oil and exchange rate

8/8/2019 Oil and Exchange Rate

1/24

Exchange Rate Fluctuations and Output inOil-Producing Countries: The Case of Iran

Mohsen Bahmani-Oskooee and Magda Kandil

AbsTrAcT: Conventional wisdom states that currency depreciation in oil-producing coun-tries is contractionary because demand e ects, limited by the prevalence o oil exportspriced in dollars, are more than o set by adverse supply e ects. Iran, however, has ex-perienced a rapid increase in nonoil exports in the past decade. Against this background,the paper tests whether the conventional wisdom still applies to Iran and concludes thatthe emergence o the nonoil export sector has made currency depreciation expansionary.The expansionary e ect is particularly evident regarding anticipated persistent deprecia-tion in the long run. Notwithstanding the varying e ects o exchange rate fuctuations onthe demand and supply sides o the economy, managing a fexible exchange rate gradu-

ally over time toward achieving stability in the real e ective exchange rate may strike thenecessary balance.

KEy words: currency depreciation, imported inputs, Iran, nonoil exports.

The e ect o exchange rate uctuations on real activity has been subject to extensive de-bate. On the demand side, there has been a common belie that devaluation or depreciationcould boost domestic production through stimulating net exports. Additional channels onthe demand side could also respond to uctuations in the exchange rate. Alexander (1952)illustrates the possibility that devaluation could lower the consumption component o

aggregate demand, as the in ationary e ect o currency devaluation redistributes incomerom workers to producers. As workers are said to have a high marginal propensity toconsume compared to producers, total consumption declines as a result.

Channels o interaction between the exchange rate and the macroeconomy are mademore complicated by developments on the supply side o the economy. Because currencydepreciation raises the cost o imported inputs, it contributes to an increase in productioncosts and thus curtails aggregate supply. I a reduction in aggregate supply more thano sets the increase in aggregate demand, depreciation results in a decrease in domesticproduction. In this case, devaluation or depreciation is said to be contractionary. Other-wise, it could be expansionary. 1

The empirical literature on the relation between currency depreciation and real outputis extensive. Di erent researchers have adopted di erent approaches. Early studies, suchas by Cooper (1971), Krueger (1978), and Edwards (1989a), use what is known as thebe orea ter approach. These studies analyze the per ormance o output be ore and a tera devaluation episode. A second group o studies compares the per ormance o outputin a devaluing country to that o a group o countries that did not devalue. These studies

Mohsen Bahmani-Oskooee ([email protected]) is the Patricia and Harvey Wilmeth Pro essoro Economics at the University o WisconsinMilwaukee. Magda Kandil (mkandil@im .org) isa senior economist in the Caribbean II division o the Western Hemisphere Department o theInternational Monetary Fund. The valuable comments o two anonymous re erees are greatly ap-preciated, without implicating them.


2/24

24 Eme ging Ma ket Finan e & T a e

use what is known as the group approach, and include Donovan (1982), Gyl ason (1987),Kamin (1988), and Edwards (1989b). A third group o studies have ollowed a macro-simulation approach. They assume a shock to the exchange rate and simulate a modelto in er the per ormance o output a ter the shock. Examples include Agenor (1991),

Barbone and Rivera-Batiz (1987), Gyl ason and Shmid (1983), Gyl ason and Risager(1984), Krugman and Taylor (1978), Solimano (1986), and Taye (1999).

The third group o studies analyzes the relation between output and exchange rateuctuations using various econometric estimation techniques. In erence is based on vari-

ous data structures that include cross-sectional data (e.g., Connolly 1983), panel data(e.g., Agenor 1991; Edwards 1989a; Nunnenkamp and Schweickert 1990), and time-series data (e.g., Bahmani-Oskooee and Anker 2001; Bahmani-Oskooee and Mirzaie2000; Bahmani-Oskooee et al. 2002; Ho maister and Vgh 1996; Kamin and Rogers1997; Kandil 2000, 2004; Kandil and Mirzaie 2002, 2003, 2005; Rogers and Wang 1995;Santaella and Vela 1996).

The bulk o the research considers the e ect o the exchange rate on output in non-oil-producing countries. Oil-producing countries, especially members o the Organizationo the Petroleum Exporting Countries (OPEC), have been excluded in this regard. Thepresumption is that oil dominates output supply. Capacity constraints or internationalcoordination may dominate decisions regarding oil exports. Oil is priced in oreign cur-rency (e.g., U.S. dollars). Moreover, many o the oil-producing countries have a peggedexchange rate system to stabilize the domestic proceeds o oil exports. There ore, oilexports could not respond to a change in the exchange rate.

Whereas decisions to produce and export oil are not likely to vary with uctuations inthe exchange rate, the supply side o oil-producing countries may urther react to uctua-tions in the exchange rate. Aggregate supply could decline because o an increase in the

cost o imported inputs due to a depreciation o domestic currency. Bahmani-Oskooee(1996), who considers the experience o Iran, demonstrated empirically that in an oil-producing country, depreciation could be contractionary, despite dominant oil exportsover the sample period 195990. More recently, however, Iran has taken serious steps todiversi y and promote nonoil exports. Such exports were a little over 3 percent o totalexports when the Iranian revolution occurred in 1979. This percent has smoothly increasedover time; in 2000, the last year or which data are available, nonoil exports stood at almost27 percent o total exports. We suspect that this ratio has increased to perhaps 40 percentin more recent years. 2 The surge in nonoil exports has introduced a new dimension throughwhich a depreciation o the Iranian rial may prove to be expansionary. On the demandside, depreciation may boost competitiveness and the demand or nonoil exports. On thesupply side, depreciation may stimulate an increase in the production o nonoil exportsas producers aim to capitalize on additional profts in domestic currency.

This study aims to contrast the e ects o exchange rate uctuations on real outputgrowth in Iran over two sample periods: a period that excludes the recent surge in nonoilexports and the longer sample period, including the 1990s. The objective is to contrastcontractionary and expansionary e ects o exchange rate uctuations over the twosample periods.

The Model and the Method

To motivate the empirical analysis, we draw on the theoretical predictions o the modelin Kandil and Mirzaie (2002). In brie , the demand side o the economy is specifed using


3/24

Ma June 2010 25

standard IS-LM equations with a modifcation or an open economy. Real consumptionexpenditure varies positively with real disposable income. Real investment expenditurevaries negatively with the real interest rate. The exchange rate measures the degree o competitiveness o oreign-produced goods and services relative to those produced

domestically. Real exports are related to an autonomous element that rises when theincome level abroad rises with relative prices. When the oreign price is higher relativeto domestic goods, exports increase. Real imports are assumed to rise with the levelo real income and to decrease with the oreign price relative to domestic goods. Realgovernment spending is assumed to be exogenous. The total expenditure by domesticresidents in real terms is the sum o real consumption expenditure, real investment, realgovernment spending, and net exports.

Substituting all equations into the equilibrium condition or the goods market resultsin the expression or real income. It is a unction o the exchange rate, the domestic pricelevel, the oreign price level, and the domestic interest rate. This expression is the ISequation, which describes the relation between real income and the real interest rate.

Equilibrium in the money market is obtained by equating the demand and supply o real money balances. The real money supply is determined by nominal balances de atedby price. The demand or real money balances is positively related to real income andinversely related to the nominal interest rate, the sum o the real interest rate, and in a-tionary expectation. It is assumed that citizens in each country must hold domestic money

or transaction purposes but may speculate by holding oreign currency. An unexpectedtemporary depreciation o the domestic currency would lead to speculation o uture ap-preciation to restore the steady-state normal trend o the exchange rate. Consequently,agents increase speculative demand or domestic currency, establishing a positive rela-tion between demand or real money balances and agents expectation o the price o the

domestic currency relative to its current value.The LM equation is determined by the equilibrium condition in the money market.

It establishes a positive relation between real income and the real interest rate. Solvingor the interest rate rom the LM equation and substituting the result into the IS equation

results in the equation or aggregate demand.On the supply side, output is produced using a production unction that combines labor,

capital, and imported intermediate goods. The demand or labor varies negatively withthe real wage and positively with imported intermediate goods. Similarly, the demand orimported intermediate goods increases with labor input. When the currency depreciates(or is devalued), it is more expensive to buy intermediate goods rom abroad, decreasingdemand or these goods. The supply o labor increases with the increase in the nominalwage relative to workers expected price.

The nominal wage solution is obtained by equating labor demand and supply. Substi-tuting the nominal wage into labor demand results in the solutions or employment andimported intermediate goods. Substituting these solutions into the production unctionresults in the specifcation or aggregate supply o domestic value added. Aggregatesupply has a direct positive relation with output price surprises. Workers decide on laborsupply based on their expectation o the aggregate price level. An increase in aggregateprice, relative to workers expectations, increases the demand or labor and, hence, thenominal wage. A rise in expected real wage increases employment and, hence, the out-put supplied. In addition, aggregate supply moves negatively with the domestic price

o oreign currency. Currency depreciation increases the cost o imported goods anddecreases the output supplied.


4/24


The combination o demand and supply channels indicates that real output depends onmovements in the exchange rate, the money supply, and government spending. In the shortrun, transitory exchange rate uctuations determine competitiveness and, in turn, produc-ers reaction to the change in the relative price o tradables and nontradables. A transitory

currency appreciation decreases demand or exports and increases demand or imports,decreasing supply o tradables. Nonetheless, cheaper cost o imports may increase outputsupply in the short run. In the long run, persistent exchange rate appreciation necessitatesa sustained adjustment in the output supply. Depending on the relative strengths o thecompetitiveness and cost channel, producers may increase or decrease the output supplywith respect to persistent uctuations in the exchange rate in the long run.

In light o the theoretical predictions, we study the e ect o currency depreciation ondomestic output using an empirical model that replicates the reduced- orm solution o the theoretical model in Kandil and Mirzaie (2002) as ollows: 3

ln GDP t = a + bln M t + clnG t + d ln RE t + et . (1)

ln is the natural logarithm o variables. Real output, GDP t , varies with three policy vari-ables. The log o the real money supply, M t , approximates monetary policy. Fiscal policyis measured by real government spending, G t , and the real exchange rate is denoted by

RE t . Other uctuations in real output growth are measured by et .Positive estimates o b and c indicate that expansionary monetary and fscal policies

could boost real output growth in the long run. In the interest o identi ying channels o interaction between the exchange rate and real output, the analysis employs two mea-sures o the exchange rate. The frst measure is the bilateral exchange rate, defned asthe number o Iranian rial per U.S. dollar. Accordingly, an increase in the exchange ratemeasures currency depreciation. The rial has been pegged to the dollar, which limits

uctuations in the real exchange rate to price movements. Given limited supply o dollarreserves, the rialdollar bilateral exchange rate uctuated in the parallel market relativeto the o fcial price. Fluctuations in the parallel market price capture the e ects o excessdemand or the dollar, relative to the limited supply. I depreciation o the rialthat is,an increase in RE is to be expansionary, an estimate o d should be positive. Otherwise,

or contractionary devaluations, an estimate o d could be negative.Reduced- orm models, such as Equation (1), in which all variables enter in levels, are

re erred to as long-run models. To get more consistent and e fcient estimates, short-rundynamics must be incorporated into the estimation procedure. The task is reduced to oneo speci ying (1) in an error-correction ormat, ollowing the suggestions o Engle andGranger (1987). In the short run, real output growth varies with its lags and distributedlags o policy variablesthat is, the money supply, government spending, and the realexchange rate:

ln ln ln

ln

GDP GDP M

G

t k t k k

n

k t k k

n

k t k k

= + +

+

=

=

0 11

1

20

2

3

==

= + + +

0

3

40

4

1

n

k t k k

n

t t RE ln .

(2)

Given that real output varies with policy variables in the long run, the gap between thetwo sides o (1), as measured by e, should decrease. To capture the speed o adjustmenttoward equilibrium, the lagged value o et is included in the empirical model (2). A nega-tive and signifcant coe fcient or et1 indicates that variables are converging toward their


5/24

Ma June 2010 27

long-run equilibriumthat is, cointegratedprovided that all variables in the model arenonstationary. O course, another way o establishing cointegration in this setting is toshow that all variables in (1) are integrated o order one, denoted by I(1), and the residual,as a proxy or their linear combination, is integrated o order zero, I(0).

In the event some variables are I(1) and some I(0), Pesaran et al. (2001) propose yetanother method or testing cointegration that does not require preunit root testing; vari-ables could be I(1), I(0), or mutually cointegrated. They basically solve or et 1 in (1) andsubstitute the result into (2). The resulting specifcation yields (3), as ollows:

ln ln ln lnGDP GDP M Gt k t k k

n

k t k k

n

k t k k

= + + +=

=

0 10

1

20

2

3

==

=

+ + + + +

1

3

40

4

0 1 1 1 2 1 3

n

k t k k

n

t t t RE GDP M G ln ln ln ln ln R RE t t +1 .

(3)

Equation (3) is somewhat di erent than a standard VAR model. It includes a linearcombination o lagged variables themselves, rather than the lagged error term rom(1). Pesaran et al. (2001) propose applying the amiliar F -test or joint signifcance o the lagged-level variables. I the lagged-level variables are jointly signifcant, there is asingle-level relation between the dependent variable and the independent variables (i.e.,they adjust jointly toward ull equilibrium).

The F -test that Pesaran et al. (2001) propose has new critical values that they tabulateusing Monte Carlo experiments. By assuming all variables to be I(1), they provide anupper-bound critical value, and by assuming all variables to be I(0), a lower-bound criticalvalue. For cointegration, the calculated F- statistic should be greater than the upper-boundcritical value. Once cointegration is established, the long-run e ects o independent

variables on the dependent variable are in erred by the size and signifcance o 1 3 that is normalized by 0. The short-run e ects are in erred by the size and signifcanceo 2k 4k .4

Results

We estimate the error-correction model outlined by Equation (3), using annual data.

Basic Model Estimation

We analyze the e ect o exchange rate uctuations on real output growth, controlling ormonetary and fscal policies. The primary hypothesis under investigation relates to thepossibility o a structural break that marks a surge in nonoil exports in the early 1990s.Iran has maintained a ormal peg o the rial to the dollar. Nonetheless, uctuations in theexchange rate can be captured by two measures. First, in the parallel market, the rial-dollar value deviated rom the o fcial rate in response to supply and demand pressures.In addition, the real e ective exchange rate uctuated in response to changes in the rialvalue relative to currencies (other than the U.S. dollar) o major trading partners.

In the shorter sample period, 195990, the export sector in Iran was dominated byoil exports. Over this sample period, uctuations in the exchange rate, regardless o themeasure, are not likely to have a ected the export sector signifcantly. Oil exports are

mostly determined by capacity constraints and/or the central governments decisions re-garding oil production in the context o international coordination. Nonetheless, currency


6/24


depreciation could have resulted in an increase in the cost o intermediate imports, shrink-ing output supply and contracting output growth. As Irans exports have become morediversifed, currency depreciation over the longer sample period, 19592003, is likely tohave increased incentives to promote and produce nonoil exports.

Be ore embarking on the estimation, a number o decisions have to be fnalized.First is the selection o the optimal lag length in Equation (3). Given the limited numbero observations using annual data, we impose a maximum o our lags on each frst-di erenced variable and rely on the Akiake in ormation criterion (AIC). Second, we test

or cointegration without invoking the requirement that all variables in the model arenonstationary. We test or the joint signifcance o lagged level variables in (3) by carryingout the F- test, using the optimum number o lags on each frst-di erenced variable. Weestimate the empirical model in (3) using several variants, 5 estimating each variant overthe short sample period that ends in 1990 and the longer sample period through 2003.The evidence contrast the e ects o the exchange rate on real output growth be ore anda ter the surge in nonoil exports.

dep e iati n in real Pa allel Ex hange rate

Table 1 reports the results o estimating the model using the real value o the dollar inrial terms in the parallel market. An increase in this exchange rate indicates deprecia-tion o the rial relative to the dollar. Fluctuations in the exchange rate are likely to a ectproducers decisions to export and import in Iran. External demand is not likely to varywith uctuations in the parallel market.

Panel A presents estimates o short-term dynamics. Monetary growth stimulates realoutput growth in the short run. The evidence is supported over the short and long sample

periods. Government spending is a major determinant o short-term uctuations in realoutput growth in both periods. This evidence attests to the success o monetary and fscalpolicies in determining real output growth in Iran in the short run. Moreover, the positivesignifcant e ect supports the procyclical nature o fscal policy in Iran. Large swings inthe oil price determine uctuations in oil reserves and budgetary resources in the shortrun, which coincide with cyclical episodes o economic booms and recessions.

In the short run, uctuations in the real parallel price o the U.S. dollar are insignifcantor real growth in the sample period 195990. Over the longer sample period, 19592003,

depreciation shrinks real output growth, as the negative and signifcant coe fcient shows.Transitory currency depreciation increases the cost o imports and shrinks the output sup-ply. Accordingly, the cost channel dominates in guiding supplier decisions with respectto currency depreciation in the short run.

The long-run estimates in Panel B support the central theme o the paper. Whereas thereal exchange rate carries an insignifcant coe fcient in the sample period 195990, itcarries a positive and highly signifcant coe fcient during the period 19592003. In thelong run, as producers anticipate sustained real depreciation and an increase in profts indomestic currency, they have higher incentives to produce and promote nonoil exports,resulting in a long-lasting increase in output growth. 6

In Panel C, the F- statistic exceeds the upper-bound critical value. Hence, cointegra-tion among the variables is supported over the 195990 sample period and the extendedsample period, 19592003. There is a common stochastic trend implying close coordina-

tion among the monetary, fscal, and exchange rate policies in the long run.


7/24

Ma June 2010 29

Table 1. Results employing the real bilateral rialdollar exchange rate, byperiod

19591990 19592003

Panel A: Short-run estimates

ln GdP t 1 0.06(0.32)

ln GdP t 2 0.42(2.09)

ln GdP t 3 ln M t 1.12 0.98

(2.53) (4.27) ln M t 1 0.55 0.21

(2.03) (1.02) ln M t 2 0.23 0.29

(0.88) (1.28)

ln M t 3 0.43(1.51) ln G t 0.35 0.37

(2.17) (3.32) ln G t 1 0.27 0.03

(1.22) (0.22) ln G t 2 0.57 0.33

(3.28) (2.61) ln G t 3 0.23

(1.50) ln rE t 0.04 0.11

(0.75) (1.35) ln rE t 1 0.11 0.13

(0.69) (1.96) ln rE t 2 0.24

(1.59) ln rE t 3 0.33

(3.64)

Panel B: Long-run estimates

Constant 10.54 7.32(0.33) (1.95)

ln M 14.59 0.71(0.31) (1.61)

ln G 13.01 0.22(0.29) (0.56)

ln rE 3.23 1.05(0.29) (2.51)

Panel C: Diagnostics

Adjusted r 2 0.91 0.72F 4.57 3.67EcM t 1 0.07 0.24

(0.33) (2.79)

Notes: Numbers in parentheses are the absolute values o the t -ratios. The upper-bound critical valueo the F -test at the 5 percent level o signifcance is 4.01. The comparable fgure at the 10 percentlevel is 3.52. These come rom Pesaran et al. (2001, p. 300, table CI(iii)Case III).


8/24


To establish robustness, we use the long-run coe fcient estimates rom Panel B to orma lagged error-correction term, ECM t 1. Having replaced the linear combination o thelagged level variables in Equation (3) by ECM t1, we reestimate the model, imposing theoptimal number o lags. A signifcantly negative coe fcient or ECM t 1 provides additional

and stronger support or cointegration. 7 The ECM t 1 in Panel C carries a signifcant negativecoe fcient only during the extended period 19592003. Consistent with the evidence inPanel B over this period, the unique signifcance o the exchange rate e ect in the longrun implies signifcant adjustment toward ull equilibrium in Panel C.

dep e iati n in N minal Pa allel Ex hange rate

Two actors underlie movements in the exchange rate in Table 1: the nominal price o the dollar relative to the rial in the parallel market and in ation in Iran relative to theUnited States. For policy consideration, however, uctuations in the parallel market rateare an important signal. They measure pressures on the rial value relative to the dollar, inresponse to supply and demand pressures in the parallel market, compared to the o fcialpegged rate. Interventions by the central bank in the oreign exchange market o ten targetdeviations in the parallel market relative to the o fcial rate. More recently, the Iraniancentral bank has unifed the o fcial and the ree parallel market rates and has olloweda strategy o gradual depreciation against the U.S. dollar. 8 To isolate the signal e ectattributed to uctuations in the parallel market on real activity, we employ the nominalrial-dollar parallel rate ( NE ) in Equation (3) and carry out the same analysis. The resultsare in Table 2.

In Panel A, uctuations in fscal and monetary policy determine real output growthsignifcantly in the short run, attesting to the procyclical nature o these policies. Depre-

ciation o the rial relative to the dollar in the parallel market has a positive expansionarye ect on real output growth. Transitory currency depreciation increases incentives toproduce and promote nonoil exports in the short run. This evidence is robust in the shortsample period, 195990, and in the longer sample period, 19592003.

In Panel B, over the longer sample period 19592003, both fscal and monetary poli-cies carry their expected positive signs and are all highly signifcant. 9 Having isolatedthe exchange rate measure rom in ation in Table 2, fscal and monetary expansionsinduce long-lasting real e ects on real output growth. Consistent with the evidence inTable 1, the nominal exchange rate does not determine real output growth signifcantlyin the long run during the 195990 period. In contrast, it carries a positive and highlysignifcant coe fcient or the period 19592003. Hence, the positive expansionary e ecto persistent exchange rate depreciation on output growth is attributed to pressure onthe exchange rate value in the parallel market, independent o the relative price channel.Depreciation increases incentives to gear production toward promoting nonoil exports.Through this channel, persistent depreciation has a long-lasting signifcant e ect in thelonger sample period that accounts or the surge in nonoil exports.

App e iati n in the Effe tive Ex hange rate

Historically, the dollar has been a major reserve currency in Iran and the rial-dollar rate wasthe most dominant exchange rate in Irans trade with the rest o the world. However, since

the Islamic revolution in 1979, the share o trade with the United States has shrunk overtime. Trade shares with countries such as China and Japan have gone up over time.


9/24

Ma June 2010 31

Table 2. Results employing the nominal bilateral rialdollar exchange rate, byperiod

19591990 19592003

Panel A: Short-run estimates

ln GdP t 1 0.29 0.36(1.15) (2.43)

ln GdP t 2 ln GdP t 3 ln M t 1.07 0.66

(2.56) (3.87) ln M t 1 0.11

(0.39) ln M t 2 0.47

(1.89) ln M t 3

ln G t 0.34 0.38(1.69) (4.51) ln G t 1 0.89 0.13

(2.73) (1.31) ln G t 2 0.85 0.19

(3.72) (2.74) ln G t 3 0.45

(2.71) ln NE t 0.12 0.08

(1.21) (4.21) ln NE t 1 0.31

(2.53) ln NE t 2 0.40

(3.13) ln NE t 3 0.44

(5.11)

Panel B: Long-run estimates

Constant 5.62 1.85(1.03) (7.60)

ln M 11.60 0.49(0.80) (3.90)

ln G 11.42 0.27(0.85) (2.35)

ln NE 2.50 0.15(0.89) (8.55)


Adjusted r 2 0.92 0.78F 5.54 6.90EcM t 1 0.14 0.58

(0.68) (4.69)

Notes: Numbers in parentheses are the absolute values o the t -ratios. The upper-bound critical valueo the F -test at the 5 percent level o signifcance is 4.01. The comparable fgure at the 10 percentlevel is 3.52. These come rom Pesaran et al. (2001, p. 300, table CI(iii)Case III).


10/24


Fluctuations in the e ective exchange rate capture the composite e ects o bilateralexchange rate uctuations with major trading partners. Whereas the central bank maydevalue the rial against a major currency, such as the U.S. dollar, the rial may gain againstthe currency o another trading partner i Irans exports to that partner exceed imports.

To incorporate uctuations o the rial against currencies o all trading partners, we in-vestigate the e ect o the real and nominal e ective exchange rates on domestic output.To that end, we construct the real e ective exchange rate ollowing the ormulation byBahmani-Oskooee (1995), as outlined in Equation (4):

REFEC

P E

P

P E

P

t j

i J

j

i j

j

=

2000

100=

J 1

21

,

(4)

where j is the total trade share (exports plus imports) o trading partner j such that j = 1. Note that in (4), E j is the nominal bilateral exchange rate, defned as the numbero units o j s currency per Iranian rial; P i is the price level in Iran; and P j is the pricelevel in trading partner j. Identity (4) basically amounts to taking a weighted averageo the twenty-one real bilateral exchange rates. Given the defnition o E j , an increasein REFEC re ects a real appreciation o the rial. The nominal e ective exchange rate( NEFEC ) is constructed in the same manner, omitting the price levels rom the ormula.Due to the lack o consumer price index data or some trading partners, the e ective ratesare constructed only or the 19662003 period. 10

In contrast to the parallel market value o the U.S. dollar, the e ective exchange rateis a better measure to assess the e ect o relative competitiveness on the demand ornet exports. 11 In addition, suppliers may be inclined to consider uctuations in the e -

ective exchange rate when calculating profts and the cost o imports in the productionunction.

Having employed the e ective exchange rate in (3), we report the estimation resultsrom the optimal models in Table 3 or the sample period 19662003. I a depreciation

o the rial, nominal or realthat is, a reduction in the real or nominal e ective rateisto be expansionary, the estimate o 3 normalized on 0 must be negative. The resultsestablish the superiority o the e ective exchange rate, relative to the parallel marketrate, to determine uctuations in real output growth in Iran. Almost all the estimated

coe fcients in Table 3 are signifcant. In Panel A, signifcant coe fcients or the laggedvalues o real output growth indicate a high degree o persistence. Both monetary andfscal policies are important determinants o uctuations in real output growth, attestingto the signifcance o procyclical policies in the short run. Appreciation o the e ectiveexchange rate has an expansionary e ect on real output growth in the short run, as thepositive and signifcant coe fcient shows. Transitory currency appreciation stimulatesgrowth through the cheaper cost o intermediate imports in the short run. This channeldominates the negative e ect o currency appreciation on competitiveness and, there ore,real growth in the short run.

The long-run evidence in Panel B indicates the long-run signifcant e ects o govern-ment spending on real growth, which is robust regardless o the exchange rate measure.More important or our purposes, an appreciation in the e ective exchange rate is


11/24

Ma June 2010 33

Table 3. Results employing real and nominal e ective exchange rate

EX = real EX = nominale ective rate e ective rate

Panel A: Short-run estimates ln GdP t 1 0.22 0.98

(1.22) (6.02) ln GdP t 2 0.49 0.47

(2.60) (5.00) ln GdP t 3 0.21 0.35

(1.41) (4.00) ln M t 1.31 0.57

(5.37) (8.25) ln M t 1 1.16

(4.15) ln M t 2 1.19

(4.12) ln M t 3 1.17

(3.39) ln G t 0.17 0.45

(1.54) (7.07) ln G t 1 0.41 0.46

(2.69) (4.58) ln G t 2 0.24

(1.98) ln G t 3 ln EX t 0.17 0.09

(2.45) (2.34) ln EX t 1 0.63 0.06

(5.26) (1.43) ln EX t 2 0.36

(3.84) ln EX t 3 0.17

(2.38)Panel B: Long-run estimates

Constant 10.83 3.56(6.50) (24.6)

ln M 0.11 0.35(0.67) (10.5)

ln G 0.57 0.47(4.71) (16.9)

ln EX 1.19 0.13

(5.69) (25.4)


Adjusted r 2 0.84 0.90F 10.33 19.62EcM t 1 0.63 1.62

(4.45) (9.46)

Notes: Numbers in parentheses are the absolute values o the t -ratios. The upper-bound critical valueo the F -test at the 5 percent level o signifcance is 4.01. The comparable fgure at the 10 percentlevel is 3.52. These come rom Pesaran et al. (2001, p. 300, table CI(iii)-Case III).


12/24


13/24

Ma June 2010 35

The observed e ects o exchange rate uctuations on real output growth may be di -erentiated according to expectations. Signifcant e ects o anticipated exchange rateuctuations would signal to policy makers the importance o managing agents orecasts

over time. Signifcant e ects o random unanticipated exchange rate uctuations would

warrant a speedy policy intervention to reduce uncertainty.

Anti ipati n f dep e iati n in the Pa allel Ma ket

We decompose the change in the rial/U.S. dollar parallel exchange rate into anticipatedand unanticipated components. Table 4 reports the results using each component, apply-ing this decomposition to each o the real and nominal rial/dollar parallel exchange ratesover the sample period 19592003. 15

The evidence generally supports the role o anticipated exchange rate movements indetermining real output growth. In Panel A, the signifcant e ects o procyclical poli-cies remain robust in the short run. However, signifcant uctuations in real growth areattributed only to the anticipated component o exchange rate movements. In the shortrun, anticipated depreciation in the parallel exchange rate increases incentives to producenonoil exports and stimulate real output growth. Random unanticipated uctuations in theexchange rate dissipate quickly, with no signifcant short-run e ects on real growth.

In Panel B, the evidence remains robust regarding the positive e ects o monetary andfscal policies on real output growth in the long run. Anticipated depreciationboth realand nominalthat is, an increase in the rial/U.S. dollar exchange rate, has a positive andsignifcant long-run e ect on output. Consistent with the supply-side channel, anticipatedpersistent depreciation o the rial value in the parallel market increases anticipated proftin the export sector and stimulates output growth.

Consistently, the error-correction term in Panel C is negative and signifcant, indicatinga tendency or output, along with independent regressors, to adjust jointly toward ullequilibrium. In contrast, unanticipated exchange rate shocks do not exhibit a long-lastinge ect on output growth. As shocks dissipate quickly, they do not permanently increaseoutput growth. Consistent with the theorys predictions, con icting channels on the de-mand and supply sides render the e ects o unanticipated depreciation o the rial/dollarparallel exchange rate insignifcant on output growth in the short run. 16

Anti ipati n f App e iati n in the Effe tive Ex hange rate

Similar results arise in an extension that tests the e ects o anticipation on the interactionbetween real output growth and the real e ective exchange rate. In Table 5, we employthe anticipated and unanticipated components o the real and nominal e ective exchangerates into the empirical model (3) over the sample period 19662003.

The results generally support the a orementioned implications. In Panel A, procyclicalpolicies, both fscal and monetary, have signifcant e ects on real output growth in theshort run. An increase in the anticipated e ective exchange ratean anticipated currencyappreciationstimulates real output growth, although with a lag. The evidence, whichis robust or nominal and real e ective exchange rates, indicates the e ect o anticipatedappreciation in decreasing the cost o imported inputs and increasing output growth. Incontrast, unanticipated currency uctuations dissipate quickly, with no signifcant e ect

on real output growth.


14/24


T a b l e

4 .

R e s u

l t s e m p

l o y i n g a n

t i c i p a

t e d a n

d u n a n

t i c

i p a

t e d b i l a t e r a

l r i a

l d o

l l a r e x c

h a n g e r a t e

E X = a n

t i c

i p a

t e d

E X

= u n a n

t i c

i p a

t e d

E X = a n

t i c

i p a

t e d

E X = u n a n t i c

i p a

t e d

r e a

l b i l a

t e r a

l r a

t e

r e a

l b i l a

t e r a

l r a

t e

n o m

i n a l

b i l a

t e r a

l r a

t e

n o m

i n a

l b i l a t e r a

l r a

t e

P a n e l

A :

S h o r t - r u n e s t

i m a

t e s

l n

G d P

t 1

l n

G d P

t 2

l n

G d P

t 3

l n

M t

0 . 9

1

0 . 7

8

0 . 8

3

0 . 7

7

( 4 . 2

5 )

( 3 . 6

4 )

( 4 . 6

5 )

( 3 . 5

3 )

l n

M t 1

0 . 2

8

0 . 3

4

( 1 . 3

3 )

( 2 . 0

4 )

l n

M t 2

l n

M t 3

l n

G t

0 . 3

6

0 . 3

6

0 . 4

4

0 . 3

6

( 3 . 5

0 )

( 3 . 3

6 )

( 4 . 8

7 )

( 3 . 4

0 )

l n

G t 1

0

. 0 5

0 . 0

2

0 . 0

2

0 . 0

2

( 0 . 5

3 )

( 0 . 2

2 )

( 0 . 2

9 )

( 0 . 2

7 )

l n

G t 2

0 . 1

8

0 . 1

8

0 . 2

2

0 . 1

8

( 1 . 9

7 )

( 1 . 8

8 )

( 2 . 9

4 )

( 1 . 9

6 )

l n

G t 3

0 . 1

3

0 . 2

0

( 1 . 3

6 )

( 2 . 4

2 )

l n

E X t

0 . 1

3

0

. 0 4

0 . 0

8

0

. 0 4

( 2 . 2

2 )

( 0 . 5

9 )

( 4 . 2

0 )

( 0 . 5

8 )

l n

E X t

1

l n

E X t

2

l n

E X t

3


15/24

Ma June 2010 37

P a n e l

B :

L o n g - r u n e s t

i m a

t e s

C o n s t a n

t

3

. 3 6

0 . 0

8

1 . 6

0

0 . 0

6

( 1 . 5

1 )

( 0 . 0

4 )

( 7 . 0

5 )

( 0 . 0

3 )

l n M

0 . 8

5

1 . 8

3

0 . 5

4

1 . 8

1

( 1 . 9

4 )

( 1 . 7

3 )

( 4 . 9

2 )

( 1 . 7

2 )

l n G

0

. 0 0 1

0

. 8 8

0 . 2

7

0

. 8 4

( 0 . 0

1 )

( 0 . 8

0 )

( 2 . 4

5 )

( 0 . 7

8 )

l n E X

0 . 5

9

0

. 3 1

0 . 1

2

0

. 3 0

( 2 . 0

4 )

( 0 . 5

8 )

( 8 . 2

5 )

( 0 . 5

6 )

P a n e l

C :

D i a g n o s t

i c s

A d j u s

t e d r 2

0 . 7

1

0 . 6

7

0 . 8

0

0 . 6

7

F

4 . 3

8

2 . 8

3

7 . 6

7

2 . 8

7

E c M

t 1

0

. 2 2

0

. 1 1

0 . 6

6

0

. 1 1

( 2 . 3

1 )

( 1 . 3

0 )

( 4 . 5

7 )

( 1 . 3

1 )

N o

t e s :

N u m

b e r s

i n p a r e n

t h e s e s a r e

t h e a b s o

l u t e v a

l u e s o

t h e

t - r a

t i o s .

T h e u p p e r - b o u n

d c r

i t i c a l v a

l u e o

t h e

F - t e s

t a t t

h e

5 p e r c e n

t l e v e

l o

s i g n

i f c a n c e

i s 4 . 0

1 . T

h e c o m -

p a r a b

l e f g u r e a t

t h e

1 0 p e r c e n

t l e v e

l i s 3

. 5 2

. T h e s e c o m e

r o m

P e s a r a n e t a l .

( 2 0 0 1

, p .

3 0 0

, t a b l e C I ( i i i )

C a s e

I I I )

.


16/24


T a b l e

5 .

R e s u

l t s e m p

l o y i n g a n

t i c i p a

t e d a n

d u n a n

t i c

i p a

t e d r e a l a n

d n o m

i n a

l e

f f e c

t i v e e x c

h a n g e r a

t e

E X = a n

t i c

i p a

t e d

E X

= u n a n

t i c

i p a

t e d

E X = a n

t i c

i p a

t e d

E X = u n a n t i c

i p a

t e d

r e a

l e

f f e c

t i v e

r e a

l e

f f e c

t i v e

n o m

i n a

l e

f f e c

t i v e

n o m

i n a

l e f

f e c

t i v e

e x c

h a n g e r a

t e

e x c

h a n g e r a

t e

e x c h a n g e r a

t e

e x c

h a n g e

r a t e

P a n e l

A :

S h o r t - r u n e s t

i m a

t e s

l n

G d P

t 1

0

. 2 5

0 . 2

7

( 1 . 3

4 )

( 1 . 6

6 )

l n

G d P

t 2

0

. 4 8

( 2 . 4

2 )

l n

G d P

t 3

0

. 5 9

( 2 . 3

2 )

l n

M t

1 . 0

9

0 . 6

4

0 . 6

7

0 . 6

2

( 4 . 5

0 )

( 2 . 7

9 )

( 3 . 2

9 )

( 2 . 6

5 )

l n

M t 1

1 . 4

4

0 . 3

9

( 3 . 9

3 )

( 2 . 0

8 )

l n

M t 2

0 . 8

4

0 . 1

4

( 2 . 7

9 )

( 0 . 7

5 )

l n

M t 3

0 . 4

4

( 2 . 2

2 )

l n

G t

0 . 3

1

0 . 4

0

0 . 3

4

0 . 4

0

( 2 . 5

2 )

( 3 . 1

8 )

( 3 . 0

3 )

( 3 . 1

7 )

l n

G t 1

0

. 1 4

0 . 1

6

( 1 . 0

4 )

( 1 . 6

5 )

l n

G t 2

0 . 2

9

0 . 2

0

( 2 . 3

4 )

( 2 . 6

8 )

l n

G t 3

0 . 2

9

0 . 1

5

( 1 . 9

4 )

( 1 . 8

3 )

l n

E X t

0

. 1 0

0 . 0

4

0 . 0

1

0 . 0

3

( 1 . 2

8 )

( 0 . 5

5 )

( 0 . 2

7 )

( 0 . 4

3 )

l n

E X t

1

0 . 4

2

0 . 1

2

( 3 . 6

6 )

( 2 . 0

3 )

l n

E X t

2

l n

E X t

3


17/24

Ma June 2010 39

P a n e l

B :

L o n g - r u n e s t

i m a

t e s

C o n s t a n

t

1 0

. 7 1

0 . 5

1

4 . 7

6

0 . 6

1

( 4 . 0

5 )

( 0 . 0

4 )

( 6 . 9

1 )

( 0 . 0

4 )

l n M

0 . 1

6

2 . 4

9

0 . 2

9

2 . 8

4

( 0 . 8

5 )

( 0 . 5

8 )

( 3 . 6

2 )

( 0 . 4

6 )

l n G

0 . 3

1

1

. 6 8

0 . 3

3

2

. 1 1

( 1 . 5

4 )

( 0 . 3

0 )

( 3 . 9

6 )

( 0 . 2

6 )

l n E X

1

. 2 5

0 . 7

8

0 . 1

4

0 . 7

1

( 4 . 5

2 )

( 0 . 4

0 )

( 9 . 3

7 )

( 0 . 3

0 )

P a n e l

C :

D i a g n o s t

i c s

A d j u s

t e d r 2

0 . 8

4

0 . 6

3

0 . 8

5

0 . 6

4

F

1 0

. 7 7

1 . 0

9

8 . 4

2

1 . 0

6

E c M

t 1

0

. 5 3

0

. 0 5

1 . 0

2

0

. 0 4

( 3 . 6

3 )

( 0 . 4

3 )

( 5 . 3

4 )

( 0 . 3

5 )

N o

t e s :

N u m

b e r s

i n p a r e n

t h e s e s a r e

t h e a b s o

l u t e v a

l u e s o

t h e

t - r a

t i o s .

T h e u p p e r - b o u n

d c r

i t i c a l v a

l u e o

t h e

F - t e s

t a t t

h e

5 p e r c e n

t l e v e

l o

s i g n

i f c a n c e

i s 4 . 0

1 . T

h e c o m -

p a r a b

l e f g u r e a t

t h e

1 0 p e r c e n

t l e v e

l i s 3

. 5 2

. T h e s e c o m e

r o m

P e s a r a n e t a l .

( 2 0 0 1

, p .

3 0 0

, t a b l e C I ( i i i )

C a s e

I I I )

.


18/24


In Panel B, signifcant e ects o fscal and monetary policies are evident on real outputgrowth in the long run. The results are unique in the model specifcation employing theanticipated nominal e ective exchange rate. 17 In contrast to the short-run evidence, ananticipated persistent currency appreciation decreases real output growth signifcantly in

the long run. As appreciation persists over time, anticipated appreciation prompts decisionsto shrink the output supply in light o concerns about deteriorating competitiveness.

Consistently, the error-correction term in Panel C is signifcant, indicating that outputmoves toward ull equilibrium closely with the right-handside variables, including per-sistent anticipated appreciation o the exchange rate. Unanticipated transitory appreciationo the e ective exchange rate, both nominal and real, is insignifcant in the long run.Consistent with the cyclical nature o these shocks, they do not produce a long-lastinge ect and, hence, do not contribute to adjusting output toward ull equilibrium.

Summary and Conclusions

Using annual data or Iran between 1959 and 2003, we have analyzed the e ects o currency uctuations on output growth over time. Previous studies have ruled out theexpansionary e ect o currency depreciation in Iran. Because the rial is pegged to theU.S. dollar and oil is priced in dollars, a depreciation o the rial does not a ect exportcompetitiveness. On the contrary, depreciation increases the cost o imported inputs,

orcing a contraction in the output supply. Over time, Iran has expanded its nonoil exportsector, which appears to have picked up in the 1990s. This diversifcation has renewedinterest in the e ect o exchange rate depreciation on the economy. To detect the possi-bility o a structural break in the 1990s, the empirical investigation studies the e ects o exchange rate uctuations over the sample period through 2003 and a truncated sample

period that ends in 1990.Employing a new test or cointegration, we test the existence o a single-level relation,

without invoking any prior test; that is, we assume the series are stationary, nonstationary,or mutually cointegrated. Be ore the surge in nonoil exports, 195990, there is no evidenceo cointegration between output growth and the parallel rial/U.S. dollar exchange rate,either real or nominal. In contrast, over an extended sample period, 19592003, there isstrong evidence o cointegration. Further, supporting export competitiveness in the lon-ger sample, currency depreciation has a positive expansionary e ect on output growth,in both the short and long runs. A reduction in the rial value relative to the dollar in theparallel market increases export proceeds in domestic currency, stimulating an increasein the supply o nonoil exports.

The a orementioned evidence is robust to variation in the exchange rate measurement.Using real and nominal measures o the e ective exchange rate, currency appreciation iscontractionary in the long run and expansionary in the short run. The latter channel sup-ports the e ects o currency appreciation in reducing the cost o imports and expandingthe output supplied in the short run. In the long run, however, currency appreciation de-creases demand, orcing a long-lasting output contraction. The demand channel respondsmore to uctuations in the real e ective exchange rate, as opposed to the rial/U.S. dollarparallel market rate, which in uences domestic suppliers decisions.

Interactions between the exchange rate and real activity may vary with agents ore-casts. Recent theoretical developments have di erentiated the e ects o exchange rate

movements on demand and supply, based on agents orecasts. To urther elaborate onthe channels governing demand and supply adjustments, we decompose movements


19/24

Ma June 2010 41

in the exchange rate into anticipated and unanticipated components. Theory predicts amore dominant e ect o anticipated exchange rate movements on the supply side o theeconomy. In contrast, unanticipated exchange rate shocks determine both demand andsupply decisions.

Using the rial/U.S. dollar parallel exchange rate and the e ective exchange rate, theevidence di erentiates the e ects o anticipated and unanticipated exchange rate move-ments. Anticipated rial/U.S. dollar depreciation increases the output supply, which issignifcant on output growth in the short and long runs. Consistently, anticipated e ec-tive exchange rate appreciation decreases the output supply in the long run. Deteriorat-ing competitiveness is a dominant actor that gauges suppliers decisions, resulting ina reduction in output growth with respect to anticipated currency appreciation in thelong run.

Regardless o the exchange rate measure, unanticipated exchange rate shocks are in-signifcant in determining real output growth, in both the short and long runs. Consistentwith theory, con icting channels on the demand and supply sides render the e ects o these shocks insignifcant on output growth in the short run. Moreover, shocks disap-pear over time, canceling their e ects on the output adjustment toward ull equilibriumin the long run.

Overall, the evidence supports the importance o exchange rate uctuations to outputgrowth in Iran, in both the short and long runs. Attesting to the countrys success indiversi ying its export base, uctuations in the exchange rate have determined demandcompetitiveness and suppliers decisions over the long sample period that accounts orthe surge in nonoil exports in the 1990s.

For policy implications, the evidence points to the need to maintain a higher degreeo exibility in managing the exchange rate, as oil-producing economies take serious

steps toward diversi ying their export bases. The recent surge in the price o oil hasdemonstrated the in ationary risk o accumulated reserves under a pegged exchangerate. Exchange rate exibility does not necessarily imply abandoning the peg in avoro a oating exchange rate. Countries in transition need to manage the exchange ratemind ul o the implications o changes in the exchange rate on the demand and supplysides o the economy, in both the short and long runs. Some trade-o s might be neces-sary, but managing a exible exchange rate gradually, moving toward achieving stabilityin the real e ective exchange rate, may strike the necessary balance.

Notes

1. For a discussion o choices o exchange rate regimes with respect to demand and supplyshocks impinging on the economic system, see Beidas and Kandil (2008).

2. In addition to petrochemical products, the list o nonoil exports includes carpets, resh anddried ruits, pistachio, raisins, all kinds o skin and leather, caviar, casings, gum tragacanth, cumin,cotton, metallic mineral ores, detergents and soaps, chemical products, ootwear, readymade clothes,

abric, cement, stones and construction materials, transportation vehicles and their spare parts,aluminum and articles thereo , copper and copper bar, cast iron, iron and steel, and others.

3. The results are robust with respect to a modifcation that accounts or other variables in themodel. Because we are interested in the interaction between domestic policies and the exchangerate, we pre er to report the results employing the money supply and government spending.

4. This bound-testing approach has a special appeal in models that include the real exchangerate. In countries where purchasing power parity (PPP) holds, the real rate is stationary, or I(0),

where other variables are I(1). For Iran, Bahmani-Oskooee (1993) has shown that, using the parallelexchange rate, PPP holds, implying a stationary real exchange rate.


20/24


5. Following Bahmani-Oskooee (1996), a dummy variable is included in the model to account orthe e ect o the Islamic revolution in 1979. It takes a value o zero or be ore 1979 and one therea ter.It is excluded i the coe fcient is not signifcant.

6. The estimated negative intercept in Table 1 (7.3) indicates that, absent other actors in theempirical model, real output growth would have been negative. This evidence supports the role o stabilization policies (monetary, fscal, and the exchange rate) to stimulate real output growth. More-over, exchange rate depreciation had a positive e ect, sustaining real growth in the long run.

7. See, e.g., Bahmani-Oskooee and Ardalani (2006). The negative and signifcant coe fcientobtained or ECM t1 is also indicative o the exogeniety o the right handside variables in the model.However, it is still possible to have eedback between the dependent and independent variables.Bahmani-Oskooee and Tanku (2006, p. 261) have argued that the issue cannot be too serious, mostlybecause lagged values in the model could be considered as instruments or current values.

8. Since 2002, the exchange rate regime is characterized by a de acto crawling peg. For moreon the oreign exchange policy in Iran, see Bahmani-Oskooee (2005).

9. The positive e ect o monetary policy on output growth in the long run indicates rigidity in wageor price adjustment in the long run. Institutional rigidity may have inter ered with nominal adjustment,necessitating that output growth absorbs the long-run e ects o monetary policy. Furthermore, thenonneutral e ects o monetary and fscal policy indicate the relevance o oil resources to sustainingreal growth in Iran. Oil resources contribute to oreign reserves and monetary expansion. Moreover,these resources support government spending on in rastructure and investment, which are evidentlyan important source o real growth in the long run.

10. Given the maximum lag structure in the model estimation ( our lags), the e ective period orestimation is 19702003. Hence, it is impossible to truncate the sample period into pre- and post-1990periods, using the e ective exchange rate.

11. Fluctuations in the parallel rialdollar exchange rate are likely to determine supply decisions.This is because the domestic value o exports varies with this value. It is unlikely that uctuations inthe exchange rate will a ect the demand side, as exports are recorded according to the o fcial ratein the world market.

12. The nominal e ective exchange rate is measured such that an increase indicates appreciation.An appreciation has a negative signifcant e ect on real growth, which supports the negative e ect o

exchange rate appreciation on nonoil exports. The bilateral exchange rate is measured such that anincrease indicates depreciation. The positive sign indicates that depreciation stimulates incentives toproduce nonoil exports. Hence, the evidence is consistent or both measures o the exchange rate.

13. Using the nominal e ective exchange rate, the coe fcient o the error correction term, 1.62,indicates aster adjustment to ull equilibrium. In contrast to the evidence using the real e ectiveexchange rate, real growth would have converged more quickly toward ull equilibrium. The speedo adjustment, albeit unrealistic, describes convergence toward ull equilibrium under a hypotheticalscenario that leaves out in ation.

14. The advent o the Islamic revolution in 1979 marks a break point in the our measures o theexchange rate. The evidence, however, does not support a signifcant structural break in the 1990s.Hence, the varying e ects o the exchange rate on output growth in the longer sample period areprimarily the result o a structural break in the export composition toward more nonoil exports inthe 1990s.

15. In generating the anticipated exchange rate, we assume that expectations are adaptive anddepend on the past behavior o the exchange rate, with more weight assigned to the recent past andless weight to the distant past. To capture this notion, we regress the exchange rate on our lags anduse the ftted values as the anticipated component and residuals as the unanticipated component. Thisis applied to the real or nominal rialdollar rate, as well as the real or nominal e ective exchangerate. Because the lagged variables o other explanatory variables are also included in the regressionto per orm the upper-bound F- test, it is impossible to invoke a rational expectation assumption in

orecasting the exchange rate, to avoid multicollinearity between the anticipated exchange rate andother regressors in the model.

16. In light o institutional rigidity, the large long-run e ect o monetary policy re ects the cu-mulative short-run e ects that build up in the long run, as nominal rigidity persists. This evidence isunique in the model with unanticipated exchange rate shocks, indicating a high correlation between

the anticipated exchange rate and the growth o the money supply.17. This evidence indicates high correlation between fscal and monetary variables and the pricelevel, embedded in the real e ective exchange rate.


21/24

Ma June 2010 43

Re erence

Agenor, P.R. 1991. Output, Devaluation, and the Real Exchange Rate in Developing Coun-tries. Weltwirtschaftliches Archive 127, no. 1: 1841.

Alexander, S.S. 1952. E ects o a Devaluation on a Trade Balance. IMF Staff Papers 2, no. 2:

263278.Bahmani-Oskooee, M. 1993. Black Market Exchange Rates Versus O fcial Exchange Rates inTesting PPP: An Examination o the Iranian Rial. Applied Economics 25, no. 4: 465472.

. 1995. Real and Nominal E ective Exchange Rates or 22 LDCs: 1971I1990 IV. Ap- plied Economics 27, no. 7: 591604.

. 1996. Source o Stag ation in an Oil-Exporting Country: Evidence rom Iran. Jour-nal of Post Keynesian Economics 18, no. 4: 609620.

. 2005. History o the Rial and Foreign Exchange Policy in Iran. Iranian Economic Review 10 (Fall): 120.

Bahmani-Oskooee, M., and P. Anker. 2001. On the Relationship Between the Value o theMark and German Production. Applied Economics 33, no. 12: 15251530.

Bahmani-Oskooee, M., and Z. Ardalani. 2006. Exchange Rate Sensitivity o U.S. Trade Flows:Evidence rom Industry Data. Southern Economic Journal 72, no. 3: 542559.

Bahmani-Oskooee, M., and A. Tanku. 2006. Black Market Exchange Rate, Currency Substitu-tion and the Demand or Money in LDCs. Economic Systems 30, no. 3: 249263.

Bahmani-Oskooee, M.; S. Chomsisengphet; and M. Kandil. 2002. Are Devaluations Contrac-tionary in Asia? Journal of Post Keynesian Economics 25, no. 1: 6981.

Bahmani-Oskooee, M., and A. Mirzaie. 2000. The Long-Run E ects o Depreciation o theDollar on Sectoral Output. International Economic Journal 14, no. 3: 5161.

Barbone, L., and F. Rivera-Batiz. 1987. Foreign Capital and the Contractionary Impact o Cur-rency Devaluation, with an Application to Jamaica. Journal of Development Economics 26,no. 1: 115.

Beidas, S., and M. Kandil. 2008. Setting the Stage or a National Currency in the West Bank and Gaza: The Choice o Exchange Rate Regime. Middle East Business and Economic

Review 20, no. 2: 125.Connolly, M. 1983. Exchange Rates, Real Economic Activity, and the Balance o Payments:

Evidence rom the 1960s. In Recent Issues in the Theory of the Flexible Exchange Rate, ed.E. Classen and P. Salin, pp. 129143. Amsterdam: Elsevier.

Cooper, R.N. 1971. Currency Devaluation in Developing Countries. In Government and Eco-nomic Development, ed. G. Ranis, pp. 472515. New Haven, CT: Yale University Press.

Donovan, D.J. 1982. Macroeconomic Per ormance and Adjustment Under Fund-SupportedPrograms: The Experience o the Seventies. IMF Staff Papers 29, no. 2: 171203.

Edwards, S. 1989a. Real Exchange Rates, Devaluation, and Adjustment. Cambridge, MA: MITPress.

. 1989b. Exchange Controls, Devaluations, and Real Exchange Rates: The Latin Ameri-can Experience. Economic Development and Cultural Change 37, no. 3: 457494.

Engle, R.F, and C.W.J. Granger. 1987. Cointegration and Error Correction: Representation,Estimation, and Testing. Econometrica 55, no. 2: 251276.

Gyl ason, T. 1987. Credit Policy and Economic Activity in Developing Countries with IMFStabilization Programs. Princeton Studies in International Finance no. 60, Princeton Univer-sity, Princeton, NJ.

Gyl ason, T., and O. Risager. 1984. Does Devaluation Improve the Current Account? Euro- pean Economic Review 24, no. 1: 3764.

Gyl ason, T., and M. Schmid. 1983. Does Devaluation Cause Stag ation? Canadian Journalof Economics 25, no. 1: 3764.

Ho maister, A.W., and C.A. Vgh. 1996. Disin ation and the Recession-Now-Versus Reces-sion-Later Hypothesis: Evidence rom Uruguay. IMF Staff Papers 43, no. 2: 355394.

Kamin, S.B. 1988. Devaluation, External Balance, and Macroeconomic Per ormance in Devel-oping Countries: A Look at the Numbers. Princeton Essays in International Finance no. 62,Princeton University, Princeton, NJ.

Kamin, S.B., and J.H. Rogers. 1997. Output and the Real Exchange Rate in Developing Coun-tries: An Application to Mexico. International Finance Discussion Paper no. 580, Board o Governors o the Federal Reserve System, Washington, DC.


22/24


Kandil, M. 2000. The Asymmetric E ects o Exchange Rate Fluctuations: Theory and Evi-dence rom Developing Countries. Working Paper no. 00/184, International Monetary Fund,Washington, DC.

. 2004. Exchange Rate Fluctuations and Economic Activity in Developing Countries:Theory and Evidence. Journal of Economic Development 29, no. 1: 85108.

Kandil, M., and A. Mirzaie. 2002. Exchange Rate Fluctuations and Disaggregated EconomicActivity in the U.S.: Theory and Evidence. Journal of International Money and Finance 21,no. 1: 131.

. 2003. The E ects o Dollar Appreciation on Sectoral Labor Market Adjustments:Theory and Evidence. Quarterly Review of Economics and Finance 43, no. 1: 89117.

. 2005. The E ects o Exchange Rate Fluctuations on Output and Prices: Evidencerom Developing Countries. Journal of Developing Areas 38, no. 2: 189219.

Krueger, A. 1978. Foreign Trade Regimes and Economic Development: Liberalization Attemptsand Consequences. Cambridge, MA: Ballinger.

Krugman, P., and L. Taylor. 1978. Contractionary E ects o Devaluation. Journal of Interna-tional Economics 8, no. 3: 445456.

Nunnenkamp, P., and R. Schweickert. 1990. Adjustment Policies and Economic Growth inDeveloping Countries: Is Devaluation Contractionary? Weltwirtschaftliches Archiv 126:474493.

Pesaran, M.; Y. Shin; and R. Smith. 2001. Bound Testing Approaches to the Analysis o LevelRelationships. Journal of Applied Econometrics 16, no. 2: 289326.

Rogers, J.H., and P. Wang. 1995. Output, In ation, and Stabilization in a Small Open Econo-my: Evidence rom Mexico. Journal of Development Economics 46, no. 2: 271293.

Santaella, J.A., and A.E. Vela. 1996. The 1987 Mexican Disin ation Program: An Ex-change Rate-Based Stabilization? Working Paper no. 24, International Monetary Fund,Washington, DC.

Solimano, A. 1986. Contractionary Devaluation in the Southern Cone: The Case o Chile. Journal of Development Economics 23, no. 1: 135151.

Taye, H.K. 1999. The Impact o Devaluation on Macroeconomic Per ormance: The Case o Ethiopia. Journal of Policy Modeling 21, no. 4: 481496.

Appendix: Data Defnition and Sources

All data are annual over the 19592003 period and come rom the ollowing sources:

a. International Financial Statistics o the IMF (International Monetary Fund).b. Direction o Trade Statistics o the IMF.c. Central Bank o Iran.d. Bahmani-Oskooee (2005).e. World Economic Outlook.

Variables

GDP = Real GDP. Nominal GDP in billions o rials (source a) is de ated by the only priceindex availablethat is, CPI (consumer price index) (source: International FinancialStatistics o the IMF) to generate real GDP.

M 2 = Real M2. Defned as M1 plus quasi-money in billions o rials. This nominal fgureis de ated by CPI to arrive at real fgures. Data come rom the International FinancialStatistics o the IMF and the Central Bank o Iran.

G = Real government spending in billions o rials. Again, nominal fgures are de atedby CPI to arrive at real fgures. Data come rom the Central Bank o Iran.

NE = Nominal rialdollar parallel market exchange rate defned as number o rials perU.S. dollar. Data come rom Bahmani-Oskooee (2005).


23/24

Ma June 2010 45

RE = Real bilateral rialdollar rate defned as ( P US NE / P Iran ), where NE is defned aboveand P is measured by CPI in the United States and Iran. The CPI data come rom theInternational Financial Statistics o the IMF.

REFEC = Real e ective exchange rate. The data on the CPI o trading partners come rom

source b. The bilateral exchange rate data between rial and each partners currency arenot directly available. Following Bahmani-Oskooee (1995), they are generated usingrialdollar and partners currencydollar rates. Whereas the data on parallel rialdollarrate come rom source b, the data on the partners currency-dollar rates come romsource a. Import shares come rom source b.

NEFEC = Nominal e ective exchange rate. Relevant data in constructing this variableare the same as those used or constructing REFEC, excluding price levels.

To order reprints, call 1-800-352-2210; outside the United States, call 717-632-3535.


24/24

Copyright of Emerging Markets Finance & Trade is the property of M.E. Sharpe Inc. and its content may not be

copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written

permission. However, users may print, download, or email articles for individual use.

oil and exchange rate

Documents