Heckscher-Ohlin Trade Model

Evidence and data

Use factor endowments, sector factor intensities, trade flows, relative goods prices, factor prices, and technology differences.

Calibration or measurement

Relative factor abundance and factor intensity define the prediction. Technology gaps, trade costs, and global value chains must be accounted for before empirical use.

Boundaries

• Identical technology is a strong assumption.
• Factor-price equalization often fails in data.
• Firm heterogeneity and scale economies are outside the two-good baseline.

Use guidance

When sufficient

Comparative-static predictions about trade patterns when factor endowments differ substantially across countries and technology is similar. North-South trade in goods with very different labor and capital intensities, the distributional effects of trade on factor prices via Stolper-Samuelson (1941), and the direction of specialization across industries with stable factor intensities are all within the model's reliable range.

When sketch only

Use as the baseline for factor-driven trade intuition, not for explaining observed trade volumes between similar economies. The model assumes identical technology across countries, no economies of scale, and perfect competition. These assumptions fail for most North-North trade, which is dominated by intra-industry exchange in differentiated goods.

When to switch

Switch to a Krugman (1980 JPE) monopolistic-competition model when intra-industry trade in similar goods between similar countries is the question. Switch to an Eaton-Kortum (2002 Econometrica) Ricardian model when technology differences, rather than factor endowments, are the primary driver of comparative advantage.

Falsification signal

The Leontief paradox (1953): US exports in the 1940s were less capital-intensive than US imports, the opposite of what the model predicts for the most capital-abundant economy of that era. More broadly, the volume of intra-industry trade between high-income countries with similar endowments is far larger than factor-proportion differences can explain.