VI Sustainability of the Exchange Rate Regime
- Howard Handy
- Published Date:
- May 1998
Over the past six years the Egyptian pound has moved in a 3 percent range vis-à-vis the dollar, and any upward or downward pressure on the exchange rate has been accommodated through passive intervention. There is no doubt that Egypt has prospered under this hard currency policy, with the exchange rate providing a nominal anchor. The benefits of this strategy include low inflation, disciplined financial policy, and a stable external environment. Capital inflows have bolstered long-term growth prospects by increasing investment and confidence. At the same time, the real effective exchange rate has appreciated by about 30 percent during 1991–96 owing to substantial inflation differentials in the early part of the period with partner countries. Clearly, this real appreciation would not represent an overvaluation if it resulted from a recovery from an undervalued exchange rate, if productivity growth in the tradable sector was substantially larger than compared with trading partner countries, or if the appreciation reflected changes in other fundamentals, such as those inducing higher confidence in the Egyptian market.
To address this issue, we estimate the equilibrium real exchange rate for the Egyptian pound, that is, the level of the real exchange rate that is consistent with the internal and external equilibrium. We find that before 1991 the real exchange rate was significantly overvalued relative to its estimated equilibrium level but has since adjusted to a level broadly consistent with the equilibrium rate.
Nominal and Real Exchange Rate Developments
Movements in the exchange rate in the 1987–91 period (prior to pegging of the Egyptian pound to the U.S. dollar) are closely linked to moves to liberalize the exchange regime.30 Until May 1987, the inter-bank foreign exchange market was organized into two official pools, the central bank pool and the commercial bank pool, each handling different foreign exchange transactions. The official exchange rates were set by the Egyptian authorities (at LE 0.7and LE 1.36 per U.S. dollar, respectively) and did not reflect market forces. In addition, a formally illegal nonbank free market was officially tolerated. In May 1987, a new bank foreign exchange market was introduced. The rate was initially set at LE 2.165 to the U.S. dollar, well below the pool rate, and subsequently allowed to slide, to reach LE 3.0 at the end of 1990. The central bank pool rate was devalued a number of times to reach LE 2.0 on July 1. 1990. The old commercial bank pool ceased to exist in March 1989.31
To simplify the exchange rate system and ensure a competitive exchange rate, the multiple exchange rate system was replaced by a temporary dual exchange system consisting of a primary market and a secondary (free) market in February 1991, and subsequently these markets were unified in October 1991. Since then the Egyptian pound has been freely traded in a single exchange market with the authorities intervening to maintain the rate in a tight band against the U.S. dollar (Figure 14).
The administered exchange rates led to specific commodity shortages prior to liberalization in 1987, indicating that the Egyptian pound was overvalued. The real effective exchange rate depreciated sharply during 1987, as an increasing share of transactions went through the new (depreciated) exchange market. In 1988 and 1989, a modest real appreciation occurred as administrative adjustments tended to lag behind inflation; subsequently, they were reversed in 1990/91 as the exchange markets were unified. In the early years of pegging to the U.S. dollar with a market-determined exchange rate, there was a substantive real appreciation. Since 1993, the real appreciation has been very modest as domestic inflation fell to single digits.
Figure 14.Nominal and Real Effective Exchange Rate, January 1987-December 1996
Equilibrium Real Exchange Rate
Macroeconomic stabilization together with liberalization of the external sector may pose a challenge to external competitiveness. As stabilization is often associated with fiscal tightening and high real interest rates, and if the capital market is sufficiently liberal, this will usually induce substantial capital inflows, putting upward pressure on the real exchange rate. This has been Egypt’ recent experience, as it has also been for Chile, Malaysia, Mexico, and Argentina for example.32 In these circumstances, the policymaker is faced with the option of intervening in the exchange market and tightening fiscal policy to counter act the upward pressure to maintain competitiveness or, alternatively, of leaving the policy unchanged, thereby raising the specter of an unsustainable current account deficit down the road. For Egypt, this policy dilemma translates into a choice between keeping the current peg or introducing flexibility in the exchange rate, hoping for a market driven correction to follow in the medium term.
Accordingly, it is important to determine whether a real exchange rate appreciation is consistent with the fundamentals of the economy after the process of liberalization or if it represents an overvaluation—defined here as a short-term misalignment of the exchange rate induced by fiscal or monetary policy.33 If the exchange rate is in line with fundamentals, there is no reason to believe that a move to a more flexible exchange rate would lead to the desired depreciation. On the other hand, if the real exchange rate is misaligned, a flexible exchange rate would tend to reduce the misalignment over the medium term.
The description of movements in the real effective exchange rate in the previous section highlights the need for a reference rate that is consistent with economic fundamentals, the equilibrium real exchange rate (ERER).
The ERER can be defined as the level of the real exchange rate that is consistent with internal equilibrium (in terms of the goods and labor market) and external equilibrium (in terms of a sustainable current account). The relatively large theoretical and empirical literature on the topic can be classified into two categories.-34 The first comprises theoretical works that identify the ERER as the exchange rate compatible with a non accelerating inflation rate of unemployment (NAIRU) (Frenkel and Razin, 1987;Clark and others (1994)) or with a normative target for the current account (Williamson, 1994). It also comprises numerous empirical studies aiming at the identification of the ERER for the major industrial economies, such as those implemented by the IMF on the basis of the multilateral exchange rate model (MEMR) or multicurrency macro econometric model (Multimod), as well as those done by Stein (1995) on the estimation of the U.S. dollar equilibrium exchange rate.
In the second category, the seminal work by Edwards (1989 and 1994) represents the first substantial endeavor to build an equilibrium exchange rate model specifically for developing countries. Within this context, the ERER is defined as the relative price of tradables to nontradables that, ceteris paribus, results in the simultaneous attainment of internal and external equilibrium. Elhadawi (1994) develops a simplified version of Edwards’ model that requires a smaller set of fundamental variables, and using this framework provides estimates of the ERER for Chile, Ghana, and India, In estimating Egypt’ ERER, we will follow Edwards’ methodology.
Edwards’ (1989) basic model is an intertemporal general equilibrium model of a small open economy in which both tradables and nontradables are exchanged. Time is limited to two periods; identifying the short-and long-run behavior of the economy, the model has also been extended to an infinite horizon without substantial differences. Internal equilibrium is defined in the model as the clearing of all nontradable markets (static equilibrium). External equilibrium is attained when the net present value of future current accounts is non-negative, given the level of exogenous long-run capital inflows (dynamic equilibrium). These two equilibrium conditions identify a unique ERER. Agents in the model are endowed with perfect foresight so that they will immediately respond to an unsustainable current account by changing their consumption and investment decisions.
Box 2.Determinants of the Real Exchange Rate in Edwards’ Model
In Edwards’ empirical study of 33 developing countries, he identifies the following set of fundamental variables affecting the equilibrium in the real exchange rate (ERER) (FUNDu):
An improvement in the terms of trade (TOT), defined as the ratio of the world price of a country’ exports over the world price of its imports, will have a positive impact on the current account and thus lead to an appreciation of the ERER.
A shift toward government consumption of nontradables (GCN) vis-à-vis tradables will improve the current account and thus lead to an appreciation of the ERER.
A liberalization of controls over capital flows (KCON) could either improve or worsen the capital account, depending on the interest rate differential between the domestic and the world economy prior to the liberalization (i.e., depending on whether the controls acted, on balance, to deter inflows or outflows). If the removal of such controls leads to a higher (lower) level of capital inflows, the ERER would appreciate (depreciate).
A relaxation of the severity of trade restrictions and exchange controls (XCON) usually leads to an increase of imports, a worsening of the current account, and thus a depreciation of the ERER.
Faster technological progress (TECH) is usually reflected in stronger productivity growth for the tradable sector (Balassa-Samuel son effect) and thus leads to an appreciation of the ERER.
An increase in the ratio of investment to GDP (1NV) will increase absorption, worsen the current account, and lead to a depreciation of the ERER.
In addition to the fundamental variables outlined above, Edwards uses the following proxies for policy variables (Z,):
The excess supply of domestic credit (EXC), defined as the increase in domestic credit that is unmatched by higher growth in the economy. Under a flexible exchange rate, excessive monetary expansion will lower interest rates, boost the domestic demand for nontradables, and thus induce an appreciation of the real effective exchange rate (REER). Under a fixed exchange rate, an excessive monetary expansion would be immediately reversed by a capital outflow, leaving the REER unchanged.
The ratio of fiscal deficit to lagged high-powered money (DEH). Under a flexible exchange rate, an increase in the fiscal deficit relative to the monetary base in the previous period (loose fiscal policy) will increase domestic demand for nontradables and thus lead to an appreciation of the REER. Under a fixed exchange rate, loose fiscal policy will initially boost domestic demand with the upward pressure on interest rates dampened by capital inflows and no impact on the REER. In the long run, the higher demand for nontradables will put upward pressure on inflation and thus lead to a REER appreciation.
For the specific case of Egypt, this data set should be augmented by the following fundamental variables in the estimation of the real exchange rate:
An increase in the debt-service ratio (DEBT), which includes private and official transfers, will worsen the sustainability of the current account and thus lead to a depreciation of the ERER.
A Middle East conflict dummy (MECON) spanning the period August 1990 to March 1991. A negative exogenous shock like the Middle East conflict of 1990–91 would put downward pressure on the REER.
The variable for the nominal depreciation (NDEP) will be measured by changes in the nominal effective exchange rate (NEER).
The kernel of Edwards’ empirical analysis is to determine the equilibrium real exchange rate by disentangling fundamental changes in the level of the actual rate from temporary influences brought about by nominal exchange rate shifts as well as monetary and fiscal policy deviations. Two equations describe the long-run behavior of the exchange rate as the function of fundamentals and the short-run influences on the dynamics of the real exchange rate. Both can be combined in the following reduced form equation for the real exchange rate:
where (FUNDit) is a vector of fundamental variables affecting the equilibrium real exchange rate (Zt - Zt*) is a vector of the deviation of policy variables from their equilibrium, and (NDEPZt) is the nominal exchange rate depreciation. The variables are defined in Box 2. The estimates of the parameters γi, can provide an estimate of the equilibrium real exchange rate and θ indicates the speed of adjustment to equilibrium. In deriving the equilibrium rate, it is assumed that the long-run elasticities of the nominal depreciation and the policy variables are zero, that is, all policy and nominal factors do not affect the equilibrium exchange rate.
Determinants of the Real Exchange Rate
While the data set outlined in Box 2 is ideal for the empirical estimation of Edwards’ model, it is clear that operationally some of these variables are not readily available. For Egypt, the limited availability and frequency of the required data necessitates estimation using total government consumption as a proxy for government consumption of non tradable goods, the lagged capital account balance as a percentage of GDP (as a proxy for controls over capital flows), and a dummy variable to account for the trade and exchange liberalization in January 1991, when most exchange restrictions were lifted. Technological progress is based on total factor productivity estimates derived by Bisat, El-Erian, and Helbling (1997). Six out of 11 variables35 in the data set are only available on an annual basis, compared with a monthly basis for the remainder; it made sense therefore to interpolate the monthly variables linearly, where necessary.36 The most important variables are presented in Figures 15–18.
Figure 15.Terms of Trade
Sources: IMF, World Economic Outlook database.
Figure 16.Capital Account Balance
Source: Data provided by the Egyptian authorities; and IMF staff estimates.
Figure 17.Estimates of Total Factor Productivity
Source: Bisat, El-Erian, and Helbling (1997).
Figure 18.Debt-Service Ratio
Sources: Data provided by the Egyptian authorities; and IMF staff estimates.
To choose the appropriate estimation procedure for equation (3), we test first for the stationarity of the fundamental variables using the augmented Diekey-Fuller test and then proceed with the appropriate estimation procedure. All variables, except for the terms of trade variable (TOT), can be considered stationary in first differences. For the variable TOT, we are unable to reject either the I(0) or I(1) process, possibly indicating that the variable is fractionally integrated. The difference stationarity of the real effective exchange rate is consistent with most other empirical studies of the real exchange rate.
The presence of a fractionally integrated process does not allow employment of standard cointegration techniques. We hence adopt Pesaran and Shin’(1995) method based on an augmented autoregressive distributed lag procedure. Augmented ARDL estimates are asymptotically consistent and provide asymptotically valid inferences on the long-run result seven in the presence of fractionally integrated processes.
We first find evidence of a long-run relation between the fundamental variables.37 Second, through the specification search procedure we eliminate theXCON and INV variables from the data set as the coefficients on these variables were insignificant.38 Finally, we apply the augmented ARDL procedure to derive short-and long-run estimates of equation (3).39
Table 14 presents the short-term estimates. Overall, the table shows that all of the estimates on the fundamental variables and the nominal depreciation have the expected sign and are significant at the 99th percentile. The error correction model estimate (ecm) is highly significant and supports the significance of the cointegration procedure.
|Terms of trade (TOT)||0.2629||3.7632|
|Government consumption in percent of GDP (GCN)||0.1447||2.3067|
|Lagged capital account balance in percent of GDP (KCON-1)||-1.3170||-3.3135|
|Technological innovation (TECH)||0.3773||1.6726|
|Middle East conflict of 1990–91 (MECON)||-0.0114||-3.1354|
|Debt-service ratio (DEBT)||-0.1592||-1.9683|
|Nominal depreciation (NDEP)||0.3502||9.9193|
|Ratio of fiscal deficit to H-Money (DEH)||0.0139||0.5597|
|Excessive credit (EXC)||-0.0012||-0.0977|
|Error correction model variable (-1)||-.20818||-3.9329|
|Sample size = 119 observations||R2 = 0.6463||s = 0.00714|
The proxies for monetary (EXC) and fiscal policy(DEH) perform poorly in the estimation however. The coefficients on DEH and EXC are not statistically different from zero and the coefficient on EXCis not consistent with Edwards’ model. The results of these variables are probably affected by the de facto change in exchange rate regime during the sample period, and thus provide limited insight into the effects of monetary and fiscal policy. Given Edwards’ assumption that their elasticity in the long run is zero, the lack of significance will not actually affect our derivation of the ERER below.
It is also worth highlighting that the estimated coefficient on the lagged real effective exchange rate (REER) implies a relatively slow speed of adjustment to shocks in the fundamental variables. In fact, a positive unitary shock would be reflected in the equilibrium rate by 50 percent after 5 months, by 75 percent after 8 months, and by 90 percent after 11 months. While this may seem surprising, this coefficient is well within the range of results obtained by Edwards’ (1989) and Elbadawi (1994) for other developing countries.40
Table 15 gives the long-run estimates of the fundamental variables, given the underlying economic assumption that the long-run elasticities on the nominal depreciation and the policy variables are zero. The derivation of the long-run estimates using the ARDL procedure is consistent with Edwards’ method outlined in the previous section. The results in Table 15 are encouraging as they show that all the long-run coefficients are statistically significant at the 99th percentile. We can therefore have sufficient confidence in the statistical significance of the ERER, as it is derived from these estimates.
|Terms of trade (TOT)||1.2633||7.1498|
|Government consumption in percent of GDP (GCN)||0.6951||2.3839|
|Lagged capital account balance in percent of GDP (KCON-1)||-6.3263||-4.3982|
|Technological progress (TECH)||1.8125||2.2885|
|Middle East conflict of 1990–91 (MECON)||-0.0545||-3.1976|
|Debt-service ratio (DEBT)||-0.7645||-2.1328|
Derivation of Equilibrium Real Exchange Rate
The coefficients provided in Table 15 are used in the derivation of the equilibrium real exchange rate. Following Edwards’, we calculate 12-month moving averages of the fundamental variables so as to smooth out temporary volatility.41
Figure 19 presents the time series of the REER, the NEER and the estimated ERER (marked with dots) for the period January 1988 to December 1996. The confidence interval around the ERER is estimated to be equal to ±1.0026. When the actual rate is above (below) the equilibrium rate, it indicates that the real exchange rate is overvalued (undervalued), because of temporary effects (the monetary or fiscal policy stance). Figure 20 presents the difference between the REER and the ERER for the same period.
Figure 19.Actual Versus Equilibrium Real Effective Exchange Rate, January 1988-December 1996
Source: IMF staff estimates.
The results shown in Figures 19 and 20 are quite striking: in the first part of the period up to the beginning of 1991 the REER is substantially misaligned with the equilibrium rate; since then, the REER has moved closer to equilibrium, and in spite of a subsequent overvaluation in the second half of 1995 and the beginning of 1996, at the end of 1996 the difference between the actual and its equilibrium rate is about 7 percent. This suggests that Egypt’ real exchange rate has on average moved closer to its equilibrium rate in the past six years. The two sharpdips in the time series of the ERER represent the estimated impact of the Gulf war from August 1990 to March 1991.
Figure 20.Difference Between Actual Equilibrium Real Effective Exchange Rate, January 1998-December 1996
Source: IMF staff estimates.
The source of this change in the relationship between the actual and equilibrium real exchange rate is due to the large appreciation of the ERER in the period 1991–95. To better understand the sources of this appreciation, Table 16 presents the breakdown in the contribution of the different fundamental variables during the last six years of the sample. The largest contribution by far is that of the reduction in the debt-service ratio (DEBT), accounting for 122 percent of the increase in the ERER. The second largest contribution is technological innovation (TECH) accounting for 27 percent of the increase, followed by the impact of the conflict (MECON) 8 percent. The other variables contributed to a depreciation in the ERER, with the worsening of the terms of trade (TOT) accounting for a 28 percent decrease, capital controls accounting for a 20 percent decrease, and finally the scaling back of government consumption of non tradables accounting for a 9 percent decrease.
|Terms of trade (TOT)||1.2633||-28.37|
|Government consumption in percent of GDP (GCN)||0.6951||-9.28|
|Lagged capital account balance in percent of GDP (KCON-1)||-6.3263||-19.57|
|Technical progress (TECH)||1.8125||26.94|
|Middle East conflict of 1990–91 (MECON)||-0.0545||8.28|
|Debt-service ratio (DEBT)||-0.7645||121.98|
Overall, these results seem to confirm the sizable impact of debt rescheduling and forgiveness on the external position of developing countries. In the case of Egypt, in the period 1991–96 the rescheduling of Paris Club debt implied a cumulative reduction in the net present value of the outstanding foreign debt stock of 55 percent, thus reducing the thus service ratio from 33 percent in 1991 to below 10 percent in1996. At the same time, the ERER appreciated by about 35 percent. Moreover, indirect support for this analysis also comes from the parallel literature on the impact of official development assistance on developing countries’ competitiveness; there it is shown that prolonged international aid, in the form of grants or concessional lending, will result in an (equilibrium) appreciation of the real exchange rate. The theoretical argument behind this result runs parallel to the impact of the reduction in a country’ debt burden indicated above.
The main conclusion from the analysis above is that the lifting of this debt overhang has had a significant impact on the external position of the economy, which is in turn reflected in the appreciation of the real exchange rate. Mongardini (1998) shows how such empirical evidence is consistent with Edwards’ model.
Summary of Results
The estimation of the equilibrium real exchange rate for Egypt suggests that while the real exchange rate was substantially overvalued at the turn of the decade (by about 20 percent above the equilibrium rate over the period between 1989 and 1991), it has since moved into closer convergence with the equilibrium rate. The overvaluation averaged less than 10 percent between 1991 and 1996, and the convergence trend brought the actual real exchange rate inline with the equilibrium line by mid-1995. Although, the REER is estimated to have appreciated by 7 percent vis-à-vis the ERER as of the end of1996, the current success in lowering inflation (4 percent during July-September 1997, a period that is outside the estimation sample) is likely to contribute to the correction of such a small deviation.
It is important to interpret these results with caution. While the econometric results are statistically significant, the derivation of the equilibrium real exchange rate is ultimately dependent upon the assumptions underlying Edwards’ model. The results should not therefore be taken as a conclusive determination of Egypt’ equilibrium real exchange rate. Rather, we intended to offer an alternative insight into the developments of Egypt’ competitiveness over the past ten years. In addition, these results should not be interpreted as suggesting an appropriate long-term growth strategy for the Egyptian economy. The theory of the equilibrium real exchange rate does not take into account the well-known benefits of a dynamic export sector, which is particularly relevant in Egypt’ case, given that non-oil exports only account for about 3’A percent of GDP.
In the course of estimating the equilibrium real exchange rate, we concluded that the Paris Club debt relief phased in during the period 1991–96 had a significant impact on Egypt’ real effective exchange rate. This result is consistent with the theoretical framework underlying the empirical investigation (see Mongardini, 1998). This is a novel result that may have important policy implications not only for Egypt but also for other developing countries that will benefit in the future from significant debt rescheduling, like those countries currently being considered for the multilateral initiative for highly indebted poor countries (HIPC).42