- Empirical Studies

The empirical literature investgatng trade efects of volatlity in the bilateral exchange rate or currency unions is mostly based on country-level data. Some studies use disaggregated data to control for diferences between industries or to emphasize the higher responsiveness of single sectors. There are no unambiguous results of empirical macro studies investgatng the impact of exchange rate volatlity, but most fnd weak and mostly signifcant negatve efect. Studies difer in the sample of countries, the covered tme period, the degree of disaggregaton of the trade data and empirical methodology.²¹ In a very comprehensive meta-regression on empirical macro-studies of the topic, Ćoric & Pugh (2010) fnd exchange rate volatlity and trade to have a modestly negatve relatonship with pronounced heterogeneity and with litle evidence of publicaton bias, but mainly positve evidence that this relatonship is an authentc empirical efect. They fnd uncertainty arising from exchange rate volatlity to be a serious concern for least developed countries, what points towards the importance of hedging instruments and thus the stage of development of fnancial markets in least developed countries.

The introducton of a common currency completely eliminates nominal exchange rate volatlity between the members. Positve trade efects of currency unions can go beyond the eliminaton of volatlity in the exchange rate as it also facilitates currency handling and lowers uncertainty in the

20 This Informaton has been obtained from surveys on exchange rate handling practce of frms for Hungary and the Czech Republic described in secton III.I.III.

21 See survey papers on the relatonship between exchange rate volatlity and trade from Côté (1994), McKenzie (1999), Ozturk (2006), Bahmani-Oskooee & Hegerty (2007) and Auboin & Ruta (2011).

long term due to the binding character of a currency union. Studies investgatng trade efects of currency unions usually fnd robust positve efects. While early studies found extremely high results of an increase in trade up to 200 percent (Rose 2000), estmates have decreased to a positve efect between 5 and 30 percent, but remain robust. Most of the later studies focus on trade efects for the Eurozone.²²

Only a few studies take advantage of data at the frm-level and usually focus on frms in a single country. Empirical studies on trade efects of currency unions and exchange rate volatlity based on frm-level data are shown in Table III.I.1. Campa (2004) investgates the case of 2188 Spanish manufacturing frms for the years 1990-1997 and their responsiveness in export behaviour to exchange rate changes by estmatng a dynamic discrete-choice model. He does not fnd a signifcant efect of exchange rate volatlity on foreign market entry and exit, instead he fnds evidence for sunk costs hysteresis to play an important role and that a depreciaton of the domestc currency increases export volumes slightly via the extensive margin.

In a similar way, Guillou (2008) employs data on french manufacturing frms for the years 1994-2004 and distnguishes between efects on the probability to export and export intensity. On the one hand, she fnds that for most industries a depreciaton of the domestc currency afects the probability to export, while there is fairly no impact on the export intensity when introducing the lagged value of the dependent variable. Exchange rate volatlity yields positve results on the probability to export for most industries.

Solakoglu et al. (2008) estmate the efect of exchange rate volatlity on real exports for the years 2001-2003 using a sample of 143 large Turkish frms. According to their results, exchange rate volatlity does not afect trade and frm size and the level of internatonal actvity do not infuence a frms' responsiveness to volatlity. They do fnd evidence for natural hedging via imported intermediaries.

In a more recent study, Héricourt & Poncet (2012) investgate the efect of real efectve exchange rate volatlity on export performance and the role of fnancial constraints. They employ export data for more than 100,000 Chinese exporters over the period 2000-2006 and fnd a negatve efect on extensive and intensive margins of trade.

22 A good overview of the literature is delivered by Baldwin (2006).

36 III - Exchange Rate Policy and Exports of Firms

Several studies analyse the efect of membership in a currency union (usually the Euro) on export behaviour of domestc frms. In general, studies using frm-level panel data fnd that Eurozone membership signifcantly lowers the exchange rate risk to which domestc frms are exposed to (e.g. Bartram & Karolyi 2006). Vicarelli & Pappalardo (2012) employ diference-in-diference estmaton techniques for a panel of around 21,000 Italian frms for the years 1996-2004 and fnd that the euro has had a positve infuence on Italian exports, mostly through the intensive margin.

Etzel et al. (2013) try to explain Germany’s export success story using data at the plant-level for the years 1996-2008. They fnd that higher plant-level compettveness due to higher productvity or lower wages is positvely correlated with export intensity at the intensive and extensive margins.

This lets them conclude that the introducton of the Euro has led to higher export demand for the relatvely cheaper German products as separate regressions for the pre- and post-Euro periods reveal that the export promotng efect of compettveness is strongest shortly afer the Euro was introduced.

In the only other empirical study that includes both variables, measures for exchange rate volatlity and currency unions, Berthou & Fontagne (2008) investgate export behaviour of French frms for the years 1998-2003. Estmatng a model that consists of both variables allows to disentangle the trade efect of eliminated exchange rate volatlity from the efect stemming from a common currency. They fnd trade deterring efects via the intensive and extensive margin of trade for exchange rate volatlity and trade boostng efects for Euro membership only via the extensive margin.

Table III.I.1: Studies Investgatng Trade Efects of Exchange Rate Volatlity or Currency Unions at the Firm-level

Sample Efect on Margin Methodology

Authors # of Firms Time Country extensive intensive

2,188 1990-1997 Spain no efect - Dynamic discrete choice Guillou (2008) 47,716 1994-2004 France positve** - RE Probit

143 2001-2003 Turkey - no efect OLS, FE, RE

? 1998-2003 France - negatve FE, RE

113,368 2000-2006 China - negatve FE

Euro ? 1998-2003 France positve no efect FE, RE

21,000 1996-2004 Italy no efect positve Diference in Diference 16,000* 1996-2008 Germany positve positve FRACP, FRACPL, Tobit, Probit Notes: *observatons at the plant level; **probability to export.

Exchange Rate Volatlity

Other studies focus on the impact of exchange rate volatlity on general economic performance of frms (Carranza et al. 2003), frm value (Hutson & O’Driscoll 2010), plant-level investment (Kandilov

& Leblebicioglu 2011) or are only descriptve (Sanderson 2009).

Recent fndings suggest that the impact of exchange rates on trade fows in general may be lower than expected due to the type of frms that export. Berman et al. (2012) fnd that exportng frms are generally bigger and more productve than non-exporters and in a situaton of depreciaton of the domestc currency tend to increase their margins rather than the volume of sales. For less productve frms, the reverse is true: they increase the volume of exports rather than prices.

Although there is a positve impact of an exchange rate depreciaton on the number of exportng frms or extensive margin, the efect is less evident on a macro level due to the smaller size of the frms startng to export.