VoxEU Column Exchange Rates International trade

Exchange rate devaluations: When they can work and why

The effect of exchange rate devaluations on a country’s exports is frequently debated, with some supporting a strong impact and others suggesting no sizeable response of exports to relative price changes. This column argues that firm heterogeneity in terms of productivity and size can explain the opposing views. The measured reaction of aggregate exports to relative price movements is largely determined by the reaction of the most productive/largest companies.

How and by what extent exchange rate devaluations can affect country aggregated exports is the source of a permanent debate among policymakers and economists. In particular, while trade economists tend to support a strong impact (with price elasticities often estimated to be above 5), macroeconomists tend to assign a low value to these elasticities, often below 1 (Fitzgerald and Haller 2014, Bas et al. 2015). What is the source of the gap between the microeconomic and macroeconomic responses of exports to relative price changes?

Using firm-level evidence on 17 EU countries from the CompNet database (Lopez-Garcia, di Mauro et al. 2015), in this column we argue that firm heterogeneity in terms of size or productivity can help explain the gap. Merging the gap, however, is not a mere econometric exercise but has critical policy implications. In particular, the overall aggregate response of country exports to exchange rate movements will strongly depend on the underlying microstructure of the respective economies. Ceteris paribus, countries characterised by a larger share of low productive and small exporters are likely to record a larger positive impact from depreciation. Though suboptimal, this is the economic structure still prevailing in several EU countries under stress. For once, this could be good news as this could mean higher aggregate response of their exports to an exchange rate depreciation.

Macro versus sector-level estimates of trade elasticities

The response of trade to relative price changes is a key parameter in analysing the adjustment of external imbalances through real exchange rate adjustments. Traditional macroeconomic estimates of trade elasticities based on time series analysis are often found to be low and not statistically significant (e.g. Hooper et al. 1998). This finding has been challenged by recent works using sector-level trade data (Feenstram1994, Imbs and Mejean 2015), which emphasise the existence of an aggregation bias in the presence of sector heterogeneity, and provide much larger estimates. For instance, Imbs and Mejean (2015) find an elasticity of around 5 in the unconstrained case when elasticities are estimated at sector-level. Within CompNet, Corbo and Osbat (2013) find that the aggregate elasticity of substitution for European countries’ imports ranges from 2.6 to 4.8 (with a median value of 3.4), while values for export elasticities are slightly larger, ranging from 3.0 to 6.5 (with a median value of 3.8). Benkovskis and Wörz (2015) find import elasticities of substitution to be between 2.4 and 3.4 for the major world economies (the US, China, Germany, and Japan).

Firm productivity and size dispersion within sectors

The role played by heterogeneity across sectors, however, potentially undermines a second source of heterogeneity due to the strong variance in terms of firms’ size and productivity within each sector. Recent empirical research using micro data at the firm level has confirmed that structural characteristics of the population of firms operating in a country or sector can modify the response of aggregate exports to exchange rate movement (Berman et al. 2012).

The micro-based dataset recently created by CompNet using firm-level information for 17 European countries (Lopez-Garcia, di Mauro and the CompNet Taskforce, 2015) shows that firm productivity is highly heterogeneous  across the 17 countries and the some 60 sectors included.

  • This implies that, since there is no ‘representative’ firm in each country and sector, the aggregate response of exports to shocks (such as a depreciation) varies in accordance with the underlying heterogenous response of the individual firms.
  • A second important finding from the CompNet research is that aggregate exports are concentrated among top productive firms in each country and sector (Figure 1).

This implies that the measured reaction of aggregate exports consecutive to relative price movements is going to be largely determined by the reaction of the most productive/largest companies.

Figure 1. Share of export by labour productivity deciles (2006-2012)

Source: Berthou et al. (2015)

Firm heterogeneity modifies trade elasticities

Two CompNet projects have provided new micro-based estimations of the exchange rate elasticity of exports at the country-sector and firm level. First, di Mauro and Demian (2015) investigate the response of exports to exchange rate fluctuations using CompNet sector-level information for a set of EU countries and look at how this response is affected by the shape of the productivity distribution. Controlling for the standard deviation of firms’ productivity in each country and sector equation significantly increases (doubles) the real exchange rate elasticity estimated in a gravity-equation framework (Figure 2). In addition, the results indicate lower elasticity in sectors with high levels of productivity dispersion, which is typical of highly productive sectors where a few top players’ production can be distinguished from the rest of the population.

Figure 2. Export elasticity with respect to real exchange rate and shape of productivity distribution

Source: Demian, di Mauro (2015).

The first bar reports the exchange rate elasticity. A negative sign implies that a depreciation of the exchange rate increases country-sector exports. The second bar reports the exchange rate elasticity, when the interaction between exchange rates movements and the dispersion of Total Factor Productivity (TFP) in each country and sector is controlled for. With the third bar, the positive sign implies a weaker elasticity when productivity dispersion is larger.

Second, in a related study using CompNet data, Berthou et al. (2015) investigate the underlying factors driving the heterogeneous response of European exporters to exchange rate fluctuations (see Berthou et al. 2015 for a detailed description of the so called ‘trade module’ of CompNet ). The data allow to distinguish between exporting firms within sectors on the basis of their level of productivity or size. In the empirical specification, the export performance of different categories of firms (differentiated by their productivity or size) is explained by changes in the real effective exchange rate (REER), while allowing the elasticity to vary across groups. After controlling for other macroeconomic determinants (foreign demand) and sector or firm-level characteristics, estimation results show substantial heterogeneity across the different categories, with large (and more productive) firms reacting much less than the average firm to real exchange rate variations (Figure 3). While the reaction of export values by large and productive firms is close to what is usually observed in macroeconomic estimations, low productive firms are by contrast much more responsive to relative price movements.

Figure 3. Export elasticity with respect to ULC-REER by size and productivity quartile

Source: Berthou et al. (2015).

Note: *** p<0.01, ** p<0.05, * p<0.10 Real effective exchange rates based on Unit Labor Costs. Each bar reports the estimated  elasticity of firm-level exports with respect to real exchange rate variations. A negative sign implies that a depreciation of the exchange rate increases firm-level exports. Size and Total Factor Productivity (TFP) quartiles computed by country and sector: 1 (small/low) to 4 (large/high).

Aggregate implications and policy conclusions

This heterogeneity in the responses of exporters facing the same exchange rate shock has a substantial influence on the aggregate outcomes at sector and country level, which helps to understand why the trade elasticities computed at the aggregate level are fairly low. In fact, as most trade flows are concentrated in the largest and most productive firms, the low trade elasticities found at the macro level may be explained by the low elasticities estimated on the far right of the size and productivity distributions.

This result has important implications for the impact of relative price adjustments on the aggregate export performance of European countries.

  • First, the reaction of aggregate exports to an exchange rate depreciation is likely to be – overall – quite limited in the short run as it is largely determined by the (low) reaction of the largest and most productive firms, which make the largest share of EU exports.
  • Second, the greater elasticity estimated for small and least productive exporters suggests that an exchange rate decline can be very effective in helping those more vulnerable exporters to succeed in international markets. As the contribution of these firms (often new exporters) to aggregate exports increases rapidly over time (Eaton et al. 2007, Berthou and Vicard 2014), the benefits from a depreciation should be larger in the medium term.
  • Third, while having a large share of small and least productive firms is by many respects suboptimal, it could provide a bonus to EU stressed countries when it comes to impacts of depreciations.


Bas, M, T Mayer, and M Thoenig (2015). "From Micro to Macro: Demand, Supply, and Heterogeneity in the Trade Elasticity," CEPR Discussion Papers 10637, London.

Benkovskis, K and J  Wörz (2015), “‘Made in China’. How does it affect our understanding of global market shares?”, Working Paper Series, No 1787, ECB, April.

Berman, N, P Martin and T Mayer (2012), “How do Different Exporters React to Exchange Rate Changes?”, The Quarterly Journal of Economics, Vol. 127(1), pp. 437-492.

Berthou, A, C-V Demian, and E Dhyne, E. (2015), “Exchange rate movements, firm-level exports and heterogeneity”, Working Paper Series, ECB, forthcoming.

Berthou, A, E Dhyne, M Bugamelli, A-M Cazacu, C-V Demian, P Harasztosi, T Lalinsky, J Merikull, F Oropallo, and A C Soares (2015), “Assessing European Firms’ Exports and Productivity Distributions: the CompNet Trade Module”, Working Paper Series, No 1788, ECB.

Berthou, A and V Vicard (2015), “Firms’ Export Dynamics: Experience Versus Size”, The World Economy 38(7), 1130-1158.

Corbo, V and C Osbat (2013), “Trade Adjustment in the European Union – A Structural Estimation Approach”, Working Paper Series, No 1535, ECB.

Eaton, J, M Eslava, M Kugler, and J Tybout (2007), “Export dynamics in colombia: Firm-level evidence”, Banco de la Republica WP003957.

Feenstra, R C (1994), "New Product Varieties and the Measurement of International Prices," American Economic Review 84(1), 157-77.

Fitzgerald, D and S Haller (2014), "Exporters and Shocks: Dissecting the International Elasticity Puzzle," NBER Working Papers 19968.

Hooper, P, K H Johnson, and J R Marquez (1998), "Trade elasticities for G-7 countries," International Finance Discussion Papers 609, Board of Governors of the Federal Reserve System (U.S.).

Imbs, J and I Mejean (2015), "Elasticity Optimism," American Economic Journal: Macroeconomics 7(3), 43-83.

Lopez-Garcia, P and F Di Mauro (2015), "Assessing European competitiveness: the new CompNet microbased database," Working Paper Series 1764, European Central Bank.

4,724 Reads