Exchange rate moves have surprisingly small effects on prices. This apparent ‘disconnect’ is one of the central puzzles in international macroeconomics. It is also a continual headache for policymakers who rely on exchange rates to accommodate the adjustment of global (current account) imbalances. The main mechanism is a change in relative prices that shifts expenditure towards countries whose currency has depreciated. If prices don’t respond sufficiently to exchange rates then neither do quantities, and the expenditure-switching role of exchange rates is diminished (see Engel 2003).
In a new study (Amiti et al. 2012), we show that in order to understand this exchange rate disconnect we have to take into account the fact that the largest exporters are also the largest importers. This is important because when exporters are hit by an exchange rate shock in their destination market, they typically face a compensating movement in the marginal costs if they are importing their intermediate inputs.
For example, an appreciation of the euro relative to the US dollar, while increasing the domestic costs of European firms (in US dollars), typically reduces their euro costs of international sourcing of intermediate inputs. This effect is most intuitive when inputs are sourced from the US but it turns out to be, in effect, more general. Why? Because empirically, the movements in the value of a country's currency are correlated across its trade partners. This natural hedging from exchange rate movements, inherent in the imports of intermediate inputs, reduces the need for exporters to adjust their export market prices.
Using Belgian firm-level data, we show that exporters that import a large share of their inputs pass on a much smaller share of the exchange rate shock to export prices. Interestingly, we find that import-intensive firms typically have high export market shares, and hence set high mark-ups and actively move them to offset the effects of changes in their marginal cost on export prices, thus providing a second channel that limits the pass-through of exchange rate shocks. In our data, a typical small exporter with no imported inputs has a nearly complete pass-through, while a large import-intensive exporter has a pass-through just above 50%, at an annual horizon1. We show that this large cross-sectional variation in pass-through is accounted for roughly equally by the two mechanisms: the offsetting movements in both the mark-up and the marginal cost by means of imported inputs. These two mechanisms reinforce each other and help insulate the international prices of major exporters from the economic conditions in their domestic market, hence limiting the expenditure-switching role of exchange rates.
A new understanding of exporters and importers
An interesting new stylised fact to emerge from our analysis is that there are large systematic differences across exporters. Previous studies have emphasised the differences between exporters and non-exporters, highlighting that exporters are bigger, more productive, and pay a wage premium, features that are also present in our data. However, even within the select group of firms that are exporters there are very large differences between high import-intensity exporters and low import-intensity exporters. Splitting the sample based on the median import intensity, we show that import-intensive exporters are 2.5 times larger in terms of employment than exporters with low import intensity, and an order of magnitude larger than non-exporters. Similar rankings are present for measured productivity, material costs, and wages, as well as extensive and intensive margins of imports and exports. These patterns are illustrated in Table 1.
Table 1 Exporting firms with high and low import intensity
|High import intensity||Low import intensity|
|Share of total imports in total cost||36.8%||17.3%||1.6%|
|Share of non-euro imports in total cost||16.6%||1.2%||0.3%|
|Employment (# full-time equiv. workers)||270.9||112.1||20.7|
|Material cost (millions of euros)||103.5||28.1||3.0|
|Average wage bill per worker (thousands of euros)||48.8||42.3||34.9|
|Export value (millions of euros)||49.6||9.4||—|
|Import value (millions of euros)||49.3||6.9||—|
|— outside Eurozone||20.8||0.5||—|
We show that these patterns are central in explaining the heterogeneity of exchange rate pass-through into export prices across firms. Furthermore, because import-intensive firms are also the largest exporters, these results help to explain the low aggregate pass-through.
Three major results
- First, our study shows that firm market share and import intensity are the two key determinants of pass-through in the cross-section of firms within a given industry and export market, explaining a wide range of variation in exchange rate pass-through.
This is consistent with the theory which, as we show, predicts that market share and import intensity form a sufficient statistic for the firm-level pass-through. In the context of a small country within the Eurozone, the relevant concept of import intensity is imports that originate outside the Eurozone, and hence we focus on this measure in our empirical analysis. Indeed, we confirm that imports from within the Eurozone have little effect on pass-through.
Table 2 summarises our findings by splitting the firms into four ‘bins’ based on whether the firm’s import intensity is above or below the median and whether it has a market share in the export market that is higher than average. Within each bin, we estimate a simple pass-through regression of the change in producer export prices (in destination currency) on the change in the bilateral exchange rate (with an increase indicating a euro depreciation) and report the pass-through rate into export prices. We find that firms with below-median import intensity and below-median market share have the highest pass-through rate, at 89%. The pass-through decreases both moving towards high import-intensity firms and high market-share firms, and the lowest pass-through rate, 61%, is found for the firms with both market share and import intensity above the respective medians. These patterns are robust to alternative cuts of the data and additional controls2.
Table 2 Pass-through by import-intensity and market-share bins
|Low import intensity||High import intensity|
|Low market share|
|Fraction of observations||30.3%||20.0%|
|Share in export value||8.8%||9.3%|
|High market share|
|Fraction of observations||19.9%||30.1%|
|Share in export value||21.2%||60.7%|
- Second, we decompose the incomplete pass-through into the marginal cost channel and mark-up channel, and find that the two channels contribute roughly equally to the variation in pass-through across firms.
This decomposition is important from a welfare perspective as the implications differ if incomplete pass-through is due to different distributions of mark-ups across firms or if it is due to the complex global sourcing patterns, which directly affect marginal costs. Half of this incomplete pass-through is due to the offsetting effects of an exchange rate change on marginal cost (which we proxy using the firm’s import intensity) and the other half is due the variation in the mark-ups (which, guided by the theory, we proxy using the firm’s market share). Firms with large market shares adjust their mark-ups more than small firms do in response to cost shocks.
- Finally, these results have implications for aggregate pass-through;.
Because of the positive correlation between import intensity and market share, most of the observations lie along the diagonal in Table 2 (30% in each diagonal bin and only around 20% in the off-diagonal bins). Interestingly, while the ‘high import intensity-high market share’ bin contains 30% of the observations, those firms account for more than 60% of the value of exports. Given that these are the lowest pass-through firms, this suggests low aggregate pass-through. Indeed, a large share of exports is that of the bigger firms which source their inputs globally and are thus only partially linked to the domestic market conditions in their home country. As a result, these firms are effectively hedged against exchange rate fluctuations and do not need to fully adjust their prices. Furthermore, these are the strong market-power firms, setting high mark-ups and actively varying them to accommodate for cost shocks.
The prices of the largest firms, accounting for their disproportionate share of trade, are insulated from exchange rate movements both through the hedging effect of imported inputs and through active offsetting mark-up adjustment in response to cost shocks. Both forces limit the expenditure-switching effect of a given exchange rate movement, but have very different implications for the allocative efficiency of global production.
These results imply that the international competitiveness effects of a euro devaluation are likely to turn out to be modest given the extensive international sourcing by major exporters. In addition, a weaker euro is likely to have limited effects on prices and quantities, with the changes largely reflected in the profit margins of major exporters.
Amiti, M, O Itskhoki, and J Konings (2012), "Importers, Exporter, and Exchange Rate Disconnect," CEPR Discussion Paper No. 9260.
Berman, N, P Martin, and T Mayer (2012), "How Do Different Exporters React to Exchange Rate Changes?", Quarterly Journal of Economics 127(1), 437-492.
Engel, C (2003), "Expenditure Switching and Exchange Rate Policy," NBER Macroeconomics Annual 2002, 231-272.
1 A 50% pass-through implies that, on average, a firm raises its US dollar price by 5% when the Euro appreciates by 10% against the US dollar, that is less than proportionally. A complete pass-through is a 100% pass-through.
2 Earlier research (Berman et al., 2012) has found that more productive firms have lower exchange rate pass-through. This is consistent with our findings since both import intensity and market share are positively correlated with productivity. However, we show that import intensity and market share remain the prime determinants of pass-through even when we simultaneously control for firm productivity and employment size.