The equilibrium real exchange rate for Israel

Transcription

1 The equilibrium real exchange rae for Israel Zvi Ecksein 1 and Ami Friedman 2, 3 1. Inroducion This paper analyses he equilibrium real exchange rae (ERER) for Israel. In a small open economy such as Israel, in which expor flows accoun for 40% of GDP, he real exchange rae (RER) has an imporan impac on growh and sabiliy. 4 RER misalignmens ha are due o medium-erm deviaions of he acual exchange rae from he ERER could cause oupu loss and cyclical, inefficien allocaion of resources, including low uilisaion of facors of producion. The fac ha a large share of expors is based on he high-ech indusry and high invesmen in human capial makes his concern an imporan policy facor. On he eve of he crisis in 2008 he Israeli currency (New Israeli Shekel, NIS) appreciaed by abou 20%. This raised concern ha he RER was overvalued due o shor-run capial flows. This, and he low level of foreign currency reserves, promped he BOI o inervene in he foreign exchange (FX) marke afer 10 years during which he exchange rae was freefloaing. The policy during he sress period of he crisis was ha of buying $100 million a day. Since Augus 2009, he Bank of Israel has adoped a policy of FX marke inervenion in case of sharp exchange rae movemens ha do no reflec fundamenal forces, and in cases of marke anomalies. Thus, he quesion of wheher he RER is aligned wih economic fundamenals has become cenral o acive policy. 5 Currenly, he low rae of growh in he developed economies and he ineres rae differenial beween Israel and hese economies, ha mainain hisorically low key raes, have creaed massive capial inflows ino Israel, and similarly ino oher growing economies. 6 These flows have resuled in episodes of RER overvaluaion; consequenly, some cenral banks have sared o defend heir exchange raes eiher by purchasing foreign currency or by resricing capial flows. The case of Israel illusraes he above: cumulaive per capia growh in Israel in was higher han in he main developed economies. The rapid recovery afer a mild recession and growing inflaionary pressures induced he BOI o raise he key ineres rae from 0.5% in April 2009 o 2% by he end of 2010, while he key raes in he US and he euro zone are a hisorical lows and are expeced o say here for mos of his year. The ineres rae differenial led o a subsanial increase in carry-rade capial inflows owards Israeli shor-run zero-coupon deb insrumens issued by he BOI (makam shor-erm borrowing), puing addiional pressure on he exchange rae. The BOI s policy response was o purchase Depuy Governor, Bank of Israel. Research Deparmen, Bank of Israel. We hank Roni Frisch for valuable advice and help hroughou he research process and Jonahan Sidi for research assisance. We also hank Sanley Fischer, Karni Flug, Avihay Sorezcky, Michel Srawczynski and Jacob Braude for discussion. Flug and Srawczynski (2007). Usually, he exchange rae analyses are used inernally and are no repored o he public. The excepions are he BOI Annual Repor 2004, which presened he ERER mehodology and repored ha he RER was aligned, and he Annual Repor 2008, which repored ha he RER was overvalued. IMF (2010). BIS Papers No

2 foreign currency in order o moderae he shekel appreciaion rend. Toal reserves increased from $28 billion in 2008 o $71 billion a he end of The jusificaions for he foreign currency purchases by he BOI are eiher an overvalued currency relaive is equilibrium level, or a non-gradual appreciaion owards equilibrium. Our analysis is divided ino wo mehods: (i) a direc mehod based on an economic model ha is derived by long-erm fundamenal economic indicaors; and (ii) an indirec mehodology based on he analysis of he curren accoun. According o his laer mehodology, he difference beween he acual and he fundamenal curren accoun indicaes RER misalignmen, which may be caused by ransiory capial flows or oher facors. The magniude of he misalignmen depends on he elasiciies of expors and impors wih respec o he RER. Our analysis includes wo measures of he RER: (i) he real effecive exchange rae (REER), ha is, he nominal effecive exchange rae (NEER) divided by he relaive price levels; and (ii) purchasing power pariy (PPP) indices, which include boh low-frequency panel daa and high-frequency ime-series daa. We follow he vas lieraure on his subjec and use several economic fundamenal variables: relaive per capia GDP, he erms of rade, he dependency raio, governmen size, foreign direc invesmen (FDI) flows, ineres rae differenials and ohers ha help o explain RER flucuaions and rends. We find he firs wo variables o be he mos imporan and robus. Our main findings are as follows. In 2006, a year during which he RER reached a hisorical low (ha is, when he shekel depreciaed he mos), he RER was undervalued. This conclusion holds for boh he direc and indirec mehods, as here was an unprecedened curren accoun surplus in he same period. This is also in line wih he claim ha he earlier financial liberalisaion was he main cause of he massive capial ouflows ha were one of he main forces behind he undervaluaion of he NIS in Thus, he rapid appreciaion of he shekel in 2008 was o some exen an equilibrium-revering movemen. However, he RER assessmen for 2008 depends on he RER measure in he sample period: on he one hand, inferences based on long samples and on he REER poin o convergence wih no overshooing. On he oher hand, inferences based on PPPs, or hose based on shorsample REER, poin o evidence of overvaluaion in In 2009, however, all models poin o undervaluaion, which could be a resul of he BOI s policy of foreign exchange inervenion, which suppored he NIS and seems o have had a sizeable influence on he nominal effecive exchange rae. 8 An indirec RER assessmen based on curren accoun analyses provides evidence ha he RER was considerably undervalued in 2006 and in 2009, hus providing furher suppor for he conclusions based on he direc ERER analysis. In 2010 he RER coninued o appreciae, and he laes assessmen, based on he direc mehod and quarerly daa, indicaes ha he RER was back o is equilibrium level by he end of he year or even slighly overvalued. An indirec assessmen based on he curren accoun poins o undervaluaion of abou 10%, even a he end of The paper is organised as follows. The nex secion presens he main rends of he RER and he curren accoun during he las decade. The hird secion presens he main economeric resuls of he RER equaions based on boh panel and ime-series daa. The fourh secion briefly discusses he fundamenal equilibrium exchange rae ha balances he curren accoun, and concluding remarks are given in Secion Friedman and Liviaan (2009). Sorezcky (2010). 202 BIS Papers No 57

3 2. The real exchange rae and he curren accoun Israel is an advanced, small open economy. Wih a relaively large segmen of highly rained workforce and innovaive engineering, expors are concenraed in several human capialinensive indusries. To a large exen, economic growh in recen decades has been led by he expor secor, which is direcly affeced by RER flucuaions. The REER and he NEER are presened in Figure 1. No surprisingly, he wo are highly correlaed, as he shekel is a free-floaing currency and inflaion in Israel in he 2000s was on a par wih ha in is rading parners, so ha he changes in he NEER direcly affec he REER boh in he shor run and in he long run. The Israeli economy has been characerised by a curren accoun surplus since 2003, as presened in Figure 2. Afer a relaively depreciaed level in he mid-2000s, he REER appreciaed rapidly beween mid-2007 and mid During a shor period, of less han a year, i appreciaed by abou 20%. This inciden raised grea concern a he BOI abou he compeiiveness and resilience of he Israeli expor secor. This siuaion, and he low level of he foreign exchange reserves ($28 billion) on he eve of he Grea Recession, riggered a change in he BOI s policy owards he foreign exchange marke. Afer allowing he exchange rae o floa freely for over a decade beween 1998 and March 2008, he BOI resumed is policy of direc exchange marke inervenion. This policy seemed o reverse he appreciaion rend of he NEER and he REER and resuled in a depreciaion of abou 10%. 9 The appreciaion rend resumed in he second quarer of 2009, and in he hird quarer of 2010 he REER and he NEER were back o heir levels of he firs quarer of 2008, prior o he period when direc inervenion had sared. The reasons for he appreciaion in 2008 were boh local, as he economy reached is poenial capaciy, and exernal, as he financial crisis ha began in he summer of 2007 caused capial inflows o he Israeli economy, which was considered o have a relaively safe and sable banking sysem, as was laer confirmed when he crisis reached is peak in lae The appreciaion rend resumed in 2009 and coninued in 2010, due o boh fundamenal and financial facors. There are several oher fundamenal reasons for he appreciaion rend in recen years. Israel has become officially a developed economy, a change ha was manifesed in is saus in inernaional organisaions eg joining he OECD and in he markes eg he classificaion change by Morgan Sanley. This may have affeced negaively he long-run risk premium of he counry and herefore added o he appreciaion rend. In addiion, in 2009 and 2010 large naural gas fields were discovered, and his will work o reduce he impors of fossil fuels in he fuure. 10 The increase in he ineres rae spread beween Israel and he main developed economies in 2010 resuled in he formaion of massive shor-erm capial inflows and in appreciaion of he RER. The REER and he NEER depreciaed considerably beween 2001 and Beween 2001 and 2003 Israel suffered from a major economic downurn caused by he inifada a period during which he counry suffered from a muliude of Palesinian error aacks on civilian arges and he high-ech crisis in he world financial markes. Consequenly, during his period GDP per capia in Israel dropped, he governmen defici increased dramaically and he deb/gdp raio reached a level of abou 100%. The counry risk premium increased 9 10 Sorezcky (2010). For a review on his subjec and he possibiliy ha he Israeli economy will suffer from Duch disease, see he BOI Annual Repor 2010, Box 7.1. BIS Papers No

4 significanly. These facors, and a drasic 2 percenage poin ineres rae cu by he BOI in early 2002, led o he depreciaion of boh he NEER and REER. Even when hese circumsances changed, however, as he inifada peered ou and he economy sared o recover rapidly from mid-2003, and he curren accoun moved ino surplus saring in 2003 afer deficis since he sae was esablished in 1948, 11 he depreciaion rend coninued in 2004 and Depreciaion during he firs sages of he expansionary phase of he cycle, when he oupu gap is sill large, may occur in he case of capial marke imperfecions. 12 In addiion, he liberalisaion process ha ook place a ha ime creaed exogenous pressure for capial ouflows, as he ceiling on foreign asses as a share of oal porfolios of insiuional invesors was raised, and he ax discriminaion agains invesing abroad was cancelled. These changes resuled in rapid porfolio adjusmens ha were probably he principal caalys behind he depreciaion in hose years. Figure 1 Figure 2 REER and NEER index Q = 1 Curren accoun and rade balance (Quarerly, SA, US$ billions) REER NEER Curren Accoun Trade Balance An increase indicaes an appreciaion. Sources: IMF; Bank of Israel. Source: CBS. An alernaive measure for he RER is he PPP index. While he REER is based on nominal exchange raes correced for inflaion differences, he (deviaion from) PPP is more relaed o he relaive price of non-radable goods. One of he advanages of he PPP measure is comparabiliy: i allows cross-counry analyses of he levels, raher han jus analysing changes in he RER, as is he case when using he REER. Figure 3 below compares he REER o he PPP indices from wo differen sources. Generally, he rends of he differen measures of he RER coincided in he las wo decades. The deviaions from PPP indicae ha on average prices in Israel are lower han in he US. However, in prices were higher han in he US, and in 2008 prices reached he US level In 1985 he dramaic fiscal consolidaion programme (he Economic Sabilizaion Plan) and foreign aid from he US resuled in a emporary curren accoun surplus. Friedman and Liviaan (2009). 204 BIS Papers No 57

5 Figure 3 Figure 4 REER and PPP indices Israeli expors vs OECD expors Volume, Q = PPP IMF PPP OECD REER Israeli Expors OECD Expors An increase indicaes an appreciaion. Sources: IMF; OECD. Sources: IMF; OECD; CBS. Afer he oubreak of he Grea Recession, world rade collapsed by 15%, and Israeli expors dipped accordingly (Figure 4). Alhough expors plunged, he effecs of he recession on he Israeli economy were mild relaive o he effecs on oher developed economies. The decrease in he volume of expors was mached by a decrease in impors of he same order of magniude, and he erms of rade improved, so ha he curren accoun surplus reurned o is pre-crisis level in less han a year (Figure 2). 3. The equilibrium real exchange rae Our direc empirical approach for esimaing he equilibrium RER is o relae i o a number of economic fundamenals ha affec is behaviour. The raionale for his approach is derived from he uncovered ineres pariy (UIP) in real erms: E (q +1 ) - q + = r - r *, where q denoes he naural logarihm of he RER, r denoes he real ineres rae in he home counry, r* denoes he real ineres rae level abroad, and denoes he counry risk premium. Thus, he level of he RER a ime is a funcion of he ineres rae differenial, risk premium and he expeced level of he RER a ime +1. The mehodology is based on he assumpion ha his expeced level, E (q +1 ), can be relaed o a se of economic fundamenals a ime, 13 while he ineres rae differenials are assumed o be ransiory and herefore he empirical assessmens of he ERER do no ake hem ino accoun. In he nex secions he acual RER is relaed o a se of economic fundamenals. The fied values are inerpreed as equilibrium values: his inerpreaion is based on he assumpion ha on average, over long sample periods, he RER is aligned. The fac ha he average curren accoun balance in Israel beween 1995 and 2010 is close o zero provides suppor for his assumpion. 13 This approach follows Frankel and Rose (1995), Driver and Wesaway (2004) and Lee e al (2008). BIS Papers No

6 3.1 Assessmens based on panel daa We sar by relaing he PPP index o per capia GDP. 14 The relaion beween price levels or he RER and per capia income is well known in he lieraure. Theoreically his correlaion holds boh in cross secions and over ime, which may be he resul of a Balassa-Samuelson effec, alhough here are oher possible reasons for his relaion. Moreover, his correlaion seems o be higher in counries wih a floaing exchange rae regime. 15 In Figure 5 and Figure 6 below, he PPP index is ploed agains he level of per capia GDP. The fac ha he level of he PPP is below he logarihmic regression line in 2006 implies ha he RER was undervalued, by 7.5%. In 2009, afer appreciaing considerably, mainly versus he US dollar, he RER is sill slighly undervalued, by 2%. The reason is ha he RER appreciaion versus he US dollar is common o oher counries, so he regression line shifed upwards. Figure 5 Figure 6 PPP and per capia GDP, 2006 PPP and per capia GDP, 2009 OECD counries excluding Iceland and Luxembourg, and including Esonia and he Russian Federaion. Sources: OECD; auhors calculaions. In order o ake oher facors ino accoun, we esimae he relaion beween RER and per capia GDP using panel daa, and conrol for oher variables ha may affec he RER. We use he variables below: Eichengreen (2008), Berka and Devereux (2010). We use he OECD PPPs as in he second sudy. Berka and Devereux (2010). 206 BIS Papers No 57

7 Terms of rade: An improvemen in he exernal erms of rade has a posiive income effec on he economy and herefore is expeced o resul in an RER appreciaion. Governmen consumpion: Governmen consumpion is usually biased owards non-radable goods and herefore may resul in an RER appreciaion. 16 Dependency raio: An increase in he child and old-age dependency raio may increase he relaive demand for non-radable goods. 17 In addiion, he increase in child dependency ends o increase invesmen, reduce savings and accordingly increase he RER. 18 The resuls of he panel regressions are repored in Table 1 below. The dependen variable is eiher an index of he REER (log), as in Lee e al (2008), or PPPs (log) from he OECD daabase, as in Berka and Devereux (2010). As he dependen variables and he explanaory variables are inegraed of order one, hese models represen long-run coinegraing vecors for he equilibrium RERs, so ha he residuals are a measure of RER misalignmen wih respec o he equilibrium RER. 19 The elasiciy of he REER wih respec o relaive per capia GDP is abou uniary and he elasiciy of PPPs wih respec o relaive per capia GDP is above one in boh cases, much higher han in comparable sudies. The dispariy is generaed because he esimaion is based on advanced economies only, and on a relaively shor sample from 1995 onwards, while he resuls when using a longer sample period (saring from 1980) are on a par wih hose repored in Berka and Devereux (2010). In addiion, we noe ha he elasiciy in cross secions is much lower, abou 0.4, suggesing ha he ime dimension of he panel increased he esimae. The elasiciy of he REER o he erms of rade or governmen consumpion is somewha lower han hose repored in Aguirre and Calderón (2005) and Lee e al (2008). Oher variables ha were esed bu evenually excluded are: (i) he ne inflow of FDI: his exogenous inflow migh be expeced o pu posiive pressure on he RER; our resuls, however, imply ha causaliy runs from he RER o FDI, ie undervalued counries arac FDI, as in Koseleou and Liargovas (2000); (ii) ne foreign asses (NFA): a higher level of NFA generaes higher income flows and herefore allows for a higher RER level as hese flows can finance a rade balance defici (Lee e al (2008)); we found ha he effec of his variable was insignifican or alernaively negligible (Aguirre and Calderón (2005)); (iii) governmen deb, which is a proxy for he counry risk, was found o be significan; however, i is available only for a limied number of counries and was herefore excluded; and finally (iv) he oil posiion was found o be insignifican Lee e al. (2008). Braude (2000). Braude e al (2009). Alhough in some cases he inegraion level of he residuals is inconclusive as he sample period is relaively shor, dynamic difference equaions poin o an error-correcion process of abou five years in line wih slow convergence usually found in he lieraure. BIS Papers No

8 Relaive per capia GDP Terms of rade Governmen consumpion Dependency raio Table 1 Real exchange rae and PPP regressions Dependen variables: REER and PPP in levels Model 1 Model 2 Model 3 Model 4 PPP REER PPP REER PPP REER PPP REER 1.32 (12.2) 0.98 (15.2) 1.12 (7.7) 0.30 (4.0) R 2 adj D-W Secions Observaions URT * EC ** 3.36 (0.00) 0.26 ( 8.1) 0.95 (0.16) 0.27 ( 8.2) 3.33 (0.00) 0.27 ( 7.2) 0.76 (9.6) 0.27 (6.7) 1.15 (0.12) 0.29 (7.6) 1.56 (8.8) 0.34 (5.0) 1.87 (2.1) 0.51 (0.3) 0.31 ( 6.1) 1.09 (9.7) 0.30 (6.8) 1.42 (2.6) 1.44 (0.07) 0.33 ( 5.9) * Tes for uni roo in residuals. Im, Pesaran and Shin W-saisic. P-values in parenheses. ** Error correcion in a dynamic equaion. Dependen variable: rae of change in RER/PPP (6.9) 0.63 (4.9) 2.91 (3.3) 2.07 (0.02) 0.27 ( 6.0) 0.95 (7.8) 0.47 (6.8) 1.14 (2.5) 0.14 (0.44) 0.27 ( 5.7) Sample: 28 OECD counries excluding Iceland, Korea, Mexico, Slovenia and Turkey. Sample period: PPPs are calculaed by OECD. REER are indices chained o All regressions include fixed counry effecs. Relaive per capia GDP is chained o 1995 PPPs. Dependency raio is he sum of children up o age 15 plus persons age 65 and over, divided by he populaion beween 15 and 65 years of age. The RER assessmens based on he panel regressions above are presened in Figure 7. When he REER was used as a measure for he RER, he conclusion is ha he rapid appreciaion in 2008 was evenually a movemen owards equilibrium, afer undervaluaion of abou 10% in 2006, when he curren accoun surplus reached is peak. Thus, depreciaion in 2009 represens undervaluaion of he RER by abou 5%. Esimaion based on PPPs, however, ells a differen sory: he appreciaion in 2008 is inerpreed as an overvaluaion of 10% on average. Ye he RER in 2009 is undervalued by 2% on average. Thus, alhough he models yielded differen conclusions wih respec o he pas, he assessmens of 2009 were quie similar. 208 BIS Papers No 57

9 Figure 7 Model-based RER misalignmen, direc mehod (Posiive figures indicae overvaluaion) PPP REER PPP REER PPP REER PPP REER Model 1 Model 2 Model 3 Model RER 2008 RER 2009 RER 3.2 Assessmen based on ime-series daa In order o provide high-frequency real-ime esimaes of he ERER, we use a coinegraing relaion beween he REER (log), relaive per capia GDP (RGDP) and he erms of rade (ToT, log, HP-filered). The equaion is esimaed using a long sample period, saring from 1980, in order o esablish a coinegraing relaion beween he variables. Addiional variables ha were checked bu excluded from he model were: ne financial asses, public deb (boh normalised by GDP), he Tel Aviv sock marke premium relaive o an inernaional benchmark, and various capial flows, including boh porfolio invesmen and FDI. We find only RGDP and ToT o be robus; however, boh elasiciies increase over ime. The elasiciies in he shorer subsample saring from 1995 are doubled relaive o he full sample, saring from The elasiciy wih respec o RGDP in he long sample (0.7) is somewha lower bu sill on a par wih he resuls based on panel esimaion, while he elasiciy of he REER wih respec o he ToT (1.0) is considerably higher han ha in a panel of counries, and higher han esimaed in he lieraure. 21 This may reflec he relaively high openness of he Israeli economy, which implies ha he income effec of a change in ToT is larger han ha in oher counries. The acual vs equilibrium RER and he RER misalignmens based on his equaion ( ) are presened in Figure 8 below This resul also emerges in he panel regression in he previous secion. In a relaed sudy, Sellin (2007) repors ha he empirical elasiciies of he real Swedish krona wih respec o RGDP and ToT are 0.7 and 0.5 respecively. BIS Papers No

10 Figure 8 Acual vs equilibrium RER and RER misalignmens, direc mehod Long sample: Shor sample: Log( REER) 0.71*( RGDP) 1.04* log( ToT) R 2 adj (4.7) (9.4) 0.48, DW , ADF( ) 3.27 Log( REER) R 2 adj 0.89, 1.48* ( RGDP) 2.00* log( ToT ) (14.4) (19.7) D. W. 0.81, ADF( ) 3.88 The conclusions from he long-sample equaion are similar o hose from panel daa assessmens based on REERs: The RER was undervalued beween 2004 and 2007 by more han 10%, and he rapid appreciaion in 2008 can be inerpreed as a reurn o he equilibrium level. According o he subsample, however, undervaluaion in 2006 was considerably lower, so in early 2008 an overshooing of he real exchange rae resuled in overvaluaion of abou 5%. When esimaing he model up o he firs quarer of 2008 and conducing an ou-ofsample forecas for wo quarers ahead, he overvaluaion is esimaed a 8%. The depreciaion in early 2009, which was apparenly he resul of he BOI s foreign exchange purchase policy, worked agains his overvaluaion, and resuled in he reversion of he REER o is equilibrium level. Thus, his case demonsraes he imporance of ERER assessmen for policy purposes. The ERER assessmen for 2010, based on he subsample, poins o a low, insignifican overvaluaion of abou 4% in lae The fundamenal equilibrium exchange rae (FEER) An alernaive, indirec assessmen of he equilibrium RER is based on he analysis of he curren accoun which esimaes he FEER. According o his approach, he FEER is he RER which is consisen wih exernal and inernal balance of he economy, ie i is he RER ha balances he curren accoun while producion facors are employed a normal capaciy. 22 The FEER concep can be formalised as follows: * CA X ( RER, Y ) M ( RER, Y ). gap gap The curren accoun, absracing from ren on producion facors and unilaeral ransfers, is given by expors less impors, which are boh affeced by he RER in he usual way. In addiion, impors increase in he domesic oupu gap, Y gap, while expors increase in he 22 A formal definiion of he FEER concep was firs inroduced by Wren-Lewis (1992). 210 BIS Papers No 57

11 * oupu gap in he res of he world, Y gap. Theoreically, he FEER is defined as he RER ha resuls in a balanced curren accoun given ha boh gaps equal zero. In order o apply he FEER concep, however, one has o ake ino accoun ha in he medium run an economy can run a persisen defici or surplus wih no endency o converge owards zero. 23 Pracically, his requires esimaion of a fundamenal curren accoun for he medium run. 24 In addiion, applicaion of he FEER concep requires esimaion of he elasiciies of expor and impors wih respec o boh he RER and he relevan oupu gaps unobserved variables ha are roughly assessed. Thus, alhough having a well defined, clean, heoreical background and definiion, RER assessmen based on he FEER concep requires vas knowledge of he srucural characerisics of he economy. As he uncerainy around he assumpions on which he FEER is calculaed is high, he resuls of he assessmen are only indicaive. The fundamenal curren accoun in Israel was esimaed using he Macroeconomic Balance approach. 25 The resuls indicae ha several srucural facors were behind he curren accoun reversal in Israel in The fundamenal curren accoun in Israel in 2009 was esimaed a beween 1 and 2.5% of GDP, 26 while he acual curren accoun surplus reached 4.0%; he fac ha he acual curren accoun was higher han he fundamenal curren accoun indicaes ha he RER was undervalued. In order o quanify he RER undervaluaion, we assumed ha he elasiciy of expors wih respec o he RER is beween 0.2 and 0.3, and ha he elasiciy of impors wih respec o he RER is beween 0.4 and These elasiciies sugges ha he RER was undervalued by 5 o 15% in When aking ino accoun he fac ha he oupu gap in Israel was lower han in is rading parners, he undervaluaion in 2009 is even higher. 5. Concluding remarks In his paper we briefly presen some of he analyical ools ha were used by he BOI in order o assess he equilibrium RER. These assessmens enhance our abiliy o disinguish wheher he RER rends could be aribued o fundamenal, long-run facors, or o flucuaions ha may be he resul of shor-run forces such as capial flows, price rigidiies and ohers. The RER flucuaed considerably during he las decade; beween 2000 and 2006 i depreciaed by abou 30%, and since hen has appreciaed by abou 20% o dae. The conclusions from our assessmens lend credence o he hypohesis ha he RER was undervalued in he mid-2000s, a resul which is in line wih he unprecedened curren accoun surplus in hose years. Thus, he rapid appreciaion in 2008 is explained, a leas parially, as a reversion owards equilibrium. In early 2008, however, here is (and was in real Obsfeld and Rogoff (1996). Lee e al (2008). Lee e al (2008). Bank of Israel Annual Repor (2009). The low elasiciies of boh expors and impors are based on Friedman and Lavi (2007) long-run elasiciies. Noe ha he sum of he coefficiens in absolue value is lower han 1, hus violaing he Marshall-Lerner condiion. An RER depreciaion, however, resuls in a curren accoun increase because Israeli expors are usually priced in foreign currency, hus leaving only limied room for a negaive value effec. The high elasiciies of boh expors and impors are based on Tokarick (2010) long-run elasiciies. BIS Papers No

12 ime) evidence of overvaluaion; a ha ime he BOI sared o purchase foreign currency, a policy ha probably conribued o he realignmen and even o he undervaluaion of he RER in 2009, and suppored he expor secor a a criical poin in ime as world rade collapsed in he fourh quarer of The slow appreciaion rend in 2010 is in line wih fundamenal facors, so he RER is close o is equilibrium level, and was even slighly overvalued a he end of Indirec ERER assessmen, however, sill indicaes ha he real exchange rae is undervalued, as a relaively large curren accoun surplus was recorded in References Aguirre, A and C Calderón (2005): Real exchange rae misalignmens and economic performance, Working Paper no 315, Cenral Bank of Chile. Bank of Israel: Annual Repor, various years. Berka, M and M Devereux (2010): Wha deermines European real exchange raes?, Working Paper 15753, Naional Bureau of Economic Research. Braude, J (2000): Age srucure and he real exchange rae, Discussion Paper no , Research Deparmen, Bank of Israel. Braude, J, A Rose and S Supaa (2009): Feriliy and he real exchange rae, Canadian Journal of Economics 42/2 (May), pp Driver, R and P Wesaway (2004): Conceps of equilibrium exchange raes, Working Paper no 248, Bank of England. Eichengreen, B (2008): The real exchange rae and economic growh, Working Paper no 4, Commission on Growh and Developmen. Flug, K and M Srawczynski (2007): Persisen growh episodes and macroeconomic policy performance in Israel, Discussion Paper no , Research Deparmen, Bank of Israel. Frankel, J and A Rose (1995): Empirical research on nominal exchange raes, in G Grossman and K Rogoff (eds), Handbook of Inernaional Economics. Friedman, A and Y Lavi (2007): The real exchange rae and foreign rade, Bank of Israel Review, no 79 (Hebrew). Friedman, A and N Liviaan (2009): The real exchange rae and he curren accoun: a rerospec, unpublished. Inernaional Moneary Fund (2010): World Economic Oulook, Ocober. Koseleou, N and P Liargovas (2000): Foreign direc invesmen and real exchange rae inerlinkages, Open Economies Review, 11, pp Lee, J, GM Milesi-Ferrei, J Osry, A Prai and L Ricci (2008): Exchange rae assessmens: CGER mehodologies, Occasional Paper no 261, Inernaional Moneary Fund. Obsfeld, M and K Rogoff (1996): Foundaions of Inernaional Macroeconomics, MIT Press. Sellin, P (2007): Using a New Open Economy Macroeconomics model o make real and nominal exchange rae forecass, Sveriges Riksbank Working Paper Series, no 213. Sorezcky, A (2010): Did he Bank of Israel influence he exchange rae?, Discussion Paper no , Research Deparmen, Bank of Israel. 212 BIS Papers No 57

13 Tokarick, S (2010): A mehod for calculaing expor supply and impor demand elasiciies, Working Paper 10/180, Inernaional Moneary Fund. Wren-Lewis, S (1992), On he analyical foundaions of he Fundamenal Equilibrium Exchange Rae, in C Hargreaves (ed), Macroeconomic Modelling of he Long Run, Aldersho, UK: Edward Elgar. BIS Papers No