I. Introduction

Policy makers in developing countries often need to balance the demand for foreign exchange for imports against the desire to maintain adequate foreign exchange reserves. As a result of these conflicting objectives, their limited foreign exchange reserves, restricted access to international capital markets, many developing countries limit imports through rationing foreign exchange, as well as through tariff and other nontariff barriers. In view of the above, one should observe a tight empirical relationship between imports and the availability of foreign exchange. In recognition of this, a number of studies of imports in developing countries have posited and/or empirically verified that aggregate imports by developing countries are determined largely by the amount of foreign exchange available to them. ^2/

The relationship between imports and the availability of foreign exchange should be particularly close in the case of oil exporting economies where sizable portions of both foreign exchange receipts and national income are determined by petroleum exports. In other words, in the case of oil exporters, an increase in foreign exchange receipts represents both a significant income effect, especially for the governments of these countries, and a relaxation of the foreign exchange constraint, thus allowing for an increase in imports.

This paper provides some evidence for the existence of a long-term relationship between imports and foreign exchange earnings in the case of Iran during the period 1961/62 to 1992/93. ^3/ Unlike previous works that measure the foreign exchange constraint in terms of official foreign exchange reserves, given that official data on Iran’s foreign exchange reserves are not available for the period since 1983, the availability of foreign exchange is measured here in terms of foreign exchange earnings in the form of exports. ^4/

We begin our analysis in 1961/62, shortly after the rise of systematic and reliable national income accounting in Iran. ^5/ Chart 1 illustrates the close relationship between imports and foreign exchange receipts, as represented by total exports, for Iran. It is evident that during the period under study, Iran’s imports have been moving closely in line with its foreign exchange earnings. Following the balance of payments problems in 1961/62-1963/64, both imports and export receipts rose steadily until the surge in oil revenues and the consequent relaxation of the foreign exchange constraint in the early 1970s. Indeed, imports rose less than exports in the period of the rapid increase in oil exports, which may be attributed to a number of factors. In the period prior to the oil boom, the notional demand of imports was restricted by the availability of foreign exchange. Hence, imports could rise only commensurately with the increase in foreign exchange earnings. With the sharp rise in oil income in the early 1970s, however, imports were less constrained by the availability of foreign exchange, and the notional level of import demand might not have increased In step with foreign exchange earnings, especially given that Iran did not have the absorptive capacity, such as port facilities, for higher import levels. ^6/

Iran - Total Imports and Exports

(In millions of U.S. dollars)

Citation: IMF Working Papers 1995, 097; 10.5089/9781451851960.001.A001

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Iran - Total Imports and Exports

(In millions of U.S. dollars)

Citation: IMF Working Papers 1995, 097; 10.5089/9781451851960.001.A001

• Download Figure
• Download figure as PowerPoint slide

Iran - Total Imports and Exports

(In millions of U.S. dollars)

Citation: IMF Working Papers 1995, 097; 10.5089/9781451851960.001.A001

• Download Figure
• Download figure as PowerPoint slide

However, the oil boom, which allowed for a significant rise in imports, dissipated in the late 1970s. The close relationship between imports and exports was interrupted during 1979, the year of the Iranian revolution; while oil exports surged on account of higher prices, imports were curtailed due partly to a decline in industrial activity and government restrictions. Since 1979, overall, Iran’s export revenues diminished sharply on account of economic sanctions, the war with Iraq, and the decline in oil revenues. As a result, Iran suffered a serious balance of payments crisis to which the Iranian authorities responded by imposing trade and foreign exchange restrictions. Imports were particularly restricted following the negative shock to petroleum revenues in 1986, ^7/ but rose steadily thereafter on account of improved oil export earnings and the end of the war with Iraq in 1988, Imports rose significantly after 1990 as a result of increased foreign exchange earnings and the partial liberalization of Iran’s foreign trade regime, which entailed the removal of a number of restrictions on imports. For purposes of comparison, Chart 2: and Chart 3: show the same close relationship between imports and exports for Algeria and Venezuela, two oil-exporters which are structurally similar to Iran.

The remainder of this paper focuses on the empirical examination of the impact of foreign exchange receipts on imports in Iran. Section II provides an empirical examination of the determinants of aggregate imports in Iran, and section III considers some of the economic consequences of the effect of the dependence of imports on foreign exchange receipts.

Algeria - Total Inports and Exports

(in millions of U.S. dollars)

Citation: IMF Working Papers 1995, 097; 10.5089/9781451851960.001.A001

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Algeria - Total Inports and Exports

(in millions of U.S. dollars)

Citation: IMF Working Papers 1995, 097; 10.5089/9781451851960.001.A001

• Download Figure
• Download figure as PowerPoint slide

Algeria - Total Inports and Exports

(in millions of U.S. dollars)

Citation: IMF Working Papers 1995, 097; 10.5089/9781451851960.001.A001

• Download Figure
• Download figure as PowerPoint slide

Uenezuela - Total Imports and Exports

(in millions of U.S. dollars)

Citation: IMF Working Papers 1995, 097; 10.5089/9781451851960.001.A001

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Uenezuela - Total Imports and Exports

(in millions of U.S. dollars)

Citation: IMF Working Papers 1995, 097; 10.5089/9781451851960.001.A001

• Download Figure
• Download figure as PowerPoint slide

Uenezuela - Total Imports and Exports

(in millions of U.S. dollars)

Citation: IMF Working Papers 1995, 097; 10.5089/9781451851960.001.A001

• Download Figure
• Download figure as PowerPoint slide

Empirical Results

In order to examine the impact of foreign exchange receipts on imports in Iran, we estimated an aggregate import function that relates imports to economic activity, the relative price of imported goods, and receipts of foreign exchange. The estimated import equation takes the following form:

where M_t is real imports of goods and services (deflated by import prices); X_t is real receipts from exports of goods and services (also deflated by import prices); RP_t is relative price of imported goods; A_t is real absorption; and D is a dummy variable set equal to 1 for 1979, the year of the Iranian revolution.

It is expected that in the above model (Ml), imports are affected positively by increases in export earnings, absorption, and the proxy for import dependence, and inversely by an increase in the relative price of imports and the revolution dummy variable. ^8/ Given that oil export earnings comprise the dominant share of Iran’s foreign exchange receipts, the above import equation was also estimated using real oil export proceeds, H_t instead of total foreign exchange receipts, which is presented below as Model 2 (M2).

In the above framework, the small country assumption was invoked whereby foreign exchange receipts and the relative price of imported goods are treated as exogenous variables.

2. Regression results

Equations 1 and 2 suggest that there exists a stable long-run relationship between imports and absorption, the relative price of imports, and exports. Given that the null hypothesis of a unit root could not be rejected for the variables entering in the regression equation, the possibility of cointegration among them was tested using Johansen’s method. ^10/ The Johansen procedure is a method of testing the presence of a long-term relationship (cointegration) among a set of variables. Consider the following general vector autoregression (VAR):

$\begin{array}{cc}{Y}_t={II}_1{Y}_{t-1}+\mathrm{.}\mathrm{.}\mathrm{.}+{II}_k{Y}_{t-k}+\mu +{\epsilon }_{t,}& \left(3\right)\end{array}$

where Y_t is an mxl vector of nonstationary variables, μ is a mxl vector of constants, and ε_t is a mxl vector of Gaussian errors. Equation 3 could be rewritten in the first difference form as follows:

$\begin{array}{cc}\mathrm{\Delta }{X}_t={\mathrm{\Gamma }}_1\mathrm{\Delta }{X}_{t-1}+{\mathrm{\Gamma }}_2\mathrm{\Delta }{X}_{t-2}+\mathrm{...}+{\mathrm{\Gamma }}_{k-1}\mathrm{\Delta }{X}_{t-(k-1)}+II{X}_{t-k}+\mu +{\epsilon }_{t,}& \left(4\right)\end{array}$

where Γ_i (mxm) is defined as -I + II₁ + ... +II_i(i=1, . . . , k-1) II is defined as -I + II₁ + ... + II_k, and Δ is the difference operator. The Johansen procedure involves the maximum likelihood estimation of Equation 4, and the testing of hypotheses about the rank of matrix II, which captures the long-term relationships among the variables. There are three possible outcomes with respect to the rank of II. First, it can have full rank (r=m) which implies that the variables in Y_t are stationary. Second, it can have rank equal to zero, in which case the variables are nonstationary but not cointegrated. Third, if it has rank r less than m (r< m), then there exist r cointegration relationships among the variables. In the latter case, the matrix II can be decomposed into two mxr matrices α and β such that II = αβ’ Matrix α contains the adjustment coefficients, and matrix β contains the coefficients of the cointegration vector, representing the long-term relationship among the variables under study. The Johansen procedure involves two tests for determining the number of cointegration relationships (the rank of matrix II) among the variables in Y_t, the trace test and the maximum eigenvalue test.

Table 1.

Dickey-Fuller and Augmented Dickey-Fuller Tests for the Presence of a Unit Root in the Variables in the Import Equations, 1961/62-1992/93

^1/All variables are in logarithms.

M = real imports of goods and services.

X = real exports of goods and services.

RP = relative price of imported goods.

A = real absorption.

H = real earnings from oil exports.

^2/DF is the Dickey-Fuller Statistic.

^3/ADF(j) is the Augmented Dickey-Fuller test statistic with lag j.

Table 1.

Dickey-Fuller and Augmented Dickey-Fuller Tests for the Presence of a Unit Root in the Variables in the Import Equations, 1961/62-1992/93

Variable. 1/	DF 2/	ADF (l) 3/	ADF(2)	ADF(3)
M	−0.56	−0.41	−0.83	−0.65
X	−0.64	−0.64	−0.81	−0.92
RP	−1.91	−1.65	−1.83	−1.61
A	−1.57	−1.82	−2.04	−1.96
H	−0.69	−0.65	−0.78	−0.93
Critical . Values (α=0.05):	−3.56	−3.56	−3.57	−3.57

^1/All variables are in logarithms.

M = real imports of goods and services.

X = real exports of goods and services.

RP = relative price of imported goods.

A = real absorption.

H = real earnings from oil exports.

^2/DF is the Dickey-Fuller Statistic.

^3/ADF(j) is the Augmented Dickey-Fuller test statistic with lag j.

View Table

Table 1.

Dickey-Fuller and Augmented Dickey-Fuller Tests for the Presence of a Unit Root in the Variables in the Import Equations, 1961/62-1992/93

Variable. 1/	DF 2/	ADF (l) 3/	ADF(2)	ADF(3)
M	−0.56	−0.41	−0.83	−0.65
X	−0.64	−0.64	−0.81	−0.92
RP	−1.91	−1.65	−1.83	−1.61
A	−1.57	−1.82	−2.04	−1.96
H	−0.69	−0.65	−0.78	−0.93
Critical . Values (α=0.05):	−3.56	−3.56	−3.57	−3.57

^1/All variables are in logarithms.

M = real imports of goods and services.

X = real exports of goods and services.

RP = relative price of imported goods.

A = real absorption.

H = real earnings from oil exports.

^2/DF is the Dickey-Fuller Statistic.

^3/ADF(j) is the Augmented Dickey-Fuller test statistic with lag j.

The results of the cointegration test are provided in Table 2 under Model 1 (M1). Using Johansen’s maximal eigenvalue test, the null hypothesis of nonexistence of any cointegration vectors was rejected in favor of the existence of one cointegration vector at the 95 percent confidence level. ^11/ The long-term relationship obtained from the cointegration test was as follows: ^12/

${\mathrm{M}}_{\mathrm{t}}\mathrm{=}\mathrm{-}\mathrm{0.00}{\mathrm{RP}}_{\mathrm{t}}\mathrm{+}\mathrm{0.20}{\mathrm{A}}_{\mathrm{t}}\mathrm{+}\mathrm{0.80}{\mathrm{H}}_{\mathrm{t}}\mathrm{,}$

The results indicate that in addition to the traditional variables (relative prices and aggregate output), foreign exchange earnings have a significant impact on the level of imports. ^13/ The overriding significance of export earnings in the determination of imports suggests that such earnings acted as a binding constraint on imports in Iran.

Given that oil revenues have, on average, comprised over 90 percent of Iran’s total exports in the period under study, the above tests were conducted using oil, H, in place of total exports, X. Again, the presence of cointegration was verified; the results of the cointegration test are reported in Table 2 under Model 2 (M2). The long-term relationship obtained from the cointegration test was as follows:

Table 2.

Johansen Test Statistics for Cointegration, 1961/62-1992/93

^1/r is the number of cointegration vectors.

^2/M1: M_t = α₀ + α₁RP_t + α₂A_t + α₃X_t + α₄D.

^3/M2: M_t = α₀ + α₁RP_t + α₂A_t + α₃H_t + α₄D.

Table 2.

Johansen Test Statistics for Cointegration, 1961/62-1992/93

		Maximum Eigenvalue Test Statistics
Hypothesis: 1/
	H0	r=0	r≤1	r≤2	r≤3
	H1	r=l	r=2	r=3	r=4
Model:
	Ml 2/	51.56	12.24	8.33	0.41
	M2 3/	50.39	11.62	9.48	1.29
Critical values α (a=0.05):		27.07	20.97	14.07	3.76
		Trace Test Statistic
Hypothesis:
	H0	r=0	r≤1	r≤2	r≤3
	H1	r≥l	r≥2	r≥3	r≥4
Model:
	M1 2/	72.54	20.97	8.74	0.41
	M2 3/	72.78	22.39	10.77	1.29
Critical values (α-0.05):		47.21	29.68	15.41	3.76

^1/r is the number of cointegration vectors.

^2/M1: M_t = α₀ + α₁RP_t + α₂A_t + α₃X_t + α₄D.

^3/M2: M_t = α₀ + α₁RP_t + α₂A_t + α₃H_t + α₄D.

View Table

Table 2.

Johansen Test Statistics for Cointegration, 1961/62-1992/93

		Maximum Eigenvalue Test Statistics
Hypothesis: 1/
	H0	r=0	r≤1	r≤2	r≤3
	H1	r=l	r=2	r=3	r=4
Model:
	Ml 2/	51.56	12.24	8.33	0.41
	M2 3/	50.39	11.62	9.48	1.29
Critical values α (a=0.05):		27.07	20.97	14.07	3.76
		Trace Test Statistic
Hypothesis:
	H0	r=0	r≤1	r≤2	r≤3
	H1	r≥l	r≥2	r≥3	r≥4
Model:
	M1 2/	72.54	20.97	8.74	0.41
	M2 3/	72.78	22.39	10.77	1.29
Critical values (α-0.05):		47.21	29.68	15.41	3.76

^1/r is the number of cointegration vectors.

^2/M1: M_t = α₀ + α₁RP_t + α₂A_t + α₃X_t + α₄D.

^3/M2: M_t = α₀ + α₁RP_t + α₂A_t + α₃H_t + α₄D.

M_t = -0.00 RP_t + 0.20 A_t + 0.80 H_t,

which is very similar to the results obtained under Model 1, except for the zero coefficient for relative prices under Model 2. ^14/ The presence of a cointegrating relationship among the variables in Equation 1 implies that the short-term relationship among the variables in the import function can be represented by an error-correction model. Hence, an error-correction import function of the following form was estimated:

$\begin{array}{cc}\mathrm{\Delta }{M}_t={\beta }_1\mathrm{\Delta }R{P}_t+{\beta }_2\mathrm{\Delta }{A}_t+{\beta }_3\mathrm{\Delta }{X}_t+\delta (M_{t-1}-={\beta }_4-{\beta }_{5}R{P}_{t-1}-{\beta }_6{A}_{t-1}-{\beta }_7{X}_{t-1}-{\beta }_{8}D)+{\epsilon }_{t,}& \left(5\right)\end{array}$

where € is a white-noise error term. The error-correction term in parentheses represents the difference between actual imports and the level predicted from the long-term import function. The above equation was estimated using nonlinear least squares, and the results are reported in Table 3. Several aspects of the results are worth noting. Export revenues and relative prices play significant roles in the equation. However, absorption did not have a statistically significant effect. ^15/

The above analysis relies on the assumption that foreign exchange receipts, the relative price of imported goods, and absorption are exogenous, and that there exists no feedback from imports to these variables. However, the results of the Johansen cointegration test pointed to the possibility that the variables entering the estimated import equation might not be exogenous. ^16/ In order to further examine the exogeneity of the regressors in the import equation, the exogeneity test developed by Wu and Hausman was used. ^17/ This test entails estimating the following equation:

Table 3.

Estimation Results of the Error-Correction Import Equation, 1961/62-1992/93 ^1/

^1/The estimated equation was:

M_t = β₁ΔRP_t + β₂ΔA_t + β₃ΔX_t + δ(M_t-1 - β₄ - β₅RP_t-1 - β₆A_t-1 - β₇X_t-1 - β₈D).

^2/The numbers in parentheses are t-statistics.

Table 3.

Estimation Results of the Error-Correction Import Equation, 1961/62-1992/93 ^1/

	Coeficients	Estimates
	β1	−0.98
		(−2.84) 2/
	β2	0.09
		(0.41)
	β3	0.44
		(6.83)
	β4	1.40
		(0.98)
	β5	−0.37
		(−2.37)
	β6	0.12
		(1.78)
	β7	0.83
		(23.07)
	β8	−0.34
		(−2.40)
	δ	−0.98
		(−7.40)
Adj. R2		0.86
Maximum Log-
likelihood		30.57
LM test for
serial
correlation		0.47

^1/The estimated equation was:

M_t = β₁ΔRP_t + β₂ΔA_t + β₃ΔX_t + δ(M_t-1 - β₄ - β₅RP_t-1 - β₆A_t-1 - β₇X_t-1 - β₈D).

^2/The numbers in parentheses are t-statistics.

View Table

Table 3.

Estimation Results of the Error-Correction Import Equation, 1961/62-1992/93 ^1/

	Coeficients	Estimates
	β1	−0.98
		(−2.84) 2/
	β2	0.09
		(0.41)
	β3	0.44
		(6.83)
	β4	1.40
		(0.98)
	β5	−0.37
		(−2.37)
	β6	0.12
		(1.78)
	β7	0.83
		(23.07)
	β8	−0.34
		(−2.40)
	δ	−0.98
		(−7.40)
Adj. R2		0.86
Maximum Log-
likelihood		30.57
LM test for
serial
correlation		0.47

^1/The estimated equation was:

M_t = β₁ΔRP_t + β₂ΔA_t + β₃ΔX_t + δ(M_t-1 - β₄ - β₅RP_t-1 - β₆A_t-1 - β₇X_t-1 - β₈D).

^2/The numbers in parentheses are t-statistics.

$\begin{array}{cc}{M}_t={\beta }_0+{\beta }_{1}R{P}_t+{\beta }_2{A}_t+{\beta }_3{X}_t+{\beta }_4{Z_t}^{RP}+{\beta }_5{Z^A}_t+{\beta }_6{Z^X}_t+{\mu }_t,& \left(6\right)\end{array}$

where z^RP, z^A, and z^X are the residuals of regressions of RP_t A_t and X_t on an intercept, and lagged values of M, X, RP, and A. μ is a white-noise error term. The Wu-Hausman test evaluates the exogeneity of RP, A, and X by testing the null hypothesis β₄ = β₅ = β₆ = 0, using the F-test. This test yielded the following results:

$\begin{array}{c}{M}_t=1.11-0.74R{P}_t+0.10A_t+0.85X_t\\ \left(0.68\right)(-4.16)\left(1.17\right)\left(22.90\right)\\ \text{\hspace{0.17em}-\hspace{0.17em}}0.34{z_t}^{RP}+0.07{z_t}^A-0.41{z_t}^X\mathrm{.}\\ (-0.75)\left(0.25\right)(-4.76)\end{array}$

All variables are in logarithms, and the numbers in parentheses are t-statistics. The above results indicate that the variable Z^X is statistically significant and that, therefore, X is not an exogenous variable in the import equation. The F-statistic obtained was 7.86, which leads to a rejection of the null hypothesis of the joint exogeneity of RP, A, and X.

In order to examine further the existence of a feedback between imports and foreign exchange receipts, the Granger test was used. ^18/ Given that the series for imports and exports were nonstationary, the first difference of these variables were used in the Granger test:

$\begin{array}{cc}\mathrm{\Delta }{M}_t=\underset{i=1}{\overset{m}{\mathrm{\Sigma }}}{\alpha }_i\mathrm{\Delta }{M}_{t-i}+\underset{i=1}{\overset{m}{\mathrm{\Sigma }}}{\beta }_i\mathrm{\Delta }{X}_{t-i}+u_t& \left(7\right)\end{array}$

and

$\begin{array}{cc}\mathrm{\Delta }{X}_t=\underset{i=1}{\overset{m}{\mathrm{\Sigma }}}{\lambda }_i\mathrm{\Delta }{M}_{t-i}+\underset{i=1}{\overset{m}{\mathrm{\Sigma }}}{\gamma }_i\mathrm{\Delta }{X}_{t-i}+w_t\mathrm{.}& \left(8\right)\end{array}$

Unless the coefficients λ_i in equation 8 are equal to zero, the possibility of imports Granger-causing exports can not be rejected. Ordinary least squares estimation of equations 7 and 8 with two lags yielded the following results:

$\begin{array}{c}\mathrm{\Delta }{M}_t=0.00-0.12\mathrm{\Delta }{M}_{t-1}-0.13{\mathrm{\Delta }M}_{t-2}+0.47\mathrm{\Delta }{X}_{t-1}+0.31\mathrm{\Delta }{X}_{t-2}\\ \left(0.10\right)(-0.52)(-0.71)\left(3.12\right)\left(1.91\right)\end{array}$

$Adj\mathrm{.}-R^2=0.32,DW\mathrm{.}=1.93,\hat{\sigma }=0.23$

$\begin{array}{c}\mathrm{\Delta }{X}_t=-0.03+0.73\mathrm{\Delta }{M}_{t-1}+0.40\mathrm{\Delta }{M}_{t-2}-0.27\mathrm{\Delta }{X}_{t-1}-0.25\mathrm{\Delta }{X}_{t-2}\\ \left(-0.47\right)\left(1.90\right)\left(1.31\right)(-1.09)(-0.95)\end{array}$

$Adj\mathrm{.}{R}^2=0.05,D\mathrm{.}W\mathrm{.}=1.86,\hat{\sigma }=0.37$

As the above results indicate, the evidence for feedback between imports and export receipts, as measured by the estimates of the coefficients λ₁ and λ₂, was not statistically significant. The existence of significant feedback from imports to exports would have suggested that it is necessary to estimate import functions in the context of a simultaneous equation framework. Indeed, it is difficult to interpret the feedback from imports to exports since Iran’s oil exports have been determined somewhat exogenously in the context of the international oil industry. Prior to the ascendancy of OPEC in the early 1970s, Iran’s oil output was determined almost entirely by an international oil consortium. ^19/ Changes in the structure of the international oil industry since the early 1970s have, however, allowed for more independence for the individual oil-producing countries such as Iran with regard to oil production and exports. Although the turbulent decade of the 1980s witnessed occasional unilateral oil supply decisions, Iran’s oil output has been set largely within the OPEC framework. In view of this discussion, empirical results suggesting the endogeneity of foreign exchange receipts should be interpreted cautiously.

III. Some of the Consequences of Dependence of Imports on Foreign Exchange Receipts

The empirical results reported above confirm the central role played by foreign exchange receipts in the determination of Iran’s imports, as well as providing support for the view that the Iranian authorities have often compressed imports whenever oil exports have contracted during 1961/62-1992/93. One implication of the above results is that attempts at estimating an import demand function for Iran without due attention to the role of the foreign exchange constraint will lead to incorrect econometric estimates for income and price elasticities of imports. Thus, attempts at gauging the impact of economic stabilization and exchange rate devaluation on aggregate imports may therefore be statistically biased. Indeed, these results should be viewed cautiously in light of the small sample used.

The presence of a close relationship between imports and foreign exchange receipts also poses important implications for the assessment of the industrialization process in Iran. Paramount among these is the issue of foreign exchange requirements of domestic industry. As indicated above, during the period under study, Iran systematically pursued a policy of import substituting industrialization. ^20/ The pronounced rise in oil revenues in the early 1970s allowed for the intensification of the import substitution strategy by significantly relaxing the foreign exchange constraint facing the Iranian manufacturing sector, and for the increased importation of intermediate goods and raw materials.

The overall decline in oil revenues, as well as various other economic adversities in the 1980s, compounded by the inability to borrow adequately on the international capital markets, led to a situation whereby Iran’s manufacturing sector, which had previously been operating under a regime of soft foreign exchange constraints, was unexpectedly forced to adapt to a system of very stringent foreign exchange controls and to a notable decrease in the importation of capital goods and raw materials. Chart 4 illustrates the decline in the share of capital goods in Iran’s total imports. This points to several issues. First, it is vital for planners to save an adequate portion of the proceeds from oil exports in the form of foreign exchange reserves in order to reduce the impact of both anticipated and unanticipated fluctuations in oil revenues. Second, the heavy reliance of domestic industry on imported intermediate goods and their foreign exchange requirements impose serious externalities on the rest of the economy. Heavy foreign exchange requirements by any sector deprives other sectors of access to scarce foreign exchange. Thus, the social cost of foreign exchange would be higher than its private cost. This problem has to be explicitly allowed for in the choice of production techniques, and the extent of the dependency of domestic industry on imported intermediate goods. The fact that Iran’s imports were sharply compressed during periods of low foreign exchange earnings, and the adverse impact of this on manufacturing sector output, points to the necessity of undertaking both near- and medium-term measures to ease the foreign exchange constraint. These include corrections in the incentive system to encourage further non-oil exports, as well as steps to gain adequate access to the international capital markets in order to smooth the fluctuations in imports during periods of decline in exports.

The Share of Capital Goods Inports in Total Inports

(in percent)

Citation: IMF Working Papers 1995, 097; 10.5089/9781451851960.001.A001

• Download Figure
• Download figure as PowerPoint slide

View Full Size

The Share of Capital Goods Inports in Total Inports

(in percent)

Citation: IMF Working Papers 1995, 097; 10.5089/9781451851960.001.A001

• Download Figure
• Download figure as PowerPoint slide

The Share of Capital Goods Inports in Total Inports

(in percent)

Citation: IMF Working Papers 1995, 097; 10.5089/9781451851960.001.A001

• Download Figure
• Download figure as PowerPoint slide