLS.
- PPML.
- Robustness specification.
Baseline equation:
\[ \begin{aligned} \log X_{ijt} &= \beta_0 + \beta_1 \log GDP_{it} + \beta_2 \log GDP_{jt} \\ &\quad + \beta_3 \log D_{ij} + \gamma Z_{ijt} + \varepsilon_{ijt} \end{aligned} \]
PPML equation:
\[ \begin{aligned} E[X_{ijt} \mid Z_{ijt}] &= \exp\left( \beta_0 + \beta_1 \log D_{ij} + \gamma Z_{ijt} + FE \right) \end{aligned} \]
Explain why each estimator is included and what changes across specifications.
D.9 Results Section
Suggested length: 1,000-1,400 words.
Do not narrate every coefficient. Focus on the variables that answer the research question.
Recommended order:
- Baseline gravity results.
- Institutional coefficient.
- Fixed-effects changes.
- PPML changes.
- Economic significance.
Template:
The coefficient on
[policy variable]is[positive/negative/unstable]across the main specifications. In the preferred model, the estimated coefficient is[coefficient], which corresponds to a conditional percent difference of[percent effect]. This association should be interpreted conditional on[controls and fixed effects].
D.10 Robustness Section
Suggested length: 700-1,000 words.
Robustness should test reasonable alternatives, not random model shopping.
Examples:
- Alternative estimator.
- Alternative fixed effects.
- Alternative sample period.
- Alternative trade-flow measure.
- Excluding outliers.
- Positive-flow versus zero-inclusive estimation.
- Robust versus clustered standard errors.
Template:
The main result is robust across
[specifications], but changes under[specification]. This difference likely reflects[sample, fixed-effect, or estimator reason]. The result should therefore be interpreted as[cautious interpretation].
D.11 Policy Discussion
Suggested length: 600-900 words.
Connect coefficients to mechanisms. Avoid treating formal membership as full implementation.
Discuss:
- Trade costs.
- Customs and logistics.
- Infrastructure.
- Institutional quality.
- Short-run versus long-run adjustment.
- Limits of causal interpretation.
Template:
The estimates suggest that
[institution]is associated with[direction]bilateral trade conditional on the model controls. This pattern is consistent with[mechanism], but the estimates do not prove that membership alone caused the change. Implementation, infrastructure, and market conditions may condition the effect.
D.12 Conclusion
Suggested length: 400-600 words.
The conclusion should:
- Restate the research question.
- Summarize the main empirical pattern.
- Explain the policy relevance.
- Acknowledge limitations.
- Identify one clear next step.
Avoid introducing new estimates or new literature in the conclusion.
D.13 Required Tables
| Table | Content |
|---|---|
| Table 1 | Sample diagnostics and data coverage |
| Table 2 | Descriptive statistics |
| Table 3 | Top exporters and importers or trade concentration |
| Table 4 | Main gravity results |
| Table 5 | Robustness checks |
| Table 6 | Coefficient comparison or preferred model summary |
D.14 Required Figures
| Figure | Content |
|---|---|
| Figure 1 | Distribution of trade flows |
| Figure 2 | Top exporters or top importers |
| Figure 3 | Trade versus distance or GDP product |
| Figure 4 | Selected coefficient comparison or trade network |
D.15 Reproducibility Checklist
- Dataset path is documented.
- Raw and cleaned files are distinguished.
- Required variables are listed.
- Sample restrictions are explicit.
- Missing values are reviewed.
- Zero-flow treatment is stated.
- Python package versions are recorded.
- Notebook runs from top to bottom.
- Tables are generated from code.
- Figures are generated from code.
- Regression formulas match the paper.
- Fixed effects are documented.
- Standard errors are documented.
- Convergence status is reported for PML models.
D.16 Submission Checklist
Submit:
- Final paper PDF.
- Reproducible Python notebook.
- Clean dataset or documented data-construction script.
- Presentation slides.
- Replication notes.
Before submission:
- Confirm table numbers and captions.
- Confirm figure numbers and captions.
- Confirm all citations appear in the bibliography.
- Confirm the paper does not claim causality unless the design supports it.
- Confirm all placeholders have been replaced.
- Confirm code and paper report the same sample size.