In the ‘currency wars’ discussion – including, for instance, Blanchard (2016) – it is almost taken for granted that exchange rate depreciations will result in non-trivial export gains. As we all know, this is not so straightforward and this column’s contribution is to show that non-price/non-exchange rate factors play a critical and often predominant role in shaping trade developments. Moreover, the size of the non-price component is proven to be highly heterogeneous across the developed and emerging economies that we have examined. This suggests more attention should be paid to the empirical evidence before engaging in discussion – or actually fighting wars – over currencies.
The analysis is conducted here using methodologies of decomposition of the export market shares developed in a series of papers by Benkovskis and Wörz (2014, 2015), within the Competitiveness Research Network (CompNet). There are two novelties we bring in this column with respect to that work: (i) calculations are updated to 2014, and (ii) selected European and Asian countries are used as a case in point.
The puzzling starting point
When thinking about the exchange rate as a driver of export results, the first obvious thing to do is to check whether increasing price competitiveness following depreciation is correlated with higher market shares, and vice versa after an appreciation. Figure 1 shows the well-known, but often forgotten fact that for most countries this is not the case. Cumulated market shares are hardly related to changes in price/exchange rate competitiveness (here proxied by relative export prices, or RXP).1 If price competitiveness were a reliable covariate of market share, there should be a clustering of points in the NW and SE quadrants, i.e. higher (lower) market shares (on the y-axis) are associated with worsened (improved) price competitiveness. This, however, is not the case as points are scattered around, indicating that a host of factors are at play. Also, this result consistently holds for large sets of countries regardless of the periods under observation (see di Mauro and Foster 2008 for an analysis pre-Crisis), as well as in a variety of set-ups (e.g. Berthou et al. 2015, and Berthou and di Mauro 2015 using firm level data).
Figure 1. Market shares and price competitiveness, 2000-2014 (annual average changes)
Note: positive values for RXP mean worsening price competitiveness
The decomposition methodology
Since it is evident that other, non-price competitiveness factors must be considered, the empirical strategy we adopt consists in decomposing the actual changes in export market shares into one component that can be explained by relative export prices (RXP) – a conventional price-competitiveness indicator – and another that cannot. This residual by construction must contain additional factors such as ‘consumer taste’ and quality of exports, as well others, such as market penetration strategy (which affects the so-called ‘extensive margin’ of trade). Unfortunately, the residual is unobservable. To be able to measure it, we adopt a methodology suggested by Benkovskis and Wörz (2014), which consists in developing a model where consumers try to optimise their expenditure (including on foreign goods), but not only in price terms. These consumers, in particular, are modelled as thinking about ‘dollars per unit of utility’ instead of ‘dollars per kg’; i.e. they take fully into account non-price considerations when selecting their purchases.
Ultimately, actual developments in export market shares are split in five parts:
- price factor;
- changes in the set of competitors;
- non-price factor;
- global demand shifts; and
- the extensive margin.
In order to actually compute this decomposition, use is made of trade data (COMTRADE) that are highly disaggregated at the product level (about 5,000 categories) and include bilateral transactions for all the above sectors across more than 100 reporting countries. The resulting huge dataset guarantees a very fine account of the complex interaction of products and destination markets. For instance, the contribution of price factors takes into account the differences in elasticities of substitution across markets. As a result, the importance of a price change in a particular market is determined by its weight in the country’s exports and by the degree of substitutability among varieties and products. This makes the derived indexes superior to the traditional aggregated real effective exchange rate (REER) since, realistically, markets that are closer to perfect competition obtain more weight.
We have identified a number of countries in two continents – Europe and Asia – as providing rather opposite and informative results, which we summarise below.
Starting with Europe, all countries considered have experienced market share losses and, for all of them, the non-price component (blue bar) was critical.
Figure 2. Europe: Decomposition of changes in (gross) export market share
This can easily be interpreted as part of the erosion of the overall weight of industrialised countries in world trade as new emerging economies acceded the global markets. The erosion could be related to decreasing quality advantage of European producers or a shift of tastes in favour of products from emerging countries. Actually, price factors per se were favourable in relative terms for all of them, with the notable exception of Italy, which saw its share of exports cut the most.
The outcome is much more diversified for Asia. For highly developed Asian countries, the story is similar to the European one.
Figure 3. Developed Asia: Decomposition of changes in (gross) export market share
Market share declines for both Japan and Singapore mostly correlate – as in Europe – to non-price effects (blue bars). For Singapore, however, in the last few years rapidly worsening export price competitiveness (red bars) appears to have played a role as well in explaining losses in export shares.2
The story for developing Asia is totally different and in fact varies across countries, which is critical for the ‘currency war’ discussion.
We can identify two very distinct groups of countries.
- On one hand, there are countries that were steadily gaining in market shares (e.g. China, India, as well as Vietnam in the Annex). For all, declining relative export prices can explain only a small fraction of their market share gains, and actually not at all for Vietnam. There must have been a huge catching-up in quality of exports – or other critical structural factors, such as the opening-up of trade starting from very low export prices – underpinning their gains in share.
Figure 4 Developing Asia 1: Decomposition of changes in (gross) export market share
- On the other hand, there are countries whose shares were stagnating (e.g. Indonesia, Malaysia). Remarkably, however, there are opposing underlining factors among these countries. In the case of Indonesia, price components were actually favourable; quality mismatch or other structural factors must have hampered a better trade performance. For Malaysia, the opposite appears to be the case, since the stagnation (or slight decline) in export market shares appears to be related to a worsened relative export price competitiveness, including therefore adverse exchange rate developments.
Figure 5. Developing Asia 2: Decomposition of changes in (gross) export market share
Adding the role of global value chains
An obvious reason for the low explanatory power of price competitiveness is that a large part of trade involves intermediates products – i.e. inputs used within rather well established global value chains (GVCs) – and is thus far less influenced by pure exchange rate considerations. Against this background, we follow the recent approach by Benkovskis and Wörz (2015). We look at value-added in exports of final products, which is defined as the (direct and indirect) value-added of a specific country in total world exports of final products. This is done by using global value added matrices out of the WIOD database (Timmer et al. 2015).3 As a result, we can observe how increasing participation in GVCs affects the market shares of emerging countries. In addition, the structure of weights of the decomposition of the non-price component is replaced by that resulting from bilateral trade measured in value-added terms.
The results are not directly comparable to the one previously presented, but what is obvious is the extent to which – yet again – there is a large cross-country differentiation, here in the role of GVCs. While for big European countries and Japan the contribution is negative, signalling outsourcing of production to developing countries, emerging Asian countries increase their role in global production. But strong heterogeneity exists within emerging Asia. For China, for instance, the role of value chains in fostering export share changes – proxied by the component ‘shift in production chains’ – appears to be much smaller than for India, where lately it accounts for almost one third of the change in market shares. The message remains clear, however: price factors are not the only driving force behind export market share changes, and there is a need to get deeper into non-price components.
Figure 6. Developing Asia: Decomposition of changes in value-added export market share of final products
By disentangling the impact of exchange rate changes on trade results, we have shown that the underlying assumption of the ‘currency wars’ discussion – that devaluations bring about substantial export gains – may be severely flawed. Non-price/non-exchange rate factors often appear to explain the lion’s share of export outcomes, and this is particularly the case when exports are measured in value-added terms. We have also provided rather compelling evidence that the extent of this non-price/non-exchange rate component is highly diversified across a sample of developed and emerging economies in Europe and in Asia. This amounts to a strong call for caution in trusting too extensively in the ‘power’ of the exchange rate.
Authors’ note: The opinions expressed not necessarily reflect those of the ECB or the ESCB.
Benkovskis, K., and Wӧrz, J. (2014) “What drives the market shares changes? Price versus non-price factors”, ECB Working Paper No 1640.
Benkovskis, K., and Wӧrz, J. (2015), “'Made in China": How Does it Affect Our Understanding of Global Market Shares?”, ECB Working Paper No 1787.
Berthou, A., Demian, V., and Dhyne, E. (2015) “Exchange Rate Movements, Firm-Level Exports, and Heterogeneity”
Berthou, A., and di Mauro, F. (2015) “Exchange rate devaluations: When they can work and why”, VoxEu.org, 24 December.
Blanchard, O. (2016) “Currency wars, Coordination, and Capital Controls”, preliminary version prepared for the Asian Monetary Policy Forum, Singapore,.
di Mauro, F., and Foster, K. (2008) “Globalisation and the Competitiveness of the Euro area”, ECB Occasional paper No 97.
Feenstra, R.C. (1994) “New Product Varieties and the Measurement of International Prices”, American Economic Review 84(1): 157-177.
Feenstra, R.C., and Romalis, J. (2014) “International Prices and Endogenous Quality”, The Quarterly Journal of Economics, forthcoming, doi: 10.1093/qje/qju001.
Lopez-Garcia, P., di Mauro, F. and the CompNet Task Force (2015) “Assessing European competitiveness: the new CompNet micro-based database”, ECB Working Paper No 1764.
Timmer, M. (ed.), Erumban, A., Gouma, R., Los, B., Temurshoev, U., de Vries, G.J., Arto, I., Genty, V., Neuwahl, F., Rueda-Cantuche, J.M., Villanueva, A., Francois, J., Pindyuk, O., Pöschl, J., Stehrer, R., and Streicher, G. (2012). “The World Input-Output Database (WIOD): Contents, Sources and Methods”, WIOD Working Paper No 10.
Timmer, M.P., Dietzenbacher, E., Los, B., Stehrer, R., and de Vries, G.J. (2015) "An Illustrated User Guide to the World Input-Output Database: the Case of Global AutomotiveProduction", Review of International Economics 23(3): 575-605.
ANNEX: Decomposition of changes in (gross) export market share – additional Asian countries
 RXP is a better measure of price competitiveness than the real effective exchange rate as it considers the export prices (including the exchange rate) of any countries with respect to their world competitors.
 The negative contribution of price factors is in line with the CPI-based REER of Singapore, which appreciates since 2005. In 2014, however, our indicator diverges from REER and indicates a sharp real appreciation. This could be related to the increase of Singapore's export prices for some narrow product groups that are traded in highly competitive markets.
 Note that actual WIOD data are available only up to 2011. We assume that the structure of global value chains did not change between 2011 and 2014. Thus, we may underestimate the contribution of ‘shifts in production chains’ during 2012-2014.