Supply effects in Danish exports?

01-07-2021

In macroeconomic models, exports are typically described by demand-based Armington equations. Recent literature, however, has argued that supply effects should be added. In this working paper, we examine the extent to which such supply effects can be found in data for Danish exports.

Abstract

In macroeconomic models, exports are typically described by demand-based Armington equations. Recent literature, however, has argued that supply effects should be added. In this working paper, we examine the extent to which such supply effects can be found in data for Danish exports.

Supply effects are expressed via an increase in the Danish production capacity which enables the scaling of total exports. Typically, one will think of supply effects as some which materializes on an extensive margin of foreign trade: Exporting new product types and / or to new markets (countries). The intensive margin, on the other hand, expresses changes in exports of existing product types and to existing countries, i.e. movements one would typically associate with demand effects. We use detailed trade data for the period 1995-2016 to first analyze the relative degree of explanation of the intensive and extensive margin in growth rates in total exports over different horizons. We find that growth in exports is primarily driven by the intensive margin in the short run, but that, in the long run, variations in the growth rate in exports is to a larger extent driven by the extensive margin. The full effect has occurred after five year, where the extensive margin explains 70% of the variation in the growth rate of exports.

We subsequently estimate the supply effect for Danish exports, i.e. the percentage change in export for a percentage change in e.g. private GVA, based on a set of Gravity equations. We use both the actual variables, their structural levels and instruments to remove business cycles and address obvious endogeneity problems. Our preferred estimate on the supply effect for private GVA indicates that the elasticity is not significantly different from 1 and that an error correction term of approximately 30% for supply effects seems to be consistent with data. Finally, we find that the extensive margin is highly significantly correlated with the structural levels in the supply variables, which we see as evidence that the supply variables create a scale effect via the extensive margin. We conclude that supply effects are significant in the description of Danish exports, especially in the long run, and should be included in the specification of exports in MAKRO.