INTELLIGENCE MEMORANDUM INTERNATIONAL FINANCE SERIES NO. 35 A MODEL TO PREDICT THE IMPACT OF THE EXCHANGE-RATE AGREEMENT ON INTERNA

• Document Type: CREST
• Collection: General CIA Records
• Document Number (FOIA) /ESDN (CREST): CIA-RDP85T00875R001700030043-0
• Release Decision: RIPPUB
• Original Classification: C
• Document Page Count: 23
• Document Creation Date: December 22, 2016
• Document Release Date: January 18, 2011
• Sequence Number: 43
• Publication Date: March 1, 1972
• Content Type: IM

Sanitized Copy Approved for
Release 2011 /01 /18
CIA-RDP85T00875R0017000
Sanitized Copy Approved for
Release 2011/01/18:
CIA-RDP85T00875R0017000
 Sanitized Copy Approved for Release 2011/01/18: CIA-RDP85T00875R001700030043-0
25X1
Confidential
DIRECTORATE OF
INTELLIGENCE
Intelligence Memorandum
International Finance Series No. 35
A Model to Predict the Impact of the Exchange-Rate
Agreement on International Trade
Confidential
ER IM 72-43
March 1.972
Copy No.
88
Sanitized Copy Approved for Release 2011/01/18: CIA-RDP85T00875R001700030043-0
 Sanitized Copy Approved for Release 2011/01/18: CIA-RDP85T00875R001700030043-0
WARNING
This document contains information affecting the national
defense of the United States, within the meaning of Title
18, sections 793 and 794, of the US Code, as amended.
Its transmission or revelation of its contents to or re-
ceipt by an unauthorized person is prohibited by law.
GROUP 7
Excluded f om "ta alit
dawn0cudinp and
d,cIoU Gcnhtnn
Sanitized Copy Approved for Release 2011/01/18: CIA-RDP85T00875R001700030043-0
 Sanitized Copy Approved for Release 2011/01/18: CIA-RDP85T00875R001700030043-0
CONFIDENTIAL
CENTRAL INTELLIGENCE AGENCY
Directorate of Intelligence
March 1972
A MODEL TO PREDICT THE IMPACT
OF THE EXCHANGE-RATE AGREEMENT
ON INTERNATIONAL TRADE
Introduction
1. This memorandum uses a trade-flow model to examine the impact
on international trade of the international realignment of parities reached
18 December 1971. The model is similar to that developed and used by
the International Monetary Fund (IMF), an approximation of which was
described in an earlier memorandum of this Office.(1)
2. The model was substantially advanced and refined for this
memorandum and now explicitly reflects the time lags in trade-balance
adjustment an d the impact of international differences in rates of economic
growth. However, the model still does not reflect international. differences
in rates of inflation, the impact of most trade restrictions, or numerous
other influences on international trade. The parameters are, except for
Japan, those used by the IMF and have not been tested. For these reasons,
the model's predictions should be used with great care, bearing in mind
the simplified assumptions on which they are based. Pending further
refinements and testing, these results should be regarded as indications of
general orders of magnitude and not as estimates of this Office.
Note: This memorandum was prepared by the Office of Economic Research.
CONFIDENTIAL
Sanitized TI_-_~
AT '_ -_
Sanitized Copy Approved for Release 2011/01/18: CIA-RDP85T00875R001700030043-0
 Sanitized Copy Approved for Release 2011/01/18: CIA-RDP85T00875R001700030043-0
CONFIDENTIAL
Discussion
.The Model
3. In the period after 15 August' 1971 the IMF hoped to facilitate
early agreement on new exchange rates by recommending a package of new
currency parities designed to achieve trade-balance targets similar to that
suggested by the Organization for Economic Cooperation and Development
(OECD). In making these recommendations the Fund used a trade-flow
model developed by its staff. A description of the model and the
assumptions made about its parameters have been published.(2)
4. We have developed a trade-flow model with the basic structure
of the Fund model but with several additions. The CIA model -- unlike
the IMF model - not only includes equations indicating the relationship
between price changes and changes in the amount of goods supplied and
demanded, but also equations reflecting the delay before the full impact
of the parity changes is felt. Our model, moreover, permits real output
in each country to vary, while the IMF model assumes real output in each
country remains constant. By varying the rate of growth of real output,
we can explicitly consider the impact of national business cycles on the
trade balances.(3)
The Smithsonian Agreement
5. The agreement reached by the Group of Ten (Belgium, Canada,
France, West Germany, Italy, Japan, Netherlands, Sweden, the United
Kingdom, and the United States) and Switzerland on 18 December 1971
ended a four-month period of floating exchange rates and increasing
monetary uncertainty. Its key features include a 7.9% devaluation of the
US dollar relative to gold; an 8.6% appreciation both of the British pound
and French franc; a 13.6% appreciation of the German mark; and a 16.9%
appreciation of the Japanese yen, all relative to the dollar. The Canadian
dollar - floating since May 1970 -- continues to trade at about an 8%
premium over its previous parity with the US dollar (see Table 1). For
comparative international accounts of these countries, see Table 2.
3. The CIA model is described in detail and evaluated in the Appendix.
2 -
CONFIDENTIAL
Sanitized Copy Approved for Release 2011/01/18: CIA-RDP85T00875R001700030043-0
 Sanitized Copy Approved for Release 2011/01/18: CIA-RDP85T00875R001700030043-0
CONFIDENTIAL
Table 1
Exchange-Rate Changes from pre-May 1970 Parities
Percent
Country
Relative
to Gold
Relative to
the US Dollar
United States
-7.9
--
Japan
7.7
16.9
Canada
Continues
Nearly 8.0
Belgium-Luxembourg
to float
2.8
France
No change
8.6
Italy
-1.0
7.5
Netherlands
2.8
11.6
Sweden
-1.0
7.5
Switzerland
4.9
13.9
United Kingdom
No change
8.6
West Germany
4.6
13.6
Other OECD a/
7.4
a. Excluding Australia and Yu osZavia; including
Austria, Denmark, Finland,. Greece, Iceland, Ire-
land, Norway, Portugal, Spain, and Turkey;
weighted by 1970 export shares.
CONFIDENTIAL
Sanitized Copy Approved for Release 2011/01/18: CIA-RDP85T00875R001700030043-0
 Sanitized Copy Approved for Release 2011/01/18: CIA-RDP85T00875R001700030043-0
Comparative International Accounts, 1971 a/
Million US $
Country
Trade
Balance
Exports
f.o.b.
Imports
f.o.b.
Current
Account
Balance
International
Reserves
as of December
United States
-2,840.
423800
45,640
-1,520
13,190
C
J
7
900
630
23
730
15
5
880
15
360
z
apan
Canada
,
2,230
3
17,670
3
153440
,
50
,
5,700
Be lg iurmt-Luxembourg
540
10,330
93790
600
3,470
France
780
22.0550
21,770
-420
7,490b/
Italy c/
-400
14,290
14.X690
850
6,790
Netherlands
-260
15,100
15,360
-390
3,800
Sweden
300
73300
7,000
140
1,110
Switzerland
-1,750
53700
7,450
-210
6,970
United Kingdom
700
21.9300
20.S600
2,150
6,580
West Germany
7,910
42,200
343290
160
18,380
a. Preliminary, except entry for International Reserves.
b. As of November.
c. Transactions basis except entry for International Reserves.
Sanitized Copy Approved for Release 2011/01/18: CIA-RDP85T00875R001700030043-0
 Sanitized Copy Approved for Release 2011/01/18: CIA-RDP85T00875R001700030043-0
CONFIDENTIAL
Its Trade Impact
6. Our detailed examination of the agreement's impact comprises
two cases. For each we make different assumptions about the time period
covered and about the rates of real-output growth. In the first case - the
comparative static (or long-run equilibrium) case - we hold real output
in each country constant. This assumption permits us to separate the effect
of price changes from the effect of income changes.(4) This, in turn, enables
us to isolate the direct impact of the parity realignment. Thus Case 1 results
are not a prediction of the actual changes that will take place in the trade
balances over any particular time period, but rather indicate how the price
effect of the Smithsonian realignment alters these balances in the long run
from what they would have been in the agreement's absence. Conceptually,
Case 1 corresponds to the analysis undertaken by the IMF staff during
the exchange-rate negotiations.(s)
7. In Case 2 - the dynamic case - we permit real output in each
country to grow at an exogenously determined rate. This assumption allows
us to consider simultaneously both the price and income effects of parity
and real-output changes. Thus Case 2 results are predictions of the changes
in the trade balances in the near term, given assumed rates of growth of
real national output.
8. In both cases the results are given in current US dollar prices.
These prices will rise because of the appreciation of most important
currencies relative to the dollar. In addition, Case 2 prices will rise because
of increasing demand pressures. In the fiyst case the price rise accounts
for approximately 12% of the increase in the dollar value of world trade;
in the second, world trade prices in dollars will increase at an average annual
rate of about 13% over the period.
Assumptions
9. The exchange rate agreement's trade impact depends, in part, on
the parameter values chosen to indicate the strength of the casual linkages
among changes in prices and changes in flows. The parameter values used
in the model are the same as those employed by the IMF; they are described
4. Parity changes induce changes in real output as well as in prices and trade balances;
these induced real-output changes, in turn, indirectly, further affect prices and trade
balances. Governments, however, can adjust the rate of real-output growth to some
desired level. We assume, therefore, in isolating the direct price effect, that real output 25X1
in each country remains unchanged and that the indirect price and trade-balance effects
are neutralized.
CONFIDENTIAL
Sanitized Copy Approved for Release 2011/01/18: CIA-RDP85T00875R001700030043-0
 Sanitized Copy Approved for Release 2011/01/18: CIA-RDP85T00875R001700030043-0
CONFIDENTIAL
in the Appendix. However, we make an adjustment in the case of Japan,
which has experienced far greater growth rates for both GNP and exports
than any other OECD country. We assume that Japan's outstanding export
performance of the 1960s will continue through 1974. We assume,
accordingly, that the Japanese supply elasticity is one and one-half times
larger than the supply elasticity of other countries - that is, for any given
percentage increase in price, the Japanese will increase their output in
percentage terms one and one-half times more than other producers. Even
with this adjustment, the assumption that all countries have the same
demand and supply elasticities is, of course, a great oversimplification. In
reality these parameters undoubtedly differ among countries as a result of
differences in the composition of trade and in their economic policies and
institutions.
10. The relatively high proportion of Japanese trade subject to
quantitative restrictions necessitates a further adjustment. Commodities
making up about 40% of Japanese exports are covered either by quotas
in importing countries or by a variety of voluntary Japanese restraints and
export controls. Japan's trade would probably not respond, therefore, to
the parity realignment in the way predicted by an unadjusted model. It
is consequently assumed that those exports covered by quantitative
restrictions grow 10% annually in value terms (beyond the 16.9% adjustment
for yen revaluation). Although some Japanese imports are also covered by
quotas or other quantitative restrictions, the proportion is probably not
high in absolute terms nor in relation to the quantitative import restrictions
maintained by several other countries included in the model. The Japanese
do make widespread use of informal import restrictions, but there is no
way to measure their impact. Even with the adjustments made, the model
may lead to some overstatement of the impact of the parity realignment
on Japanese trade. Failing to take the restrictions of other countries
explicitly into account may also cause some overstatement of impact.
11. The exchange-rate agreement's trade impact also depends, in part,
on the non-OECD countries' actions. Estimates of the US trade-account
improvement made with the assumption that these countries do not, on
the average, further alter their exchange rates or import policies are likely
to understate substantially the size of the US improvement. Developing
countries - accounting for the bulk of non-OECD country trade - in
particular, are unlikely to allow a considerable improvement in their trade
balances to occur. Instead, they are likely to increase their imports by means
of other policy measures such as easing of import controls. We assume,
therefore, that the manufactured imports of all non-OECD countries taken
as a group rise by one-half of the improvement in their aggregate trade
balance implied by the Smithsonian agreement. These additional imports
are distributed according to each OECD country's share in total OECD
manufactured exports to non-OECD countries, with the United States
accounting for about 25% of the total.
- 6 -
CONFIDENTIAL
Sanitized Copy Approved for Release 2011/01/18: CIA-RDP85T00875R001700030043-0
 Sanitized Copy Approved for Release 2011/01/18: CIA-RDP85T00875R001700030043-0
CONFIDENTIAL
Case 1 - Equilibrium Impact
12. The long-run equilibrium impact of the exchange rate
.agreement - with real output in each country remaining constant - is
shown in Table 3 for each of the Group of Ten countries and Switzerland,
for other OECD countries as a fxoup, and for the non-OECD countries
as a group. The values in the table indicate in current prices, on an annual
basis, by how much the price effect of the Smithsonian realignment would
alter each country's total trade balance and its bilateral balance with the
United States, as compared with what these would have been in the
agreement's absence.
13. Using the model, we calculate that the price effects of the
Smithsonian agreement would improve the US trade account by around
$11.5 billion on an annual basis, or nearly the amount sought. Higher
exports account for most of the improvement; they increase about $9.5
billion in the period over what they would have been in the agreement's
absence. Imports would decline about $2.0 billion. On a commodity basis,
the greatest improvement would be in US exports of manufactures, which
should increase about $8.0 billion; US imports of these products should
decline by about $2.1 billion (see Table 4). Trade-balance improvements
of about $650 million each are also projected in US trade in food products
and crude materials.
14. The model indicates that Japan would experience sharp
trade-account losses because of the agreement. Japan's imports would
increase about $4.1 billion while its exports are expected to increase only
$1.3 billion. About 40% of the expected deterioration in Japan's total trade
account is accounted for by the deterioration in its bilateral trade with
the United States. Japanese exports to the United States, according to the
model, would hardly increase at all while Japanese imports from the United
States would increase about $1.2 billion.
15. Among European countries the sharpest deterioration in trade
account would occur in West Germany - about $3.3 billion and in the
United Kingdom - about $1.7 billion. Although the French franc
appreciated relative to the dollar by the same amount as the British pound,
the French trade account would deteriorate only slightly because, relative
to the currencies of their trade partners, the franc appreciated much less.
A large proportion of French trade is with West Germany, while most UK
trade is with the European Free Trade Area (EFTA) countries,(6) the United
States, and Canada.
6. Including Austria, Denmark, Finland, Ireland, Norway, Portugal, Sweden,
Switzerland, and the United Kingdom.
- 7 -
CONFIDENTIAL
Sanitized Copy Approved for Release 2011/01/18: CIA-RDP85T00875R001700030043-0
 Sanitized Copy Approved for Release 2011/01/18: CIA-RDP85T00875R001700030043-0
Case I: Static Equilibrium Impact of Agreement
Million US $
Change in Bilateral
Change in Total
Account
Trade Account
With the United States
Country
Balance
Exports
f.o.b.
Imports
f.o.b.
Balance
Exports
f.o.b.
Imports
f.o.b.
United States
11,504
9,485
-2,019
--
--
--
Japan
-2,798
1,317
4,115
1,140
15
1,155
C7
Canada.
-1,464
276
1,740
2,091
-503
1,588
O
Z
Belgium-Luxembourg
-898
1,034
1,932
512
-122
390
France
-128
2,028
2,156
513
-111
402
Italy
-109
1,457
1,566
467
-149
318
1~1 co
zI
Netherlands
-1,141
1,092
2,23.3
561
-65
496
Sweden
-130
702
832
198
-53
145
Switzerland
-1,054
235
1,289
331
-85
246
United Kingdom
-1,684
1,383
3,067
928
-268
660
West Germany
-3,308
1,903
5,211
1,423
-565
858
Other OECD b/
-1,192
1,599
2,791
623
-106
517
Rest of world
2,403
6,517
4,114
2,717
-7
2,710
a. Change in world trade prices (in US dollars) = 11.6%.
b. Excluding Australia and Yugoslavia.
Sanitized Copy Approved for Release 2011/01/18: CIA-RDP85T00875R001700030043-0
 Sanitized Copy Approved for Release 2011/01/18: CIA-RDP85T00875R001700030043-0
CONFIDENTIAL
Equilibrium Impact of Agreement:
Commodity Composition of Change in US Trade
Million US $
Food
Crude
Materials
Mineral
Fuels
Manu-
factures
Balance
680
637
76
10,111
Exports f.o.b. 718
643
145
7,979
Imports f.o.b. 38
6
69
-2,132
16. The Canadian dollar's appreciation - assuming it continues tc
trade at near parity with the US dollar - results in a worsening of Canada's
trade account by about $1.5 billion from what it would have been in the
agreement's absence. The deterioration in the bilateral account with the
United States more than accounts for the change in Canada's total trade
account; Canadian exports to the United States decline about $0.5 billion,
and Canadian imports from the United States increase about $1.6 billion.
Case 2 - Trade Impact in 1972-74
17. The speed with which the full impact of the parity changes is
felt depends on two factors. First, there is a lag in the response of producers
and consumers to new market conditions. Second, there is a delay between
the time contracts are negotiated and prices set and the time merchandise
coveied by these contracts enters the importing country's market. We
assume that the delay between contract and import averages six months,
during which time the new parity changes have no impact on the trade
balances (in terms of dollars). The parity changes influence the trade
balances in subsequent periods, but their full impact is still not felt. This
is so because it is assumed that suppliers and consumers make their
production and consumption decisions in light of past, as well as current,
prices and incomes (for details see the Appendix).
18. The exogenously determined rates of -Pal-output growth used in
the model are derived primarily from OECD estimates. For 1972, we assume
that real output of each country grows at the GNP (or GDP) growth rates
estimated by the OECD. For the first half of 1973, we assume real output
grows at a rate midway between the growth rates in the second halves
of 1972 and 1973. For the second hall of 1973 and for 1974, we assume
that real output grows at rates the OECD believes are required to achieve
9 -
CONFIDENTIAL
Sanitized Copy Approved for Release 2011/01/18: CIA-RDP85T00875R001700030043-0
 Sanitized Copy Approved for Release 2011/01/18: CIA-RDP85T00875R001700030043-0
CONFIDENTIAL
full employment by 1974-75. We also assume that parities do not change
during the period; thus assumed differences in rates of real-output growth
have a very strong influence on the trade balances, particularly in the latter
part of the period, when much of the impact of the Smithsonian realignment
would have already been felt.
19. The dynamic trade impact of the exchange-rate agreement is
shown in Table S. The values in the table are projections in current prices
of changes in trade balances for the given real-output growth rates, assuming
that the basic structure of international trade relations is not altered.
Changes in tariff and non-tariff barriers, EC enlargement, majv- new
innovations, and other structural changes would alter these predi:,,ions.
20. The model projects a substantial improv-ment in the US trade
account during 1972-74 despite the assumed rapid growth in US output.
In the first half of 1972, however - before the parity changes have an
impact on the trade balances - the trade account deteriorates by about
$900 million. Such a deterioration, although anticipated by most analysts,
could increase uncertainties in foreign exchange markets and precipitate
further speculative capital movements. In the second half of 1972, the trade
account, improving by about $2.3 billion, would show a slight deficit.
Overall, in 1972 Cie US merchandise trade account would show a deficit
of about $3.6 billion, or about $800 million more than in 1971. The model
shows that the US trade account would swing into surplus in the second
half of 1973, and by the end of 1974, on a semiannual basis, it would
be in surplus by about $1.6 billion, or $2.8 billion for the year. Thus the
swing in the US trade balance would be about $5.6 billion (at an annual
rate) over the three-year period.
21. Although the introduction of time lags and a real-output variable
makes the model more realistic, some important variables continue to be
omitted. There is in pai?iicular no attempt to project the rate of inflation,
apart from those price increases that directly stem from the parity and
income changes themselves. Past experience indicates that different
economies respond differently to demand pressure and suffer from cost-push
inflation to different degrees, depending on their institutions and policies
and on the growth of productivity. In the late 1960s, US export prices
grew substantially more rapidly than those of our competitors. A continued
relatively poor export price performance in the future would reduce the
impact of the parity realignment, and the improvement in the US trade
account would be smaller than predicted.
22. Although Japan's trade account is expected to deteriorate during
the period because of the large yen revaluation and the economy's very
rapid growth, the Japanese would continue to run a large surplus. The model
predicts that during the first half of 1972 Japan's record trade surplus will
10 -
CONFIDENTIAL
Sanitized Copy Approved for Release 2011/01/18: CIA-RDP85T00875R001700030043-0
 Sanitized Copy Approved for Release 2011/01/18: CIA-RDP85T00875R001700030043-0
CONFIDENTIAL
Came 2s Dynamic Impact of Agreement J
1972
1973
1974
Country and Indi
t
Jul-Dp
V
ca
or
-r -
1971
Jan-Jun
Jul-Dec
Jan-Jun
Jul-Dec
Jan-Jun
Jul-Dec
United States
Assumed annual real growth of GNP
Trade balance (Million US S)
(percent)
5.75
6.00
0.25
8.50
6.50
6.50
Change
Total
-900
2,301
463
594
785
453
Japan
-2,070
-2,970
-669
-206
388
1,173
1,626
Assumed annual real growth of GNP
Trade balance (Million US 5)
(percent)
6.170
7.00
8.50
10.00
10.00
10.00
Change
Total
-310
-531
-06
-145
-317
-361
Canada
4,490
4,172
3,641
3,555
3,410
3,093
2,732
Assumed annual real growth of GNP
Trade balance (Million US S)
(percent)
6.25
6.50
8.50
6.50
6.50
0.50
Change
278
-106
108
214
184
167
Total
790
1,068
962
1,070
1,284
1,460
1,635
Belgium-Luxembourg
Assumed annual real growth of GNP (percent) 2/
3.25
5.75
4.32
4.90
4.90
4.90
Trade balance (Million US 5)
Change
Total
251
-30
64
105
172
163
France
270 d/
521
491
555
660
032
995
Assumed annual real growth of GNP
Trade balance (Million US 5)
(percent)
5.00
5.00
5.37
5.75
5.75
5.75
Change
Total
-27
-87
-90
-74
3
-46
Italy
330
303
216
118
44
47
1
Assumed annual real growth of GNP
Trade balance (Million US 5)
(percent)
3.00
5.00
5.50
6.00 ,
0.50
6.50
Change
Total
49
-386
-169
-168 -255
-214
-200
-151
-537
-706
-974 -1,129
-1,343
- 11 -
CONFIDENTIAL
Sanitized Copy Approved for Release 2011/01/18: CIA-RDP85T00875R001700030043-0
 Sanitized Copy Approved for Release 2011/01/18: CIA-RDP85T00875R001700030043-0
CONFIDENTIAL
Cese 2s Dynamic Impact of Agreement J
(Continued)
1972
1973
1974
Jul-Dec
Country and Indicator
1971 b Jan-Jun Jul-Dec
Jan-Jun
Jul-Dec Jan-Jun Jul-Doc
Netherlands
Assumed annual real growth of GNP (percent)
3.00
3.00
3.00
4.25
2/ 4.90
4.90
Trade balance (Million US $)
t ~.Ige
2
-144
-35
-26
-45
24
i.'ota1
-130
J -120
-272
-307
-333
-370
-354
Sweden
Assumed annual real growth of GNP
(percent) J
3.00
3.25
3.70
4.10
4.10
4.10
Trade balance (Million US $)
Chance
137
24
23
46
i0
74
Total
150
/ 207
311
334
300
458
532
Switzerland
Assumed annual real growth of GNP (percent)
3.50
4.00
3.80
3.70
3.50
3.00
Trade balance (Million US S)
Change
-157
-281
-70
-69
-67
-79
Total
-880
/-1,037
-1,310
-1,396
-1,465
-1,532
-1,611
United Kingdom
Assumed annual real growth of GNP
(percent)
3.75
3.25
3.50
3.75 f
4.00
4.00
Trade balance (Million US $)
Change
-91
-00
23
96
56
126
Total
610
519
439
462
558
614
740
West Germany
Assumed annual real growth of GNP
(percent)
2.00
3.50
3.50
4.35 ,
5.20 j
5.20 l/
Trade balance (Million U5 $)
Change
2,067
-128
621
537
368
702
Total
4,930
6,997
6,869
7,490
8,027 8
,395
9,097
Other OECD
Assumed annual real growth of GNP
(percent)
4.00
4.75
5.00 g/
5.25 g/
5.50 g/
5.50 y./
Trade balance (Million US $)
Change
-1,139
-972
-584
-675
-760
-76l
Reat of world
Assumed annual real growth of GNP
(percent)
5.00
5.00
5.25
5.50
5.75
5.75
Trade balance (Million US $)
Change
-152
419
-253
-433
-835
-963
a- Value data are seasons y adJunted on a semi-annual basin, peroentage data on an annua aoiu.
b. Preliminary; for full year, ese Table 2.
a. Entries for 1872 are oomponitee of OECD and national estimates. Entries for the oeoond half of 1973
through the aeoond half of 1974 are OECD Zong-run estimates of growth.
d. Assumed to be one-half of the 1971 total.
o. CIA assumption for transitional period to long-run growth path.
f. OECD estimated of Zong-run growth rates.
g. CIA assumption.
12 -
CONFIDENTIAL
Sanitized Copy Approved for Release 2011/01/18: CIA-RDP85T00875R001700030043-0
 Sanitized Copy Approved for Release 2011/01/18: CIA-RDP85T00875R001700030043-0
CONFIDENTIAL
be reduced by about $320 million on a semiannual basis. In the second
half of 1972 the trade account would deteriorate an additional $530 million.
In 1972, Japan's overall merchandise trade account would show a surplus
.of aoout $7.8 billion, or nearly the same as in 1971. According to the
model, this trade surplus will continue to deteriorate, and it will be reduced
to about $5.8 billion in 1974.
23. Despite West Germany's large effective revaluation, the model
predicts that its trade surplus will increase substantially because of the
economy's relatively slow anticipated growth during most of the period.
The increase in the German trade surplus in the first half of 1972, when
the economy is expected to grow at an average annual rate of only 2%,
or less rapidly than any other major trading nation, would be unusually
large - about $2.1 billion. By the end of the three-year period, the German
trade surplus would have increased about $9.6 billion. If German economic
growth is more in line with recent experience and so is more rapid than
assumed, the trade account improvement, of course, would be substantially
less than predicted.
24. None of the European countries except Italy and Switzerland is
expected to experience a sharp trade-account deterioration during the
1972-74 period. The French trade account would deteriorate by about $730
million. The British trade account, according to the model, will improve
by about $650 million because of the economy's relatively slow anticipated
growth. EC entry and poor export price performance, however, could offset
the expected UK trade-account improvement. The Swiss, who, despite a
large trade-account deficit in 1971, revalued their currency by a greater
amount than any country except Japan, are expected to experience
increasing deficits through the period.
25. Canada's trade surplus - assuming the Canadian dollar continues
to trade at near parity with the US dollar - is shown as increasing on
an annual basis by about $870 million over the period. Although the
Canadian surplus would increase about $280 million in the first half of
1972, it would decline in the second half of the year. The annual surplus
for 1972 - $2 billion (based on Canadian export and import data) - would
be slightly smaller than in 1971. Canada's trade account, on a semiannual
basis, would improve by about $320 million in 1973 and by about $350
million in 1974. The improvement is due, in part, to the assumed rapid
economic growth in the United States, Canada's most important trading
partner.
A Note on the Balance of Payments
26. The CIA trade-flow model does not permit the user to relate the
parity changes to elements of the balance of payments other than the trade
CONFIDENTIAL
Sanitized Copy Approved for Release 2011/01/18: CIA-RDP85T00875R001700030043-0
 Sanitized Copy Approved for Release 2011/01/18: CIA-RDP85T00875R001700030043-0
CONFIDENTIAL
account. The current-account and basic balances of a devaluing country
are likely to improve, however, by more than the improvement just on
merchandise trade account. Particular elements of the service account are
very sensitive to parity changes. Expenditure on travel, for example, is
probably abc'it as responsive to changes in exchange rates as is expenditure
on traded goods. Investment earnings are probably relatively insensitive in
the short run to parity changes, In the long run, however, the realignment
may increase profitability in the devaluing country, thereby increasing the
outflow of investment remittances. Certain elements of the capital account
are quite sensitive to parity changes. Production will probably become more
attractive in a devaluing country because of the exchange rate change.
Multinational companies will therefore choose to "source" more of their
output there, which will tend to increase the net inflow of new direct
investment.
27. Portfolio investment and short-term, capital flows are probably
relatively insensitive to parity changes. Short-term capital moves primarily
in response to interest-rate differentials and expectations as to future
exchange-rate changes. A devaluing country's short-term capital inflow will
consequently increase only if that country also has higher interest rates
or is expected to appreciate its currency.
Conclusions
28. The CIA trade-flow model predicts that the 18 December 1971
agreement to realign exchange rates will substantially strengthen the US
merchandise trade account and the US balance of payments. Although the
trade account will deteriorate in 1972 by about $800 million, an annual
surplus of $2.8 billion can be expected by 1975, for a swing of about
$5.6 billion in a three-year period. The very large Japanese, German, and
Canadian trade-account surpluses will continue through the period. Indeed,
Germany's trade surplus is expected to increase substantially because of
its relatively slow economic growth.
- 14 -
CONFIDENTIAL
Sanitized Copy Approved for Release 2011/01/18: CIA-RDP85T00875R001700030043-0
 Sanitized Copy Approved for Release 2011/01/18: CIA-RDP85T00875R001700030043-0
CONFIDENTIAL
APPENDIX
The Model: Its Assumptions and Limitations
Description of the Model
The revised CIA trade-flow model is designed to forecast the effect of
real-output and exchange-rate changes on the pattern of trade balances. It is
based on the modified-share approach-that is, apart from the effect of price
changes, each exporting country is assumed to maintain its share of trade by
value with each importing country.
The model identifies 13 countries or groups of countries: Belgium-Luxem-
bourg, Canada, France, West Germany, Italy, Japan, Netherlands, Sweden,
Switzerland, United Kingdom, United States, rest of OECD (excluding Australia
and Yugoslavia), and the rest of the world; and five classes of goods: food,
beverages, and tobacco (SITC 0-1); crude materials (SITC 2-4); mineral fuels
(SITC 3); and manufactures (SITC 5-9); and a class of nontraded goods and
services. A good produced by a particular country-here named a "product"-is
assumed to have special characteristics that dil"ferentiate it from similar goods
produced elsewhere. In total, the model includes 65 (13 x 5) different products,
each supplied by one country.
World demand for a particular product is related to three factors: to the
distribution of trade, to changes in each of the 13 countries' or areas' total mone-
tary expenditure on all goods and services, and to changes in relative prices
among similar products.
CONFIDENTIAL
Sanitized Copy Approved for Release 2011/01/18: CIA-RDP85T00875R001700030043-0
 Sanitized Copy Approved for Release 2011/01/18: CIA-RDP85T00875R001700030043-0
Sanitized Copy Approved for Release 2011/01/18: CIA-RDP85T00875R001700030043-0
 Sanitized Copy Approved for Release 2011/01/18: CIA-RDP85T00875R001700030043-0
CONFIDENTIAL
Data Inputs
Three types of numerical inputs are required for the model: first, a matrix
of trade data showing commodity flows within and from all countries or areas;
second, a list of parameters indicating the strength of the causal linkages among
changes in various flows; and third, the input-output coefficients for the "cost-
push" equations. Data on foreign merchandise trade in the CIA model arc in
the form of a 5 x 13 x 13 export matrix for 1970; the data arc expressed in annual
averages in millions of US dollars. Thn matrix is derived from OECD market
summaries and includes (along the diagonal) values representing internal trade
in each country. It was assumed that each country's internal trade in each product,
by value, was equal to twice the value added in that product's production, less
exports. The parameter values used in the model are the same as those employed
by the IMF, except in the case of Japan. The demand parameters used are
summarized in Table Al, the supply elasticities are summarized in Table A2.
The input-output coefficients are derived from standardized input-output tables
prepared by the UN Economic Commission for Europe.
Sensitivity of the Model
While the trade data are quite solid, the problem is that the parameter values
have to Le econometrically estimated or assumed. Cler.ly, caution must be
exercised in using the model. The results arc sensitive to the particular parameter
values chosen, and the parameter values themselves are subject to wide margins
of error.
To test the sensitivity of the model, several simulations using alternative
elasticity values were run. In each simulation, 20 sets of elasticity values for
a particular set of parameters were drawn from a modified normal distribution,
with the mean approximately equal to the value assumed for the set of parameters
in the tables, and with the variance equal to twice the mean.' The results of
The normal distribution was modified-by taking the absolute value of the selected
number-so that the distribution contained no value with a sign reversed from the parameter
values originally assumed.
CONFIDENTIAL
Sanitized Copy Approved for Release 2011/01/18: CIA-RDP85T00875R001700030043-0
 Sanitized Copy Approved for Release
Food
Crude
Materials
Fuels
Manufactures
\on-traded
Price elasticity of demand for the good...........
-0.50
-0.50
-0.50
-1.00
-1.10
Elasticity of substitution between similar products.
-1.50
-1-50
-1.50
-2.50
-
Expenditure elasticity of demand for the good....
0.50
0.50
0.50
1.00
1.10
Foed .........................................
Crude materials ...............................
Fuels ........................................
Manufactures .................................
Non-traded ...................................
Food
Crude
Materials
Fuels
Manufactures
Non-traded
2.0
0.0
0.0
-0.3
-0.7
0.0
1.0
0.0
-0.3
-0.2
0.0
0.0
0.0
0.0
0.0
-0.1
-0.05
0.0
3.0
-1.2
0.0
-0.02
0.0
-0.7
3.0
Assumed to be the same for each country, with respect to quantity. For the rest of world, the direct price elasticity of supply for
rood was taken equal to 1.0; and the cross-elasticities of supply with respect to a change in the price of manufactures and non-traded
goods were taken equal to -0.2 and -0.3, respectively. For Japan the direct price elasticity of supply for food was taken equal to
3.0; for crude materials. 1.5; for manufactures, 4.5; and for non-traded goods, 4.5.
Sanitized Copy Approved for Release 2011/01/18: CIA-RDP85T00875R001700030043-0
 Sanitized Copy Approved for Release 2011/01/18: CIA-RDP85T00875R001700030043-0
CONFIDENTIAL
Hennltivity to Parameter AHHUnnpiionM
(Million US $)
Mean Impact
with New
Vnluen
Itango of
Impact with
Now Valunn
Standard Devia-
ti me of Impact
with New Values
United HtotcH ...........................
1(1,242
(1,5(10
I H25
.lapona .................................
10,(1:3(1
8,253
I ,404
Canada ................................
. 2 ,313
I 5124
430
France .................................
- 141)
I , 520
3)17
United I(ingdonn ........................
:3,()153
2,(112
698
West, Germany .. . ........ . ............. .
8,020
314(1
H115
United Htnt(H ...........................
1(1057
21 031
5,267
1015:^a .................................
0,11IO
1(1,74(1
2,991
Canada ................................
2,400
3,657
I 031
Franc(.e .................................
217
I 048
3511
United Kingdom .. . ........... . ........ .
3 ,277
4 ,340
1 ,128
West, Germany ..........................
5 ,(35(3
(1,144
I 687
United States .......... ................
111,147
(1,008
I ,:31)7
Japans ........... ....................
-1),1()))
4,183
1,004
Canada ................................
2 234
1 , 51(15
:33H
France .................................
182
1,608
:385
United Kingdom ........................
2 845
1 ,4117
:33:3
Went Germany ..........................
0,I 50
5,8411
1 204
I Twenty Hetn of price and expenditure elasticity of demand values are drawn from it modified
normal distribution with it mean of 0.75 and It variance of I.8.
2 I~ Xeludlog quota effect.
s'Twenty nets of Plasticity of nubntito(ion values are drawn from it modified normal distribution
with it mean of 2 and It variance of 4.
Twenty sets of pricy elasticity of supply values on, drawn from it modified normal distribution
with it mean of '2 and it varianee of 4 for the diagonal elements nod it mean of 0,05 and it variance
of ((.1 for the off-diagonal (lementn.
the simulations are summarized in Table A3, For six major countries, it presents
the mean trade balance impact with the new parameter values, the range of
the trade-balance impact with the new values, and the standard deviation of
the trade-balance impact with the new values. If the true elasticiti( s are charac-
terized by the assumed distribution, which seems roughly reasonable, then in
at least 75 of 100 cases the actual equilibrium trade-balance impact will be
within a range, plus and minus two stands rd deviations, around the mean.3
The model d&zs not appear to be especially sensitive to the values chosen
for the price and expenditure elasticities of demand or the price elasticities of
' Derived, using Tchebycheff's inequality, assuming the sample mean and variance are
accurate estimates of the population mean and variance. If the trade-balances impact is normally
distributed the corresponding Intervals are 95% confidence intervals.
- 19 -
CONFIDENTIAL
Sanitized Copy Approved for Release 2011/01/18: CIA-RDP85T00875R001700030043-0
 Sanitized Copy Approved for Release 2011/01/18: CIA-RDP85T00875R001700030043-0
CONFIDENTIAL
supply. The 75% confidence interval for the US equilibrium trade-balance im-
provement ranges from $12.6 billion to $19.9 billion for various price and ex-
penditure elasticities of demand, and from $13.4 billion to $18.9 billion for the
various price elasticities of supply.
Unfortunately, the CIA trade-flow model, and the IMF model from which
it is derived, are very sensitive to the assumptions made about the elasticities
of substitution. The 75% confidence interval for the US equilibrium trade-balance
improvement ranges from $6.4 billion to $27.5 billion for various elasticities of
substitution.
Accuracy of the Model
The ultimate test of the model and of the parameter values chosen is the
model's ability to forecast accurately the effect of ' hanges in real output and
exchange rates on the trade balances. There has been no opportunity to test
the model.
- 20 -
CONFIDENTIAL
Sanitized Copy Approved for Release 2011/01/18: CIA-RDP85T00875R001700030043-0