Melitz Heterogeneous Firms Model: How Trade Reshapes Industries from the Inside

For most of the twentieth century, trade theory treated industries as collections of identical firms. The Ricardian model had countries with different technologies. The Heckscher-Ohlin model had countries with different factor endowments. Even Paul Krugman’s New Trade Theory, which finally introduced firms with market power, still assumed those firms were identical to one another within a sector. When economists looked at actual trade data in the 1990s, this assumption looked increasingly indefensible. Within narrowly defined industries, some firms exported, and most did not. Exporters were systematically larger, more productive, and paid higher wages than non-exporters in the same sector. No existing trade model could explain this.

The Melitz heterogeneous firms model, published by Marc Melitz in Econometrica in 2003, closed that gap. The model takes the Krugman framework of monopolistic competition and adds a single twist: firms within an industry differ in productivity, and trade selects which firms export, which survive in the domestic market, and which exit altogether. The result was a theory that finally matched the firm-level data and changed how economists think about the gains from trade. Within twenty years, it became the workhorse of modern international trade research.

The Importance of Firm Heterogeneity

Trade-flow statistics had long shown that exports were concentrated among a small number of firms. The 1990s and early 2000s brought a wave of new evidence from firm-level datasets in the United States, France, Colombia, and elsewhere, all pointing to the same patterns.

In manufacturing economies, only a minority of firms export, often fewer than 20 percent within most industries. The firms that do export are not randomly distributed: they are larger, more productive, more capital-intensive, and pay higher wages than firms in the same industry that sell only domestically. Exporters were already different from non-exporters before they began exporting, suggesting that exporting is an outcome of underlying firm capability rather than the cause of it. And when trade barriers fall, the largest gains in industry productivity come not from firms expanding output but from market share shifting away from less productive firms toward more productive ones.

None of this fits a model with identical firms. If all firms in an industry are the same, they should all make the same decision about exporting. If they all face the same costs and the same prices, they should all earn the same profits. The fact that they do not is the empirical fact that the Melitz model was built to explain.

Core Mechanics of the Model

The Melitz model uses three building blocks. Each is borrowed from earlier work; the contribution is the way they are combined.

Monopolistic competition with differentiated products. Each firm produces a single variety of a differentiated good, has some market power, and faces downward-sloping demand. This part is taken directly from Krugman.

Iceberg trade costs. Exporting is expensive. To deliver one unit of output to a foreign market, a firm must ship τ ≥ 1 units, with the difference “melting” in transit. This represents transportation, tariffs, and other variable costs of cross-border trade.

Fixed costs of entry, production, and exporting. To enter the industry, a firm pays a sunk cost and draws a productivity level from a probability distribution. To produce for the domestic market, it must pay a fixed cost per period. To export, it must pay an additional fixed cost per period, separate from the variable iceberg cost. The fixed export cost can be interpreted as the cost of setting up foreign distribution, navigating customs and regulations, learning the foreign market, or maintaining a sales presence abroad.

The fixed export cost is what does the heavy lifting in the model. Even with low variable trade costs, a firm that exports must clear a hurdle: its expected operating profits from foreign sales must be enough to cover the fixed export cost. Less productive firms cannot clear this hurdle. More productive firms can.

Productivity Cutoffs and Sorting

The model generates three sorting outcomes that occur simultaneously inside every industry. Each is driven by a productivity threshold.

The exit cutoff (φ*). Firms with productivity below this level cannot generate enough operating profits to cover even the domestic fixed costs. They exit the market immediately. In the data, this is the explanation for the constant churn of firm entry and exit, even in industries that look stable in aggregate.

The export cutoff (φ*_x). Firms with productivity above this threshold export. Their operating profits from foreign sales cover the fixed export cost. Crucially, the export cutoff lies above the exit cutoff, because exporting requires not just survival but additional revenue sufficient to cover the additional fixed cost.

The domestic-only zone. Firms with productivity between φ* and φ*_x survive in the domestic market but do not export. These are the majority of firms in most industries.

The productivity cutoff diagram captures all three regions in a single picture. Productivity is measured on the horizontal axis; expected per-period profits from each activity on the vertical axis.

Reading the diagram from left to right traces a firm’s possible fates by productivity. A firm with very low productivity (left of φ*) earns negative expected profits in the domestic market and cannot pay the fixed cost. It exists. A firm with productivity between φ* and φ*_x earns positive domestic profits but cannot cover the additional fixed export cost. It is produced for the home market only. A firm with productivity above φ*_x earns enough additional revenue abroad to cover the export fixed cost. It exports.

Key Parameters of the Model

The full Melitz model is more elaborate than the diagram suggests, but a small number of parameters do most of the analytical work.

Table 1. Key Parameters of the Melitz Heterogeneous Firms Model. Source: Melitz (2003), Econometrica.

Symbol	Parameter	Role in the model
φ	Firm productivity	Drawn from a probability distribution upon entry. Determines unit cost and all subsequent firm outcomes.
f	Domestic fixed cost	Per-period cost of producing for the home market. Sets the exit cutoff.
f_x	Export fixed cost	Per-period cost of exporting. Sets the export cutoff above the exit cutoff.
f_e	Sunk entry cost	One-time cost paid before drawing the productivity level. Balanced against expected discounted profits.
τ	Iceberg trade cost	Variable cost of shipping output abroad. τ > 1 means part of the shipment “melts” in transit.
σ	Elasticity of substitution	How easily consumers substitute between varieties. Determines the markup over marginal cost.
g(φ)	Productivity distribution	The probability distribution from which firms draw productivity. Usually assumed Pareto in applications.

Three of these parameters drive most of the comparative statics. Lowering τ (variable trade costs) increases the foreign revenue earned by any given firm and lowers the export cutoff, expanding the share of firms that export. Lowering f_x (the fixed export cost) does the same. Raising σ (elasticity of substitution) reduces firm markups and tightens both cutoffs, pushing more firms out of the market and reducing the share that exports.

The choice of productivity distribution matters more than it might appear. Empirical work consistently finds that firm productivity follows a Pareto distribution: a small number of highly productive firms and a long tail of less productive ones. Under Pareto productivity, the Melitz model produces analytical predictions that match the data closely, including the concentration of exports among the largest firms and the way export participation responds to trade liberalization.

The Within‑Industry Reallocation Effect

The most important prediction of the Melitz model concerns what happens when trade costs fall. In classical models, the gains from trade come from countries specializing in goods where they have a comparative advantage. In Krugman’s New Trade Theory, the gains come from access to more varieties of differentiated goods. In the Melitz model, a third source of gains appears: within-industry reallocation.

When trade costs fall, three things happen simultaneously inside every industry. Less productive firms see their profits squeezed by foreign competition and exit. Already-exporting firms see their foreign sales grow. Some firms previously just below the export cutoff cross it and begin exporting. Industry output reallocates from less productive to more productive firms. Aggregate industry productivity rises mechanically, not because any individual firm becomes more productive, but because output shifts toward firms that were already more productive.

This is the reallocation effect, and it has been central to interpreting the empirical literature on trade liberalization. The big productivity gains observed after Canada’s free trade agreement with the United States, Mexico’s NAFTA accession, and various European integration episodes all show the same pattern: within-industry productivity rises sharply, driven mostly by market share shifting between firms rather than by individual firms becoming more productive. The Melitz model gave economists a framework to predict, measure, and interpret this effect.

Empirical Validation of the Model

The Melitz model is unusual in trade theory for how directly its predictions match firm-level evidence. Four patterns in the data line up with the model’s predictions in a way earlier theories could not match.

Only a minority of firms export. In most industries, exporters are a small fraction of total firms. The model predicts this exactly: only firms above the export cutoff export, and the productivity distribution is skewed enough that this is a minority.

Exporters are larger and more productive. Because the export cutoff lies above the exit cutoff, exporters by definition have higher productivity than non-exporters. The model also predicts that exporters should have larger domestic sales than non-exporters, because higher productivity means lower prices and higher quantities sold even at home.

Trade liberalization triggers exits. When trade costs fall, the model predicts that the least productive firms exit, the productivity cutoff rises, and industry productivity increases through composition. Empirical studies of trade liberalization episodes consistently find this pattern.

Most exporters serve only a few markets. Subsequent extensions of the Melitz framework, particularly the work of Eaton, Kortum, and Kramarz on French exporters, document that even among exporters, productivity sorting continues: the most productive firms export to many countries, while less productive exporters serve only one or two. The cutoff logic extends naturally to multiple destinations.

Extensions and Theoretical Influence

The Melitz model became the foundation for an entire wave of research that came to be called the “new new trade theory.” Several extensions deserve mention.

Multi-product firms. Research by Bernard, Redding, and Schott extended the framework to firms that produce multiple products. The same cutoff logic applies within firms: trade liberalization induces firms to drop their less productive product lines and concentrate on their best products.

Quality differentiation. Extensions by Verhoogen, Khandelwal, and others added quality choice alongside productivity. Higher-productivity firms tend to produce higher-quality goods and charge higher prices, helping to explain why richer countries’ exports command higher unit values.

Trade in services. The same logic applies to services, with even sharper sorting effects given the larger fixed costs of cross-border service delivery.

Gravity model microfoundation. The Melitz framework, particularly when combined with the Pareto productivity distribution, generates the empirical gravity equation directly from firm-level decisions. This connected micro firm behavior to the most successful empirical workhorse in trade.

Trade and inequality. The Melitz framework predicts that opening to trade widens the wage gap between workers at exporting firms and workers at non-exporting firms, a prediction confirmed in firm-worker matched datasets. This connects to the broader literature on trade and wages, including Stolper-Samuelson and the China shock literature.

Model Limitations

The model is not a complete theory of trade. Three limitations are worth flagging.

It assumes monopolistic competition and a fixed elasticity of substitution. Industries with strong oligopoly structures, where strategic interactions between firms matter, are not well captured.

It treats firm productivity as a one-time draw from a distribution. The model has been extended to allow firms to invest in productivity, but the basic version does not endogenize where productivity comes from.

It explains within-industry reallocation but tells us little about which industries a country specializes in. The classical comparative advantage logic of Ricardo and Heckscher-Ohlin still applies to between-industry questions. Modern trade theory uses the Melitz framework alongside, not instead of, the older models.

The Model’s Contribution to Trade Theory

The Melitz model changed trade theory in three ways. It made firm heterogeneity central rather than peripheral. It identified a new source of welfare gains from trade, distinct from comparative advantage and variety expansion. And it gave empirical trade economists a tractable framework that could be matched directly to firm-level data, which had not been possible before.

The Krugman framework explained why countries trade similar goods. The Melitz framework explained why, within those flows, only a minority of firms actually do the exporting and why opening to trade has such large effects on industry composition. Together, they form the core of modern trade theory and shape how economists interpret nearly every trade liberalization episode of the last twenty years.

Explains

Conclusion

The Melitz heterogeneous firms model reframes international trade as a process that operates inside industries, not just between them. By introducing productivity differences across firms within a single sector and adding a fixed cost of exporting, the model produces sharp predictions: only the most productive firms export, less productive firms exit when trade costs fall, and aggregate industry productivity rises through within-industry reallocation. Each of these predictions has been confirmed across firm-level datasets in many countries, which is why the model became the dominant framework in trade research within a decade of its publication.

The deeper contribution is conceptual. Classical trade theory treated countries as the unit of analysis and asked which goods each country should specialize in. The Melitz framework treats firms as the unit of analysis and asks which firms within each country can survive and export under different trade conditions. Both questions matter. Modern trade economics uses them in parallel, with the Melitz model providing the firm-level microstructure that connects trade policy to the actual decisions made by the producers it affects.

Frequently Asked Questions

Thanks for reading! If you found this helpful, share it with friends and spread the knowledge. Happy learning with MASEconomics