3 Trade Research
3.1 Introduction
Applied trade research begins with a policy question, but a gravity paper requires that question to become measurable. The researcher must define the trade outcome, the countries, the years, the institutional variable, the comparison group, and the estimator.
The Post-Soviet project asks how institutional alignment is associated with bilateral trade among fifteen post-Soviet economies. The institutional variables are \(wto\_joint\), \(EU\_joint\), and \(EAEU\_joint\).
3.2 From policy concepts to variables
Institutions are broad concepts. Gravity models require specific indicators.
\(wto\_joint\) equals 1 when both countries in a dyad are WTO members in year \(t\). It measures shared multilateral trade-system membership.
\(EU\_joint\) equals 1 when both countries are EU members in year \(t\). In the Post-Soviet sample, this mainly captures Baltic EU membership.
\(EAEU\_joint\) equals 1 when both countries are EAEU members in year \(t\). It captures Eurasian regional institutional alignment after 2015.
These variables do not prove implementation quality or causality. They measure shared institutional status.
3.3 Why regional cases matter
Regional gravity papers are useful because institutional history, geography, and trade costs differ across regions. A result from the Post-Soviet sample may not generalize directly to Africa, GCC, MENA, COMESA, ECOWAS, ASEAN, or the Western Balkans.
Students use the Post-Soviet paper as a template, then adapt the structure to another region. The adaptation requires new institutional variables, a justified country sample, and clear policy motivation.
3.4 Policy question map
| Policy question | Gravity variable | Possible interpretation |
|---|---|---|
| Does shared WTO membership matter for intra-regional trade? | \(wto\_joint\) | Association between common multilateral membership and bilateral trade |
| Is deep regional integration linked to higher trade? | \(EU\_joint\) or regional equivalent | Association between shared deep-integration status and trade |
| Does Eurasian regional integration matter after 2015? | \(EAEU\_joint\) | Association between joint EAEU membership and trade |
| Do neighboring countries trade more? | \(contig\) | Conditional association between common border and trade |
| Does language reduce trade costs? | \(comlang\_off\) | Conditional association between common official language and trade |
| Does distance still matter in a historically connected region? | \(distw\) | Distance elasticity or trade-cost association |
The table turns broad policy language into variables that can be estimated.
3.5 Adapting the template
Students can adapt the Post-Soviet design to other regions:
- Africa: AfCFTA, WTO membership, regional economic communities;
- GCC: customs union, energy trade, logistics, common institutions;
- MENA: regional agreements, border frictions, food trade;
- COMESA, ECOWAS, or EAC: regional economic community membership;
- ASEAN: regional integration and production networks;
- Western Balkans: EU integration, CEFTA, WTO accession.
The structure remains the same:
region -> policy question -> gravity variables -> dataset -> estimator sequence -> interpretation3.6 Research question discipline
A good research question is narrow enough to estimate and important enough to motivate.
Weak question:
Does trade policy matter?
Better question:
Is joint EU membership associated with higher bilateral trade among post-Soviet economies after controlling for economic size, distance, and exporter/importer heterogeneity?
The better question identifies the institution, outcome, region, controls, and empirical comparison.
3.7 From replication to extension
Replication teaches students what must be specified before running a model:
- country sample;
- year range;
- dependent variable;
- policy indicators;
- control variables;
- estimator sequence;
- fixed effects;
- standard errors;
- interpretation limits.
An extension should change the research question or region, not the standards of evidence.