For Staff Use Only
EXCHANGE RATES: CONCEPTS AND MEASUREMENT ISSUES
Riccardo Faini
Division Working Paper No. 1986-3
September 1986
Economic Analysis & Projections Department
Country Analysis & Projections Division
Division Working Papers report on work in progress and are circulated
for Bank staff use to stimulate discussion and comments.    The views and
interpretations in this document are those of the author.    The comments
and suggestions of Jean Baneth greatly improved both the content and the
presentation of the paper.     I am also very grateful to Nick Nteireho
for his valuable research assistance.     I am of course responsible for
any remaining errors.



TABLE OF CONTENTS
INTRODUCTION                                                        1
EXCHANGE RATES: A BRIEF SURVEY OF POSSIBLE DEFINITIONS              3
1.  Nominal Exchange Rate                                   3
2.  Nominal effective exchange rate                         3
3.  Real exchange rate                                      4
4.  The Choice of the Price Index                           5
MEASUREMENT                                                         8
A.  THE PPP REAL EXCHANGE RATE                               9
B.  THE REAL EXCHANGE RATE AS A RATIO OF NON-TRADEABLES
TO TRADEABLES PRICES                                   12
MULTILATERAL EXCHANGE RATES
1.  Assessing the importance of Trade Partners             13
2.  An Illustration: Colombia and Korea                    15
EQUILIBRIUM AND DISEQUILIBRIU' REAL EXCHANGE RATES                 25
MONITORING THE EXCHANGE RATE                                       31
SELECTED REFERENCES                                               33
APPENDIX A                                                        34
APPENDIX B                                                         37
FIGURE 1: COLOMBIA: DOMESTIC CPI TO FOREIGN WPI                    16
FIGURE 2: KOREA: DOMESTIC CPI TO FOREIGN WPI                      17
FIGURE 3: COLOMBIA: PPP REAL EXCHANGE RATES                       20
FIGURE 4: KOREA: PPP REAL EXCHANGE RATES                          21
TABLE 1: REAL EXCHANGE RATES: COLOMBIA AND KOREA                  19
TABLE 2: A DECOMPOSITION OF THE REAL EXCHANGE RATE WEIGHTS        23
TABLE IA: MANUFACTURING SECTOR WEIGHTS: COLOMBIA                  38
TABLE 2A: MANUFACTURING SECTOR: KOREA                             39



TABLE OF CONTENTS (contd)
TABLE 3: REAL EXCHANGE RATES FOR THE MANUFACTURING SECTOR:
COLOMBIA                                                 26
TABLE 4: REAL EXCHANGE RATES FOR THE MANUFACTURING SECTOR:
KOREA                                                    26
TABLE 5: DOMESTIC NON-TRADED TO FOREIGN MANUFACTURING
GOODS PRICES: COLOMBIA                                   27
TABLE 6: DOMESTIC NON-TRADED TO FOREIGN MANUFACTURING
GOODS PRICES: KOREA                                      27



RFE. PEXCRATE-2
EXCHANGE RATES: CONCEPTS AND MEASUREMENT ISSUES
INTRODUCTION
The management of the exchange rate is a central issue in
recent economic policy debates dealing with developing countries.     The
concern about exchange rate policies can be attributed to a number of
factors.
1.    Current account imbalances are a major issue for many
LDCs.   Empirical evidence shows that exchange rate variations have a
sizeable impact on the current account balance in LDCs.      As a result,
exchange rate devaluation is now often indicated as an effective remedy
to correct current account imbalances.
2.    Exchange rate variations also affect inflation.    Indeed,
the reluctance to devalue has often been attributed to the policy-maker
fears of contributing to inflationary pressure.       In some cases (for
instance, in the Southern Cone countries of Latin America), the exchange
rate has been directly used as a tool to control inflation.            The
unwelcome side effect of such a policy has been the real over-valuation
of the currency.
3.     Finally, the real exchange rate is deemed to have an
important impact on the efficiency of resource allocation in the long-
run.    It thus plays a central role in any policy package designed to
achieve growth as well as adjustment.
It is important, therefore, that the concept of the exchange
rate be carefully and precisely defined.    The purpose of this paper is
to survey the main issues related to the definition and the empirical
computation of an index of the real exchange rate. Both theoretical and



RFE.PEXCRATE-2                        - 2 -
measurement issues are indeed quite relevant.     The best definition of
the exchange rate may vary according to both the purpose of the analysis
and the specific economic conditions of the reporting country.      Also,
due to the poor quality of the available statistical information and the
difficulty to fit the data into theoretical constructs, it is
unavoidable to resort to second best indices.        We must, therefore,
determine how well available indices fit theoretical constructs and
provide relevant and accurate information for policy purposes.
Section 1 of this paper presents a brief survey of the
differenr concepts of the exchange rate and an analysis of the criteria
underlying  the   choice  of  a  specific   index.    Section  2   tackles
measurement problems, and assesses the meaning, as well as the
drawbacks, of the available indices of the real exchange rate. For the
sake of simplicity in these first two sections we pretend that the
reporting country trades with only one large ("Rest of the World")
partner.
In section 3 we dispense with this assumption and examine how
to assess the importance of various foreign trade partners.           Not
suprisingly, the conclusion is that the choice of the optimal weighting
system will depend to a large extent on the structural characteristics
of trade flows.
Section 4 discusses how to interpret, both for policy and
analytical purposes, the information provided by the behaviour of the
real exchange rate over time.    Finally, the concluding section reviews
the main issues and discusses how the monitoring of the exchange rate
can contribute to a better design of policy.



RFE.PEXCRATE-2                        - 3 -
Exchange Rates: A Brief Survey of Possible Definitions
1.       Nominal Exchange Rate.   The nominal exchange rate measures the
price of domestic currency in terms of the foreign currency.         Each
country has as many nominal exchange rates as it has trading partners
with their own national currencies.    For expository simplicity in this
and in the following section, we shall assume that the reporting country
trades only with one "large" foreign partner.      As a result, only one
nominal exchange rate will need to be considered. I/
2.       Nominal effective exchange rate.     In a world where tariffs,
subsidies, and export taxes are present, the nominal exchange rate does
not fully capture the economic incentives facing economic agents engaged
in foreign transactions, exporters, importers and borrowers of foreign
capital.   To this purpose we need to consider the nominal effective
exchange rate,2L   i.e. the nominal exchange rate corrected for taxes,
other trade restraints and subsidies, reflecting the effective amount of
domestic currency an exporter gets fir one unit of foreign exchange and
an importer pays for one.     To compute the nominal effective exchange
rate, information is needed on the full structure of tariffs, subsidies
and  other  trade   restraints.    The  data  requirements  may  be   very
1/  Our definition implies that a devaluation in which the number of
foreign currency units per domestic currency unit decreases results
in a lower exchange rate.    This terminology is used throughout the
paper.
2/  This is standard terminology in international trade theory.    It may,
nonetheless, be the source of some confusion insofar as, in the
field of international monetary theory (and in IMF publications),
the effective exchange rate is defined as a weighted average of the
nominal   exchange   rates  of   the   reporting   country's  trading
partners.   It is used, therefore, to assess the relative importance
of different foreign partners and not the impact of trade policy.
In Section 3 we refer to this measure as the muLtilateral exchange
rate.



RFE.PEXCRATE-2                        - 4 -
demanding.    The existence of quota3     further compounds the problem
insofar as it becomes necessary to incorporate into the nominal
effective exchange rate a measure of the import premium.
Even in the absence of quotas, given that export subsidies and
tariff rates will generally differ, the average nominal effective
exchange rate for imports and exports will also differ, as will indeed
the nominal effective exchange rates for different imports and
exports. As a result, the relative incentives to produce importable and
exportable goods are modified by the operations of trade policy.      The
ratio of the average nominal effective exchange rates for importables to
the one for exportables will define the bias of the trade regime.       If
there are QRs or tariffs on imports, but no subsidies given to exports,
the system is biased:    firms will have a greater incentive to produce
import substitutes for the domestic market than to produce for export.
3.       Real exchange rate.     The real exchange rate is an index of
relative prices expressed in terms of a common currency.XL    It provides
a measure of the incentive for domestic producers to supply foreign
markets and of the international competitiveness of domestically
produced goods.    In expressing relative price in terms of a common
currency we can use alternatively either the nominal exchange rate or
the nominal effective exchange rate.      On a priori ground the latter
would be more appropriate, insofar as it better reflects the opportunity
1/ Unlike the nominal exchange rate which in terms of any one partner
currency can be expressed as a price, say 2.6 DM per dollar, the
real exchange rate must normally be expressed as an index.        This
index measures the changes, relative to a base period of the ratio
of domestic to foreign goods prices.       For simplicity's sake, we
shall nonetheless in this paper refer to the real exchange rate as
the ratio of two price levels.



RFE.PEXCRATE-2                        - 5 -
cost of a unit of foreign currency.    However it is not unique and may
even vary widely, due to a distortive trade policy, across goods. While
it would be possible to compute an average measure of the nominal
effective exchange rate, the data requirement might be, as we argued
earlier, too demanding.   A simpler, even though not fully satisfactory
approach, is to use the nominal exchange rate.       The impact of trade
policy must then be assessed separately.
4.       The Choice of the Price Index.         When computing   the real
exchange rate we have to determine which price index to use.          The
empirical literature supplies a number of measures of the real exchange
rate based on wholesale prices, consumer prices, GDP deflators, unit
labour costs and unit values of imports and exports.     IHowever, before
evaluating the merits and the demerits of each index, it is essential
that we clarify what we are trying to measure.          It should become
apparent that there is not a uniquely preferable choice of the relative
price index for the purpose of defining the real exchange rate. It will
vary according to the purposes of the analysis and the economic features
of the situation being analysed.    Suppose we are concerned with export
behaviour.   We know that there has been some disagreement both in the
literature and in policy circles on whether the main constraint to
export growth in LDCs comes from international demand or domestic supply
conditions.   The policy implications of this issue are quite relevant.
If world market conditions are the main factors inhibiting exports, a
reduction of the trade bias of the economy would not be very effective
as a way to promote exports.     If, however, domestic supply conditions
are perceived to be the main constraint to export expansion, policy
recommendations should move along opposite lines. It is not always easy



RFE.PEXCRATE-2                        - 6 -
to assess at a given point of time and for a given country whether
demand or supply conditions represent the main binding factors.        The
point to be stressed here is that the relevant definition of the real
exchange rate will differ in the two cases.
(a) Consider the case where the country is small and goods are
relatively homogenous across countries.     Demand no longer matters, as
the country is able to sell abroad any quantity at a given price. Only
supply conditions matter.      The incentives for domestic producers to
supply foreign markets will be measured by the ratio of exportable to
other domestically produced goods prices.     This ratio is the relevant
measure of the real exchange rate.
(b) Suppose now that goods are not homogenous across
countries.    Foreign demand is now downward sloping and effectively
constrains export growth.    The failure to provide convincing empirical
evidence in favor of the "law of one price" would suggest that this is a
more realistic case to consider. In general foreign demand will depend
on world trade volume and, provided it is not fully price inelastic, on
relative prices as well. The latter is the relevant measure of the real
exchange rate.    It will be equal to the reporting country price of
export relative to the export price of foreign competitors expressed in
terms of a common currency.
In general, both demand and supply conditions will matter.     As
a result, both definitions of the real exchange rate will have to be
considered.
The analysis does not substantially change if the focus is set
on the current account balance instead than on exports. We only need to
consider the aggregate basket of tradeable goods, i.e., of all those
commodities which can be potentially exchanged on international markets.



RFE.PEXCRATE-2                        - 7 -
Consider again the case of a small country, which canno't affect
the foreign currency price of traded goods.    The real exchange rate is
now equivalent to the ratio of non-tradeable goods to tradeable goods
prices.      This provides a measure of the evolution of the price
incentive for domestic producers to supply tradeable goods.      When the
relative price of non-tradeable goods decreases, i.e. the real exchange
rate depreciates, the domestic supply of tradeables will increase, while
domestic  demand  will  go down.     Foreign  consumers  will  absorb   at
unchanged prices, the domestic excess supply of tradeables.       This is
the mechanism by which depreciation of the real exchange rate will bring
an improvement of the current account. Or again, suppose that the price
of an imported natural resource (say, oil), which is not being produced
domestically, suddenly increases. At the old equilibrium exch-nge rates
a current account deficit will emerge.     Only if the relative price of
tradeables increases will the economy be able to induce the higher
relative supply of tradeables to pay for the higher oil bill.     The oil
shock would have caused a real depreciation, or, if it does not, either
a balance of payments deficit will emerge (increase), or will be
contained only by fiscal or monetary policies which reduce domestic
income below full employment levels.l/
1/ The ratio of the price of non-tradeable to tradeable goods is a
long-run measure of the price incentive facing domestic producers of
tradeables relative to those facing producers of non-tradeables. It
may be argued that, at least in the short-run, when resources are
relatively immobile, what matters most is the absolute profitability
of producing tradeable goods.      The ratio of unit labour costs to
traded goods prices may give a better indication of this.          Of
course, if unit labour costs move together with non-traded goods
prices, it would not make any significant difference which measure
is being used.



RFE.PEXCRATE-2                        - 8 -
When goods are not homogenous and markets are not perfectly
competitive, the index of the relative incentives for domestic producers
to supply foreign markets may move somewhat differently from the index
of competitiveness of domestic goods in foreign markets.      This may be
important -when, as is the case for manufacture, domestic and foreign
products are somewhat differentiated. The ratio of the index of foreign
to domestic tradeable goods prices expressed in terms of a common
currency is sometimes called the Purchasing Power Parity (PPP) real
exchange   rate.     It   is  a  measure   of   the  international   price
competitiveness of domestic goods.
Measurement
We shall discuss in this section how an index of the real
exchange rate can be empirically computed.     From an empirical point of
view measurement is a crucial issue.       Even if a theoretically ideal
measure of the real exchange rate could be identified, we would still be
left with the problem of implementing it empirically.      In the previous
section we have identified two possible measures:
(a)  the Purchasing PowEc   Parity real exchange rate which is,
basically, an index of the price competitiveness of the
economy
(b) the ratio of non-tradeable to tradeable goods prices which
measures the supply incentive for domestic firms to
produce tradeable goods.



RFE.PEXCRATE-2
A.   The PPP Real Exchange Rate.
Five possible candidates for the construction of the real
exchange rate have been suggested in the literature.     The controversy
focussed mostly on whether any of these five indices constituted an
adequate approximation of the PPP real exchange rate.   However, many of
the insights gained in such discussion can be applied to the measurement
of the relative price of tradeable goods as well. The five measures of
the real exchange rate are based alternatively on:
(a) the consumer price index (CPI)
(b) the wholesale price index (WPI)
(c) the unit value of imports and exports
(d) the GDP deflator
(e) unit labour costs
The main drawback with all these indices is that none of them
fit in unambiguous way the concept of tradeable (or non-tradeable)
goods.   This is, of course, an essential requirement if we aim at
providing an adequate measure of the price competitiveness of domestic
goods on foreign markets.    Consider the case where the real exchange
rate is measured as the ratio of domestic to foreign CPI.         The CPI
however includes many goods and services which are not traded (or even
tradeable) on international markets.    An increase of the relative CPI
for the reporting country may only reflect the fact that non-tradeable
goods prices have increased there at a faster rate than in the trading
partner country (and also faster than tradeable prices in the home
country). This does not represent a decline in competitiveness. It may
represent a decline in the relative incentive to produce tradeables, but
it need not do so if in fact the relative price change corresponds to



RFE.PEXCRATE-2                       - 10 -
the faster growth of productivity in tradeable than in non-tradeable
sectors.   Price competitiveness, as measured by relative traded goods
prices, may even have moved in the opposite direction from the real
exchange rates deflated by the CPI.    Thus, while the CPI represents a
readily available index, it may give misleading results and its
indications should be interpreted with great caution.
The WPI overcomes many of the previous problems, as it is much
more heavily weighted by tradeables.      However, the use is also not
without problems. First we have to tackle the fact that the WPI are not
strictly comparable across countries insofar as they are based on
different weights.      Furthermore,  in a country with a major export
product heavily represented in the WPI, a rise in the latter may reflect
improved foreign demand and not reduced price competitiveness of
domestic goods. Or again, consider the case of a sector whose domestic
market is protected, say, by a tariff, but is nevertheless able to
export part of its production. Possibly domestic producers will be able
to discriminate between domestic and foreign markets. As a result they
will charge higher prices in domestic market. Their monopoly power will
not be constant however but will vary according to domestic demand
conditions, tariff and commercial policy, etc. WPI may not, under those
circumstances, be a good indicator of the price of domestic tradeables
in foreign markets.    Presumably a unit value index (c) would be better
suited to capture the existence of price discrimination between domestic
and foreign markets. The drawbacks of unit values indices are, however,
all well-known.    These are computed as a ratio between the value of
exports  and  a   given  quantity  index.    If  expOrt   aggregates  are
sufficiently homogenous no problfem arises.  Variations of the uiit value
index will reflect only price changes.    if however, as it is genIerally



RFE.PEXCRATE-2                        -
the case, the export aggregate is not homogenous, variations in unit
values will reflect both price and commodity composition changes.     One
possible way to disentangle these two effects would be to define a fixed
weight (Laspeyres) unit value index on the basis of highly disaggregated
unit values.
The CDP deflator and the unit labour costs index have been
relatively less popular deflators in the empirical literature aiming at
measuring the real exchange rate.      This may seem at first somewhat
surprising. The GDP deflator has indeed some desirable properties. It
is a genuine price index. It is less subject to the distortions imposed
by the presence of price controls.     However, its most severe drawback
stems from the fact that, for many LDCs, it is available only on a
yearly basis and often with a considerable lag. Similar considerations
apply to the use of relative unit labour costs as a measure of cost
competitiveness   across  countries.     Unit  labour   costs  depend   on
productivity and wage costs. Productivity, however, varies widely over
the  cycle.    Attempts  to correct   for this effect have led to the
definition of normalised unit labour cost which are based on the trend
value of productivity.     This index is, however, available only for
industrial   countries.    For   developing  countries   the  problem   is
compounded by the poor quality of the wage data.       Overall it is not
surprising that measures of the real exchange rate based on unit labor
costs have been found to be often out of line with all the other
measures.



RFE.PEXCRATE-2                       - 12 -
B.   The Real Exchange Rate as the ratio of Non-Tradeables to
Tradeables Prices
The discussion up to this point has focussed on how to measure
the PPP real exchange rate. No attempt has been made to approximate the
relative price of non-tradeable to tradeable goods prices. However, as
we noticed earlier, this is the relevant measure if it is believed that
domestic supply constraints play a significant role in determining
expc:rt performances.  The discussion in the earlier section suggested
that we cohnsider two polar cases:
(a) exportable and foreign goods are homogenous commodities;
home markets are highly competitive; and the reporting
country is relatively small, i.e., it behaves as if traded
goods prices are given.    Under those circumstances only
supply  conditions  matter.    Arbitrage  on   foreign  and
domestic markets will ensure that the prices of foreign
goods, domestic exports and domestic exportables are
equal. The index of foreign traded goods prices adjusted
for variation of the nominal exchange rate represents
therefore the average revenue from foreign sales to
domestic  producers.    When compared with the index of
domestic non-tradeatle goods prices, it provides an
adequate measure of the real exchange rate.
(b) exportable and foreign goods are imperfect substitutes..
The reporting country faces a negatively sloped demand for
its own exports.   Both demand and supply conditions will
determine export performances. We know that, under those
circumstances, supply will depend on the ratio of



RFE.PEXCRATE-2                       - 13                     l
exportable to non-traded goods prices while (foreign)
demand will be a function of the ratio of foreign to
domestic traded goods prices. It should be intuitive that-
the   index of the domestic non-traded to foreign traded
goods prices should be able to capture both demand and
supply factors.
In both cases, therefore, the index of the ratio of domestic
non-traded goods to foreign traded goods prices is a good measure of the
price incentives influencing the export (or the current account)
performance of a given country.      While no available index measures
directly the prices either of tradeable or of non-tradeables, the ratio
of domestic CPI to foreign WPI may provide a good empirical
approximation of this measure. The WPI is heavily weighted by tradeable
goods and its movements can be deemed to approximate fairly well the
price that domestic producer will receive on foreign markets.     The CPI
at the same time is heavily weighted with non-traded goods.
Multilateral Exchange Rates
1.   Assessing the importance of Trade Partners
For simplicity's sake we have stuck so far to the hypothesis
that the reporting country trades with only one large ("Rest of the
World") partner. In this section we relax this hypothesis and analyse
the way to assess the relative importance of different foreign partners
in computing the real exchange rate.      If tradeable prices could be
measured directly and did not exhibit any significant difference across
countries, the problem would remain simple. We could use alternatively
an index of domestic traded goods prices or, if the latter was not



RFE.PEXCRATE-2                        - 14 -
available, an index of tradeable prices for only one foreign country.
Unfortunately, as we noticed in the previous section, for practical
purposes the movement in the price of tradeables may have to be replaced
by a proxy. Also goods may not be perfectly homogenous across countries
and may fail therefore to command a unique price in international
markets.   Finally price equalisation may not be instantaneous, because
of informational lags, transaction and transportation costs and of the
impact of the numerous non-price factors (punctuality in delivery,
product reliability, etc.) which affect demand.     While there may be a
definite trend toward equalisation of traded good prices across
countries, still, if this process is not instantaneous, it is useful to
consider the behaviour of tradeable goods prices for a sufficiently
large number of foreign competitors.
There are many ways to gauge the relative importance of foreign
competitors.   Often in the empirical literature, a proxy for the index
of prices of tradeable goods for each foreign country is simply weighted
according to the share of domestic exports going to that country.
Import  shares  and/or   trade  shares  may   be  used  instead.    These
procedures, however, are not fully satisfactory. They implicitly assume
that the reporting country exporters compete only with domestic
producers   in  the   importing   country.     This  assumption   may   be
unwarranted.   Consider, for instance, the case of two countries which
trade little with each other (perhaps because of a fairly similar
composition of export) but which compete in the same third markets.      A
double weighting scheme, where, say, export shares are in turn weighted
by the importance of third country suppliers to the given importing
country, might capture this effect but would implicitly assume that the



RFE.PEXCRATE-2                        - 15 -
reporting country exporters do not compete with the importing country
producers, but only with third country suppliers.       Allowing for both
effects may be difficult insofar as, except at a very disaggregated
level, the domestic producer's share of the importing country market may
be overwhelming and, therefore, cancel any third country effect.     Some
arbitrary compromise needs, therefore, to be made.         This issue is
tackled again in the appendix.
2.   An Illustration: Colombia and Korea
We argued in a previous section that the ratio of domestic CPI
to foreign WPI may represent, both from a theoretical and an empirical
point of view, an adequate approximation to the real exchange rate.    We
have computed an index of the real exchange rate based on such measure
for Colombia and Korea.     The results are presented     in Figures 1-2.
In assessing the importance of various foreign trade partners,
multilateral weights based on trade shares have been used and compared
with simple bilateral measures of the real exchange rate to the U.S.
dollar.   In computing trade shares the ten largest partners for each
country were used. In all cases more than 80 percent of total trade is
accounted for.   The results highlight the importance of the weighting
scheme.   For both Korea and Colombia the multilateral and the bilateral
exchange rates heve sometimes moved in opposite directions.            For
instance, from 1980 to 1982 the multilateral exchange rate for Korea
appreciates while the bilateral measure moves the opposite way. This is
mainly due to the recent strong nominal appreciation of the U.S.
dollar.   The problem is not circumscribed to Korea or Colombia.     Given
that many countries used the dollar as the reference currency, their
multilateral exchange rate turns out to exhibit a stronger trend towards



WiI
- I
- 9128



FIGURE 2
KOREA
DOMESI CP K0        WP
/ "/
-* U



RFE.PEXCRATE-2                        - 18 -
appreciation   (a weaker   trend  towards  depreciation)  than  what  the
bilateral exchange rate would suggest.   Thus policy-makers watching only
the bilateral relationship might have been misled into believing their
currency was depreciating, while in fact it was appreciating.
To provide a further check on these results, we have also
computed in Table 1 the real exchange rate using the GDP deflator.    The
latter, while not suitable for monitoring purposes insofar as it is
available for many LDCs only on a yearly basis and with a considerable
lag, can still provide some ex-post insights on the behaviour of the
real exchange rate.    As an approximation to the tradeable goods sector
we have   considered  the  aggregate  of   the  agricultural,  mining and
manufacturing  sectors.      Construction, trade,   business and private
services have been taken as representing the non-tradeable goods
sector.   We have first computed the ratio of domestic non-tradeable to
foreign tradeable goods prices.       For both Colombia and Korea, this
measure shows a stronger tendency to appreciate than the one based on
the comparison between domestic CPI and foreign WPI.      Had we compared
the PPP exchange rates, i.e. the ratio of domestic to foreign traded
goods GDP deflators and the ratio of domestic to foreign WPI (Fig. 3-4),
the results would not have differed that much, particularly for
Colombia.    The previous puzzle, therefore, is very much a function of
the fact that the GDP measure of non-traded goods prices grows at a
faster rate than the CPI.    While it is true that the CPI contains many
tradeables and may also be influenced by price control measures, the
result is nonetheless somewhat surprising.



RFE.PEXCRATE-2                               - 19
TABLE 1:      UEAL EXCHANGE RTES:       COLOMBIA AND KOREA
I                I    DOMESTIC     i CPI TO FOREIGN t      DOMESTIC       CPI TO FOREIGN |
i  NON-TRADED TO  | WPI COLOMBIA    | NON-TRADED TO    |    WPI KOREA
t                  FOREIGN TRADED |                   | FOREIGN TRADED i
I GOODS PRICE     j                 |      GOODS
COLOMBIA      I                 j   PRICE KOREA    |
| s975           |      100.       I      100.        |     100.        |     100.
1 1976           I      107.419    I      104.018     I     115.323     ;     110.07      j
| 1977           |      119.199    j      120.751     |     122.351     j     110.447
| 1978           |      126.012    i      119.643     I     125.671     |     109.367
I 1979           i      132.963    I      120.803     1     155.986     I     118.688
| 1980                  137.812    |      121.317     |     147.981     |     106.69
| 1981           |      151.19     |      132.778     |     147.225     I     110.647
I 1982           I      161.371    I      141.154     I     153.176     I     113.855
I 1983           I      155.435           1 !37.914.        147.074     I     108.495
===-===-I=====-===I                   I========-==-========-===



FIGURE 3
447
R    WCOLieA
PPREAL EXCANG RATES
r _    _  _ _ _ _ _ _ _ _ _ _ _ _  _ _ _ _
2   t\               //I
i                  ,-1  ,"    1  " -
_____Relative WPI ----------Relative GDP DEFL.



FIGURE 4
KOREA
PPREX CHANE RATB
I                            \
-                  /      \N
-S/                        N\
4/ /
-Relative WPI  - - - - - - Relative GDP DEFI.



RFE.PEXCRATE-2                        - 22 -
It is useful to pursue the issue a bit further and decompose
the GDP based measure into its components: the ratio of domestic non-
traded to traded goods prices (the 'internal' real exchange rate) and
the domestic to foreign traded goods price ratio (the PPP real exchange
rate).     The results of this decomposition are presented in Table 2.
They  sharply differ    between  the  two countries.       For Korea   the
international competitiveness of the traded goods sector does not
deteriorate too much (Column 4), but the internal exchange rate exhibits
a strong upper trend (Column 3).      The reverse holds for Colombia: the
PPP real exchange rate appreciates, while the ratio of non-traded to
traded goods prices actually decreases until 1977 and goes up only
moderately afterwards.
This new puzzle is probably easier to explain.        First, the
Colombian index is likely to be influenced by the oscillation of coffee
prices.    However, as we shall see in the next paragraph, even if we
remove this effect, the results do not change much.     Second, we may not
have a good index of traded goods prices.         Due to the existence of
prohibitive tariffs, many manufacturing goods in Colombia may be
virtually non-traded.       In other words,   the loss of international
competitiveness of Colombian traded goods has not made its impact felt
in domestic markets because of the high rate of effective protection
accorded to domestic producers.       Also, our measure may overstate the
loss of international competitiveness of Colombian goods.        Under the
assumption that Colombian producers could act as discriminating
monopolists between domestic and foreign markets, the GDP deflator of
traded and/or manufacturing goods would over estimate actual export
prices.    For Korea, however, given its different trade policy stand, a



RFE.PEXCRATE-2                              - 23 -
TABLE 2: A DECOMPOSITION OF THE REAL
EXCHANGE RATE WEIGHTS
i jINTERNAL REAL                  |    PPP REAL     | INTERNAL REAL  I    PPP REAL
|  EXCHANGE RATE  | EXCHANGE RATE  | EXCHANGE RATE   | EXCHANGE RPrE
l               l    COLOMBIA     |    COLOMBIA     |    KOREA       I      KOREA
1975                 100.             100.        i     100.             100.
1976                 95.5179    I     112.46      I     101.753    I     113.336
1977                 90.7901    I     131.291     I     103.549    I     118.158
I 1978          I      104.081    I     121.071           101.107    I     124.295
| 1979          I      112.154    |     118.554     |     114.486    |     136.248
I 1980          I      109.831    I     125.476     I     115.337    I     128.303
I 1981           |     116.617    |     129.647     |     119.259    |     123.45
1982          I      115.466    I     139.756     I     124.947          122.593
1983           I     112.952    I     137.612     I     ¶27.732    I     115.A42
I=====-===      =I==--=====I==-====-==========-===:====



RFE.PEXCRATE-2                     "    24 -
different explanation is called for and probably wages should play an
important role there.   The index of re,.l wage shot from a base-value of
100 in 1975 to 181 in 1979 and declined af Ietward.    This behaviour is
mirrored by the internal real exchange rate.    Prtsumably, the real wage
push strongly affected non-traded goods prices '' while" the increase of
tradeables prices was instead checked by international competition.    We
should  also   try  to  take   into  account  the   effect  of  technical
progress.     If productivity growth is higher in the tradea'Dle goods
sector, then the increase of the relative price of non-tradeables in
Korea would represent an equilibrium appreciation.    ;By the same token,
the more   stable behaviour of    the  internal srea'! exchange  rate  for
Colombia may reflect smaller inter-Eectoral dif.ference!s in productivity
growth.    Labour productivity, in t4e manuac.th.iring, and in   tertiary
sectors grew respectively at an average annual rate of 6.3% and 2.5% in
Korea between 1975 and 1979, nut declineNd tat,' rate of 2.6% and 1.6% in
Colombia between 1973 and 2978e
These explanations <are, to a larL e externt, mere conjecture.
Any attempt to validate themn empirically, would, take us well beyond the
scope of this paper.     They,, however, highligh.t the danger of relyins.
exclusively on one measure     of the real   exchange rate.       This  is
particularly true if we recall tbi!.t none of the empir'ically .available
indices can fit our theoretical constraints.
We have not yet allowed for t4ie DoIssibility that the tradeable
7/
price index may be influenced by t'h  erratic behaviour of some commodity'
price.  A further check on our results can be obtained if we wor'k at a.
more disaggregated level and cons4de: instead of the whole tradesable
only the manufacturing sector.     'In 'the case of Colombia, this woul.d
7* ,
;e.7,



RFE.PEXCRATE-2                        - 25 -
help in removing the effect of coffee prices.         The results do not
change significantly, though.    Indeed, given that for both Colombia and
Korea prices increase less for manufacturing than for other tradeables,
the divergence of behaviour between the PPP and the internal real
exchange rate is even more pronounced (Tables 3-4).        Nor can these
results be attributed to the choice of the weighting scheme.            In
addition to simple bilateral and trade weights, we have computed export
and composite trade weights where third country suppliers effects can be
taken into account (Tables 5-6).     With the exception of the bilateral
measure, our findings are not much influenced by the choice of the
weighting scheme.
It would be unwarranted to draw any general conclusion from
this simple exercise whose aim was only to illustrate some of the
problems encountered when building an index of the real exchange
rate.    Still, it is difficult to avoid the temptation to highlight the
need to use more elaborate schemes than the one based on simple
bilateral relationships and to avoid relying on only one measure of the
real exchange rate.
Equilibrium and Disequilibrium Real Exchange Rates
After   a) having selected the relevant price index to define
the real exchange rate, b) having chosen the weighting system to assess
the relative importance of the reporting country foreign partners, the
policy-maker is eventually confronted with the behaviour over time of a
given index of the real exchange rate.        The problem now is how to
interpret, both for analytical and policy purposes, the information
provided by such an index.       Probably if all measures of the real
exchange rate did exhibit a pronounced upward trend, one would tend to



RFE.PEXCRATE-2                              - 26 -
TABLE 3:     REAL EXCHANGE RATES FOR THE
MANUFACTURING SECTOR: COLOMBIA
I  DOMESTIC NON   |  INTERNAL      | PPP EXCHANGE
TRADED TO     | EXCHANGE RATE  |     RATE
FOREIGN      I                I
I MANUFACTURING  i                 i
GOODS PRICE   I
=-  I
1975          I     100.        I     100.             100..
1976           |    107.97      |      94.3324   |     114.457    I
1977          I     119.918     I      91.706    I     130.763    1
1978          I     126.889 s         103.363    i     122.761    1
1979          I     136.156     I     110.911    I     122.761
I1980         I     143.077     1     103.005    I     138.903
1981          j     158.883     |     111.094    |     143.017
1 1982          I     169.565     I     110.455    I     153.515
1 1983          |     160.599     |     107.211    I     149.797
I------------I---------------------= -======------I   ====-====
TABLE 4:     REAL EXCHANGE RATES FOR THE
MANUFACTURING SECTOR: KOREA
DOMESTIC NON                         I iNTERNAL   | PPP EXCHANGE
TRADED TO     | EXCHANGE RATE  |     RATE
FOREIGN
| MANUFACTURING  |                I                 I
I   GOODS PRICE   |
| 1975          I     100.       I      100.       |     100.
I 1976          I     116.833     |     102.882    |     113.561
| 1977          I     124.722     |     108.725    |     114.713
I 1978          |     128.        |     116.649    I     109.731
1979                161.009     I     128.777    I     125.029
1980          I     152.654     I     127.023    I     120.178
1981          I     154.568     I     129.712    I     119.162
1982          I     163.297     I     131.726    I     123.967
1983          I     158.273     I     134.703    I     117.498
I               I                 I                I-= 1= = = = -= = = = == = = == = - = 1= = = = -= = = =



RFEE.PEXCRATE-2                            - 27
TABLE 5: DOMIESTIC NON-TRADED TO FOREIGN
MANUFACTURING GOODS PRICES:
COLOMB IA
TRADE WEIGHTS   I  COMPOSITE     I COMPOSITE TRADE|
| EXPORT WEIGHTS      WEIGHTS
I 1975          j     100.       I     100.       I     100.
| 1976          |     107.97      |    107.537     |    108.102
| 1977          |     119,918    1     118.84      |    120.225
1 1978          I     126.889     1    127.246    I     126.405
I 1979                136.156    I     134.868    I     136.24      I
I 1980          I     143.077    I     140.183    I     143.52
1981          |     158.883     |     154.268    |     160.176
I 1982          I     169.565    1     165.864    I     171.409
I 1983          1     160.599    I     156.455    I     162.862
TABLE 6: DOMESTIC NON-TRADED TO FOREIGN
MANUFACTURING GOODS PRICES:
KOREA
|                 TRADE WEIGHTS  |   COMPOSITE     | COMPOSITE TRADE|
I        I         EXPORT WEIGHTS |    WEIGHTS
1975          |     100.     .       100.       I     100.
1976          |     116.833     |    116.895     |     117.338
1977          I     124.722    I     126.519    I      125.707
I 1978          I     128.       I     134.258    I     128.637
| 1979          |     161.009    |     161.962     |    160.426
I 1980          I     152.654    I     147.682    I     151.482
i 1981          I     154.568    1     152.829    I     155.688
1982          I     163.297    I     158.425     I     165.429
1983          I     158.273    I     154.605    I      161.22



RFE.PEXCRATE-2                        - 28 -
conclude that domestic prices have grown out of line with respect to
foreign prices and that a devaluation is called for to restore both
international price competitiveness of domestically produced tradeable
goods and the incentive to domestic firms to shift production toward
tradeable goods.    This conclusion will often be correct; it is useful
nonetheless to exert some caution in this respect. First one must check
that the base year, i.e., the benchmark period with respect to which
exchange rate movements are computed, reflected a situation of overall
equilibrium.   This means that the exchange rate prevailing in the base
year should have been compatible with full employment and an adequate
investment-consumption balance, and with a long-run desirable current
account balance. It would be undesirable to return to a base year rate
which itself was undesirable or unsustainable.      We also want to know
whether the equilibrium value of the exchange rate is the same as in the
base period.   Only if we are convinced that the initial situation was
one of equilibrium and that conditions did not change significantly
afterwards, can we assert that a given variation of the real exchange
rate would bring it back to equilibrium. One should therefore identify
the main factors which may cause the equilibrium value of the exchange
rate to vary. One should consider at least the following:
(a) exogenous variations of the terms of trade
(b) the discovery of a natural resource (the so-called Dutch
Disease)
(c) variations of capital flows
(d) interest-rate shocks



RFE.PEXCRATE-2                        - 29 -
(e)  changes in the stance of commercial policyl-
(f) productivity trends
Consider the economic impact of the discovery of a natural
resource, say oil. On the one hand, as real income increases, part of
the consequent increase in domestic spending goes into non-traded goods
whose relative price therefore increases. On the other hand, resources
are pulled into the newly discovered activity.2- Factor prices increase
because of higher demand and this has a negative impact on the other
sectors of the economy.      On both counts the production of non-oil
tradeable goods declines and     their relative price decreases.      The
discovery of a natural resource causes therefore an equilibrium
appreciation of the real exchange rate.     From the balance of payments
point of view, this may be an equilibrium appreciation, but it may
nevertheless cause some concern in particular if the resource is fastly
exhaustible. Indeed the squeeze of the non-oil tradeable sector may be
undesirable as it involves economic losses (like closing down of firms,
lower export market shares) which may not be easily recouped when the
real exchange rate returns to its long-run normal value.
To take another example, an exogenous increase in the foreign
price of an export will have a very similar effect to the one we have
just described. The increase of coffee price in the second half of the
seventies caused a strong appreciation of the real exchange rate of
1/ Commercial policy is thus treated as having an independent impact on
the real exchange rate.     This is unavoidable insofar as we rely,
when expressing domestic and foreign prices in terms of a common
currency, on the n¢minal exchange rate instead of on the nominal
effective exchange rate.
2/ The resource movement effect may not be of great quantitative
importance for oil, however.



RFE.PEXCRATE-2                        - 30 -
Colombia and had a pronounced negative effect on manufacturing. It was
not easy for manufacturing exporters to regain their market shares
afterwards.   Colombia was confronted at that time with a difficult
policy dilemma. A nominal devaluation of the exchange rate wQuld have
preserved the incentives for domestic producers to supply foreign
markets, but would have added to the inflationary pressure brought by
increased coffee prices. What was not fully realised was the existence
of a third alternative.    If devaluation had been accompanied by a full
liberalization of trade (the latter made easier by the coffee bonanza),
the squeezing of the manufacturing sector may have been avoided without
adding to inflationary pressure.
This example illustrates some of the dilemmas which the policy
makers must face.     It also highlights the fact that the equilibrium
value of the exchange rate may move considerably in response to
exogenous shocks.   For a debtor country, a rise in world interest rates
or a change in lenders' attitudes would also lower the equilibrium
exchange   rate.     More  generally,   caution  will   be  necessary   in
interpreting the significance of a given movement in the real exchange
rate, however defined, and in calculating the direction and magnitude of
the necessary corrections.    It should be clear however that if no major
exogenous shock has occurred, or if it is deemed that variations of the
real exchange rate are too pronounced to be related only to the
exogenous shock, it could be safely concluded that the real exchange
rate is now out of equilibrium and an adjustment of the domestic
currency value is called for.



RFE.PEXCRATE-2                       - 31 -
Monitoring the Exchange Rate
We argued in the introduction that the exchange rate plays a
crucial role both as a determinant of the current account and in the
resource allocation process.   As a result close monitoring of the real
exchange rate could p....ide essential information for policy purposes.
We can distinguish three kinds of situation:
a)   countries where the current account balance represents the
main policy concern
b)   countries where the current account does not represent a
major concern, but where the isst-e of trade liberalization is at the
forefront of the economic policy debate
c)   countries where neither the current account balance nor
trade liberalization constitute a major policy problem, but where policy
makers are concerned about maintaining the relatively high level of
efficiency achieved in the economy.      In all those cases, the real
exchange rate will play an essential role.     Appreciation of the real
exchange rate will signal a process which may jeopardise the improvement
of the current account, the success of the trade liberalization process
and the preservation of a set of sound economic incentives to the
private sector.
Of course, monitoring of the exchange rate will have to be
supplemented by the analysis of other indicators. For instance, if the
main policy concern relates to the. successful implementation of a
liberalization program, the focus should also be set on the behaviour of
-an  index  of the trade bias.     A comparison between the ratio of
exportable to importable prices at home and abroad would provide
significant insights in this respect.    Even if the current account is



RFE.PEXCRATE-2                        - 32 -
,
the main issue, the policy maker may still not be indifferent as to
whether a current account improvement comes from an increase of export
or a reduction of imports.      Indeed,  imports often include a strong
investment component and import reduction could imply an undesired cut
of investments.     It would be useful, under     such circumstances,   to
closely follow the behaviour of the effective exchange rate for imported
capital goods.   Finally a widespread concern about devaluation is that
it will fuel inflation and may, if price increases are passed into
wages, be totally ineffective.     The close monitoring of the ratio of
wholesale price to unit labour costs could indicate whether such concern
is well-grounded.



RFE.PEXCRATE-2                       - 33 -
SELECTED REFERENCES
Edward,  S.    (1985a).  '"Trends in real exchange rate behaviour in
selected developing countries," CPD Discussion Paper, 1985-16.
Edward, S.   (1985b).   "Real exchange rate misalignment in developing
countries:    analytical   issues  and  empirical  evidence,"  CPD
Discussion Paper 1985-43.
Maciejewski, E.   (1983).   "Real effective exchange rate indices," IMF
Staff Papers, pp. 491-541.
Morawetz, D.   (1981).   Why the Emperor's New Clothes Are Not Made in
Colombia, Oxford University Press for the World Bank.
Rhomberg, R. (1976). "Indices of effective exchange rates," IMF Staff
Papers, pp. 88-112.



RFE.PEXCRATE-2                         - 34 -
APPENDIX A
The purpose of this appendix is to show that, when foreign and
domestic   tradeables   are  differentiated   commodities,   the  ratio   of
domestic non-tradeables to foreign tradeable goods prices represents the
relevant measure of the real exchange rate.
Consider the following simple model of the trade balance:
X (PT/PN' PT /N' E)(1)
X     XD (P /P*)                           (2)
M  =  M (PT 'EN' PTI N
p
B  =         x - M                          (4)
T
The prices of domestic tradeables, domestic non-tradeables and
foreign tradeables   are denoted respectively by PT'       pN and   PT     E
represents real expenditure, while X, M and B denote respectively export
and import flows and the trade balance.          Equation 1 repreents the
excess   supply   of   domestic   tradeables,   i.e.   the   export   supply
function.     In equilibrium it is equal to foreign demand for domestic
tradeables (Equation 2).     Notice that domestic and foreign tradeables
are by hypothesis differentiated products and, therefore, command
different prices.     Equation 3 represents domestic demand for foreign
tradeables, i.e. imports.    Equation 4 defines the trade balance.



RFE.PEXCRATE-2                         - 35 -
Taking log derivatives of Equations 1 - 4, we get:
X    y    (P  - P)-y     (P*  -)             (5)
xx   T    N      xm  T     N
X  -   Tdx (P -p     )                      (6)
I            .    I                I
N      y    (P  -P)     y   (     -P)       (7)
mm   T     N     mm   T     N
dB=   M [X  + (P      T       ]            (8)
T    T
The dominance of substitution effects implies that all         y'sare
positive.    For instance, the income and substitution effect on domstic
demand   for   domestic   tradeables   of  an   increase   of    P  /P  are
T   N
respectively negative and positve.       The stronger substitution effect
implies that domestic demand increases and, as a result, export supply
decreases.
Solving for dB/M:
dB/M  = ----------- IYdx       -        )  -Y)                   (9)
xx~ dx
+(Yml YX    Ym   YX )      YO  (Yml  YXr   ( T     N )
The trade balance, therefore, will be driven by the ratio of
foreign tradeable to domestic non-tradeable goods prices.      An increase
of  P /P   s  i.e. a real appreciation, will have a negative impact on B
N T
if the generalized version of the Marshall Lerner condition embedded in
Equation 9 holds.
Let us look at this condition more closely.         Suppose.  PT /PN
increases with   P /P    unchanged.    Export demand goes up while export
T N



RFE.PEXCRATE-2                        - 36 -
supply  decreases.      Then   PT /IN   must   increase  to re-establish
equilibrium.    If own price effects dominate cross price effects (if
Y   > Y  ), then P /P   will not have to increase as much as PT 1PN'
xx    xm         T N'.
ie. PT    T   will decrease.    As a result, exports will increase and,
provided that    YMM > y  ,  imports will decrease.        For the trade
balance to improve, we only need these effects to be strong enough to
compensate for the terms of trade loss.



RFE.PEXCRATE-2                         - 37 -
A,
APPENDIX B
The purpose   of this   appendix  is to describe the weighting
schemes used in this paper.
Let X, M, Xi and Mi denote respectively total exports, total importz,
exports to and imports from country i of the reporting country.       Then
X   = X/X and mi = Mi /M represent the export and import shares of
country i.
Third couIntry (j) effects can be expressed as:
= .x. m(i)
where  m(i)     denote the share of country j in country i imports.
We now define
1     a.   m      X   Xi  M+X       trade share
ii+
[2]     b.       (S. + x.)            composite export share
2    1     1
3]  c.m.           +  b
[3]       i    i= M+X      i   X+M         composite trade share
The choice to give equal (1/2) weight to importing country producers
(xi) and third country competitors       (d.) effects is arbitrary.       It
would be possible to importing country market share of both third
country and importing country producer.        Generally, this last share
would dominate and we would in the end revert to simple trade share.
Tables 1A and 2A present the weights for manufacturing in Colombia and
Korea respectively.      Aggregate trade weights have been borrowed from
Edwards (1985a).



RFE.PEXCRATE-2                             .-8
TABLE !A:    MANUFACTURING EStC'TC,R IWEIGHTS: COLOMBIA
TRADE WEIGHTS       COMP.3SITE    | C fiP0 ITE TRADEI
1 |                                  EXPORT WEIGHTS I    'WEIGH.T     I
''-I == == '---
USA            I      43.2       I      29.      41 3.2?4
l.CANADA         |       2.7       |       4775'
| JAPAN                lto.6    ' I "      9.8      1       11 72     i
| GERMANY V          .s1.7    i           10.35             12 .23
|UK              |    A  3.8       1       2.9              4.34
| ITALY          I       3.7               5.15    i        3.91
I FRANCE               6           ' 6.                      14
MEXICO                 0.                0.65  t   I       0.13
I 0. A$IA NES*           0 °.   ¢I         0.75             0 1 5
H HONK KONG      I       0.        I       0.55  \          0.11
|KOREA           I       °-       t4               I|       0.13
IBELGIUM        I       0.        T                I        0'28
ISPAIN A3.9                       I       0 .      I       4. 0?2
B 8RAZIL*                0.                0o,G5    I      1 0.13
I NETHERLANDS      .,    0. '               .35    I        Q0.27
i SWEDEN         I '     1.8                                 1 ;'5  .'j  1.93
| SW`TZERLANDE   I       2.3              C, 65       |      Z ;45
| VENEZUELA      I       3.4        '      8.                I ^ 1,'2
|ECUADOR         |       3.-       I  -         5            1.5
| PERU          '        2                          I        1,14
| PANAMA                 0.0       I       2,,15    i -      0.43
; DOM. REP.t        .    0.'              . 1.8 A        .   036
==  = = =-  == . 1   ==I-=  =-=      ---=-   ==  -    = ==  =  =,   -:
*The Wholesa'le POrice Inde "'was. uKad -ivn fe~~h           J
deflator when computinig-bte r ali ax6iminge raLe,
I             ,
r             ,r



RFE.PEXCRATE-2                            - 39 -
TABLE 2A:     MANUFACTURING SECTOR:     KOREA
I ==ZZ===== =-I=========I            =3=-s====|=t==-===
|               | TRADE WEIGHTS  |   COMPOSITE    | COMPOSITE TRADEt
I |                EXPORT WEIGHTS |    WEIGHTS     I
| USA           |      31.3      1      32.1      |      25.9
| JAPAN         |      42.1      I       19.75    I      39.98
| GERMANY       I       7.        |      10.6             8.11
| CANADA        |       4.1              9.55     |       6.28      I
|UK             |       3.8      I       5.8       |      4.63
| HONK KONG     I       1.8       |      2.55     I       1.35
| NETHERLANDS   I       1.9       |      2.7      |       1.43
| FRANCE        |       1.8       |      3.3      I       3.54     I
| ITALY         I       0.4      I       2.8      I       .1.86
| MEXICO        I       0.        |      0.9 g            0.48      I
| 0. ASIA NES*          0.        |       1.6             0.85      |
S SWITZERLAND* R        0.4       |       1.2     1       1.06
| BELGIUM               0 0.4     |       1.35     |      1.09
| SWEDEN        I       0.        |      0.6      I       0.32
| CHINA *       I       0.       |       0.8       |      0.42
| IRAN t                1.9      I        1.8      |      0.95      I
| LIBERIA t     |       1.4      I        1.35            0.71
t SAUDI AR.     |       1.3       |       1.25     1      0.66
| PANAMA                0.4              0 0.      I      0.38
* See Table 1A
.