The IMF's Statistical Systems in Context of Revision of the United Nations' A System of National Accounts
Chapter

6 Treatment of Exchange Rate Differentials in the National Accounts

Author(s):
Vicente Galbis
Published Date:
September 1991
Share
  • ShareShare
Show Summary Details
Author(s)
Jan van Tongeren

The United Nations' A System of National Accounts (SNA) does not explicitly deal with the question of exchange rate differentials—that is, income generated by banks and government authorities holding foreign exchange as a result of differences in the purchase and sale prices of foreign currency. This paper offers suggestions on how to deal with this question, covering three situations: stable monetary conditions, volatile changes of exchange rates over time, and multiple exchange rate regimes.

On the basis of the general principles of the SNA, and taking into account practices in a limited number of countries, the paper develops the following proposals.

  • Distinctions should be made within the exchange rate differentials among three components—bank service charges, capital gains or losses, and taxes or subsidies.

  • Global adjustments for bank service charges and taxes or subsidies under multiple exchange rate regimes should be included in the external account of the SNA, so that individual categories of transactions (exports, imports, and so on) reflect the actual transaction values, whereas the balances of the external account reflect the adjustments for differences between purchase and sale prices of foreign currency.

  • Adjustments should be applied to all external transactions, except those in kind; these adjustments should be based on an equilibrium exchange rate that would have existed if no government monetary controls had been in effect.

  • Taxes or subsidies, including the bank service charges on foreign exchange, should be allocated directly to the government sector; bank service charges on foreign exchange should be treated as payments by the government to banks for services received from banks in implementing the government's exchange rate policy.

  • No adjustments should be made for exchange rate change over time. Revenues derived from those changes should be treated as capital gains or losses.

  • Because the exchange rate differentials do not accrue only on goods and services but also on other external transactions, the adjustment for taxes or subsidies should distinguish between indirect taxes or subsidies, direct taxes (including negative direct taxes), and capital transfers.

The SNA and recent discussions during its review have focused on the role of banks in the redistribution of financial capital and on the service charges implicit in the interest flows to and from banks. Until now no concerted attention has been paid to the service charges implicit in exchange rate differentials, which derive from the difference between banks' purchase and sale prices for foreign currency. The same questions raised about the bank service charges implicit in interest flows—that is, how they are to be measured, treated, and allocated—need to be answered with regard to the role of banks in the foreign currency market. Other questions unique to exchange rate differentials also merit attention.

Until several years ago, questions about exchange rate differentials were relevant, but quantitatively unimportant, and therefore the SNA does not mention this factor explicitly. There was a difference between the purchase and sale prices of foreign currency, but the difference was very small because foreign exchange rates were generally stable over time and did not differ for transactions that took place at the same time. In addition, national accountants found it difficult to identify the differentials separately, since information was often hidden in data on other revenues of banks and allocated together with those revenues. The situation has changed considerably in recent years. Fixed exchange rates have been replaced by floating rates that, in some countries, change frequently and considerably over time, and other countries have introduced multiple exchange rate regimes that maintain different exchange rates for different transactions at the same time. As a result, substantial exchange rate differentials are affecting the profits and losses of banks or government agencies that maintain foreign currency reserves. They also influence the profits or losses and retained earnings of other nonfinancial enterprises that are engaged in the export or import of goods and services or that make or receive foreign payments for the transfer of profits, external debt, and so on.

This paper supplements two parallel papers written by staff members of the IMF: one on multiple exchange rates (Chapter 5 in this volume), and another on changes that occur in exchange rates over time (Chapter 4 in this volume). These two papers focus primarily on the exchange rate conversion of transactions in the balance of payments and external account. The present paper deals with the treatment of exchange rate differentials, not only in the external account but also in the accounts of other sectors included in the national accounts (particularly the banking sector), and with the effects on gross domestic product (GDP) and its components. The conclusions reached in the paper are in line with those of the cited papers: to use a uniform rate of exchange in circumstances of multiple exchange rate regimes (Chapter 5) and not to include changes in exchange rates over time (Chapter 4).

The remainder of the paper is organized as follows. Section I reviews country practices in foreign exchange controls and shows how they would affect the national accounts aggregates if no adjustments were made. It distinguishes among the three situations earlier identified: stable monetary conditions with minor differences between sale and purchase prices of foreign currency; considerable changes over time in the local value of foreign currency; and, finally, multiple exchange rate regimes. Section II includes proposals to adjust the national account aggregates for exchange rate differentials in order to avoid anomalies. It distinguishes among three components of exchange rate differentials that should be treated differently in the national accounts—bank service charges, taxes or subsidies, and capital gains or loses—and shows what consequences these treatments have for the valuation of stocks and flows of foreign exchange and other national account transactions that are supported by the purchase and sale of foreign currency. Section III discusses the measurement of exchange rate differentials and the derivation of a uniform or accounting exchange rate. Section IV includes a quantitative example based on adjusted 1983 national accounts data for Venezuela. This example illustrates the proposed treatment of foreign exchange differentials in the national accounts and assesses the sensitivity of GDP and other macroeconomic aggregates to different assumptions made in the national account compilation regarding the value of exchange rate differentials.

I. Foreign Exchange Policies Affecting National Account Aggregates

To date there has been limited experience with the treatment of exchange rate differentials in the national accounts. The experiences of three countries—El Salvador, Colombia, and Jamaica—have been described by Pinot de Libreros in two papers.1 A fourth experience was described by officials from Venezuela, also in the context of review of the SNA.2

The four practices, which refer to the mid-1980s, are different and can be summarized as follows.

•El Salvador had three exchange rates: an official rate, a free bank rate, and a black market rate. The government had direct control over the first two. The official rate of the U.S. dollar was the lowest and was fairly stable. The free bank rate was higher and varied according to the supply of and the demand for foreign currency for transactions that the government allowed to take place in the foreign currency market. The black market rate varied considerably over time, and the government had very little control over it.

•In Colombia, two rates were in effect: an official rate controlled by the Banco de la República and a black market rate. The official rate was flexible and had changed considerably over time, resulting in a peso value of the dollar that was, by the mid-1980s, more than ten times higher than it was in 1970. As a result of the flexible central bank exchange rate policy, the black market rate was not significantly different from the official rate.

•In Jamaica, according to Pinot de Libreros' description, three rates existed in the mid-1980s: two official rates—one for trade with Caribbean Community (CARICOM) countries and another for foreign exchange transactions with other countries—and a free or black market rate. Subsequently, the government abandoned the official rates and tried to control the market rate through a system of auctions for foreign exchange.

•Venezuela had a dual market for foreign exchange. The first was a low rate for the U.S. dollar, which was fully controlled by the government. This rate held for the exchange of dollars obtained from oil exports and for the use of those dollars by the oil industry and some other sectors importing essential goods. It also was used by the government and public enterprises to repay the foreign debt principal and interest and to provide education grants for studies abroad. The dollars obtained from, or needed for, these transactions were directly sold to or purchased from the central bank. A higher exchange rate for the dollar held for the remaining transactions, which could be either exports of industries that the government tried to stimulate or imports that it wanted to discourage. The higher dollar rate was not a black market rate. In effect, the Central Bank of Venezuela was the main supplier of dollars, providing private banks and other foreign exchange dealers with dollars in order to influence the free market price.

If no adjustments were made, there would be three main effects on the local currency value of external transactions and on other national account aggregates as a result of multiple exchange rate regimes, rapid changes of exchange rates over time, and regular differences between the purchase and sale prices of foreign currency.

The first effect is on the foreign trade balance. If expressed in local currency, it may be smaller or larger than the balance in foreign currency relative to the size of exports or imports. The same may hold for the current balance of the external account. It may show that the current surplus that supplements domestic savings in financial capital formation is lower or higher in terms of local currency than the value of current revenues from or disbursements to other countries, when this relative value is compared with its foreign currency equivalent. When the difference between purchase and sale values of foreign currency is large, the foreign trade or current external balance may even have a different sign when expressed in foreign exchange units compared with the balance in local currency units. This may occur when imports or other disbursements to the rest of the world take place at exchange rates that are much higher in terms of local currency than the exchange rates used for the local currency conversion of exports or other revenues. Under multiple exchange rate regimes, this effect on the foreign trade balance may be a direct consequence of the government's foreign exchange policy. In the case of rapid changes over time in the rates of exchange for foreign currencies, the effect may happen because export revenues and other revenues may be received at the beginning of the year, when the exchange rates are still low, whereas import payments or other disbursements abroad may be made toward the end of the accounting period, when exchange rates have already increased considerably.

Another effect is on the value added in industries that export or import goods and services and on the profits and retained earnings of enterprises that pay interest, dividends, and other property income to nonresidents or receive revenues on investments abroad. The value added, profits, and retained earnings calculated in local currency are affected by the same discrepancies between the value at which foreign exchange receipts are converted to local currency and the value at which foreign currency is purchased in order to import goods and services and to pay dividends, interest, and so on to nonresidents.

Finally, the exchange rate differentials may cause inconsistencies in the national accounts. One reason would be because the accounts of the central bank or any other government authority that holds foreign currency would record exchange rate differentials as a revenue, while there is no explicit counterpart in accounts of other sectors in the national accounts. A further inconsistency may arise because the foreign exchange authority calculates the exchange rate differentials by using a valuation of foreign exchange balances that is different from the stock valuation measures used by other sectors in the national accounts or recommended in the national accounts.

II. Components of Foreign Exchange Rate Differentials

Exchange rate differentials refer to the income accruing to banks or government agencies holding foreign currency reserves that arises from differences between the sale and purchase prices of foreign exchange. More precisely, exchange rate differentials are an aggregate of three components: a bank service charge, a tax or subsidy, and a capital gain or loss. The three components can be associated with the three types of exchange rate circumstances cited earlier in this paper. The distinction among the three components is necessary, since each refers to transactions that are treated differently in the SNA.

Bank Service Charges

The bank service charge is based on the difference between sale and purchase values of foreign currency; the charge is always maintained, even under circumstances of monetary stability. Under stable conditions, the difference between sale and purchase values is generally small and includes only the implicit bank service charge. This was the situation in most countries before the fixed exchange rate regime was abandoned.

The implicit service charge is imposed in the domestic territory by resident financial institutions. Implicit charges are reflected in all foreign exchange transactions: a higher price is charged for foreign exchange to those that need to make payments to entities abroad and a lower rate is applied to those who receive foreign currency from abroad. This implies that receipts from abroad expressed in local currency are net of the implicit bank charges, whereas payments to entities abroad expressed in local currency include the charges that banks impose through their higher sale rate for foreign currency. Because exports are valued f.o.b. (free on board) in the national accounts and imports are recorded c.i.f. (cost, insurance, freight), however, the implicit bank service charges thus imposed by domestic institutions should not be deducted from exports and should not be included in imports. For reasons of consistency, a similar treatment should be accorded to other receipts from and payments to other countries. The implication of this suggestion is that payments to and receipts from abroad should be recorded at the same rate of exchange if the bank service charge is the only difference between the purchase and sale values of the foreign exchange. This uniform exchange rate will be referred to here as the “accounting exchange rate.”

Taxes or Subsidies

The tax or subsidy component exists under multiple exchange rate regimes when the central bank or government monetary authorities stipulate that different rates of exchange must be paid or received for different transactions, favoring some sectors and discouraging transactions carried out by other sectors. The exchange rate differential caused by multiple exchange rate regimes should be considered a tax or subsidy because the exchange rate policy influences the prices or values of transactions that are expressed in local currency.

The tax or subsidy component should be calculated in a way similar to the treatment of the implicit bank service charge, described above. Assuming that there are no exchange rate changes over time and no explicit bank service charges, the total tax or subsidy component should be based on the difference between the sale and purchase values of the foreign currency. The difference with the bank service charge is that, under multiple exchange rates, the banks' sale prices of foreign currency for some imports and other payments to entities abroad might be low because the government considers these transactions essential, whereas the banks' purchase prices of foreign currency for some exports and other receipts from abroad, which the government wants to promote, might be high. The government's total tax receipts may even be negative, and thus turn into a subsidy, when the external receipts and payments based on ''promotional” exchange rates dominate the external transactions of the country. Because the government is a domestic institution that imposes the tax or grants the subsidy, the implicit taxes and subsidies should not affect the foreign trade and other balances of the external account for the same reasons that held for bank service charges. This implies that exports and other receipts from abroad should be recorded in the external account before taxes are deducted or subsidies are granted, and that imports and other payments to entities abroad should be registered net of taxes to be imposed or subsidies to be granted.

Several complications should be considered before this treatment is adopted. One question that needs to be resolved is how to treat the exchange rate differential caused by foreign currency purchases and sales that are not channeled through the central bank or other government exchange authority but are carried out by private banks and foreign exchange dealers. For example, in Venezuela during the period of multiple rates in the mid-1980s the private banks operating in the free exchange market purchased dollars at a much higher rate than the Central Bank. To stimulate production in selected industries, the government allowed certain exporters to change their dollars in private banks at the much higher rates. This implied a subsidy for these exporters, and the local currency price of those exports is thus a consequence of a government policy. Similarly, importers of nonessential goods were forced by the government to purchase their dollars in the free exchange market. Those imports were thus implicitly taxed. The difficulty was that private banks cannot really impose taxes or grant subsidies. But the problem seems more difficult than it really is. First of all, the private banks do not pay the subsidy or receive the indirect tax because they establish only a small difference between purchase and sale prices of foreign currency and generally maintain small foreign exchange balances. Implicit indirect taxes and subsidies therefore cancel out. If, however, banks would benefit from the exchange rate policies and would not have to transfer a part of the differentials to the government, a subsidy would have to be imputed from the government to the private banks, which would reflect the extra benefits such banks receive from the government's exchange rate policy.

The second problem concerns the taxes and benefits (subsidies) that are a consequence of the government's policy. In the example above, it was assumed that only exports and imports were affected. In practice, however, the policy also affected other external transactions. In Venezuela, it also affected payments of the external public debt principal and interest, and foreign currency grants for Venezuelans studying abroad. This immediately raises the question of whether all implicit taxes and benefits should be considered indirect taxes and subsidies. There are three ways of approaching this matter.

One way would be to channel implicit taxes and benefits through the bank service charge. Doing this would imply that the government's exchange rate policy affects the market value of bank services and, through those services, the prices of exports and imports and the value of other transactions in foreign currency. The consequence of this treatment is that all taxes and benefits could be considered indirect taxes and subsidies, since they would be levied or granted on (bank) services. With this treatment, however, the value of bank services could become negative, particularly if the service charges were allocated to sectors. This is so because the size of subsidies in some sectors might be much higher than the value of the foreign exchange bank service charges excluding the subsidies. Hence this alternative is not very attractive.

Another possibility is to deal with bank services and taxes and benefits separately and to allocate the taxes minus benefits directly to the government. The consequence is that taxes and benefits levied or granted on external transactions other than exports or imports of goods and services cannot be considered indirect taxes or subsidies. They are levied or granted on other revenue or expenditure components that do not form part of the production cost, and those taxes and benefits therefore do not immediately affect the prices of goods and services. It is suggested that they be treated as direct taxes when levied on transactions of the income and outlay account (for example, payment of interest on the public debt or the transfer of education grant funds) or as capital transfers when they are levied on transactions of the capital (accumulation or finance) account (for example, repayment of the public debt principal). There are some disadvantages to this treatment. Contrary to what is now recommended in the SNA, one would have to deal with negative direct taxes when implicit benefits are paid on transactions in the income and outlay accounts. Such a concept, however, is not completely unknown at present, because some countries (for example, the United States) have introduced tax systems whereby explicit tax benefits are paid to taxpayers when their income is below a minimum level. The SNA may therefore have to be revised on this point.

A third possibility is to allocate the taxes minus benefits, including the bank service charge, to the government. This treatment implies that foreign exchange bank services are charged to the government, since banks help the government to implement the exchange rate policy. This treatment has several advantages: the bank services need not be allocated to sectors, and thus the problem of measuring those services for each sector is avoided. Negative bank service charges could not occur, and one also avoids changes in the actual values of external transactions as a result of the inclusion of the implicit bank service charges.

The last question that should be dealt with here is the allocation of the implicit taxes and subsidies to the activity or institutional sectors of the economy that are affected by the exchange rate policy of the government. There is much analytical and policy interest in this allocation. From an accounting point of view, however, there are some disadvantages of trying to build the allocation into the national accounting framework. One would be that all transaction values in the external account would be changed, and the change is dependent on the accounting exchange rate selected. Also, input-output relations would be distorted, since both gross output and intermediate consumption of those sectors that are involved in the export or import of goods and services would be changed by the amount of indirect taxes minus subsidies. Because these sectors base their costing procedures on the actual prices that they pay or receive in local currency, the analysis of technical behavior of those industries would be distorted.

To avoid the distortions, it is preferable to treat the adjustments for taxes and benefits as global adjustments in the external account and to allocate foreign exchange bank service charges directly to the government sector, as suggested above. For analytical uses, the global adjustments could be supplemented, however, by an alternative presentation of data in which the taxes and benefits and bank service charges are allocated to activities and sectors.

Capital Gains or Losses

With the abandonment of the fixed exchange rate regime, exchange rates started to fluctuate, and in some countries they rose or fell continuously over time. Thus banks accrued revenues that were in fact realized capital gains net of losses. The capital gains or losses should not be included in the flow accounts of the SNA, however, but should be dealt with in the reconciliation accounts, where all valuation changes in assets are recorded.

Before this treatment in the SNA accounts is illustrated, an explanation is needed about what is meant by capital gains and losses, which are only summarily treated in the SNA (paragraph 6.109) or defined in the Provisional Guidelines (paragraph 6.21) on balance sheets.3 The explanation is given with the help of Table 1. The table presents an illustrative example showing the bank's balances (b) of U.S. dollar holdings at the end of 1985, the purchases (p) and sales (s) that took place during 1986, and the final balance of U.S. dollar holdings at the end of that year. The transaction value of balances, purchases, and sales is presented in U.S. dollars in the second column of the table. This information is converted to local currency values in the third column using the actual exchange rates at which dollars are purchased and sold during the year, as shown in the fourth column. It is assumed here that there is no difference between the sale and purchase values of U.S. dollars and that there is therefore no implicit bank service charge. The capital gains accruing to the banks are calculated each time dollars are purchased or sold, and they are presented in the last column of the table. For example, the capital gains of 1,160 accruing on March 3, 1986, are calculated as the change in the value of the balances of U.S. dollar holdings before the purchase on that date, as a result of changes in the exchange rate between February 1 and March 3, 1986 (1,160 = 290 (9 – 5)). The subtotals of capital gains and net purchases minus sales in terms of U.S. dollars and local currency are presented at the bottom of the table. The net total of the capital gains (5,350) is equal to the difference between the closing and opening values of the stock of U.S. dollar holdings converted to local currency at the beginning and end of 1986 (4,900 = 6,000 – 1,100) minus the difference between purchases and sales of foreign currency (—450) valued in local currency at the subsequent exchange rates at which sales and purchases were carried out. Capital gains and losses defined in this manner are consistent with the valuation of purchases and sales of foreign currency at their actual exchange rates.

Table 1.Illustration of Capital Gains or Losses on Transactions in Foreign Exchange
Exchange
Rate (Local
Transaction ValuecurrencyCapital
In U.S.In localperGains or
DateTypedollarscurrencyUS$1)Losses (-)
31 December 1985b2201,1005
1 February 1986p703505
b290
3 March 1986s−60−54091,160
b230
4 April 1986p201407−460
b250
5 May 1986s−50−750152,000
b200
6 June 1986p6090015
b260
7 August 1986s−30−600201,300
b230
9 November 1986p2030015−1,150
b250
5 December 1986s−10−250252,500
31 December 1986b2406,00025
Subtotal20−4505,350
Note: b = stock of U.S. dollars; p = purchases of U.S. dollars; s = sales of U.S. dollars.
Note: b = stock of U.S. dollars; p = purchases of U.S. dollars; s = sales of U.S. dollars.

This example illustrates a situation whereby the net purchases of foreign currency by the central bank in terms of U.S. dollars is positive (20) and, when converted to local currency at the exchange rates prevailing at the time of the transactions, the net value becomes negative (−450). This change in sign of the balancing item occurs because the exchange rate of the dollar has increased over time and a larger portion of the foreign exchange sales is concentrated in the latter part of the year.

The treatment of this effect is further illustrated in Table 2, which presents the treatment of these types of foreign exchange rate differentials in the national accounts. The figures in Table 2 are consistent with those in Table 1. It is assumed that all foreign currency transactions in Table 1 concern imports and exports of goods and services in Table 2. The net purchases of U.S. dollars in terms of local currency (−450) show up as a deficit in the external account in Table 2 and reduce the value of GDP in the account for the national economy. No adjustments have been made in the current account part of the table.

Table 2.Illustration of Treatment of Capital Gains or Losses on Foreign Exchange in the National Accounts
National EconomyExternal Transactions
OtherLocalU.S.
Bankssectorscurrencydollars
ItemRDRDRDRD
Consumption/capital
formation5,000
Exports1,6901,690170
Imports2,1402,140150
GDP/foreign trade
balance4,550(450)20
Gross savings (net
lending)4,550(450)
Cash
Local currency450(450)
U.S. dollars(450)(450)
Other assets5,000
Balancing item4,5504,550(450)(450)
Note: R = revenue; D = disbursement.
Note: R = revenue; D = disbursement.

The capital gain or loss component does, however, affect banks' balance sheets, other sectors of the national economy, and the external balance sheet as it relates to the national economy. But not all changes in the value of these assets are included in the capital finance accounts of those sectors, only those changes that are an immediate consequence of the real transactions mentioned before. These are the reduction in the local currency cash or other asset balances of other sectors of the economy (−450) and the counterpart increase of those balances in the capital finance account of the banks (+450); also included are the reductions in the U.S. dollar balances of the banks in local currency (−450) and the counterpart reduction of the U.S. dollar balances in the external account in local currency (−450). Not included are the capital gains accruing to the banks as a result of the increases in the value of their foreign currency holdings. These capital gains are to be dealt with in the reconciliation accounts, as explained above. The net effect of the treatment shown in Table 2 thus maintains the deficit of the external sector, which is a consequence of changes in exchange rates over time, while compensating it with an adjustment factor in the external capital finance account and in the capital finance account of banks.

III. Measurement Issues

The three components of exchange rate differentials rarely exist in isolation. Only under stable monetary conditions might one expect to find the implicit bank service charge for foreign currency transactions alone. Under the present conditions of volatile movements of exchange rates over time, however, capital gains and losses and bank service charges are included in the exchange rate differentials. Furthermore, many countries that have multiple exchange rate regimes also experience considerable changes in exchange rates over time, so that under these circumstances all three components of exchange rate differentials are present: bank service charges, capital gains or losses, and taxes or subsidies.

To separate the three components, one would first need to determine the accounting exchange rate, or a series of accounts exchange rates if there are considerable changes in the exchange rates over time. On the basis of these accounting exchange rates, a distinction can be made between bank service charges and taxes or subsidies on the one hand and capital gains or losses on the other. The next step is to determine the bank service charge to be imputed to the government as a cost of the exchange rate policy.

The accounting exchange rate—which can only be approximated because the actual value is not recorded—is an equilibrium rate that would exist at one point in time if there were no government control of part or all of the foreign currency market. The equilibrium rate should exclude the bank service charges. On the basis of this general criterion, the different possibilities of measurement will be reviewed under conditions varying from stable exchange rates to multiple exchange rate regimes combined with changes in the exchange rates over time.

Stable Monetary Conditions

The simplest situation would be one in which there are fairly stable rates over time and, correspondingly, only minor operating margins charged by banks. Under these circumstances one might assume that there is an equilibrium between supply and demand and that purchasers and sellers of foreign exchange share the operating cost equally. The accounting exchange rate might be estimated as the unweighted average of the purchase and sale prices of foreign exchange.

Changing Exchange Rates Over Time

The approximation of an accounting rate is more difficult when the foreign currency value is fluctuating over time or increases or decreases all the time. This is the most common situation at present. Under those circumstances information will be available on the margin between the purchase and sale prices of the foreign currency, but the margin may actually be much larger than in the case of stable monetary conditions because it not only reflects the operating cost but also includes a compensation for past and expected future losses in the value of foreign currency between the moments of purchase and sale. The difference attributable to the latter factor, if realized, is a capital loss and should in principle not be incorporated in the bank service charge. Because the separation of the capital loss element is, however, difficult to carry out in practice, the actual margin at one time between the purchase and sale values may be taken as an approximation of the bank service charge—that is, the accounting value for conversion will again be the unweighted average of purchase and sale values. Under these circumstances there would not be one accounting rate but a time series that reflects fluctuations in rates over time.

Multiple Exchange Rates

The situation would be more difficult to deal with in national accounts when there are multiple exchange rates existing for different transactions that might take place at the same time. Most countries that have introduced these multiple rates are also faced with exchange rate changes over time.

If there is a free or parallel (black) market, it may be possible to observe the differences between the purchase and sale values of foreign currency, which might be taken as an approximation of the bank service charges. It should be noted, however, that as in the previous case, in which exchange rates change over time, the margin may include compensation for realized and expected losses in the value of foreign exchange over time. Furthermore, the estimated margin may refer only to a small market where the margin between the sale and purchase values of foreign exchange is not representative for margins in other foreign currency markets that are government controlled.

Several considerations should be taken into account in determining the accounting or equilibrium rate under multiple exchange rate conditions. The accounting exchange rate should probably be somewhere between the lower rates, generally controlled by the government, and the much higher rates, which dominate the free or parallel market. Presumably, if the government did not control the market, some who purchased foreign exchange at the low official rates would not exercise their demand; this would increase the supply of foreign exchange at the free market, which might then reduce the free market rate. A reasonable approximation of the accounting rate might be the weighted average of actual rates—after the estimated bank service charges are deducted. The weights would be the foreign currency values of all external transactions that took place during the period of account, including exports and imports as well as other transactions registered in the external account.

In determining this average, one should be aware that the external account includes some transactions that are actually carried out without any foreign currency. These include loans tied to purchases of goods and services from the country providing the loan at a prearranged exchange rate, or current and capital grants-in-kind for which not even an explicit foreign exchange conversion is considered. These transactions should be excluded from the weighted average.

Two approaches for deriving an estimate of the accounting exchange rate, suggested by Pinot de Libreros, are less appropriate. One suggestion is to choose an actual exchange rate as the rate that would come closest to the equilibrium accounting rate.4 Such a choice would be reasonable only if the amount of foreign exchange transactions through this market were a very large percentage of all foreign exchange transactions. It is not sufficient, as Pinot de Libreros suggested, that the selected exchange rate be determined in a free market without government intervention because, although the government does not explicitly enter this market, it does influence the price in that market through a foreign currency market of its own that has attracted foreign exchange transactions otherwise channeled through the free market.

Pinot de Libreros also suggests that, under multiple exchange rate conditions, the accounting exchange rate should be determined on the basis of the foreign exchange revenues, which the central bank says that it has accumulated as a result of the foreign exchange controls.5 The weakness of this approach is that the central bank can only determine these revenues on the basis of an accounting exchange rate, and in a manner similar to that illustrated in Table 1. This accounting rate may be different from the one that should be used in the national accounts calculations, and this therefore makes it difficult to use the central bank calculations as a point of departure for the national accounts conversion.

IV. A Quantitative Example

The following quantitative example of the treatment of multiple exchange rates is intended to illustrate and clarify the previous discussion and proposals. The example is based on national accounts data for Venezuela for 1983. Because no unpublished, detailed national accounts information was available at the time this study was written, aggregate data as published by the United Nations in its annual national accounts publication6 were used and somewhat modified to fit the adjustments for exchange rate differentials.

The Venezuelan national accounts figures are presented in Table 3 in the T-account format. The table is divided into two groups of accounts: one account for the national economy (the first two columns), and three alternative accounts covering external transactions in three modes of valuation—U.S. dollars (the last two columns), local currency based on the conversion using actual unadjusted (transaction) exchange rates (the third and fourth columns), and local currency based on the conversion using an accounting rate of exchange (the fifth and sixth columns). Each account has a revenue (R) and disbursement (D) side.

Table 3.National Accounts Aggregates in U. S. Dollars and Local Currency, Including an Adjustment for Exchange Rate Differentials: Illustration with Venezuelan Data
External Transactions
Local
currencyLocal
transac-currency
tionaccounting
NationalexchangeexchangeU.S.
Economyrateratedollars
TransactionsRDRDRDRD
Final consumption221.4
Gross capital formation34.1
Exports
Oil33.033.055.05.5
Others11.211.27.00.7
Imports
By oil industry3.03.05.00.5
Others49.649.631.03.1
Exchange rate differential
Banking service charge6.86.8
Indirect tax/subsidy29.629.6
283.528.026.02.6
GDP/foreign trade
surplus/deficit283.528.026.02.6
Factor income (interest on
external debt)3.68.43.68.46.014.00.61.4
Current transfers1.61.61.00.1
Final consumption221.4
Exchange rate differential
Direct taxes(3.3)52.4(3.3)18.317.01.7
Gross savings/current
surplus (+)/
deficit (-)52.418.317.01.7
Gross capital formation
Capital transfers34.1
Exchange rate differential
Capital transfers(2.1)(2.1)
16.216.217.01.7
Net lending16.216.217.01.7
Acquisition of financial
assets14.414.49.00.9
Repayment of external
debt principal(6.6)(6.6)(1.1)
Incurrence of liabilities4.84.83.00.3
Exchange rate differential
(increase in U.S.
dollar balances due to
changes over time in
exchange rates)
Source: Banco Central de Venezuela, “Experiencias y dificultades de Venezuela en el tratamiento de los cambios multiples en el Sistema de Cuentas Nacionales” (Experiences and Difficulties in Venezuela in the Application of Multiple Exchange Rates in the SNA), paper presented at the Regional Seminar on National Accounts, UN Economic Commission for Latin America and the Caribbean (ECLAC), Santiago, Chile, April 1986.Note: R = revenue; D = disbursement. Calculation of the exchange rate differentials is based on an accounting rate of 10 local currency units per US$1.
Source: Banco Central de Venezuela, “Experiencias y dificultades de Venezuela en el tratamiento de los cambios multiples en el Sistema de Cuentas Nacionales” (Experiences and Difficulties in Venezuela in the Application of Multiple Exchange Rates in the SNA), paper presented at the Regional Seminar on National Accounts, UN Economic Commission for Latin America and the Caribbean (ECLAC), Santiago, Chile, April 1986.Note: R = revenue; D = disbursement. Calculation of the exchange rate differentials is based on an accounting rate of 10 local currency units per US$1.

The transaction breakdown presented in the table covers the main aggregates—GDP, gross savings, and net lending—together with the external transactions equivalents—foreign trade surplus or deficit, current surplus, and net lending. The transactions in the external accounts are recorded from the point of view of the national economy; that is, exports are presented as revenues, imports as disbursements, and so on.

The accounting exchange rate used in the table is 10 local currency units per US$1. It was calculated as the weighted average of actual exchange rates, using the U.S. dollar values of the transactions presented in the last two columns as weights. All external transactions are included in the weights, except for the balancing items. By using all transactions as weights, it is assumed that transactions-in-kind are not included in the external account nor are linked transactions, such as foreign loans that are tied to imports of goods from the countries providing the loans.

Five types of global adjustments for exchange rate differentials, which were discussed in Section II, are presented in different parts of Table 3. Banking service charges and indirect taxes or subsidies are presented as adjustments to exports and imports in order to affect GDP and the foreign trade balance. Net direct taxes are included before gross savings and the current external balance; net capital transfers are included before net lending. A final adjustment item, called net increase of U.S. dollar balances due to changes of exchange rates over time, is included as a final balancing item.

The banking service charges (6.8 in the table) are assumed to be 5 percent of the total local currency value of all external transactions—revenues and disbursements—converted on the basis of the actual unadjusted (transaction) exchange rates (third and fourth columns). The implicit taxes or subsidies on external transactions are calculated separately for the production, income and outlay, and capital transactions. For each group of transactions, they are equal to (1) the difference between the local currency value of revenues converted on the basis of the accounting exchange rate minus the local currency value converted according to the actual exchange rate, plus (2) the difference between the local currency value disbursements converted on the basis of the actual exchange rate minus the local currency value converted on the basis of the accounting exchange rate, and minus (3) the implicit banking service charges. Based on this formulation, indirect taxes minus subsidies are 29.6

net direct taxes are −3.3
and net taxes on wealth are −2.1

The increase of foreign currency balances due to changes of exchange rates over time has a value of zero because it has been assumed that no exchange rate changes have taken place during the period of account.

The two accounts that ultimately enter into the national accounts are those for the national economy (the third and fourth columns of Table 3) and for external transactions, converted on the basis of the actual (unadjusted) exchange rates. Both columns include adjustments for exchange rate differentials. This means that GDP and the foreign trade balance are increased with the amount of indirect taxes (6.8) and banking service charges (29.6). Gross savings are thereafter reduced with the negative net direct taxes (−3.3), and net lending is subsequently reduced with the negative net taxes on wealth (−2.1). Because there are assumed to be no exchange rate changes over time, the external account (the third and fourth columns) balances without further adjustment. The case of Venezuela in Table 3 is an example of a country whose foreign trade balance in U.S. dollars is positive (+ US$2.6); if no adjustments had been made, the foreign trade balance in local currency would have been negative (−8.4).

The effects on GDP, gross savings, foreign trade, and other external balances depend on the choice of the accounting exchange rate (which was 10 local currency units per US$1 in Table 3). If a different accounting exchange rate were selected, all macro aggregates would be different. The difference is analyzed in Table 4. The table presents alternative values in local currency of the balancing items—GDP, foreign trade balance, gross savings, current external balance, and net lending—and the components of exchange rate differentials—bank service charges, indirect taxes net of subsidies, net direct taxes, and net taxes on wealth. Also reflected are the effects of changes in the accounting exchange rate on GDP and per capita GDP in U.S. dollars, which have been derived on the basis of the accounting exchange rate conversion.

Table 4.Value of Main Aggregates Calculated on the Basis of Different Accounting Exchange Rates: Illustration with Venezuelan Data
Alternative Accounting Exchange Rates
(Local currency units per US$1)
Item6.010.012.016.0
Value in local currency
GDP/foreign trade balancea273.1283.5288.7299.1
17.628.033.243.6
Gross savings/45.652.455.862.6
current external balancea11.518.321.728.5
Net lending16.216.216.216.2
Exchange rate differential31.031.031.031.0
Bank service charges6.86.86.86.8
Indirect taxes/subsidies19.229.634.845.2
Net direct taxes0.3(3.3)(5.1)(8.7)
Net taxes on wealth4.7(2.1)(5.5)(12.3)
In U.S. dollars
GDP45.528.324.118.7
GDP per capitab2,630.71,638.61,390.51,080.5
Source: Based on Table 3 under alternative exchange rates.

Surplus (+); deficit (−).

Based on a population of 17.3 million (Venezuela).

Source: Based on Table 3 under alternative exchange rates.

Surplus (+); deficit (−).

Based on a population of 17.3 million (Venezuela).

The first effect that should be noted is that if the accounting exchange rate is increased, the local currency value of GDP also increases. This is because a shift takes place within the total of exchange rate differentials, between the value of indirect taxes or subsidies and direct taxes and net taxes on wealth. The total of exchange rate differentials itself does not change, because it has been assumed that no changes take place over time in the accounting exchange rate. Gross savings also increase, but net lending remains the same because of the unchanged total of exchange rate differentials. The shift toward indirect taxes or subsidies when the accounting exchange rate increases can be clearly observed from the trend in the breakdown of the total exchange rate differentials in the table.

Another effect of the change in the accounting exchange rate is on GDP and per capita GDP in U.S. dollars. Assuming that the accounting exchange rate is the most appropriate one for converting local currency GDP to U.S. dollars,7 the U.S. dollar data are much more sensitive to a change in the accounting exchange rate than is the local currency GDP. The change in the U.S. dollar value of GDP is the result of two opposite effects. On the one hand, an increased accounting exchange rate increases GDP in local currency through the conversion of exports and imports; on the other hand, it decreases GDP and per capita GDP when converted back into U.S. dollars. The net effect is a considerable decrease in the U.S. dollar value of GDP and per capita GDP.

Marion Pinot de Libreros, “Effects of Devaluation on the Calculation of Flow of Funds Accounts” paper presented at the Nineteenth General Conference of the International Association for Research in Income and Wealth (IARIW), Noordwijkerhout, Netherlands, August 1985; and “Los mercados múltiples de cambio y su tratamiento en Cuentas Nacionales” (Multiple Exchange Rate Markets and Their Treatment in the National Accounts), paper presented at the Regional Seminar on National Accounts, UN Economic Commission for Latin America and the Caribbean (ECLAC), Santiago, Chile, April 1986.

Banco Central de Venezuela, “Experiencias y dificultades de Venezuela en el tratamiento de los cambios múltiples en el Sistema de Cuentas Nacionales” (Experiences and Difficulties in Venezuela in the Application of Multiple Exchange Rates in the SNA), paper presented at the Regional Seminar on National Accounts, ECLAC, Santiago, Chile, April 1986.

United Nations, Provisional International Guidelines on the National and Sectoral Balance-Sheet and Reconciliation Accounts of the System of National Accounts, Statistical Papers, Series M, No. 60 (New York, 1977).

Pinot de Libreros, “Los mercados multiples de cambio,” pp. 9–11.

Pinot de Libreros, “Effects of Devaluation,” pp. 14–16.

United Nations, National Accounts Statistics: Major Aggregates and Detailed Tables (New York, 1984).

The adjustments to the exchange rates suggested in the present paper make the accounting exchange rate more attractive than any other exchange rate for the conversion to U.S. dollars of GDP and per capita GDP. However, not all limitations in the use of exchange rates are removed through the suggested adjustment procedure to arrive at an equilibrium rate, as the equilibrium rate applies in principle only to external transactions and remains less appropriate for the conversion of domestic transactions.

    Other Resources Citing This Publication