Income and Price Elasticities of Foreign Trade Flows: Econometric Estimation and Analysis of the U.S. Trade Deficit

International Finance Discussion Papers Number 324

June 1988

INCOME AND PRICE ELASTICITIES OF FOREIGN TRADE FLOWS: ECONOMETRIC ESTIMATION AND ANALYSIS OF THE US TRADE DEFICIT

Jaime Marquez

NOTE: International Finance Discussion Papers are preliminary materials circulated to stimulate discussion and critical comment. References in publications to International Finance Discussion Papers (other than an

acknowledgment by a writer that he has had access to unpublished material) should be cleared with the author or authors.

This paper builds, estimates, and simulates a world trade model to provide a quantitative analysis of the behavior of the U.S. trade deficit. A key feature of this model is that international trade imbalances add up to zero. The analysis estimates income and price elasticities for bilateral import equations, tests for the properties of the error term, for parameter constaricy, and for the choice of dynamic specification. The paper also reexamines the structural asymmetries in elasticities noted by Houthakker and Magee and tests whether the Marshall-Lerner condition holds. The reliability of the model as a whole is assessed with residual-based stochastic simulations. The paper finds that changes in relative prices account for the bulk of the deterioration of the US trade account, that reliance on either foreigri or domestic growth to eliminate the US external imbalances entails significant changes in real income, and that the speed with which US net exports: respond to exchange rate changes is sensitive to minor changes in own-

price elasticities.

Income and Price Elasticities of Foreign Trade Flows: Econometric Estimation and Analysis of the US Trade Deficit

Jaime Marquez!

1.Int:roduction

That the rest of the world will absorb the reduction of the US trade deficit is a matter of accounting: because this deficit is the trade surplus of the

rest of the world, eliminating one means eliminating the other. What is not

so evident is which countries will absorb this reduction, an issue of growing interest in view of the significant re-allocations of productive factors that are likely to emerge. That the US trade deficit will improve in response to a real depreciation of the dollar is not in doubt. What is not so clear is how long it will take, an important question given the constraints that external

imbalances impose on the design of macroeconomic policy. Given the significance of these issues, it is not surprising to find an

increasing interest in explaining the behavior of the US trade deficit.”

I want to thank Ralph Tryon for his comments and encouragement in this project. I have also benefited from the comments of Russell Cooper, Frank Diebold, Jonathan Eaton, Hali Edison, Sebastian Edwards, Neil Ericsson, Richard Haas, William Helkie, David Howard, Joe Gagnon, David Gordor,, Peter Hooper, Maureen Kilkenny, Edward Leamer, Ross Levine, Cathy Mann, Paul Masson, Mare Noland, Adrian Pagan, and Peter Tinsley. I am also grateful to Carolyn Litynski: and Michael Mabry for their exceilent research assistance and to Linda Matthews for typing several drafts of this paper. The empirical tests are performed with the GIVE computer software developed by David Hendry and installed into TROLL by Ralph Tryon. Earlier versions of this paper have been presented in seminars at the Federal Reserve Board, the U.S. Department of Agriculture, the 1987 and 1988 meetings of the Society for Economic Dynamics and Control, and the 1°87 Summer School in Dynamic Optimization, Dublin, Ireland. I am resporisible for any remaining errors. This paper represents the views of the author and should not be interpreted as reflecting the views of the

Board of Governors of the Federal Reserve System or other members of its staff.

2 See Krugman and Baldwin (1987), Bryant and Holtham (1987), Helkie and Hooper (1987), Feldstein (1986), Hooper and Mann (1987), Eichengreen (1987), Cline (1988), Marris (1987), Hooper (1988).

However, as currently developed, the literature is not well suited to addressing the issues raised here. First, existing analyses do not account for all world trade and thus cannot identify how foreign deficits respond to the reduction of the US trade deficit. By focusing exclusively on US multilateral trade flows, the literature implicitly assumes that policies to reduce this deficit are independent of the distribution of countries absorbing it. Second, the trade elasticities are estimated with multilateral trade flows, a feature that introduces an aggregation bias, implies that the crossprice elasticities are zero, and contradicts the empirical evidence against multilateral trade equations found by Thursby and Thursby (1984) .3 Finally, as required by the literature on the persistence of the trade deficit, the adjustment period of US net exports to a given change in exchange rates is assumed to be both fixed and known. However, to the extent that this adjustment period is a function of trade elasticity estimates, which are random variables, existing analyses rest on a very restrictive assumption. This paper eliminates these restrictions by building and estimating an econometric trade model to explain bilateral trade flows among all trading partners in the world: Canada, Germany, Japan, the United Kingdom, the United States, other industrial countries, OPEC, non-OPEC developing countries, and

the rest of the world.” The (non-diagonal) entries in the resulting 9x9

3 Exceptions include Bergsten and Cline (1985) and Haynes, Hutchison, and Mikesell (1986). These studies focus on US-Japan trade without considering other US bilateral trade flows. Houthakker and Magee (1969) are among the first to study the behavior of bilateral trade for the United States. This paper extends their analysis by considering all countries.

4 This selection of country disaggregation is designed to match tnat of the FRB Multicountry Model (MCM).

trade matrix are modeled as bilateral import equations and the econometric specification used to estimate the associated trade elasticities is developed in section 2. Although the econometric estimation of these elasticities has a history of its own, the estimation of bilateral trade elasticities has received considerably less attention, especially for bilateral trade between incustrial and developing countries, and between the latter and OPEC.>

To relate these estimates to previous studies, the paper re-examines the asymmetries in income elasticities noted by Houthakker and Magee (1969) and tests the Marshall-Lerner condition. Finally, because of their importance for the issues raised here, the paper tests the properties of the error term, the choice of dynamic specification, and the assumed parameter constancy. Section 3 presents the econometric estimates of income and price elasticities. The analysis finds a substantial dispersion of bilateral elasticities across countries, confirms the asymmetries found by Houthakker and Magee (1969), and cannot reject the Marshall-Lerner condition.

Section 4 assembles the bilateral trade equations into a world trade model. To determine the reliability of the model as a whole, the paper uses residual-based stochastic simulations and multiplier analysis. Based on the est:imated model, section 5 uses counterfactual simulations to address two key questions: which countries will absorb the US trade deficit and how sensitive is the adjustment period of net exports to alternative parameter realizations.

The evidence reveals three findings of interest. First, the response of

> See Goldstein and Khan (1985) and Magee (1975) for surveys of the literature. Other models estimating bilateral trade flows are Houthakker and Magee (1969), Hickman and Lau (1973), Armington (1969), and Edison et al (1987).

foreign trade accounts to a reduction in the US trade deficit is sensitive to the manner in which this deficit is reduced--that is, through dollar depreciation or income contraction. Differences in bilateral income and price elasticities across countries determine the type of response. Second, reliance on either foreign or domestic growth to eliminate the US external imbalance entails significant changes in income levels, both foreign and domestic. Finally, the adjustment period of the US trade deficit in response to changes in exchange rates is highly sensitive to minor perturbations of the

elasticity estimates .°

6 Existing Computable General Equilibrium models (Whalley 1985, Fretz et al 1986, Deardoff and Stern 1986, Srinivasan and Whalley 1986, and Schwartz and Krissoff, 1987) are not well suited to address the questions raised in this paper. First, they introduce a real-financial dichotomy that negates the role of changes in nominal exchange rates in af<ecting the structure of trade. Second, they lack dynamic considerations and thus leave unspecified both the period required for adjustment and the properties of the associated adjustment path. Finally, paramete:: values are not estimated but drawn from literature surveys, which precludes statistical inferences. This paper also relaxes some of the assumptions embodied in existing econometric trade models. For example, the World Trade Model of the International Monetary Fund uses multilateral trade flows for industrial countries disaggregated across commodities but it pays little attention to trade vis-a-vis developing countries (see Spence 1984, Deppler and Ripley 1978). The World Model of Project Link explains bilateral trade flows across commodities and with developing countries, but it assumes constant trade shares.

2. Empirical Analysis

2.1 Econometric Formulation

To estimate bilateral

trade elasticities, the analysis relies on the

imperfect-substitute model (Goldstein and Khan, 1985), in which imports of

country. k from country s are modeled as

5 * P P (1) In(M, ./e Po), %9 + Mesh Wie + Weg (1M Ee ° Inyye)

Mest

+

2523453 IPPs tj

+ ZA 4Ksj InPyat-j

* + As in(My /e Pe) ey + 6D + Use?

dollar value of imports of country k from country s,

potential income of country k,

‘real income of country k,

p* * es s/ er yk?

* * - (1 te?) Ve P Vowk, qs,

fixed export-share of the pth country in world exports,

export price of country s, in domestic currency

= price for domestic output, in domestic currency

exchange rate, (dollar/foreign),

dummy variable for one-time events,

2 ~ NCO, O16) E(tyge: Ups gg) 7 0 VI-

According to (1), bilateral imports are homogeneous of degree zero in

prices and imperfect substitutes both for domestic products and among

themselves. The response of imports to income has two components: a secular effect, measured by the parameter a> and a cyclical effect captured by the parameter ay.’ The own-price elasticity is a3, and the effect o= competing prices from third countries is captured by the cross-price elasticity a,. If a,>0, then Ms and imports from a third country are said to be substitutes; if a,<0, then these imports are complements. The choice of a logarithmic formulation is based on the results from Box-Cox tests reported in appendix A. Finally, (1) also includes dummy variables to control for the effects of onetime events on imports.°®

Equation (1) recognizes that the response of imports to changes in income

or prices is subject to delays that might arise from contracts and delivery

lags.” This paper assumes that these delays follow an Almon lag:

2 345; ~ so + Asij + As2j . for j-0,...,J,

2 AKsy 7 Aso + Agil + Agel , for jnO,...,5q.

The (long-run) income and own-price elasticities associated with (1), "es and

? Haynes and Stone (1983) obtain a frequency decomposition for U.S. data and estimate income elasticities for each frequency. Their method is not applied here because it would require generating a series of trade matrices--one for each frequency, which is an overwhelming task given the size of the trade matrix used in this analysis (9x9).

8 In addition to being bilateral, eq. (1) is not included in the list of models that Thursby and Thursby (1984) examine. In their notation, eq. (1) is Qenf CY, Y./XT_, Almon on P, and P,, Dummies for one-time events, 4):

9 Equation (1) also assumes that the length of these lags is fixed; Gagnon (1987) and Husted and Kollintzas (1984) develop models that allow the speed of adjustment to change.

Oks respectively, are constructed as 10

(2) (3)

Mes ~ Mg /61-25K5) »

fxs 7 =j%3nsj /(b-*5ks?

The corresponding elasticities for multilateral imports are ™% 7 =sFesks ° Si - = .bxs ks ,

where brs - M/E Vi,k,s; imkss.

Finally, the residual Use is assumed to be white noise and the

stochastic structure associated with country k’s bilateral trade equations is

yuo (Uys: Wk URL Wen?’ ~ NCO, a),

2 2

2 2 where a. - diag(o), A es On?

and n is the number of trading partners of country k, 9 in this case.

10 The cross-price elasticity is constructed as ¢,. = = Oaks 4 /(1-acy,)-

2.2 Hypothesis Testing 2.2.1 Error Properties and Choice of Dynamic Specification The reliability of (1) depends on whether the residual exhibits serial independence, normality, and homoskedasticity. To test the first of these properties, the paper applies an F-test to the hypothesis that all the coefficients of an AR(4) for the residual are equal to zero. Tne hypothesis of normality is tested with the Jarque-Bera statistic (Jarque and Bera, 1980) and homoskedasticity is tested with an ARCH test (Engle, 1982).

To evaluate the validity of the Almon restrictions, the paper estimates (1) with and without the Almon restrictions and then performs an F-test. The

associated F-statistic is

F- [(SSRp - SSR,)/d,)]/[SSR,/dg] - F(d,, d2),

where SSR,y is the sum of squared residuals with the Almon lag (Hg), SSR, is the sum of squared residuals without the Almon lag (H,), d,= number of additional parameters under the alternative hypothesis, d.— number of degrees of freedom. To determine whether the overall dynamic specification of (1) is violated

by the data, the analysis compares the sum of squared residuals for both (1) and an "unrestricted" dynamic specification. The latter is constructed by eliminating the Almon restrictions and including, as additional regressors, all predetermined variables lagged one period. If these additional regressors do not contribute significantly towards the reduction of the sum of squared

residuals of (1), then it is not possible to reject the choice of dynamic

specification embodied in (1). The test-statistic is constructed as F_ above, a

- 11

and it is denoted Fue

Finally, this paper tests for the constancy in structural parameters, an important consideration in view of both the need for accurate forecasts and the vclatility in prices and exchange rates. This test is also useful for

determining the extent to which the hysteresis hypothesis is supported by the

data. !* To test for parameter constancy, (1) is first estimated with data

through 1983Q2 and then used to forecast imports through 1985Q2. Testing for

parameter constancy amounts to applying an F-test to the hypothesis that the

expected forecast error is zero (Chow 1960, p. 595) .43

2.2.2. Trade Elasticities The paper tests for the presence of structural asymmetries in income elasticities--namely, " > nt and the validity of the Marshall-Lerner

condition--that is, whether &x + st < -1, where nt - 25% ok"sk and

gt - 25% oe sk: Given the relevance of these two tests to evaluating the

effect:iveness of changes in exchange rates in correcting world external

i It ts evident that the selection of a dynamic specification entails a

certain amount of data mining, a process that requires an increase in the size o% the significance level for significance tests. As a result, this paper uses the 99 percent significance level.

12 According to the hysteresis hypothesis, exchange rate changes produce asymme=ric changes in the speed of response of trade flows. Losses in export markets, import penetration, and displacement of domestic production account for a slow adjustment in response to a depreciation (see Krugman and Baldwin, 1987). Krugman and Baldwin (1987) test hysteresis with dummy variables, which is formally equivalent to the Chow test performed here.

13 Appendix B presents sequential Chow tests for the one-period ahead forecast errors.

10

imbalances, they are applied to each country's multilateral trade flows and to world trade. Testing these hypotheses for world trade serves as a consistency check on the country estimates given that world exporzs and world imports have the same income and price elasticities.

Because the elasticities relevant for these tests are constructed as the

ratio of normal random variables (see eqs. 2 and 3), their distributions are not generally known in advance. 14 As a result, this paper generates these distributions empirically using the Monte Carlo procedure developed by Krinsky

and Robb (1986) .1° To implement this procedure, the paper assumes that

a A A A

(4) (AoKg +7: %3Rg 6h) --- Sk5)! =m a(L) - N(ay es Bs)

A

and uses this assumption to generate a random sample of a, the jth drawing of which is denoted as a , jal...4000. Substitution of each a | into (2) and (3) generates random samples for the long-run income and price

elasticities, (nd) and (eo) respectively, from which it is possible to

obtain the associated cumulative density functions, Fl. and Fees The paper

presents the median, the scaled median absolute deviation, and the 99 percent

14 The conditions (Greenber and Webster, 1983) that permit deriving the exact density function of this ratio are not met in this paper.

15 It is possible to re-parametrize (1) in order to estimate directly both the long-run elasticity and its associated standard error. However, this procedure gives only asymptotic results. The advantage of the Monte Carlo approach is that it avoids reliance on asymptotic properties.

16 The reliability of this procedure depends on the assumption that (2) and (3) are continuous functions, an assumption that is violated if a,=1. Based on the evidence presented below (table 6), the continuity assumption is not violated by the data.

ll

significance level for each of the empirically-generated elasticity

distributions. !”

2.3 The Data Data on trade flows for all countries, except data on trade flows with respect

to the Rest of the World (ROW), come from the Direction of Trade, published by

the International Monetary Fund. Bilateral imports of country k from ROW are

est:imated as the difference between total imports of country k and the sum of

country k’s imports from countries other than row: +8

Me - M. - EM, Vs =r, k.

Similarly, data for bilateral exports of country k to the ROW bloc are constructed as the difference between total exports of country k and the sum

of its bilateral exports to countries other than ROW:

Xe = x, - EX Vs =r, k.

7 The critical values associated with the 99 percent confidence interval are derived to ensure that the cumulative density at the tails of the distribution is the same for both critical values. An alternative precedure would be to derive these critical values to ensure that they have equal density.

18 One implication of the residual nature of the ROW sector is that bilateral trade between ROW and another country bloc (e.g., OPEC) includes intra-trade of the latter country bloc. For example, exports of OPEC to ROW include exports of OPEC to OPEC; similarly, imports of OPEC from ROW include imports of OPEC from OPEC. Although treating ROW as a residual biases the estimate of its trade flows vis-a-vis other country blees, this bias is the same for exports and imports and thus cancels out in the computation of trade accounts.

12

Given data on all bilateral trade flows with respect to ROW, total exports of

this bloc are Xx. = ry

s sriisr’ where Por is an adjustment factor that accounts

for differences in recording practices, delivery lags, and CIF/FOB differentials. Total imports of ROW are ML - zXee:

Real income for Canada, Germany, Japan, the United Kingdom, and the United States is defined as real GNP measured in domestic currency. For both industrial and developing countries, real income is measured as a geometric mean of industrial production for selected countries. Finally, in view of data difficulties, the paper assumes that OPEC’s income equals OPEC's exports. !? .

Data on potential output for Canada, Germany, Japan, the United Kingdon, and the United States are generated using Cobb-Douglas production functions. These functions include labor (L), capital (K), oil, (0), and imports (M) as inputs and the associated parameters are estimated econometrical.ly. Potential

value added (yb) equals

yR = £(RE LE Oy, My) - (PM /P yy)

where f(.) is the predicted value of gross output, KP is the level of the ‘ P capital stock at full capacity, Ll is the labor force, and (Pane / Py) equals

the value of imports in terms of domestic output. Data for potential output

19 As a result, the model determines OPEC’s income endogenously. The

countries included in the aggregate of developing countries (fixed weights in parentheses) are South Korea (32 percent), Mexico (36 percent), and Taiwan (32 percent). The countries included in the bloc of other OECD countries are Austria (5 percent), Belgium (15 percent), France (26 percent), Italy (18 percent), Netherlands (17 percent), Norway (4 percent), Sweden (8 percent), and Switzerland (7 percent).

13

of LDZs and the bloc of industrial countries are generated as a trend of actual output.

Trade prices are measured as multilateral export unit values with 19721. By being both multilateral and unit values, these export prices are not ideally suited for estimating bilateral price elasticities, but data on bilateral export prices are not available. Finally, the analysis uses only four bilateral exchange rates (against the US dollar): the Canadian dollar, the DM, the pound sterling, and the yen. Data for prices of LDCs, OPEC, and the bloc of other industrial countries are denominated in US dollars.

An issue relevant to the construction of the data is the statistical discrepancy between total world imports and total world exports. These discrepancies arise for several reasons: different recording practices across countries, incomplete coverage, asymmetric valuation of imports and exports for countries with incomplete reporting, and lags in recording a given trade flow as both an export and an import: because of shipment delays.7° Figure 1 presents the behavior of this discrepancy for 1976-1984 as a percentage of world imports. As the figure reveals, the average discrepancy fluctuates arourd 1 percent with a relatively large standard error. Rather than

distributing this discrepancy across individual countries, this paper treats

20 See International Monetary Fund (1987), pp. v-vi. Note that, despite treating the ROW sector as a residual category, world exports are not equal to world imports in this analysis because it is not assumed that Xk - Ms: In other words, there are two trade matrices: one for exports

and one for imports, which for the reasons mentioned in the text, accounts for the discrepancy in world trade figures.

13a

va6t ¢86l 2e6él 1861 O86t 6/61 B26t LL6t 961

vt

cl

vO

2 OF

+O!

c0 SLYUOdWI QTYOM JO LNA0YAd INNOODDV JGVYL GIYOM NI AONVdSYOSIG

T aanbia

14

it as an exogenous variable.

An examination of the evolution of the direction of trade (table 1) reveals several features of interest. First, the share of US imports in world trade has grown from 12 percent in 1970 to 15 percent in 1984 whereas the share of US exports has declined from 12.5 percent to 10.7 percent during the sane period.?? Second, OPEC’s exports represented 6 percent of world trade in 1970, but this share increased to 15 percent in 1980 as a result of the large oil price increases and the small short-run price elasticity for oil consumption. By 1984, however, this share had declined to 10 percent. Third, between 1970 and 1984, Japan increased its share of the import market in all countries. Finally, the proportion of trade among developing countries, OPEC,

and Centrally Planned Economies has shown a substantial increase.

a1 To the extent that the analysis focuses on trade responses, this

exogeneity assumption ensures that the adding-up constraint holds because changes in world exports are still identical to changes in world imports.

2 Exports of the United States do not include exports of products made by US multinetionals operating in foreign countries.

l4a

Table 1 The Structure of International Trade: Trade Matrices for Selected Periods

(percent)

EXPORTING

co

ROW Sum

ind LbCs OPEC

3

1970Q4

OoODOOOCOOOCOOON

wmrmMongdad MINAMOONS 5 OA ANNMOOAG

IONAOOOOD. ° NO ONwWONOMAAN

ANDANRAMMUO AD ONAAANN TAO

eo © © & @ «@

COCONDCOOOM

n

Germany Japan

Canada US

UK Ind LDCs OPEC ROW

sot.

1975Q4

oo 8 © eo ee an ONNONONMAN

oar oO rNIAHN AMNNNOMSTON

NOOHOANHOG

noo InNnwoonm st NOHOORNONN

ANAMNNE ANA FoGAGN asad

COCO ONCOON

NONNSTORN Oe CADADNNN ©

COCONCDCOOOMm

J

n

Germany Japan

Canada US

UK Ind LDCs OPEC ROW

1980Q4

o 6« © © © an ON SOHONMAN

NItOOo AN OW AttTONONE

~MNIMNnnrwoo sts COOAODAANAD

COC OND COON

wn

Germany Japan

Canada US

UK Ind LDCs OPEC ROW

1984Q4

AomeANnNIRE Or DOONNNANO TO COCONCOCOMm

> n 43 [oj

oe ¢ fa dome C CO ee 91 uO SSSSSERSE

for

percent; ght not add up to 100 percent

mports, expressed in

the sum of shares across countries mi

because of rounding.

Source: Direction of Trade, International Monetary Fund.

Entries are bilateral imports as a share of world i

a given period,

15

3. Estimation Results

3.1 Trade Elasticities

This paper applies ordinary least squares to estimate the parameters associated with ay .23 The estimation sample is based on quarterly data for 19711Q1 - 1985Q2 for imports of non-OPEC countries and 1973Q1 - 1984Q4 for imports of OPEC. Appendix C presents all of the estimated equations along with several summary statistics.

The evidence on secular income elasticities (table 2) offers several features of interest. First, bilateral income elasticities exhibit a large range of variation, a dispersion that weakens the case for aggregate import demand equations. Second, countries fall into one of two categories depending on whether their secular income elasticity is below or above one. The “low” income elasticity countries are Japan and LDCs; the "high" income elasticity countries are Canada, Germany, the United Kingdom, the United States, and

other industrial countries; the income elasticities for OPEC’s bilateral

23 To simplify the analysis, this paper treats trade flows with respect to the rest-of-the-world (region r) as exogenous; as a result, there are

2

56 import demand equations (9 -9 -2*8). Estimation with both OLS and 2SL3 gives similar parameter estimates. This finding would arise if the demand-supply system were recursive because in this case FIML and OLS give the same parameter estimates. To explore this possibility, the analysis estimates an import supply equation and finds (Appendix B) that the data do not violate the conditions needed to support a recursive system in 45 out of 56 cases. For the remaining 11 cases, reliance on ordinary least squares could introduce a simultaneity bias. Whether this bias actually exists depends on whether the residuals are correlated with all predetermined variables or with a subset of them (see Johnston, 1984 p.202, problem 5-13). Finally, this appendix presents Brown-Durbin-Evans results to test for parameter constancy in the processes determining the exogenous variables. Based on the notion of super exogeneity (Engle, Hendry, and Richard, 1983), the results support taking income and prices as exogenous for estimation for the majority of the equations.

15a

UO} IMP ISS Uy OZ 02 P9I9412901 Sy JUS}DTZJ209 SUL ,

*$210dx@ p[10A pue si10dwy plI0AM 103

SPpPIPIFASeT[S awoouy usanjaq aouaiezyyp aya saayzB ArQue ys0mIYyF}2 ayYL 3

*s Kaqunod #013 4 A23UN0D jo s2210duy 1073 A2JD}ASeTe eYI Jo UOT IVTASP

aINtosqe vel}pam peyless aya jo aivnbs — oC Wade hoa ueoe —- my

ee Ok: - (1a) 180 Se pagndsod Ss} Que; IVA SIqT pue

Sy5 - Ay Se pauyjyap sy} Ny ‘AQpOyAsSeyTs amoouy [er9IeTFI[NW UL

alayn

SH,

*s00°0 “a1 a1 ayn ar Aq uaay? sq} punoq aaddn & v

v

ay? ‘s00°'0 i 213A Ty Aq uaay3 sy} punog 19M0] BYL ‘pueq aouapyyzuO0D. & .

86H FD ABZ spunog ay) are sazrQUa YIINOJ pue przyyd BNL “QOO” °° °° ‘1 =

( ‘Suyavap yof aya yom parezoosse Ardypop sepa un1r-Buoy ey sy mht e104n

“o9l9 o/t ("fay ue ppaw- $%y] ue;pam [UOT IeTASP SyNTosqe Ue;paw peyers

f

oud squasaidai Atjua puoosas ay Sly JO UOFINGJIIS|P ay? Jo ue;paN

v

mr sauasatdas Azqua asxpy aya ‘moTy apead Jerazey1q uaays e© 104 r

S320N

(190)

£000'°0

(Z'0) 41 °0-

(€'0) 92%

(z'0) .y an |

(2°90) est

(€°0) 1) Be 4

(7°0) 861

(Z'0) 981

(Z'0) 69't

(9z°0) (%€°0) l2-1- cel (2'0) (1'0) dol 6€ 0

so €'0- (2"0) en) Gn!) zl zo (Z°0) €s'0 oor 1 oe 4 6°0 to 10 (7°0) (2'0) 90'T 90 el el z'0 T0- (e°0) (20) 960 "S°0 ir 4 €°0 10 T'0- (90) (10) 2 Gi | 80°0 ae | zl 00 9°0 (€'0) (20) s9°0 09°0 Te 9°0 20 T'0 (s 0) (to) ssl oc 0 072 80 < 0- T'0- (» 9) (2°0) 26 0 seo 33d0 Etoria |

(€°0) Sz'0-

(z'0) 10°Z

et

et- (s'0) st-o-

a ~~

wvor

Noam ~

f

(s€'0) s<'0-

(€°0) Caz

°

Saal wns COAG Anon

nligad NOAM ~

9°”

(4°0) 49-7

€'9

zl (o'1) 9s'€

(279'0) 86°0-

(s°0) 60'€

(99°00) (4€°0)

"91 — 6Z°0- (Z°0) (€"0) "0 suz

eu

Lat-

(¢°0) z¥'0- = ,0°0 tz re

90 et (7°0) (7°0) zit 622

Lo 9'€

8'0- ol (€°0) (so) 70'0- ST

ut cz

t'0- ZU (€°0) (€°0) 62:0 S61

vz Us

2°0- le

(so) (2°0)

z8°0 199

“9

sil

wD

“> 6BTE zt

l0-

(0) 9S°0 a £0 us

10 $0 (10) (60) se'0 €L°2

oer 135

SaPIPOPAsSeTy aswoouy avpNdsas

Zz 21qeL

(€°0) 02'0-

(20) 6e'l

.~) AOON NOT ©

orowo i=}

an ~~

~

aa NOMA

——-_«

OM ma

NnoOawnN ~

vy ° °

wa an

Bt Nan =meoon

noon Moan ~ .

n

1 :purg 266

101133 «plas 393d0

n

1 [purg 066

10133 ‘pIs soql

.

T :Purd 966

101134 "pis “pur

a .

T [purg 066

10123 "pas sn

n

T ‘purg ws:

10339 "Pi5 an

n

T spueg 866

201234 ‘pis uvder

n T :purg 066 10129 pis huveia9

n 1 :puvg 066 101139 "pis

epeu ,

om ¢: h(t) ONT LYOAXA

16

imports are near unity. Third, the income elasticity for multilateral imports

ranges from 0.3 for Japan to 3.1 for the United Kingdom; multilateral exports have an income elasticity ranging between -0.1 for OPEC to 2.3 for LDCs. ** The negative income elasticity for multilateral exports of OPEC might be the result of a reduction in oil dependence by non-OPEC countries or a bias in favor of non-OPEC oil suppliers. *> Finally, based on the 99 percent confidence interval, the evidence suggests that 44 of the 56 elasticity

estimates are statistically significant.

One of the questions more commonly analyzed in the trade literature is whether imports and exports have different income elasticities. These elasticity differentials can be used to determine whether a ceteris paribus world-wide expansion deteriorates the US trade account. To that end, the

pape:: applies the one-tail test * * Hy: mh “Nh = O versus H,: m “Ny < QO.

Thus if My MS 1.65 [var (nm, +var(m, 14, then reject the null hypothesis. “°

24 The estimates of Warner and Kreinin (1983) range from 0.2 for the United Kingdom to 2.95 for the United States; those of Thursby and Thursby (1984) vary between 1.1 for the United Kingdom to 1.7 for the United States. The secular income elasticity for LDC multilateral exporits is smaller than the cyclical estimate of Cline (1984).

25 Note that some of the secular elasticities for imports from OPEC are constrained to zero. These constraints are needed in view of the collinearity between the real price of oil and potential GNP observed in the data. A change in the real price of oil affects imports from OPEC through two channels: the substitution effect and the effect on potential outpu:. Unfortunately, the estimation results do not permit identifying the separate influences of these two channels.

aA “x 26 This test assumes that both Ny and nN, converge in distribution to independent normal variables.

17

The last row of table 2 presents the elasticity differential between exports and imports for each country and for the world. Based on the results, the income elasticity for U.S. imports is significantly greater than the income elasticity for US exports--a result first noted by Houthakker and Magee (1969) .2” The results also indicate that, for the world as whole, the income elasticity of world imports is not significantly different from the elasticity for world exports, an equality not imposed in estimation.

The pattern of cyclical income elasticities (table 3) is very similar to that exhibited by secular elasticity estimates. There are, however, some important differences worth noting. First, based on the 99 percent confidence interval, half of the elasticity estimates are not Statistically significant, whereas 78 percent of the secular elasticity estimates are statistically significant. Second, the cyclical elasticity for multilateral exports of OPEC is 1.75, which is considerably larger than the corresponding secular elasticity estimate. Third, the cyclical elasticity estimates for imports from LDCs are smaller than their secular counterparts, a finding of interest for studying the outlook for LDC debtors. Finally, with the exception of Germany and OPEC, the elasticity differential between imports and exports is smaller than that exhibited by secular elasticities. Note also that world exports and world imports do not have statistically different cyclical

elasticities.

27 Although this result implies that the US trade account would

deteriorate as a result of a world expansion, the actual path for the trade account depends on the paths of both relative prices and relative incomes, not just the latter. Consequently, it does not seem valid to argue against the possibility of a differential in trade elasticities on the basis of assumed income paths without specifying the paths for prices. -

l7a

f

Table 3 Cyclical Income Elasticities

IND LDCs OPEC

4 J

4

EXPORTINi COUNTRY

Std. Error 99% Band: 1

Canada

u

Germany

On

nOanO NOON” ~~

Std. Error 99% Band: 1

u

1.1 -0.4 2.7

-0.19

(0.7) -2.0 1.6

Std. Error _. 99% Band: 1

Japan

Std. Error 99% Band: 1

UK

Std. Error 99% Band: 1

US

Std. Error 99% Band: 1

Ind.

Std. Error 99% Band: 1

LDCs

OPEC

1.75 (0.4)

0.44

(0.2) 1.0

-0.1

Std. Error 99% Band: 1

0.14 (0.02)

0.21 (0.1)

1.21 (0.1)

1.54 (0.3)

1.69 (0.3)

0.62 (0.4)

1.97 (0.3)

2.24 (0.4)

70.54 -0.16 0.63 1.61 0.01 (0.37) (0.20) (0.19) (0.45) (0.38)

-0.64 (0.33)

-0.15 (0.46)

-1.00 (0.32)

-0.02 (0.68)

f For notes see table 2.

18

The results for the own-price elasticity (table 4) indicate that the majority of these elasticities are negative (55 out of 56), but also that many of them are not statistically significant (27 out of 56) at the l percent significance level. The own-price elasticities for multilateral imports, be are twice as large as their standard errors and range from -0.5 for the United Kingdom to -1.1 for OPEC. The own-price elasticities for exports, ae are also highly significant and range from -0.4 for the United Kingdom to -1.0 for the United States.

The magnitudes of the multilateral own-price elasticities are of interest because they determine the stability of the trading system. Specifically, the response of trade to changes in prices is said to be stable if it satisfies the Marshall-Lerner condition: &. + é < -1. This condition is tested with a

one-tail test: * *

If é. + ex <-1-1.65[var(é,) + var(é*)]", then reject the null hypothesis. “® The last row of table 4 computes the value for this condition and, based on the evidence, it is not possible to reject the Marshall-Lerner condition for any of the countries considered here. Furthermore, the entry in the lowerright corner represents the difference in the price elasticities for world

imports and world exports, and as the evidence suggests, it is not possible to

28 This test assumes that the own-price elasticities converge in distribution to independent normal variables.

18a

“sqiodxa prion pus sjiodwy pl3I0M IOJ saz QzPO;ISeTe

ad}1d-umo uaanjaq a2uai13azz}p 242 Saazd A1QUa YsowIYFTI JUL 5

‘s A1qunos

wo1zy 4 Arqunos jo sjiodmy 103 AQypoyaseyps ayy Jo uozIVyZasp aanjosqe

ueypam papeos aya yo aienbs - “ve pW ay uvau - Sy a1aya

mai - Cp) a8, SB paqoniQsuod Sy 9ouUeTIEA SIz

pur Saar adel Sitcd - Ay Se pauyzap sy AazoyAsSeTa Voyid-umo [er9IeTJFI[NM sy] 5

-¢00°0 “agit a1 29ya "y hq ueayd sy} punog saddn

‘s00'0 “Pu asaya '3 fq uany3 sy} punog 1zaMoy ayy “pueq aouepyjuoz 3 .

Q66 942 JOJ Sspunog ayy are sazy1}_Ua YIAINOJ pus PAY? F4UL “O0N0Y

aya

ar a

s { ‘Suynertp yuvf aya yaya payeyoosse AQypoyaseya uni-Zuo], aya sy} Ay a13ya

.

‘one O/1 0849) ue tpaa- “As ]usypau [UOFIVPASP sanyosqe usypse payeos

. .

aya sjuasaidas A1Qua puosas ay] BAy jo uozInqy3zISTp ay? Jo uvypam

aya squasaidaz KArqua ysrty ayd ‘MOTZ SPlAD [OIFIST Iq uaata e@ 103 S

1S320N

Eee

(120) S2'l-

(1€°0) z0'0-

(Z'0) Ss‘0-

(t'0) €9°0-

(t'0) €8'0-

(9°90) 66 °0-

(2°90) 97°0-

(2°90) 76 '0-

(20) $9°0-

(z"0) €8'0-

(820) 69 T-

(Z'0) oT:

(so) “U1

v0 91- (0) zs 0-

70 €2- (9°0) 18 '0-

€°0- Ta- (€°0) 16 '0-

€'o- 6 °2- (so) 6t'T-

T0- 6'€- (s'0) co'l-

(zt°0)

9'T

(¢°0) 18°0-

z°0

s°0- (t'0) It'o-

oo

9°e- (€°0) £9°0-

z°st °9€ (aw S97 1-

z'0

9°0- (1'0) st'o-

v0

zt: (90) 81°0-

9°0- €°9- (s°0) ao l-

€°0o

€°9- (9°90) S6'0-

ber 9 |

(810) cet

(10) 67°0-

v0

9°0- (10) 61'0-

€'0- 9 1- (Z°0) 09'0-

0'0

6°2- (9°90) zL°0-

To

s'I- (€‘0) ss'0-

4

6°9- (8°0) 94'0-

v0 8°0- (2°00) 92°0al: 0°2- (90) 16°0-

pur

S

(€7°0) 16 1-

(z°0) z6'0-

el- 6°6- (80) 62 °1-

s°o- 8°9- (€°0) $7 0-

9°0le: (9'0) LUT-

(020 16'0

(10) 87 0-

9 T- S2- (Zo) "6°1-

€°0

8°0- (z'0) c1°0-

£0

L0- (z°0) €1'0-

6°0col (90) 88 °0-

(9°0) 62°0-

6°01

“il (€°0) 67°0-

7° 6t 0°S2 (cae ®) 29°:

9 aTqeL

)

SIFIPOPISLA azIAg

(€€°0) 98'1-

(Z'0) £6'0-

coat 8°9- (Z'0) 92 °0z°0- 8z- (so) wI-

(z°0) 9€°0-

wer

(to) 09°0-

€°0

8'0- (20) Ssz'0-

€'0

£'0- (z'0) 9t'0-

€°0- 9°T- (20) €2'0-

€ o0- 8 t- (€ 0) 68 '0-

9°0

9°0- (Z°0) It0-

97°0- €'¢€- (9°0) IS ‘1-

90 Z°2- (so) 49 °0-

135

(ceo rae |

(¢'0) $6 °0-

OSI

oot (8°0) zs'0

%°0- 6'7- (97°0) iut-

sl

9°S- (e710) €L't-

€°0- %2- (¢°0) 66'C-

00

9I- (20) 9”°0-

z°0l9 (9°0) 82 I-

€'0- , ae a (€°0) 78 °0-

) - "3 ety Fy

n 1 :pueg 86s 49119 pas o3d0

n I :purg to 10219 “pay

soq1

T :purg 066 10119 "pas , pur

1 :pueg tcf AOAIG “pas sn

T :pued 066 10119 "pas an

n T i puwg 066 10139 "pas

ueder

n T :purg 066 30119 "PIs Auewre

n

T - Puree toc 10119 “pes

- eprurg

' “INNOS Diet. YOUNG

19

reject the hypothesis that world exports and world imports have the same price elasticity.

Finally, table 5 presents the estimated cross-price elasticities. As the table reveals, the econometric estimation of these elasticities proved extremely difficult and some of them are constrained to zero. According to these results, US exports to Germany, to the United Kingdom, to other OECD countries, and to developing countries will increase if third-country prices increase. Similarly, US imports from both Japan and Germany will increase if third-country prices increase. Finally, an increase in third-country prices

lowers exports of LDCs to Germany and raises LDCs exports to Japan.

3.2 Test Results for Parameter Reliability

To evaluate the reliability of these estimates, tables 7-9 present test results for the assumptions of normality, serial independence, and homoskedasticity in the error term. The results from the Jarque-Bera test (table 7) indicate that the data do not violate the normality assumption for any of the 56 bilateral trade equations considered here. Based on the evidence of the F-tests for serial correlation (table 8), it is not possible to reject the assumption of serial independence for the residuals in 54 out of 56 cases. Finally, the results from the ARCH-tests (table 9) indicate that it is not possible to reject the assumption of homoskedastic errors for any of the 56 trade equations.

To examine further the reliability of the estimates, the paper tests for the choice of dynamic specification and for parameter constancy. On the basis of F-tests, it is not possible to reject the restrictions associated with the Almon distributed lag (table 10) for the equations in which they were used (30

in total). Furthermore, F-tests comparing the dynamic specification of (1)

19a

“% @7[Qge2 vas sajou 104 Sf

n 1 :pueg 866 4o0i1i1g pas 00°0 00°0 00'0 00°0 = 000 00°0 = 000 000 9340 sz 0'0 n 0°0 Ut: 1 :pueg 266 (s'0) = (0) 10119 “pis uo} eBp,ISa Uf 01aZ 07 pauyez3suOd 1a}0MNINd | 000 “77 00°0 00°0 00°0 180 19'0- 00°0 $301 a 6't 9°97 s'0 a) n 6'0- Ol” 2 0- t'€- 1 :purg 266 (ws) (eo (9 ao (Ss) (oD 00-0 no pote ooo (aio) re gio ror7d "Pas (60 0 (90 ,000 1190 090 20 © 2340 put (9t (eo (92 Gs) CD Ore) oe one a. ae T -pueg 46 90 0 --- 120 @0 SEO 950 to 20 som . : 866 (wd (6°90) (9°90) (to) 10113 pas (2 0) (9S) (yw) (ao) (vs) (90) (9) coro GE O60 “7 TOT 00'0 = 8s 0 000 sn 00 19 0 --- 620 950 £60 600 ¥%S°0 pul ot . amen (eu (4) (99) (06) (62%) (oD sn (0) U gpund tos "0 (a 0 zs 0 660 S2o Sto. €t0 000 00'0—s«i'0- «00 _. 00'0 000 00°0 (@ %t) (7° 2) (s ©) (U's) (e's) (©) . . . . Ley z1 61 n Le wn 60 "0 770 ©6880 790 «©,00'0 = $70 | Ite: et. . T :pueg v66 . (9't) = (s"0 (70) 10134 "P2s ; : ; (9's) (29) . : . . . . . : . a1 9°0 ve n (cet) (£6) (Za) 6 (et) TD (9'S) a. . “te . wpe wee : Auew129 6°0 €o- zl ; 1 :purg 066 nl 0 69 0 t€ 0 et 0 65 0 $9 0 "S°0 20 (2'0) (8°0) 30129 ‘pis (vv) (oe (9°S) (cs) (s'2) (07%) oo 6 HO SEO tk 80'O 00°0 “77 00°0 Auew399 : eee epeuey 0 170 (60 ,000 #90 0 870 zt UL n E35 Ct: 1 :pueg 066 x 357 set Lit ie) XSINTIOS zw (an 9) 10113 ‘pis STG $907 put sa ma J 795 ONLLWOaXa 000 «= 000s t- = i000~—S—«iK'T 00°0 = 000 --+ epeue eee TALNNOT ONT IHOART 35 OTT UT mi aT = 1H i) KBINTOS ONTLYOdX

(2fasyaPas-9) arqetieg yuapuadag padaeq] $9}339]18€19 @2}14-$5029 JOP OMT SZ waLolyyrog tS 6 21981 9 aTdPL

19b

“Crar(&p “g)d)Ad Juasaizdar atqea oyr uy soz3QUg WoOpertZ, Jo saar8ap yo szaqunu ayy sy %p [Apaayyoedsaa ‘sysayyodXy aayqeuze3ye

“yd pue Pp ayy aapun stenpysai pazenbs jo wns ayq aie lass pue G5 a1aya

"Cp yy a ~ Ces Uissy/osChass - MSs)) ed

VER paisay Sy YopYyR ‘Q = "% - = \y 9, S} UO}Ie] 32109

“[enpysal uoyssai3a1 aya sy wn a1eyr

.

[vbsas ou yo sysayqyodséy pynu ayz

Fastyle - nm

v ’

pe F 20g

Uo paseq S$} UOTZIVT9II09 [VJ198 103 1802 eYL

SESE eee ee eee ee

(O79) (299) (EE'9) (CZE7) (LEM) (C29'y) Con’) aa 26°0 610 020 92°0 €c'o 190 18'0 94340 (st'9) (27°9) (299) (99) (99) (z9"9) (299) 600 aa 690 Te 0 10 080 s8 0 760 sol (se'o) (@9'9) C299) (299) C199) (29'9) (299) orn O £10 ad 670 40'0 190 elo SL'0 “pur (se'o) (299) (29'9) (279) (ly) (299) (za y) SS 0 09'0 oc 0 tte 6”'0 €9°0 sto 0c 0 sn (SE'o) C29'o) (29%) (294) (t9'9) (@9"9) (29' 9) a6 0 990 0s 0 (sO a z2'0 €e'0 "SO yn (SE'O) (299) CBM) (C279) (Za) (29°94) (29'9) nO 00 0 940 96°0 £8°0 a ot '0 9S 0 ueder (GE°7) C299) (C299) C29) (C299) Ce) (29° 9) clo 69 0 7 at) Tani) 100 S6'0 toe 62°0 Auem2a5 CoCo) C29 9) (29a) (CUM) (299) Cte) (aa) o>” 0 0s 0 so 00 0 60°0 090 z1'0 a epeur) Fd0 S571 buy sn Wi wet 135 ue5 860 XWINNOS ONT LYOdXS

oe ~ RMINGGT ONTLAOIRT

(wopaaiz yo saaaZap) STenpysay aya uy uozrwpaz109 [eyIaS quvisal tof} sptaaay aouesryrudys 8 eT1qgeL

‘(ar > (2) %)3d quasaidal a1qe) ay Uy Sozaqug ‘“syenpysar

Paz VWIIS2 Vy Jo UOFINGTIISTP ayQ Jo JyUuswow yerjUaD YIf ayr sy fy a1ayan

18°0

6

92°

16

69"

97°

ol

‘0

0

0

A 7 . Cy, lw Cw - (2) * [pC 2"/ AM) (92/1) + & 9/7 lL-oer

Sv pazonaqsuod sy 379S43eIS VIeg-anbive auL

1e'0 690 92°0 O10 62°0 St'O0 22:0 2340 --- WO 80°00 48°00 690 800 400 5907 80°0 --- 090 “2L°O €'O0 Seo te'0 “pur 4s°0 = 9S'0 --- £44°0 %9°0 0S6°0 — SE€'0 sn 91:0 %5'0 29°0 --- @4°0 -1L'0) th 0 yn 0270 «6780 = 720s S00 “7-10 «O20 ueder 190 «6©€4'00«(8tT 0 COL 0890 toe te 0 Auvaiay Sto =6bt'O) SC 2@t 0) 980) 870990 te epeue) Lilia | puy ti} pT] avr 135 oes On NTS

2JISJIVIS wieg-enbler :sTenpysay eya uy Ary TemION Buyasay_ 103 seas] aourszzyusys <4 214981

19¢

area

ay)

("4 > (2p ‘Upyayaa quesaidex

Up saysuy suo yd113Sa1 uowTy Jo 1aqunu TEST ayy sq Up

pue “suoy191118a1 uowly aYy2 InoyIyR STenpysalr paienbs jo ams oy? 6] ‘lss

“SuOpIoTIisat uowty aya Buysn syenpysaa paienbs jo wns aya Ss} Suss alayn

“Cp "py a ~ (Copslassys"pscluss - Cassy) =P a

uo paseq S} $Uu0f}2419Sa1 UOMTY 103 3822 aUL

(0€'9) apo (o€'9) 98 OO (62'9) 70 (o£ '9) “e0 (oe '9) oF 0 (ot '9) 95 0

(ott) 100

(ce'e€) (et'ot) (eal'ot) (6z'9) (ee'z) (eee) 180 70 £90 £60 780 z0°0 9340 (se°9) (4E°S) (Ze 'e). oo 190 zs'0 "70 soe “oe SsoOgl (mt'9) (4E'S) (st‘%) a aso 510 a aa 950 “pur (Te'or) (“e'9) (1€'6) -- -oe 6” 0 94°0 ose 82 °0 sn ee ee -- eee eae eee > TA] (4€'°S) : - 990 oe a aa ueder (62°9) (4€°S) (4e'4) . (60 89°0 69 0 -oe Does Auewsay (6€°1) (Te'e) (Ze') (6e'D . t£°0 69 0 900 06°0 “oe epeuey pit 7) wn weet Ta5 ve> =” XWINNOS ONT LYOdXa

MMT ONT THOT a

(wopaaty yo sx.aidap) SUOPIIELISAY UoMTY aur ysag AO} SPoraayT aouvorzprudis

Ol FTaer

ay2 uy sazzque ey

Dpasepaysomoy jo syseyjodXy [Nu ey. r2epuy

“tCleyreap p/h = ya azeyn *(xa>14 43] ag moys atqea

°3S92-2 © YIM P|zse. sy YOTYA OC ~ ry *slenpysai

‘Tenpyse1 uozsse1Za1 aya

s} on ai1aya Tr3yk + % ~ 3n uo peseq sy Ary oJ asepayxsomoy 30z 3892 ayy

.

ze z.

0” 0

020

600

es 0

Te 0

090

1€ 0

69'0

760

€7'0

"80

080

6€°0

s8'0

1¢'O0 2€°O 06°0 9l'0 $90 99°0 93d0 68°0 8s'O €t'0 1y 0 76°0 65°0 sat woe £6'0 €z2'0 270 $9'0 180 “pur

770 a €2'0 "40 0S'0 c2°0 sn 990 SO aa 92°00 620 82 0 mn e810 8699°O 8 $6°0 “T° 64°0 SLO ueder 4770 8092'°0) =6€S°0 40°0 a 91°0 Auewiay ze'0 = 6T OT 0 z0'0 €°0 aan epeued purl a wi aeT 135 ae5 ~XHINAGS

ONT LYOdX4

2335843035 -HOUV rsyenpysay eya uz AIpPITIsepaysowoy Buyyasa] 10zJ STaaa] aouedtyyusys

6 F19FL

20

against an unrestricted dynamic specification (table 11) reveal that including additional lagged variables and eliminating the Almon restrict:ions does not provide a significant increase in the explanatory power of (1)--that is, the data do not reject the dynamic specification associated with (1) for any of the 56 cases considered here. Finally, the results associated with the Chowtest (table 12) indicate that it is not possible to reject the hypothesis of

: : 29 parameter constancy in 53 out of 56 trade equations.

29 The equations failing the Chow test are U.S. imports from toth other

industrial countries and from OPEC, as well as U.K imports from OPEC.

20a

Casda- "yet aygpaa quesaidar ofqe) oud Us S2;19Uq “s108001902 yO taqunu 2y2 Sf y pus ‘poziad ysvraI103 oy) UT SUOFIFAIaBGO Jo 1equnU ey2

6] ca ‘pojiad uoj ews asa 242 UP SUO}IFAIIEGO Jo 2equmu ey? 8] u ry ‘LT = 4 203

PSoH ISPITIOZ ay Sy "ny ty “LE = § 107 Lenpyse: perewy asa oyr sy fn o10yn

‘a pa-tor Aan : balan - 74

S} pasn 349972098-280) u Olez Sf 20533 YsUdIIDy parsadua ya ‘A2ueIsUu02 2a,0mVINd Jo sysayodhy [HU 1) tapug 706860 UIn01Y sawnton apes 38052103 02 pasn ueyd pus ¢g6t uy Buypue

PIeP HEA perempass isazy ase suopyaenbs [es ‘houv2sUu02 1e20m"IEd 203 3802 OL

a

(TE'o) (ZE°G) (9Z*8) C9z'B) Cez'O) (zE'e) Cece)

sso (60 96°0 610 80 760 0*'0 9340 (iz'@) (7¢°e) (ve'e) (ve'e) (ze'e) (zEe'e) (ae'e) zo ote 197°0 08 0 z” 0 60'0 oto S40 ®001 (ez'@) « (9E'8) (ze°e) (vc'e) (ze'e) (ze'e) § (tEe‘e) of 0 iz 0 ott 66°0 000 "C0 40:0 760 “pul (ez'e) (ected (eee) (cc's) (Cece) (ze'e) (ze's) (60 16 0 0.0 soe 60 0 sto 02'0 880 sa (éz'9) «(ve's)) Cee’e) § (ez‘s) (ee'e) (ez*s) (6Ee’e) 160 tho 6» 0 | B8£'0 toe €9'0 ceo 980 mn (ez (ee'a) (tee) (Te'e) (vee) (ze'8) « (4e'8) 18 0 910 270 seo co tte £70 "L°0 ueder (ez'9) (ee‘s) (zEe‘e) (ez's) (re'e) (eE‘s) (4€'8) "90 610 "90 06°0 000 2/0 ++ “70 Kuvaies CnzZ'e) (CCB) is B) (C20'@) Cre’@)) Cece) 6 (ze'e) : SC 0 sto 10 Si 0 950 §69'0 Sz0 coe epeury 37IO E sorii| puT yi] wi oer 15 os YSYNs ONILBOdXa HINGIS ONTLIOINE

(wopaaiy jo saaiBap) (18aL aoyn) £2ue ,8u09 Aad see108g Bupasay 107 STeasy] aouesyzyusys ZU 8ngel

“(Pas(ep Ubyayaa queseidas e7qea ey uy s0s22Ug ‘@opeel; jo seeiZap jo requmu ey? sy ay pue ‘sirejeavied [euc;IIppe jo 1eqanu eyd s} Py ‘syseyiodsy say 2euz02[e O42 JepuN Bl enpyse1 pelenbs jo wns eyd sy

y2oddy [ [Mu 8yQ JepUN s[enpysez peienbs yo ums ey? 875 CnNss e204An

*(p/"ussy/"pyclass - Oussy) = "g :2802-4 ue uo peseq 61 uoj2e2;;;20G"

ayueukp 103 3802 syL ‘pozyized evo peSBe, satqejien peuswiarepaid pus quepusdap ey2 yI0q ‘s108s01301 sv ‘Buypntouy Aq paroni3su0> 6} spsayiodky aapaeuzerye EYL “(T) UOTIeNbe hq pequeserda: sy sysoyodhy TyNU eYyL

-eoo ee

(se'o) (EEL) (STOUT) CST'6T) (9Z'6) (E'S) (seo)

oT (to 480) | 6f€O)6690 0060608 a340 (92°) Coc’e) (ZE‘OT) (EES) Cez'ET) (vE'9) (eE'e) 9°0 ““- 680 890 EZ OBO CBO OO 6901

(42'6) = (6€'O) (oc’ot) (ee'6) (SE's) (vE'9) (oe's)

69'0 ceo cee 92'0 sto e.'0 "9 0 "1°0 ‘pul (62'6)) (4e°o)) (LE) (9z°St) (ee's) (re'9) Cez'E1)

160 £90 €2'0 a 7 rt Se!) 120 680 sn (sz‘ou) (ze's) (Cee’y) (zE*e) ° (4e'9) (92's) Cee’)

sso 99'0 si'0 9t'0 toe 98°0 "tO oto mn (4z'6) (LE'o) (Lee) (vets) Ceo) (yt'9) (ee'9)

9S 0 0 zs'0 08 0 180 + z7'0 "1-0

(éz'6) (LE *e) (SES) Coz —at) Cee’o) Cee’ (ec*9)

920 zo 760 680 zt'o ss‘0 te co'0 Auewze9

C4Z°6) (LE'e) (LE'o) (9E%%) C9z' ELD COE’ (set's) d7o 8 680 680 80 FO OCS 0 oot

box B 0) Tl par on wm avr ¥35 ow) =” XWINWOS ° ON1 L800

FS TR —S

(wopeexy jo see1Z0p) UosIwr;zpIeds a;meukg Bupasay 103 stone] e5ues;3;uB1S UU erges

21

4. Model Validation

To study the dynamic response of trade to changes in prices and income, this section assembles the estimated equations into a world trade model in

which each country’s trade account, denominated in current dollars, is explained as the difference between multilateral exports and multilateral imports. Country k’s multilateral imports are determined by the sum of 8 bilateral imports, 7 of which are explained with 7 import demand equations. The eighth bilateral trade flow is with respect to Centrally Planned Economies and is taken as exogenous.

In principle, multilateral exports of country k equal the sum of bilateral imports-of other countries from country k. In practice, however, differences in reporting practices across countries, shipment delays, and CIF/FOB differentials introduce a discrepancy between the value of exports of country k to country s and the value of imports of country s from country k. This discrepancy is the more serious the higher the frequency of observation. To take into account these measurement problems, the model includes estimated bridge equations that link X, to Moy!

(5) Ink, , - Bors + Biy,1nM., + error tern,

where M.,, is determined endogenously as in (1). In the absence, of systematic

measurement errors, one would expect that Bo = .0 and Ay - 1,39

30 The model, which is shown in Appendix C, includes only twenty equations such as (5). Of these, Bo = 0 in seven equations and Ay - lin

eleven equations.

22

Given equations (1) and (5), the trade account of country k is

Xest = exp (Bor. + Brg MM oe + error term),

(6) Mest Mase Peer Ssh)» Use) NX

A

xt 7 LeXkse Leese 7 Yer Peer Yes M(L)), vswk,

where Ye - (ie );

Yat Pre ™ (Pri oe Pest i Pent)? a, (L) - vec(a,, (L) wae a fh) a) (L) wae a 6b)?’ ,

where L is the lag operator and a (L) is defined by equation (4). To ensure equality between world exports and world imports, the trade account of the

bloc of other OECD countries (k=i) is determined as

(7) Ree Mie t DiMe” kee + Vewt

A

it 7 LMist ™ Ys Mise Vie: Pit? a, ,(L), Yast)? Vsmi.

where M

Although (7) guarantees that a reduction in the US trade deficit is absorbed by the rest of the world, it does not determine the extent to which a given

coun=ry absorbs this deficit reduction.

4.1 Historical Tracking

Traditionally, econometric models are evaluated by deterministic methods--that is, model predictions are compared against actual values under the assumption that the error terms take their means--namely, zero. However, when applied to

nonlinear models, deterministic methods lead to biased model predictions

23

because the expectation of a function of a random variable is not, in general, equal to the function of the expectation of that variable. In other words,

. <a . : tha because Mest is a nonlinear function of Use it follows t

A

ETM se Yue Peer My (L), Mest) * Moe lYpe Peer My (L), E(u ts:

the latter being the deterministic prediction. Biased model predictions produce, in turn, biased measures of model performance.

To avoid these biases, the paper relies om residual-based stochastic simulations as developed by Brown and Mariano (1984). Implementing their approach requires simulating the model over the historical period for alternative values of the residuals. These alternative values are generated

following McCarthy's method (McCarthy 1972)--that is,

A

Unet = T 4 Yost ME ? -N(0,1) Vk,s,t. Under the assumption that both ECP. Ugg e-5)70 Vj and E(t Un ee) = E(u) oe) E(u, ,)=0 Vp, q, t, it is possible to show (McCarthy, 1972) that u- NCO, Q,.), which is the distribution of the original disturbances (see equation (1)). One advantage of residual-based stochastic simulations is that the number of drawings is equal to the sample size used in parameter estimation,

which in this case equals 50,32

For the purpose of generating random numbers, the seed for the normal distribution of » is indexed by the replication number--that is there are 50 seeds, one for each replication. Fair (1986) reviews alternative procedures for evaluating model performance with stochastic simulations.

24

Based on these 50 stochastic simulations, the paper computes the (in-

sample) Mean Absolute Percentage Error (MAPE) as MAPE = (Del Myee - Myst /50) 1/1/36, t=1976Q1-1984Q4, r=1,...,50,

' where Micees is the simulated value of Mest for the rth drawing. According to table 13, the model exhibits relatively small errors in dynamic simulations with a MAPE ranging from 0.8 percent for total imports of other OECD countries to 3.5 percent for total imports of OPEC. For the United States, total exports have a MAPE of 2.8 percent and total imports have a MAPE of 2.5 percen:. With respect to the trade account, the model exhibits a Mean Absolute Error (MAE) ranging from $7.9 billion for the United States to $2.1 billion for Canada. Note that multilateral trade flows exhibit lower MAPE’s than the associated bilateral trade flows, a result that suggests that overpredictions in some trade flows are offset by underpredictions in other flows.

As an additional criterion for model evaluation, the analysis examines a regression of actual on mean predicted values: /50) +¥v

“ 2 Mest = + oi (2M ~N(O, o.,)-

kstr kst ’ Vest

If model predictions do not deviate systematically from actual values, then it

must be true that both yp )=0 and y¥,=1, a joint hypothesis tested with an F-

statistic distributed as F(2,34). Based on the significance levels associated with Hy: WoO and ~,=1 (table 13), there are 13 (out of 56) bilateral trade equations for which it is not possible to accept the above null hypothesis.

Equations failing this test include multilateral imports of Canada, LDCs and

24a

Table 13 Model Validation: Measures of Performance in . Stochastic Simulations

1976-1984 ——___________IMPORTING counTRy. EXPORTING Total COUNTRY Can Ger Jap UK wus Ind LDCs PEC Export Canada MAPE*..- 9.5 5.9 8.5 4.4 6.2 6.8 11.9 3.2 F-Test 0.63 0.99 0.41. 0.44 0.10 0.02 0.01 0.02 Germany MAPE 8.7 --- 4.9 4.6 4.7 1.9 3.1 9.5 1.6 F-Test 0.01 0.34 0.62 0.56 0.15 0.21 0.02 0.09 Japan ' MAPE 9.2 4.7 --- 6.2 5.6 9.5 4.2 6.2 3.1 F-Test 0.01 0.85 0.22 0.43 0.06 0.92 0.38 0.98 UK MAPE 9.7 3.9 14.6. --- 6.3 4.3 3.6 4.3 2.9 F-Test 0.82 0.61 0.06 0.01 0.62 0.06 0.31 0.94 US MAPE 3.0 4.6 5 6.3. --- 4.4 8.4 2.8 F-Test 0.12 0.87 0.22 0.95 0.09 0.00 0.43 0.00 Ind. MAPE 5.4 1.7 4.7 3.8 41. «-- 3.1 5.1 1.3 F-Test 0.01 0.23 0.95 0.11 0.19 0.03 0.06 0.71 LDCs MAPE 7.2 2.6 3.6 6.4 3.5 3. «ee 7.0 1.0 F-Test 0.00 0.53 0.07 0.20 0.35 0.85 0.61 0.53 OPEC MAPE 18.7 6.6 43 17.5 9.6 4.8 5.9 -- 2.6 F-Test 0.00 0.16 0.10 0.00 0.16 0.13 0.02 0.06 Total MAPE 3.0 1.5 2.8 3.5 2.5 0.8 2.1 3.5 Imports F-Test 0.04 0.52 0.09 0.18 0.82 0.29 0.02 0.04 Trade MAE 2.1 3.0 3.1 3.7 7.9 5.7 6.7 4.7 Balance F-Test 0.03 0.07 0.15 0.19 0.54 0.00 0.14 0.66

Mean Absolute Percentage Error for 1976Q1-1984Q4 using the mean of simulated values from 50 stochastic simulations.

F-Test for Hy :an0 and bel in the regression Yj 7arbYs where Yee is the actual

value of the ith variable and Yie is the mean associated with 50 stochastic

simulations. The entry represents Prob(F(2,34)>F-value).

Mean absolute error, expressed in US S$ billion.

25

OPEC; multilateral exports of Canada and the United States; amd the trade

balar.ces of Canada and the bloc of other OECD countries.

4.2 Multiplier Analysis To evaluate the properties of the model as a whole, the analysis examines the respense of trade accounts to changes in exogenous variables: Case 1: A l percent decline in the annual growth rate for the United States.

Case 2: A l percent increase in the annual growth rate of foreign countries.

Case 3: A 50 percent depreciation of the US dollar.

Case 4: Cases 1-3 combined.

These simulations use historical data over 1980Q1-1984Q4 as the baseline and the results are shown in table 14. Note that, by construction, they guarantee that the sum of trade account responses across countries adds up to zero.

A reduction in the U.S. growth rate of 1 percent produces, after five years, an improvement in the the US trade account of $31 billion. 22 This improvement stems from a reduction in US imports that translates into a deterioration of the trade accounts for the remaining countries. The countries absorbing the bulk of the improvement in the US trade deficit are

Canada (28 percent), Japan (25 percent), and the bloc of other industrial

countries (15 percent). The results also indicate that LDCs experience a

32 This shock is applied to the growth rate of actual US GNP; US

potential output remains on its baseline path.

Case 1: Slower U.S. Growth? Canada ..................00. United Kingdom............. Germany. ...............000. Japan... .... ee ee eee eee

Case 2:-Faster Foreign Growth? Canada... ........ eee United Kingdom............. Germany............. 002 eee Japan... cee eee

Canada............ 0.002008. United Kingdom........ eee Germany................ hee Japan... . eee ee ee eee

United Kingdom............. Germany................00., Japan... . cee eee eee

80Q4

Table 14 Response of Trade Accounts to Exogenous Changes in Income and Prices (Deviations from Baseline, billions of USS)

.25 .35

. 68 .89 .13 .54 .28

. 68 19 32 .32 . 80 06

.81

.67 .97 .43

.12 .35 .72 .63

.08 .85

74 . 60 .90 .68 .70

8194

-2. -0. -1. -1. -2. -0.

11.

-3. -2 -3. -3. -1.

99

.64

53 24 23

82 .00 .81

12 .10 .38 94 .10 .40 .51 .53

Year

82Q4

-3. “1. -1. -2. -2. -0. -1. 14.

-5. -3.

<4, -1.

-13. -1.

-8. -21. -67.

67.

-25. -8. -3,

-17.

-27.

-64, 50. 96.

s

76 64 58 58 76 49 91 71

83Q4

-42. -13. -10. -22., -17. -58.

42 122.

25a

Notes: U.S. growth is reduced by 1 percent per year; foreign growth rate increases by 1 percent per year; dollar depreciates 50 percent against all currencies.

84Q4

2 -8.88 8; 2.65 56 -3.66 96. -7.7 16 -4.58 79 -1.00 04. -2.85 .80 31.35 1& = -12. 86 7% -6.70 97 -5.66 00 -9.04 1é -0.39 03 5.61 76 8.47 87 20.57 97 -23.56 27 -6.22 31 6.18 _5€ -5.97 7120.14 .02 -68.51 22 29.36 .0C = 88.86 65 -52.40 64 -19.51 12 -5.83 29 = -24.92 62 -25.35 20 = -61.78 134 =: 39.78 24 150.02

26

small reduction in their trade account given the small cyclical income elasticity for US imports from LDCs (see table 3).

An expansion in foreign economic activity (Case 2) raises US exports and reduces) the US trade deficit by $21 billion after five years.>> The effect of this foreign expansion on foreign trade accounts depends on the magnitude of the income elasticity for exports relative to the magnitude of the income elasticity for imports. For OPEC and LDCs, the cyclical income elasticity for their exports is greater than the income elasticity for their imports, and therefore, their trade accounts improve in response to an expansion in non-US countries. For the remaining countries, the income elasticity for imports is greater: than the income elasticity for exports, and therefore, their trade accounts deteriorate.

A 50 percent sustained real depreciation of the dollar against all currencies is effective in improving the US trade account .?* There is a short-lived J-curve, and the U.S. trade account improves by $89 billion after five years. The results also indicate that Canada, the bloc of other OECD countries, and LDCs show deteriorations in their trade accounts, with LDCs experiencing a deterioration of $69 billion in their trade account. There are several reasons for this result. First, the dollar depreciation induces a

revaluation of the dollar value of LDCs’ baseline trade deficit. Second, the

33 The income shock is applied to actual foreign GNP, leaving potential foreigr. GNP on its baseline path; US real income also remains on its baselire path.

34 The analysis assumes that the "pass-through" coefficient is equal to one. Under the assumption that the prices of LDCs, OPEC, and other industrial countries remain constant in their currencies, P(fx), a dollar depreciation raises their prices denominated in dollars, P($), because P($)=P(fx) e($/fx), where e($/fx) is the bilateral dollar exchange rate, which by assumption is increasing in value.

27

depreciation of the US dollar relative to LDCs’ currency increases US real net exports to developing countries, which reinforces the deterioration induced by the revaluation effects. Finally, the dollar depreciation has a depressing effect on third markets for LDC exports--that is, the export gains arising from complementarity are more than offset by the export losses arising from the substitution effects.

To the extent that the events discussed in cases 1-3 can take place at the same time, it is of interest to determine their effects when operating simultaneously. The results reveal that after five years, the US trade account improves by $150 billion. Because international imbalances add up to zero in this model, the improvement in the US trade account implies a deterioration in foreign trade accounts and the countries experiencing the

largest deteriorations are Canada ($52 billion) and the LDCs ($62 billion) .>°

33 LDC exports have a cross-price elasticity of -0.41 with respect to

Germany but 0.81 with respect to Japan (see table 5).

36 If the model were linear, then the trade account responses from case 4 would exactly equal the sum of the trade account responses for cases 1-3. For the United States, the sum of these responses is $141 billion whereas the response in case 4 is $150 billion. The comparisons for other countries reveal small, but not negligible, differences between the sum of the trade account responses and the response of case 4. Based on these results, it is possible to conclude that the model is nonlinear.

28

5. Model Applications 5.1 The Absorption of the U.S. Trade Deficit To study how foreign trade accounts respond to the elimination of the US trade deficit, the paper endogenizes the level of real income in the United States and exogenizes the US trade deficit, which is then set to zero both at once and permanently. >’ The resulting pattern of foreign balances depends on both the associated change in US real income and the assumed level of foreign income. For example, if the United States adjusts the deficit on its own, then foreign economies experience a deterioration of their external accounts because of the contraction in US demand for their products. By expanding their economic activity, foreign economies also experience a deterioration of their trade accounts, but in the process they both reduce the degree of adjustment that is needed by the United States and avoid the contraction in their own income that would arise if the United States were the only country adjusting. To examine this sensitivity, the analysis considers two possibilities: no foreign-income response (Case 1) and a foreign-income response (Case 2). These simulations are performed over 1980Q1-1984Q4 and, by construction, they guarantee that the improvement in the US trade account is equal to the historical trade deficit.

The results for Case 1 (table 15) exhibit several features of interest.

First, reliance on US income alone to eliminate the US trade deficit requires

37 Although the assumption of an immediate return to external balance is extreme, it is adopted here because it avoids the ambiguities associated with selecting any other target path for the trade account. The framework could also be used to determine the change in the exchange rate required to eliminate the trade deficit, but it is not presented here because of space considerations.

Lowering the US Trade Deficit to Zero

Table 15 Trade Account Effects of

(Deviations from Baseline, billions of US$)

Case 1; Forei ncome Unchange

Trade Account Responses

Canada....................4. United Kingdom............. Germany................000. Japan........ ee eee Industrial................. LDCS... .... ee eee eee OPEC... . eee eee eee United States.............. US real GNP?(% deviation)..... Case 2: e

Trade Account Responses Canada.................0000, United Kingdom............. Germany............ 0.0.00 ee Japan................00008.

80Q4

72. -2. -4,

-0 -8

26.

“4, -3. “4.

-5

-5. 0.

-3

26.

-2.

.76 27 17 20 . 96 .82 .08 26

73 28 26 15 28 33 .90 26

42

8194

~10. “5. -7. -10. -8. QO. -0. 42.

-5.

.64 71 75 .15 .22 045 .87 .78

96 83 60 15 24 61 62 78

37

Years

82Q4

See NS

-11. -5. -5. -9, -9. -1. “5. 47,

-13

-7. -9, -1l. -9, .O1 .27 44

05

24 01 46 78

35

44

-48

94 07 71 51

.76

83Q4

-24., -6. 243 -19. -12. -3. -10. 87.

-9

+29, -10.5 -13.5 -23, -10.1. 1. -0.6 87.1.

“13.5

78 98

97 4G 47 ()7 14

eee

Note: * Entries for US real GNP represent percent deviation from baseline.

28a

84Q4

- 26

-12

-15.

-32. “14, -15. -34, -13.

108.

-11.

.18 -9, .07 -30. -16.

~4, 52 108.

82 43

11

29

a 15 percent fall in US real GNP after five years. Even this sizeable reduction in income is unrealistic because it assumes that prices, exchange rates, and real income in other countries remain unchanged in the face of contraction in US real income--that is, these are partial equilibrium results. To the extent that income in the United States does not adjust to such a level, the model predicts that the process of eliminating the US trade deficit will require a combination of further dollar depreciation, further expansion abroad, higher foreign prices, and lower US prices. °° Second, the model indicates that the countries absorbing most of this reduction in the US trade def'icit are Japan (29 percent), Canada (24 percent), the bloc of other OECD countries (15 percent), and Germany (11 percent).

The case of a foreign- income response (Case 2) is modeled as an increase of 1 percent per year in the growth rate of all non-US countries while US real income adjusts in order to support a zero trade deficit. Based on the simulation results, this foreign expansion raises the level of US real income consistent with balanced trade by 4.5 percent after five years. Intuitively, a foreign expansion raises US exports and improves the trade account. To avoid that surplus, US imports have to increase which implies an increase in US real income. Note that, because income elasticities differ across countries, the trade account effects of faster foreign growth are not the same for all non-US countries: the trade account improves for some countries but deteriorates for others. On balance, however, the non-U.S. balances must

weaken. Finally, the foreign expansion alters the degree to which different

8 Tais paper does not address the question of what combination of dollar depreciation and growth adjustment would be most effective or desirable for reducing the U.S. trade deficit.

30

countries would absorb the reduction in the US trade deficit. Specifically, the deterioration in the combined trade accounts of Japan, Germany, Canada, and other OECD countries would account for 89 percent of the improvement in the US trade account, compared with 79 percent in the absence of a foreign

. 39 expansion.

5.2 Quantifying the Uncertainty in the J-Curve. To quantify the uncertainty surrounding the response of net exports to changes

in exchange rates, the analysis defines the J-curve as

A A aA

= Tye (GCL), Vee 4), where Vie - Dg (exp(u,..) - exp(u..)), A is the shock to exchange rates, and the dynamic nature of the J-curve is embedded in the lag distribution of the Parameter estimates, a, (L). Inspection of (8) reveals that the uncertainty associated with the J-curve stems from parameter uncertainty, a, (L), and model uncertainty, Vig Uncertainty in the response of net exports to changes in exchange rates

translates into uncertainty about the length of time required to ‘testore

external balance. Specifically, let T, be the number of periods that it takes

A technical note describing the analytical solution associated with these simulations is available on request. Also, it is possible to examine the effects of a change in the distribution of world income with the level of world income fixed. These simulation results are also available on request.

31

for the J-curve to cross the "zero" line--that is, Jy r = 0. For a fixed Ty

* - rf 2 ry ° 2 ° Ty as assumed in the literature, the notion of deficit persistence is

formally expressed as J,. < J, _* = 0 for some t > T’. Unfortunately, only kt k,TL k ’

. * Ak . . . Ak A A the estimate of Ty, ; Ty. , is known and is given by (8) as Tat CA,@(L) v0)

Therefore, the adjustment period is a random variable and a comparison between

A

Ty, and t is not informative for determining whether the deficit is persistent

A

* ‘ sas * “40 because t 2 Ty. depending on the realizations of a. and Vee: Frem @ policy standpoint, however, the question is not whether qT, is

literally fixed, but how sensitive it is to alternative realizations of elasticity estimates and residuals. If the crossing date were to exhibit a large range of variation in response to these alternative realizations, then this evidence would indicate that the data do not permit identifying the adjustment date with any precision and, therefore, that it is not possible to esiablish whether the Us trade deficit is persistent.

One difficulty in addressing this issue is that iy is a nonlinear furction, which complicates the derivation of analytical expressions for its moments. To bypass this difficulty, the paper performs Monte Carlo simulations that permit estimating the relative importance of parameter uncertainty and model uncertainty in accounting for the overall uncertainty of Jy. Specifically, to evaluate the uncertainty associated with the residuals,

this paper applies a 50 percent exchange rate shock (A=1.5) while

simultaneously drawing random values for the residuals from their joint

A A

Ak ‘ * . 40 Note that E(T,. ) = TD, E(a@, (L)), E(v,.)] because Ty is a nonlinear

function of its arguments.

32

density function. ‘2 - The values of Use are drai:. collowing the discussion in section 4; there are 50 drawings for each residual in the model and for each period from 1980Q1 to 1984Q4. The value of Jee associated with the rth

drawing of the residuals is

A

(9) Seer - JQ, a, (L), Yetr? ™ Jace A)» and the associated mean and 95 percent confidence interval, for each period,

are

A

*e Je QQ) = [EJ A) }/50, (10)

(Te) - 2/var(J.. I, Tee) + 2Wvar(J,. Od),

respectively. Figure 2 plots the time series for both the mean and the 95 percent confidence interval for the J-curve of the United States. Fer comparison purposes, the figure also shows the deterministic J-curve--that is, 32 = 3,0, a,,(L), 0).

The evidence reveals several features of interest. First, the initial deterioration of nominal net exports ranges from $30 billion to $90 billion,

at annual rates. After one year, nominal net exports improve and the range of

uncertainty declines considerably, but it does not vanish and generally

41 : te ta . For this purpose, it is important to note that the estimator of the

variance of the residuals is statistically independent from the estimator of a; this independence justifies shocking both the residuals and the parameter estimates.

32a

Set2IuaIND 8 YSuUIGOR ZOG Sapwloasoep sBYOP ‘SN

ya6l £861 2861 1861 086! Oct 06 09 os a) + of 09 06 Oct (2PB1})P}SZ-9AINI-— _31}SBYI0}S G@AINI-( D1¥SBYDIO}S OAINI-( 31ySIuIWIE}EQq Osi Wag }UNOIIY apeil asipueydaW “SN |BUIWON eu} Jo asuodsay S$19Aa!1179 ariopw affupuaya in Aivuipiiaaun ' a Ae | ~7rTwwu vey dy z~ rm.nrevw~w~il Z eanbta

33

involves several billion dollars. Second, despite this uncertainty in the initial response, model uncertainty does not induce a large dispersion of crossing dates, which range between four and five quarters. In other words, 1 is robust to alternative realizations of model residuals. Finally, there are important differences between the deterministic J-curve and the mean of the stochastic J-curve, a difference due to the presence of nonlinearities.

To quantify the uncertainty associated with a, the paper applies an exchange rate shock while drawing alternative values for the parameter estimates from their distribution. Specifically, let Sere be the value of Jee associated with the eth drawing of a, a, where a, (L) - (Lay (L), with W=-0.1, 0.2, -0.1, and -0.2. All of these elasticity shocks are within one standard deviation of the point estimates of these elasticities ‘see table 4). Figure 3 shows the effects of these parameter changes for the J-curve of the United States. *¢ Based on the evidence, a 20 percent reduction in the own price elasticity (an increase in the associated absolute value) could increase the effect of a 50 percent depreciation by as much as $60 billion, at an annual rate. A 20 percent increase in the own-price elasticities, on the other hand, dampens the improvement of the depreciation by approximately $60 billion, at an annual rate. This Symmetry of the J-curve to parameter changes is also present for a 10 percent change in parameters.

Associated with these parameter changes there is a dispersion of crossing

A

* dates, Tus: To estimate this dispersion, figure 4 presents the J-curves

42 The baseline for these comparisons is the J-curve for a 50 percent

depreciation without parameter changes--that is, figure 3 shows Jee OW) - Jee GQ, W=0). This method allows for uncertainty about both

initial and terminal conditions.

33a

ool

08

09

Ov

Oe

+O!

Oe

Ov

og

oe

oot

vesél £e6l c@él 1861 Ov6él

Ayianseyea acid seyBry Z01

AyionsBya O211d sOMO} ZO}

oom - -—o eT oO -

AyiatysBje e21d JemMo| X02

uonlag | aBuey? sayaweseg OU YIM YIOYS ayeYy aHueYIXA XOG wos} SuO!zelAag Sd}k}S PaylUN saAIND-f ayy UO AjpUIeyPIaIUA, JAYaWRIRY JO 399354 £ aanbtg

34

associated with each of these parameter changes. The evidence reveals two features. First, the adjustment period of net exports is subject to a wide range of variation with crossing dates varying from 1 to 12 quarters depending on the parameter shock. ¢? Second, the response of the crossing date to the change in the own-price elasticity is not symmetrical. Relative to the case of no parameter change, a 20 percent reduction in the (actual) price elasticity shortens the adjustment period by 2 quarters whereas a 20 percent increase lengthens it by 8 quarters. To the extent that an appreciation of the dollar displaces domestic production, there is a reduction in the ownprice elasticity for imports and exports which, based on these results, lengthens the adjustment period associated with a dollar depreciation.

As it stands, the evidence suggests that the crossing date is more sensitive to parameter uncertainty than to model uncertainty and that the assumption of a fixed crossing date, as required by the literature on deficit

persistence, is not supported by the data.

43 This dispersion of crossing dates assumes that prices and incomes are

exogenous, Allowance for endogenous responses of income and prices to the dollar depreciation is likely to exacerbate the associated uncertainty. Note that the probability that all of the US own-price elasticities change simultaneously by a given magnitude might be small. There is ongoing research to relax this limitation via stochastic simulations.

34a

vOvegel jiyun PSeUTB}UIBW BIB PUB LO0G6I u! uiveyq seouBYyd 49}0Wuu0

veél £e6l c@él 1861 Og6l

Ocl

08

Ov

+O!

Ov

-_—- = _— -_—

0g

Oel ee aBueyd sajawesed ynoyyim oauno-p -.—.—_. — 091 A¥9SBja add-uMo sayBIy xoz = ———_— AyI91}Se18 adId-umo sayBiy xo, = -—-—-~—- - A¥IDIYSBl9 BD1d-uMO JAMO] yO? = .—-—-—-——_ AyIDNVSBla |adId-UMO 49M0/ XOL = ————_—

002 uonlas

4BII0Q 84} JO UO!}BIDaIdG9ag xQG SOPEPS POPUL :A}IDIVSBLA 3I!dd-UMO Ul SaBueYD pueR BAIND-F ByL Pp eanbty

35

6. Conclusions

This paper builds, estimates, and simulates a world trade model to study how the pattern of external imbalances responds to changes in income and prices. The econometric estimation of the associated income and price elasticities reveals a substantial dispersion of bilateral elasticities, confirms the findings of Houthakker and Magee (1969), and cannot reject the Marshall-Lerner condition.

Based on simulations of this model, the analysis finds that reliance on foreign expansion to eliminate a trade deficit of the size recorded by the United States in the mid 1980s requires extremely implausible increases in foreign growth rates. Similarly, eliminating this deficit through US growth alone would entail substantial losses in income. Finally, and in line with earlier studies, the paper finds that changes in relative prices have played the central role in the development of external imbalances during the 1980s. As a result, it seems that a depreciation of the dollar must play a major role in the process of eliminating US external imbalances.

The paper also quantifies the uncertainty associated with the response of US net exports to exchange rate changes. The findings indicate that the initial response of US net exports is subject to a wide margin of variation ranging between $30 and $90 billion, at annual rates. In addition, the paper finds that small changes in own-price elasticities induce a large range of variation in the adjustment period of US net exports to a given exchange rate shock. Uncertainty about the adjustment period translates, in turn, into

uncertainty about the persistence of the US trade deficit.

36

Appendix A

Tests of Functional Form

Equation (1) is formulated in terms of the logarithms of the variables of interest. This formulation is very convenient because it provides a quick inte:pretation of the parameter estimates and because of its relatively good fit of the data. .

To determine whether the data support the use of a double- logarithmic formulation, the paper performs Box-Cox tests in which the variables of (1) are

re-defined as

Mee 7 (MS - DAs,

> d

Pes 7 (Peg DA, x

% "OR? DA,

where \ is the Box-Cox parameter. If \ is zero, then (1) should be expressed in terms of the logarithms of the variables; if \ = 1, then (1) should be expressed as linear function of the levels of the variables involved.

To test for ’A=0, the paper substitutes the above definitions into (1) (without logs) and performs a series of log-likelihood ratio tests where the log-

likelihood function is constructed as (Khan and Ross, 1977) AQ L(A) = -(T/2)1no(A) + (A-1)Y, logm, 5.) tel,..., T. The log-likelihood ratio test is constructed as § = Lax (>) - L(A=0) where Lax

is the maximum value of likelihood function over the possible values of \ and

2 6 ~(1/2)x with one degree of freedom.

37

The results (Table Al) are based on a search over values of’ \ ranging from minus one to plus one with a step size of 0.1; out of 56 cases, only 13 specifications reject the hypothesis that A=0. Despite this strong support for the logarithmic formulation, some of these tests are not particv.larly powerful. For example, table Al presents the value of \ that maximizes the log-likelihood function. The results reveal very flat likelihood functions, as evidenced by the large values for the maximizing \ that cannot reject the hypothesis of \4*0. Of the 43 cases that support the logarithmic formulation, 17 have seemingly large values for lambda, which is indicative of low statistical power to discriminate

between alternative functional forms.

38

Table Al Box-Cox Tests for Functional From: : Log-Likelihood Ratio

EXPCRTING C COUNTRY Can Ger Jap UK US Ind LDCs OPEC Canada -- 0.38 1.28 1.43 4.60 1.05 0.73 3.82 maxr -- 0.45 1.05 0.85 -0.75 0.95 0.45 0.75 Germany 1.55 -- 0.00 0.00 0.04 0.87 0.57 0.05 maxr 0.75 -- 0.00 0.00 -0.15 -0.45 0.75 -0.15 Japan 0.00 0.43 -- 0.05 0.17 4.05 3.75 0.23 maxr 0.00 0.15 -- 0.15 -0.15 0.95 -0.55 0.15 UK 0.29 3.22 0.00 °° 0.97 0.17 0.00 0.39 maxr 0.45 -0.45 0.00 - 0.45 -0.25 0.05 -0.55 us 0.52 2.70 0.84 0.50 -- 0.00 2.42 0.54 maxr 0.05 -0.95 0.95 -0.35 -- 0.00 -0.95 -0.45 Ind 0.00 0.20 2.48 0.03 2.52 -- 0.00 0.64 maexr 0.00 0.35 1.05 0.15 -0.45 -- 0.00 -0.35 LDCs: 0.00 0.00 2.22 0.26 0.30 0.21 °- 0.08 maixr 0.00 0.00 0.75 -0.75 -0.15 -0.15 -- 0.15 OPEG 21.34 0.76 2.34 2.37. 1.33 0.21 1.55 -méxr 0.75 0.25 1.05 0.15 0.45 -0.15 -0.95 --

For a given bilateral trade flow, the top entry represents the loglikelihood ratio test and the second entry represents the value of \ that maximizes the likelihood function. If the likelihood ratio test is greater than 1.92, then reject hypothesis that \ = 0.

39

Appendix B

Simultaneity Biases and Single Equation Estimation

B.1 Recursiveness in Trade

To explore the reason for the similarity between the 2SLS and the OLS estimates, the paper examines whether the demand-supply system is recursive because in this case FIML and OLS give the same parameter estimates. To that eni, the analysis

estimates a simple import supply equation as

* * * (BL) In Plo. = Bot Byln(P, Hey Py) e+ Baln(PS / Piyde.y + Bs M,

ke * “est °

The first relative price captures supply shocks whereas the second relative price measures the opportunity cost of selling abroad relative to selling at home. If estimated simultaneously with (1), these two relative prices would identify the import demand schedule.

Based on this import supply equation, the system is recursive if 8,-0 and cov(uy., Veg) 0: The results (Table Bl) indicate that the data co not violate these two conditions in 45 out of 56 cases. For the remaining 11 cases, reliance

on ordinary least squares might introduce a simultaneity bias.

B.2 Super exogeneity B.2.1 Processes for Conditioning Variables Greatly simplified, a conditioning variable is said to be super exogenous both if the parameters characterizing its marginal density are not constant over time and the parameters of the conditioning model (the bilateral import equation) are constant (Engle, Hendry, and Richard, 1983).

In this paper, the conditioning variables are the export prices, GNP

deflators, the exchange rates, and the levels of income. To determine whether

40

their marginal densities exhibit parameter constancy, the following processes are

postulated for Canada, Germany, Japan, the United Kingdom, and the United States:

Iny, Ty Ing 2 +p IML ee2/ Pa ey) *

In Poot ying 2 th IMCL eso! Pe gig) *

In Pye 7 TING te MCL ceo/ Pe g.y)>

Ine, = 7 Ing +h IMCL eo Peel):

where Gy is the level of real government purchases, and L, is the level of money supply, which is being deflated by the overall absorption price index. For the three country blocs (OPEC, LDCs, and other industrial) there are no aggregate data for either government purchases or the money supply. Thus their government purchases are proxied by US government purchases and the money supply is prexied by the average value of the dollar.

The }arameters of these processes are estimated with the Brown-Durbin-Evans recursive estimation algorithm Starting in 1975Q1 and ending in 1985Q2: the first 8 observations (1973Q1-1974Q4) are reserved for the first estimate of these Parameters. The time-series for each parameter (7 and #) as well as its 95 percent confidence interval are shown in figures B.1-B.12. The evidence confirms the preserice of structural changes in the processes determining the conditioning variables. This evidence, coupled with the results on Chow tests for the conditioning model (Table 12), supports the notion that these conditioning variables are super exogenous and thus can be taken as exogenoxs for estimation

of bilateral trade equations.

B.2.2 Sequential Chow Tests The Chow tests of Table 12 might be sensitive to the choice of 1973 as the cutoff date. To examine this possibility, the analysis estimates the parameters of

(1) with the Brown-Durbin-Evans recursive estimator and uses the one-period ahead

4l

forecast error to perform sequential Chow tests. The estimation uses 1973Q1- 1976Q4 as the initial sample.

The probability of rejecting the null hypothesis of parameter constancy is F “2 “2 : prob(F(1, T)-K)<Foy)» where Fy ESO ST, + 11/2545 for J=1977Q1...T,, and

Tj-1977Q2. . .1985Q2. Figures B.13-B.20 display the time series for this probability for each of the bilateral trade equations.

Given the number of parameters estimated and the presence of response-delays in trade flows, a one-period ahead forecast error might not be powersul enough to detect changes in the parameters. Thus this analysis uses a 0.25 significance level as a criterion to reject the null hypothesis of parameter constancy. Out of the 56 equations, 14 fail to exhibit constant parameters (5 in Japan, 4 in Canada, 2 in LDCs, and 1 in Germany, the United Kingdom, and the United States). Although this evidence does not constitute a widespread rejection of super exogeneity, the number of violations is sufficiently large to suggest: that these

results are preliminary.

42

Table Bl f “* Evidence on Recursiveness: Bilateral Trade Equations

IMPORTING COUNTRY —_——_——— En COUNT

EXPORTING COUNTRY Can Ger Jap UK us Ind LDCs OPEC . Canada --- -1.23 -0.95 -0.41 0.28 0.21 0.51 -0.57 0.21 -0.01 -0.26 0.23 -0.11 -0.11 -0.11 Germany -2.24 --- 71.74 -2.78 -1.01 -3.78 -2.67 -0.71 -0.001 0.17. -0.14 -0.07. 0.05 -0.11 0.26 Japan -1.62 -1.35 --- 70.22) -0.70 -1.15 0.15 -2.78 -0.07 -0.33 -0.24 -0.13 0.09 -0.22 0.01 - UK -0.50 0.28 0.62 --- 1.67 -1.57 -1.27 -3.19 -0.13 -0.20 -0.40 -0.34 -0.12 0.02 -0.24 US -3.32 9 -1.28 -2.57. -0.87__ --- 0.22. -0.67 -1.31 0.089 0.15 0.12 -0.19 -0.18 -0.19 0.27 Ind. -0.54 1.14 -1.01 -2.45 -0.66 --- -0.85 -4.34 ~ 0.33 0.04 -0.09 0.05 0.10 -0.08 -0.09 LDCs 71.65 -0.75 -0.69 0.32 -0.83 2.22. --- 0.44 0.33 0.13 -0.14 0.33 0.37. -0.12 -0.17 OPEC 0.22 0.33 1.246 1.04 0.62 -0.22 -0.68 ---

-0.14 0.02 -0.13 -0.02 -0.13 0.01 0.23

eee eee

f For a given bilateral trade flow, the top entry is the t-statistic for the volume of exports in the price equation (Bl); the bottom entry is the correlation coefficient between the estimated residuals of (1) and those of (Bl).

42a

S06t tees £268 so64 teen £2461 Se61

Le

ol! 7ol

d-odx3 40 uBpEeURD d-Voduy 4O uepeues

90

27o1

90

a1

et

eol

JOWeyeQ IND ue—peuey Od Yodxy uBpeuey

Ayddng Aeuow feey uepeues vo 1U8}DYjE0F 40} FBAIO}UY BOUEPYUIOD %SE

7d Fansty

1968 4161 seer 1968 £168

ewooU; (BEY UB—PEUeD

JOWBYEG ND uBpEUED O24 WOdxy uepeuey

SOSBYIING JUBWIUOAOS) UBIPBUBD UO

WUO[IIJJOOD 104 faALEIU] BQUEPYUOD %~GE

Td ean3Ty

42b

or

se61 us) _ $868 test

40181109 dN uBuLED

Addins Aeuoy feey ueuueER UO WOEPIHEOSD 40) FBAIEW GOUEPYIOD %SE

pq ain3ty

4465

226s __ Sees e018

SOSRIUING JUCUALICADLD) UBLUIGE) UO JUGPIBIEOD 10) (BAIOW BOVEPYUOD %GE

€q aan3ty

4261

42c

sii

st

sz

$20

so

szo

seer sees 4268

1018110 dND eseueder 80pqg Wodxy eseueder

Aiddng Aeuoy peey eseusder uo WWEPHECD 10) FEAJEWI| EDUEPIFUOD %SE

QM AaNn3yTy

seek

ues sees $068 4268

eng yodxg eseueder

SesBYyaNg JUEWWEADE) ESeuBdEr UO WUOEPIHOECD 40) FAAIEWU] EIUEPIUCD %S6

Cq@ ean3ty

42a

S06! test L161

Se61 beet 4461 vO vo

*o1

90 vo v0 zo

eo Swosu] Fey *41'N

eo!

et zo

20

*o!

40188 dNO “WN e214 Yodxy ‘yw e20g YOdxy 4h

20

SSSSYTiiig jUGiiiLIEADL) 4 4) UO

A\ddng Aeuow reey “yn uo JUEPYjOOD 40} feALEW] BOUEPIIOD %GE

1Ue}DYjEOD JO} [BALE\U] EQUEPYUOD %GE

RH oaNdTy Ld aan3ty

42e

soot te6t 4464

Aiddng Aeuow jeey ‘s(n uo

1U81914J8OD 10) [eAse]U] BQUEPYUOD % SE Sulosu (Bey *S°N

Se6t 1e61 2268 t

801g Yodx3 ‘s'n

OLg e4nsTy

UOJ YOOD Joy PBAS@WU] BOUEPYUOD %GE st

80

S@SBYoINY JUEWWEADD ‘S‘N UO

se6t

1061 2161

407e178Q dNO 'S'N

co 90 60

at

6 Ansty

4264

Se6! 1861 L164 $061 $061 £461

“

ot

=-

4 eor - *o1

Oo

JB OQ 8y) JO ENIBA ey) UO SOSLYIING RIGWLWEADD UO

1U8}91jjO0D JO} JBAIEU] BQUEPUCD %SGE oz JUOPYJOOD 40) BAIOWU} CDUEP}UOD %S6 »

@o1g HodxZ $901 s

S$e6) 196! 4161 9061 1061 4166

col

+o!

0

eWwOdU] ;BEeY “SNPU;

ou 891g YOdxy ‘snpu| OWOIU] BEY “SNpUy

sio

ve

92'0

eo

NO 30 eoud ewodU (BEY S907 HO 10 89d

og L. JI oot _| oo —! gol _I

ZI eansty TIq@ oanaty

Figure B13

Probability of Rejecting Parameter Constancy Sequential One-Period Ahead Chow-Tests: Canada

a9

LDCs 0.75

045

O15

a9

OPEC 0.75

O45 O15

as

Industrial 075

O45

O15

ag Qa75 United States

O45

QS

ag Japan O75 O45

O15

a9 0.75

045

O15

United Kingdom

RSS on

0.45

0.15

1977 1978 1979 1980 1981 1982 1983 1984 1985

42h

Figure B14

Probability of Rejecting Parameter Constancy Sequential One-Period Ahead Chow-Tests Germany

ag OPEC 0.75

Q45

Q1S

ag LDCs Q7S

Q45 Q1S

as Industrial Q75 O45

O15

ole] United States 075

10.45

Quis

Japan O6

United Kingdom : O6

Canada ag

= F A A Ep y

1977 1978 1979 1980 1981 " 1982 1983 1984 1985

Figure B15

Probability of Rejecting Parameter Constancy Sequential One-Period Ahead Chow-Tests: Japan

ag OPEC 0.75 O45

Q1s

a9 LDCs O75 045

Q1S

O9

Industrial 075

O45

O18

0.9 Q75

United States

O45

ay ay bas ne

Oo

O18

United Kingdom

. ag Canada 075 O6 + 045 O3 Q1S 0

1977 1978 1979 1980 1981 1982 1983 1984 1985

423 Figure B16

Probability of Rejecting Parameter Constancy Sequential One-Period Ahead Chow-Tests: United Kingdom

a9 75 OPEC a6 O45 a3 OS 0 ag LDCs Q75 06 O45 03 O15 0

Industrial

L

United States a9

ag Japan ea O45 a3 O15 * 0

Og Q75

German ~ y O45

EA s 1S (9) a9 Q75 Canada O6 a4S O15 —,_ 1977 1978 1979 1980 1981 1982 1983 1984 1985S

42k Figure B17

Probability of Rejecting Parameter Constancy Sequential One-Period Ahead Chow-Teats: United States

a9 OPEC O75 O45

O15

9 O75

LDCs O45 O15

a9

Industrial 075

O45 O18

2

; ag

; 075

Japan 06 O45

O3

O15

S 0

3

ag O75

Germany

O45

O15

United Kingdom

Canada

1977 1978 1979 1980 1981 1982 1983 1984 1985

421 Figure B18

Probability of Rejecting Parameter Constancy Sequential One-Period Ahead Chow-Tests: Rest of OECD

, a9 OPEC 075

O45

O15

LDCs ag O75

O45 Q1S . Qs United States . 075 O45

Q1S

O9 Japan ae Q45

O15

Germany Qa75S

United Kingdom . 06

Canada Q75

7

a

wae

S .

-

x

1977 1978 1979 1980 1981 ~— 1982 1983 1984 1985

42m

Figure B19

Probability of Rejecting Parameter Constancy Sequential One-Period Ahead Chow-Tests: Non-OPEC LDCs

0.9 OPEC Q75 O45

O15

o9 Industrial O75

O45

Q1S

ag 075

United States

O45

O.1S

Germany

United Kingdom

Canada

° CERPREB

1977 1978 1979 1980 1981 1982 1983 1984 1985

Figure B20

Probability of Rejecting Parameter Constancy Sequential One-Period Ahead Chow-Tests: OPEC

. ag LDCs O75

—_ O45

Quis

~ as O75

Industrial O6 O45

O03

= O15

ag United States a O45

O15

a9 O75

Japan O45

ONS

O75

Germany

United Kingdom a4s

Canada

1977 1978 1979 1980 198i 1982 1983 1984 1985

43

References

Armington, P., 1969, Adjustment of trade balances: Some experiments with a model of trade among many countries, IMF Staff Papers, 27, 488-526.

Bergsten, F. and W. Cline, 1985, The US-Japan economic problem, Policy Analyses in International Economics, No. 13 (Institute for International Economics, Washington, D.C.).

Brown, B. and R. Mariano, 1984, Residual-based procedures for prediction and estimation in a nonlinear simultaneous model, Econometrica, 52, 321-344.

Bryant, R and G. Holtham, 1987, The US external deficit: Diagnosis, prognosis, and cure, Brookings Discussion Papers No. 55 (The Brookings Institution, Washington D.C.).

Chow, G., 1960, Tests of equality between sets of coefficients in two linear regressions, Econometrica, 28, 591-605.

Cline, W., 1984, International debt (MIT Press, Cambridge).

Cline, W., 1988, Medium-term prospects for the US external current account, Institute for International Economics, mimeo.

Deardorff, A. and R. Stern, 1986, The structure and sample results of the Michigan computational model of world production and trade, in Srinivasan

and Whalley (eds.).

Deppler, M. and D. Ripley, 1978, The world trade model: Merchandise trade, IMF Staff Papers, 25, 147-206.

Edison, H., J. Marquez, and R. Tryon, 1986, The structure and properties of the FRB MultiCountry Model, International Finance Discussion Papers, No. 293

(Federal Reserve Board, Washington D.C.).

44

Eichengreen, B., 1987, Trade deficits in the long run, paper presented at the conference "The U.S. Trade Deficit-Causes, Consequences, and Cures," Federal Reserve Bank of St. Louis, October 23-24, 1987.

Engle, R., 1982, Autoregressive conditional heteroskedasticity with estimates of the variance of the United Kingdom inflation, Econometrica, 50, 997-1008.

Engle, R. D. Hendry, and J. Richard, 1983, Exogeneity, Econometrica, 51, 277-304.

Fair, R., 1986, Evaluating the predictive accuracy of models, in Z. Griliches and M. Intriligator (eds.), Handbook of Econometrics, vol. 3 (North-Holland, Amsterdam).

Feldstein, M., 1986, Correcting World Trade Imbalances, mimeo

Fretz, D. T. Srinivasan, and J. Whalley, Introduction, in Srinivasan and Whalley (eds.).

Gagnon, J., 1987, Adjustment Costs and International Trade Dynamic, Ph.D. dissertation, Stanford University.

Goldstein, M. and M. Khan, 1985, Income and price elasticities in foreign trade, in R. Jones and P. Kenen (eds.), Handbook of International Trade, vol. II (North-Holland, Amsterdam).

Greenberg E., and C. Webster, 1983, Advanced econometrics: A bridge to the literature (Wiley, New York).

Haynes, S. and J. Stone, 1983, Secular and cyclical responses of U.S. trade to income: An evaluation of traditional models, Review of Economics and Statistics, 65, 87-95.

Haynes, S., M. Hutchinson, R. Mikesell, 1986, Japanese financial policies and the US trade deficit, Essays in International Finance, No. 162, Princeton

University.

45

Helkie, W. and P. Hooper, 1987, The US external deficit in the 1980s: An empirical analysis, International Finance Discussion Papers, No. 304 (Federal Reserve Board, Washington D.C.).

Hickman, B. and L. Lau, 1973, Elasticities of substitution and export: demand in a world trade model, European Economic Review, 4, 347-380.

Hooper, P. and C. Mann, 1987, The U.S. External Deficit: Its Causes end Persistence, "The U.S. Trade Deficit-Causes, Consequences, ard Cures," International Finance Discussion Papers, No. 316 (Federal Reserve Board, Washington D.C.).

Hooper, P., 1988, The dollar, external imbalances, and the US economy, Journal of Economic and Monetary Affairs, 2, 30-53.

Houthakker, H. and S. Magee, 1969, Income and price elasticities in world trade, Review of Economics and Statistics, 51, 111-125.

Husted, S. and T. Kollintzas, 1984, Import demand with rational expectations: Estimates for Bauxite, Cocoa, Coffee and Petroleum, Review of Economics and Statistics, 66, 608-618.

International Monetary Fund, 1987, Direction of Trade, (International Monetary Fund, Washington D.C.)

Jarque, C. and A. Bera, 1980, Efficient tests for normality, homosceclasticity,

and serial independence of regression residuals, Economic Let:ters, 6

255-259. Johnston, J., 1984, Econometric methods, Third Edition (M°Graw-Hill, New York). Khan, M. and K. Ross, 1977, The functional form of the aggregate import equation, Journal of International Economics, 7, 149-160. Krinsky, I., and L. Robb, 1986, On approximating the statistical properties of

elasticities, Review of Economics and Statistics, 68, 715-719.

Appendix C Estimated Structural Equations

4.6

Krugman, P. and R. Baldwin, 1987, The persistence of the US trade deficit, Brookings Papers on Economic Activity, 1, 1-56.

McCarthy, M., 1972, Some notes on the generation of pseudo-structural errors for use in stochastic simulation studies, in B. Hickman, ed., Econometric Methods of Cyclical Behavior (Columbia University Press, New York).

Magee, S., 1975, Prices, income, and foreign trade, in P. Kenen (ed.),

| International trade and finance: Frontiers for research (Cambridge University Press, Cambridge).

Marris, S., 1987, Deficits and the dollar: The world economy at risk, Policy Analyses in International Economics, No. 14 (Institute for International Economics, Washington, D.C.).

Spence, G., 1984, The world trade model: revised estimates, IMF St:aff Papers, 31, 469-498.

Schwartz, N. and B. Krissoff, How strategies to reduce US bilateral trade deficits in manufactures affect US agricultural exports, US Department of Agriculture, Economic Research Service.

Srinivasan, T. and J. Whalley, 1986, (ed.) General equilibrium trade policy modeling, (MIT Press, Cambridge).

Thursby, J. and M. Thursby, 1984, How reliable are simple, single equation specifications of import demand? Review of Economics and Statistics, 66, 120-128.

Warner, D. and M. Kreinin, 1983, Determinants of international trade flows, Review of Economics and Statistics, 65, 96-104.

Whalley, J., 1985, Trade liberalization among major world trading areas, (MIT

Press, Cambridge).

1. MCEV: (MCEV} CANADIAN IMPORTS FROM UK ($)

OG(MCEV/(EEI * JAIME_EXUVI)) = -.906852 + 1.75187 * (LOG(CGNP) - LOG(CGNPPOT }) (2.39578) (2.57314)

+ LAGCOEF1] * (LOG(EEI(I) *% JAIME_EXUVI(I)/(CEI(I) * CPGNP(I))))

+ =.449787 * LOG(MCEV(-1)/(EEI(-1) * JAIME_EXUVI(-1))) - .070393 * Ql + -045138 * Q2 (3.70857) (1.68978) (1.09861)

- .085620 * Q3 + MCE_ERR * BETA (1.97727)

CRSQ = .685 S.E.R. = .103 MEAN LHS = -3.68 RANGE: 1973 Q1 TO 1985 q2

DW = 2.004 NOB = 50 ESTIMATED: 9/87

LAG LAGCOEF1 T-STAT

(-0) -0.051 2.303

(-1) -0.077 2.303

(-2) -0.077 2.303

(-3) -0.051 2.303 SUM -0.256

3. MCGV: (MCGV) CANADIAN IMPORTS FROM GER($)

wee em Pm men wm eee en ee ww wn we ee eee

LOGCUMCGV/(GEI * GPXGUV)) = -4.8314¢ + .999954 * LOG(CGNPPOT ) (4.48663) (4.50994)

+ 2.93139 * (LOG(CGNP) - LOG(CGNPPOT)) - 33a? |* LOG(GEI * GPXGUV/(CPGNP * CEI))

(3.74673)

+ .538413 * LOG(MCGV(-1)/(GEI(-1) * GPXGUV(-1))) ~- .057965 * Ql + .060084 * Qz (5.36549) (1.55924) (1.62976)

~ .147455 * Q3 + MCG_ERR * BETA (3.68265)

CRSQ = .82 S.E.R. = .091 MEAN LHS = -.562 RANGE: 1973 Ql TO 1985 Q2

OW = 2.032 NOB = 50 ESTIMATED: 9/87

4. MCG: (MCG) CANADIAN IMPORTS FROM GER 1972 PRICES

5. MCJV: (MCJV) CANADIAN IMPORTS FROM JAPAN( $)

LOG(MCJV/( JEI * JPXGUV)) = -5.8522 + 1.24267 * LOG(CGNPPOT) (3.95521) (3.98426) + 1.44395 % (LOG(CGNP) - LOG(CGNPPOT)) - .517518 * LOG(JEI * JPXGUV/(CPGNP * CEI)) (1.62195) (3.11929) + .601615 * LOG(MCJIV( -1)/(JEI(-1) % JPXGUV(-1))) - .067089 * Ql - .031598 * Q2 (6.23995) (1.93616) (.681272)

- .057539 * Q3 + MCJ_ERR * BETA (1.2193)

CRSQ = .886 S.E.R. = .115 MEAN LHS = .206 RANGE: 1973 Q1 TO 1985 Q2 OW = 1.948 NOB = 50 ESTIMATED: 9/87

-2-

6. MCJ: (MCJ) CANADIAN IMPORTS FROM JAPAN 1972 PRICES

MCJ = MCJV/(JEI * JPXGUV)

7. MCUV: (MCUV) CANADIAN IMPORTS FROM US ($)

LOG( MCUV/UPXGUV ) = 76.01742 + 1.79566 * LOG(CGNPPOT) + 2.49483 % (LOG(CIGNP) - LOG(CGNPPOT)) (6.21352) (6.8207) (5.45721) + LAGCOEF1 * (LOG(UPXGUV(T )/(CPGNP(I) ® CEI(I)))) + 230863 |* LOG(MCUV! -1)/UPXGUV( -1)) -039 - .047536 * Ql + .071571 * Q2 - .113498 * Q3 + MCU_LERR * BETA (2.48321) (4.12381) (4.13567) CRSQ = .927 S.E.R. = .039 MEAN LHS = 2.904 RANGE: 1973 Q1 TO 1985 Q2

OW = 1.925 NOB = 50 ESTIMATED: 9/87

LAG LAGCOEF1 T-STAT

(-0) -0.132 .1.712 (-1) -0.090 2.009 (-2) -0.058 2.595 (-3) -0.035 2.019 (-4) -0.021 0.843 (-5) -0.016 0.534 (-6) -0.021 0.670 (-7) -0.035 1.271 (-8) -0.058 2.499 (-9) -0.090 3.023 (-10) -0.131 2.567 (-11) ~0.182 2.204 SUM: -0.868

9. MCLV: (MCLV) CANADIAN IMPORTS FROM LOC ($)

LOG( MCLV/LPXGUV ) = -7.56801 + 1.63611 * LOG(CGNPPOT) + 1.94802 * (LOG(CGNP) - LOG(CGNPPOT)) (4.768) - (4.7416) (2.71499)

~ _.687005 * LOG(LPXGUV/(CPGNP * CEI)) + .419033 * LOG(MCLV(-1)/LPXGUV(-1)) - .026230 * Ql (3.98157) (3.39263) (. 798332)

+ .055501 * Q2 + .079723 * Q3 + MCL_ERR * BETA (1.61789) (2.37223)

CRSQ = .913 S.E.R. = .081 MEAN LHS = .35 RANGE: 1973 Q1 TO 1985 Q2

OW = 1.871 NOB = 50 ESTIMATED: 9/87

10. MCL: (MCL) CANADIAN IMPORTS FROM LDC 1972 PRICES

MCL = MCLV/LPXGUV

11. MCIV: (MCIV) CANADIAN IMPORTS FROM OECD ($)

LOG( MCIV/IPXGUV ) = -2.50705 + .603119 * LOG(CGNPPOT) + 1.78058 * (LOG(CGNP) - LOG(CGNPPOT )) (2.08602 ) (2.36937) (3.05668)

+ LAGCOEF1 * (LOGCIPXGUVIT )/(CPGNP(T) * CEI(I)))) + LAGCOEF2 * (LOG(CFPXFTW(I)/(CPGNP(I) * CEI(I)))) ¢* .53592 * LOG(MCIV( -1)/IPXGUV(-1))

(4.61324) - .065427 * Ql + .027939 * Q2 - .029637 * Q3 + MCI_ERR * BETA (2.40423) (1.01093) (1.06237) = .067 MEAN LHS = .481 RANGE: 1973 Q1 TO 1985 Q2

CRSQ = .844 S.E.

R. DW = 2.18 NOB = 50 ESTIMATED: 9/87

LAG LAGCOEF1 T-STAT LAGCOEF2 T-STAT

(-0) -0.163 1.472 -0.632 2.095 (-1) -0.244 1.472 0.642 1.999 (-2) -0.244 1.472 0.749 2.274 (-3) -0.163 1.472 -0.311 0.979

SUM: -0.814 0.448

12. MCI: (MCI) CANADIAN IMPORTS FROM OECD 1972 PRICES

MCI = MCIV/IPXGUV

13. MCOV: (MCOV) CANADIAN IMPORTS FROM OPEC ($)

MCOV/(CPMGOL * CER) = -.841979 + 1.1109 * EXP(LOG(CGNP) - LOG(CGNPPOT )) (1.15981) (1.51128)

+ LAGCOEF] * ((CER(I) * CPMGOL(I)/(CEI(I) *% CPXGOL(I))))

+ 816719 * (MCOV(-1)/(CPMGOL(-1) * CER(-1))) ¢ .078139 * Ql + .008562 * Q2 (10.2592) (2.78829) (.306368)

+ .057502 * Q3 + MCO_ERR * BETA (2.05813)

CRSQ = .909 S.E.R. = .068 MEAN LHS = .466 RANGE: 1973 Ql TO 1984 Q4 OW = 2.145 NOB = 48 ESTIMATED: 9/87

LAG LAGCOEF1 T-STAT

{-0) -0.021 0.802

(-1) -0.032 0.802

(-2) -0.032 0.802

1-3) -0.021 0.802 SUM -0.106

14. CFPXFTW: (CFPXFTW) CANADIAN IMPORT PRICES FROM MCM COUNTRIES

CEPXFTR = UPXGUV=*0.328571 * (JPXGUV * JBI )e%0.2 % (GPXGUV * GET )##0.3 # (EPXGUV * EEI )**0.17

15. MCO: (MCO) CANADIAN IMPORTS FROM OPEC 1972 PRICES

M20 = MCOV/(CPMGOL * CER)

16. MCTV: (MCTV) TOTAL CANADIAN IMPORTS ($)

MCTV = MCEV + MCGV + MCUV + MCUV + MCIV + MCLV + MCOV + MCZV

17. CTRADE: (CTRADE) CANADIAN TRADE BALANCE

CTRADE = XCTV - MCTV

18. XCTV: (XCTV) CANADIAN EXPORTS TOTAL ($)

XCTV = XCEV + XCGV + XCJV + XCUV + XCIV + XCLV + XCOV + XCZV

19. XCEV: (XCEV) CANADIAN EXPORTS TO THE UK ($)

LOG( XCEV) = -,088145 + .943772 * LOG(MECV) - .010230 * Ql - .030187 * Q2 (5.03177) (48.8788) (.51611) (1.52525) - .037929 *« Q3 + XCE_ERR * BETA (1.90014) CRSQ = .959 S.E.R. = .072 MEAN LHS = .385 RANGE: 1960 Ql TO 1986 Q2

OW = 1.878 NOB = 106 ESTIMATEO: 6/87

20. XCGV: (XCGV) CANADIAN EXPORTS TO GERMANY ($)

LOG( XCGV ) = -.311306 + .913194 * LOG(MGCV) - .052766 * Ql + -001985 * QZ . (12.0773) (60.1072) (1.5055) (.056918) - .043083 * Q3 + XCG_ERR * BETA (1.22433) CRSQ = 9728 S.E.R. = .127 MEAN LHS = -.869 RANGE: 1960 Ql TO 1986 Q2 OW = 1.5 = 106 ESTIMATED: 6/87

21. XCUV: (XCJV) CANADIAN EXPORTS TO _JAPAN ($)

LOG( XCJV ) = -.226135 + 1.03328 * LOG(MJCV) - .039684 * Ql + -001677 * Qv (5.46939) (53.424) (693691) (.029317) - .006317 * Q3 + XCJ_ERR * BETA (.109409) CRSQ = .965 S.E.R. = .208 MEAN LHS = .116 RANGE: 1960 Ql TO 1986 Q2

DW = .53 NOB = 106 ESTIMATED: 6/87

22. XCUV: (XCUV) CANADIAN EXPORTS TO THE US ($)

LOG( XCUV ) = -.170¢92 + 1.03189 = LOG(MUCV) + .013725 * Ql + .013254 * Q2 (11.4945) (241.205) (1.14658) (1.10777) + .012761 * Q3 + XCU_LERR * BETA (1.05582) CRSQ = .998 S.E.R. = .044 MEAN LHS = 2.715 RANGE: 1960 Ql TO 1986 Q2

OW = .718 NOB = 106 ESTIMATED: 6/87

23

24. MEC:

MECV:

(MECV) UK IMPORTS FROM CANADA--NOMINAL $

LOG(MECV/(CEI * CPXGUV))

LAG

pettus WMhwune od weer

S

+

+

+

Cc D

= -.047241 (.023288)

+ .043925 * LOG(EGNPPOT ) (.090471)

1.51671 * (LOGLEGNP) - LOG EGNPPOT))

(1.82552)

LAGCOEF1 * (LOG(CEI(I) * CPXGUV(T)/(EEI(I) * EPGNP(T)))) LAGCOEF2 * (LOG( ECPCOMP(T )/(EEI(I) * EPGNP(T))))

.676115 * LOG(MECV( -1)/(CEI(-1) % CPXGUVI(-1))) - .035666 * Ql + .117444 * Q2 (5.34856) (1.02228) (3.42915) .096808 * Q3 + MEC_ERR * BETA (2.6315) RSQ = .786 S.E.R. = .084 MEAN LHS = .19 RANGE: 1973 Q1 TO 1985 Q2 W= 2.106 NOB = 50 ESTIMATED: 9/87 LAGCOEF1 T-STAT LAGCOEF2 T-STAT -0.060 1.836 0.243 1.108 -0.099 1.836 0.125 1.215 -0.119 1.836 0.044 0.273 -0.119 1.836 0.000 0.001 -0.099 1.836 -0.007 0.117 -0.060 1.836 0.022 0.112 -0.557 0.427

(MEC ) UK IMPORTS FROM CANADA --REAL $

MEC

25. MEGV:

26.

MECV/(CEI * CPXGUV)

(MEGV) UK IMPORTS FROM GERMANY--NOMINAL $

LOG( MEGV/(GEI * GPXGUV))

+

= -8.87822 (3.71818)

+ 2.27815 * LOG(EGNPPOT) (3.76233)

1.33089 * (LOG(EGNP) - LOG(EGNPPOT ))

(2.37969)

+ LAGCOEF1 * (LOG(GEI(I) * GPXGUV(I)/(EPGNP(I) * EEI(I)3))

+ .591065 % LOG(MEGV(-1)/(GEI(-1) * GPXGUV(-1))) - .015747 * Ql - .012944 * Q2z (5.31417) (.659868) (540618) - .056127 * Q@3 + MEG_ERR * BETA (2.24409) CRSQ = .973 S.E.R. = .059 MEAN LHS = 1.416 RANGE: 1973 Ql TO 1985 Qz DW = 2.046 NOB = 50 ESTIMATED: 9/87 LAG _LAGCOEF1 T-STAT (-0) -0.022 2.214 (-1) -0.036 2.214 (-2) -0.043 2.214 (-3) -0.043 2.214 (-4) -0.036 2.214 (-5) -0.022 2.214 SUM: -0..202

MEG: (MEG) UK IMPORTS FROM GERMANY --REAL $

MEG

MEGV/(GEI % GPXGUV)

27. MEJV: (MEJV) UK IMPORTS FROM JAPAN--NOMINAL $ LOGCMEJV/(JEI * JPXGUV)) = -8.0167 + 1.9178 *® LOG(EGNPPOT) . (3.25723) (3.20669) - .194255 * (LOGCEGNP) - LOG( EGNPPOT )) (. 281266) + LAGCOEF1] * (LOG(JEI(I) * JPXGUV(I)/( EPGNP(I) * EEI(I)))) + .598201 * LOG(MEJV(-1)/(JEI(-1) * JPXGUV(-1))) + .09000¢ * Ql + .113102 * Q2 (5.01832) (2.56423) (3.5553) + .127013 * Q3 + MEJ_ERR * BETA - (4.04346 ) CRSQ = .963 S.E.R. = .077 MEAN LHS = .339 RANGE: 1973 Q1 TO 1985 Q2 OW = 2.28 NOB = 50 ESTIMATED: 9/87 LAG LAGCOEF1 T-STAT (-0) -0.013 0.79% _€-21)0 -0.022 0.79% (-2) -0.027 0.79% (-3) -0.027 0.7% (-4) -0.022 0.79% (-5) -0.013 0.794 SUM: -0.12¢

28. MEJ: (MEJ) UK IMPORTS FROM JAPAN --REAL $

MEJ

29. MEUV:;

= MEJV/(JEI * JPXGUV)

(MEUV) UK IMPORTS FROM US--NOMINAL $

LOG( MEUV/UPXGUV ) = -6.40587 + 1.68229 % LOG(EGNPPOT) + .99023 * (1.0G(EGNP) - LOG(EGNPPOT))

(3.15483) (3.25679) (1.40674) + LAGCOEF1 * (LOG(UPXGUV(I)/(EPGNP(I) * EEI(I)))) + LAGCOEF2 * (LOG(EUPCOMP(T )/(EPGNP(I) * EEI(I)))) + .589266 * LOG(MEUV( -1 )/UPXGUV(-1))

(4.6091) + .03466¢ * Ql + .008671 * Q@2 - .091886 * Q3 + MEU_ERR * BETA (1.07658) (.255069) (2.62238) CRSQ = .855 S.E.R. = .078999 MEAN LHS = 1.445 RANGE: 1973 Q1 Ti) 1985 Q2 OW = 2.165 NOB = 50 ESTIMATED: 9/87 LAG LAGCOEF1 T-STAT LAGCOEF2 T-STAT (-0) -0.040 1.987 0.045 1.860 (-1) -0.066 1.987 0.074 1.860 (-2) -0.080 1.987 0.089 1.860 (-3) -0. 1.987 0.089 1.860 (-@) -0.066 1.987 0.074 1.860 (-5) -0.0460 1.987 0.045 1.860 SUM: -0.371 0.416

30. MEU: (MEU) UK IMPORTS FROM THE US --REAL $

MEU

= MEUV/UPXGUV

31. MELV: (MELV) UK IMPORTS FROM LDCS --NOMINAL $

Cwm we wee eee wee wee ee wen wee www we ewe ewe

LOG( MEI.V/LPXGUV ) = ~2.88219 + .947819 * LOG(EGNPPOT) + 1.53729 *® (LOG(EGNP) - LOG(EGNPPOT)) (2.1166) (2.74739) (2.21121) + LAGCOEF1 * (LOG( LPXGUV(T )/CEPGNP(I) * EEI(I)})) + - 352323 * LOG(MELV( -1)/LPXGUVE-1)) (2.56709) + .062674¢ * Ql + .067952 * Q2 + .050369 * Q3 + MEL_ERR * BETA (1. 91548 ) (2.12739) (1.549) CRSQ = .348 S.E.R. = .079 MEAN LHS = 1.609 RANGE: 1973 Ql TO 1985 Q2

Di = 2.094 NOB = 50 ESTIMATED: 9/87

LAG LAGCOEF1 T-STAT

(-0) -0.012 1.080 (-1) -0.020 1.080 (-2) -0.024 1.080 (-3) -0.024¢ 1.080 (-@) -0.020 1.080 (-5) -0.012 1.080 SUM: -0.113

32. MEL: (MEL) UK IMPORTS FROM LOCS --REAL $

MEL = MELV/LPXGUV

33. MEIV: (MEIV) UK IMPORTS FROM OTHER OECD--NOMINAL $

LOG( MEIV/IPXGUV ) = -%.96903 + 1.50997 * LOG(EGNPPOT) + 1.12319 *® (LOG(EGNP) - LOG(EGNPPOT)) (3.67087) (3.87846) (2.60929)

+ LAGCOEF1 * (LOG( IPXGUV(I)/(EPGNP(I) * EEI(I}})) + Partita LOG( MEIV( -1 )/IPXGUV(-1))

- _.042124 * Ql - .015128 * Q2 - .111669 * Q3 + MEI_ERR * BETA (2.11553) (. 752988 ) (5.02822) CRSQ = .943 S.E.R. = .047 MEAN LHS = 2.861 RANGE: 1973 Q1 TO 1985 Q2

DN = 2.166 NOB = 50 ESTIMATED: 9/87

LAG LAGCOEF1 T-STAT

(-0) -0.006 0.779 (-1) -0.011 0.779 (-2) -0.013 0.779 (-3) -0.013 0.779 (-4) -0.011 0.779 (-5) -0.006 0.779 SUM -0.060

34. MEI: (ME) UK IMPORTS FROM OTHER OECD --REAL $

MEI = MEIV/IPXGUV

35. MEOV: (MEOV) UK IMPORTS FROM OPEC --NOMINAL $

LOG( MEOV/OPOIL 72 ) = 2.65209 + 7.0711 * (LOG(EGNP) - LOG(EGNPPOT )) (10.8416) (4.6483 }

+ LAGCOEF] # (LOG(OPOIL72(I )/(EEI(I) % EPGNP(I})}) ¢ .105823 # Ql - .016073 * Q2 (1.13479) (.175679)

+ .027703 * Q3 + MEO_ERR * BETA (307369)

cRSQ = .896 S.E.R. = .201 MEAN LHS = -.558 RANGE: 1976 Q2 TO 1986 Ql DW = 1.4631 NOB ‘= 40 ESTIMATED: 9/87

LAG LAGCOEF1 T-STAT

(-0) -0.034¢ 9.231 (-1) -0.064 9.231 (-2) -0.090 9.231 (-3) -0.112 9.231 (-4) -0.130 9.231 (-5) -0.144 9.231 (-6) -0.154 9.231 (-7) -0.160 9.231 (-8) -0.162 9.231 (-9) -0.160 9.231 (-10) -0.154 9.231 (-11) -0.144 9.231 (-12) -0.130 9.231 (-13) -0.112 9.231 (-14) -0.090 9.231 (-15) -0.064 9.231 (-16) -0.034¢ 9.231 SUM: -1.939

36. MEO: (MEO) UK IMPORTS FROM OPEC --REAL $

MEO = MEOV/OPOIL72

37. METV: (METV) TOTAL IMPORTS OF THE UK --$

METV = MECV + MEGV + MEJV + MEUV + MEIV + MELV + MEOV + MEZV

38. XETV: (XETV) TOTAL EXPORTS OF THE UK -- NOMINAL $

mem ww eww wwe em ewww wee wwe wm we we we wm wee eer ese

XETV = XECV + XEGV + XEJV + XEUV + XEIV + XELV + XEOV + XEZV

39. ETRADE: (ETRADE) UK TRADE BALANCE

ETRADE = XETV - METV

40. XECV: (XECV) UK EXPORTS TO CANADA --NOMINAL $¢

LOG(XECV) 2 -.053247 + .929451 *% LOG(MCEV) - .073810 # Ql - .043554 ® Q2 (3.65609) (63.2753) (3.62568) (2.13973) - .064986 * Q3 + XEC_LERR * BETA (3.16478) CRSQ = .975 S.E.R. = .073999 MEAN LHS = -.033 RANGE: 1960 Ql TO 1966 Q2

DW = 2.126 NOB = 106 ESTIMATED: 6/87

G1. XEGV: (XEGV) UK EXPORTS TO GERMANY --NOMINAL $

LOG(XEGV) = .040672 + .942512 * LOGIMGEV) - .001625 * Ql - .007510 * Q2 (2.87178) (171.312) ( 087364 ) (.403736) - .036305 * Q3 + XEG_ERR * BETA (1.93283) CRSQ = .996 S.E.R. = .068 MEAN LHS = .828 RANGE: 1960 Ql TO 1986 Q2

OW = 1.983 NOB = 106 ESTIMATED: 6/87

42. XEJV: (XEJV) UK EXPORTS TO JAPAN --NOMINAL $

LOG(XEJV ) = .900496 * LOGIMJEV) - .274376 * Ql - .227413 * Q2 (45.415) (6.27471) (5.22701) - .305266 * Q3 + XEJ_ERR * BETA (6.88463) CRSQ = .967 S.E.R. = .219 MEAN LHS = -.782 RANGE: 1960 Ql TO 1986 Q2

OW = 1.415 NOB = 106 ESTIMATED: 6/87

43. XEUV: (XEUV) UK EXPORTS TO THE US --NOMINAL $

LOG(XEUV) = -.0¢1784 + 1.01765 % LOG(MUEV) - .046306 * Ql - .027039 * Q2 (2.66803) (149.418) (2.49477) (1.52331) - .028481 * Q3 + XEU_ERR * BETA (1.58942) CRSQ = .995 S.E.R. = .06¢999 MEAN LHS = 1.28 RANGE: 1960 Ql TO 1986 Q2

DW = 1.716 NOB = 106 ESTIMATED: 6/87

4%. ECPCOMP: (ECPCOMP) THIRD COUNTRY PRICE FOR UK-CANADA TRADE

ECPCOMP = (UPXGUV=*0.34¢3284 * (JPXGUV # JEI )#*0.2 % (GPXGUV * GEI) 40. 313433 HL 1/01 - 0.13))

45. EUPCOMP: (EUPCOMP) THIRD COUNTRY PRICE FOR UK-US TRADE

EUPCOMP = ((JPXGUV * JEI)#*0.2 * (CPXGUV * EEI )¥#0.134328 * (GPXGUV * GEI) HHO. 313433 EL I/01 - 0.3433))

- 10 -

G6. MGEV: (MGEV) GERMAN IMPORTS FROM THE UK --NOMINAL $

LOG(MGEV/(EEI * JAIME_EXUVI )) = -33.7207 + 4.61177 * LOG(GGNPPOT) (27.2951) (20.4508)

+ 2.90971 * (LOG(GGNP) - LOG(GGNPPOT ))

(5.07679) + LAGCOEF1 * (LOG(EEI(I) * JAIME_EXUVI(I)/(GEI(I) * GPGNP(I)))) + .518098 * 07731 (7.67487) ~ .018409 * Ql - .026938 * Q@2 - .090364 * Q3 + MGE_ERR * BETA (622883) (1.00672) (3.00664) CRSQ = .978 S.E.R. = .06 MEAN LHS = -2.278 RANGE: 1973 Ql TO 1985 Q2 DH = 2 NOB = 50 ESTIMATED: 9/87 LAG _LAGCOEF1 T-STAT (-0) -0.263 1.064 (-1) 0.172 0.530 (-2) 0.375 1.167 (-3) -0.144 0.441 (-4) -0.197 0.586 (-5) 0.192 0.588 (-6) 0.321 1.020 (-7) -0.312 0.990 (-8) -0.445 1.426 (-9) 0.583 1.799 (-10) 0.210 0.565 (-11) —_-0.601 2.348 SUM: -0.109 47. MGE: (MGE) GERMAN IMPORTS FROM THE UK --REAL $ MGE = MGEV/(EEI * ( JAIME_EXUVI/100)) 48. MGCV: (MGCV) GERMAN IMPORTS FROM CANADA --NOMINAL $ LOG(MGCV/(CEI * CPXGUV)) = -9.79574 + 1.41943 * LOG(GGNPPOT ) (3.21896) (3.14813) + 2.70389 * (LOG(GGNP) - LOG(GGNPPOT }) (1.89251) + LAGCOEF1] * (LOG(CEI(I) * CPXGUV(IJ/(GPGNP(I) * GEI(I)))) + .482806 * LOG(MGCVI-1)/(CEI(-1) * CPXGUV(-1))) - .188369 * Ql - .040741 * Q2z 4.02967) 14.01762) (1841971) - .017828 * Q3 + MGC_ERR * BETA (.374) CRSQ = .725 S.E.R. = ..116 MEAN LHS = -.535 RANGE: 1973 Ql TO 1935 Q@2

DW = 1.916 NOB = 50 ESTIMATED: 9/87

LAG LAGCOEF1 T-STAT

(-0) -0.613 2.404 (-1) 0.318 1.585 (-2) 0.378 1.871 (-3) -0.433 1.612 SUM: -0.349

49. MGC: (MGC) GERMAN IMPORTS FROM CANADA --REAL $

MGC = MGCV/(CEI * CPXGUV)

-ll-

50. MGJV: (MGJV) GERMAN IMPORTS FROM JAPAN --NOMINAL $

LOG(MGIV/( JEI * JPXGUV)) = -12.6377 + 1.87856 * LOG(GGNPPOT ) (4.01652) (4.0361)

+ 1.01407 * (LOG(GGNP) - LOG(GGNPPOT)) - .737357 * LOG(JEI * JPXGUV/(GPGNP x GEI)) (1.29627) (3.65073)

+ .377091 *% LOG(GJPCOMP/(GPGNP * GEI)) +¢ 5 1eAcs * LOG( MGJV(-1)/( JEI(-1) % JPXGUVI-1))) (2.12976) (5.4449

- .118525 * Ql - .027497 * Q2 - .098918 * Q3 - .348105 * D7411 + MGJ_ERR * BETA (4.16268) (1.03949) (3.62728) (4, 75398)

CRSQ = .975 S.E.R. = .066 MEAN LHS = .589 RANGE: 1973 Ql TO 1985 Q2

OW = 2.192 NOB = 50 ESTIMATED: 9/87

51. MGJ: (MGJ) GERMAN IMPORTS FROM JAPAN --REAL $

MG.) = MGJV/(JEI * JPXGUV)

52. MGUV: (MGUV) GERMAN IMPORTS FROM THE US --NOMINAL $

LOG: ( MGUV/UPXGUV } = -8.7571 + 1.46237 * LOG(GGNPPOT) + 2.15389 *® (LOG(GGNP) - LOG(GGNPPOT?) (6.43253) (6.87777) (3.42584)

- .676056 * LOG(UPXGUV/(GPGNP * GEI)) +¢ .431814¢ * LOG(GUPCOMP/(GPGNP * GEI)) (4.02952) (1.95219)

+ .249013 # LOG(MGUV( -1)/UPXGUV(-1)) - .039787 * Ql - .053619 * Q2 - .148261 * Q3 (2.8968) (1.52089) (2.07693) (5.61832)

- .461141 * 07731 + MGU_ERR * BETA (7.61598)

CRSQ = .891 S.E.R. = .057 MEAN LHS = 1.478 RANGE: 1973 Ql TO 1985 Q2

DW = 2.078 NOB = 50 ESTIMATED: 9/87

53. MGU: (MGU) GERMAN IMPORTS FROM THE US --REAL bd

54. MGLV: (MGLV) GERMAN IMPORTS FROM LDCS --NOMINAL $

LOG} MGLV/LPXGUV ) = -11.6676 + 1.94801 * LOG(GGNPPOT) +¢ 1.74064 * (LOG(GGNP) - LOG(GGNPPOT)) (6.39503) (6.5288) (3.61565)

- .137308 ‘a LOG( LPXGUV/( GPGNP *% GEI)) - .352286 * LOG(GFPXFTW/(GPGNP * GEI)) (1.03589 (2.30592)

+ .144077 * LOG(MGLV(-1)/LPXGUV(-1))} + .030580 * Ql - .020542 * Q2 + .004527 * Q3 (1.22422) (1.97839) (1.34788) (.292689)

- .199818 * 07411 + MGL_ERR * BETA (4.6877)

CRSQ = .947 S.E.R. = .038 MEAN LHS = 1.96] RANGE: 1973 Q1 TO 1985 Q2

OW = 1.819 NOB = 50 ESTIMATED: 9/87

55. MGL: (MGL) GERMAN IMPORTS FROM LOCS --REAL $

MGL. = MGLV/LPXGUV

- 12 -

56. MGIV: (MGIV) GERMAN IMPORTS FROM OTHER OECD --NOMINAL $

LOG( MGIV/IPXGUV ) =_ -10.1431 + 1.94502 *% LOG(GGNPPOT) + 1.77874 % (LOGIGGNP) - LOG(GGNPPOT)) (5.44176) (5.64768) (4.47704)

- .664719 * LOG(IPXGUV/(GPGNP * GEI)) + .109773 * LOG(GFPXFTW/(GPGNP * GEI)) (4.39183) (1.15903)

+ .091571 * LOG(MGIV(-1)/IPXGUV(-1)) - .010899 * Ql - .006130 * Q2 - .087469 * Q3 (.607951) (.036164) (.409743 ) (5.11077)

- .094282 * D7411 + MGI_ERR * BETA (3.28364) .

CRSQ = .976 S.E.R. = .026 MEAN LHS = 3.468 RANGE: 1973 Ql TO 1985 «2

OW = 1.663 NOB = 50 ESTIMATED: 9/87

57. MGI: (MGI) GERMAN IMPORTS FROM OTHER OECD --REAL $

MGI = MGIV/IPXGUV

58. MGOV: (MGOV) GERMAN IMPORTS FROM CPEC --NOMINAL $

LOG( MGOV/OPOIL72 ) = .645121 + 2.9872 * (LOGI GGNP) - LOG(GSNPPOT)) (2.41378) (2.48648 }

+ LAGCOEF1 * (LOG(OPOIL72(I )/(GEI(I) * GPGNP(I}))) + .603227 ® LOG(MGOV( -1)/NPOIL72(-1))

(5.33792) - .084049 * Ql - .0246%43 * Q2 - .001984 * Q3 + MGO_LERR * BETA (2.31607) (.6918% ) (055266) CRSQ = .949 S.E.R. = .087 MEAN LHS = .352 RANGE: 1973 Q2 TO 1985 Q2 DW = 2.182 NOB = 49 ESTIMATED: 9/87 LAG LAGCOEF1 T-STAT (-0) -0.198 3.326 (-1) 0.016 0.589 (-2) 0.104 2.380 (-3) 0.068 2.589 (-4) -0.093 1.664 SUM: ~0.102 59. MGO: (MGO) GERMAN IMPORTS FROM OPEC --REAL $ MGO = MGOV/OPOIL72 60. MGTV: (MGTV) TOTAL IMPORTS OF GERMNAY --NOMINAL $ MGTV = MGCV + MGEV + MGJV + MGUV + MGIV + MGLV + MGOV + MGZV 61. XGTV: (XGTV) TOTAL GERMAN EXPORTS -- NOMINAL $ XGTV = XGCV + XGEV + XGJV + XGUV + XGIV + XGLV + XGOV + XGZV

62. GTRADE: (GTRADE) GERMAN TRADE BALANCE

GTRADE = XGTV - MGTV

-13-

63. XGCV: (XGCV) GERMAN EXPORTS TO CANADA --NOMINAL $

LOG! XGCV } = -.103284 + .978974 * LOG(MCGV) + .005263 * Ql - -076495 * Q2 (6.87515) (125.174) (. 258594) (3.7707) - .038044 * Q3 + XGC_LERR * BETA (1.85576) , CRSQ = .993 S.E.R. = .073999 MEAN LHS = -.709 RANGE: 1960 Ql TO 1986 Q2

DW = 1.997 NOB = 106 ESTIMATED: 6/87

64%. XGEV: (XGEV) GERMAN EXPORTS TO THE UK --NOMINAL $

LOG(XGEV ) = -,018881 + 1.00673 * LOG(MEGV) - .048344 * Ql - .031613 * Q2 (1.3918) (204.69) (2. 76064 ) (1.80527) - .008397 * Q3 + XGE_ERR * BETA (47493) CRSQ = .998 S.E.R. = .06¢ MEAN LHS = 1.018 RANGE: 1960 Ql TO 1986 Q2

OW = 1.321 NOB = 106 ESTIMATED: 6/87

65. XGJV:; (XGJV) GERMAN EXPORTS TO JAPAN --NOMINAL $

LOG(XGJV) = - .064936 + .988247 * LOG(MJGV) - .063165 * Ql - .087822 * Q2 (3.81611) (121.179) (2.66369) (3.70361) - .060361 * Q3 + XGJ_ERR * BETA (2.52155) cCRSQ = .993 S.E.R. = .086 MEAN LHS = -.348 RANGE: 1960 Ql TO 1986 Q2

W = 2.086 NOB = 106 ESTIMATED: 6/87

66. XGUV: (XGUV) GERMAN EXPORTS TO THE US --NOMINAL $

LCG( XGUV) = + .009884 + .99695 * LOG(MUGV) - .085561 * Ql - .083864 * Q2 ( .602979) (153.441) (4.68036) (4.58803) - .060274 * Q3 + XGU_LERR * BETA (3.2656) CRSQ = .9% S.E.R. = -067 MEAN LHS = 1.449 RANGE: 1960 Ql TO 1986 Q2

OW = 2.398 NOB = 106 ESTIMATED: 6/87

67. GFPXFTW: (GFPXFTW) GERMAN IMPORT PRICES FROM MCM COUNTRIES

GFPXFTW = UPXGUV#O.396552 * (CPXGUV * CEI 1340.155172 *% (EPXGUV # EEI) ##0.206897 * (JPXGUV * JEI #0 .241379

68. GEPCOMP: (GEPCOMP) THIRD COUNTRY PRICE FOR UK-GERMANY TRADE

GEPCOMP = (UPXGUV#EO.396552 * (CPXGUV * CEI )##0.155172 * ( JPXGUV * JEI) ¥*0. 241379 HC 1/01 - 0.20689))

69. GJPCOMP: (GJPCOMP } THIRD COUNTRY PRICE FOR GERMAN-JAPAN TRADE

GJPCOMP = (UPXGUVI460.396552 % (CPXGUV * CEI )##0.155172 % (EPXGUV * EET) #0. 206897 HHL 1/01 - 0.2414))

- 14 -

70. GUPCS1P: (GUPCOMP) THIRD COUNTRY PRICE FOR GERMAN-US TRADE

GUPCOMP = ((CPXGUV * CEI )0.155172 * (EPXGUV ® EEI )#*0.206897 % (JPXGUV * JEI )##0. 241379 eH 1/01 - 0.3965))

71. MJEV: (MJEV) IMPORTS OF JAPAN FROM THE UK --NOMINAL $

LOG(MJEV/(EEI * JAIME_EXUVI)) = -3.92249 + .313387 * LOG( JGNPPOT)

(1.74985) (1.76135)

+ 2.38657 * (LOG(JGNP) - LOG(JGNPPOT)) - .28811 * LOG(EEI * JAIME_EXUVI/( JPGNP * JEI)) (1.43452) (1.51905)

* 618213 * LOG(MJEV{ -1)/( EEI(-1) * JAIME_EXUVI(-1))) - .054207 * Ql - .024¢031 * Q2 (5.31699) (. 939302) (415304)

- .052443 * Q3 + MJE_ERR * BETA . (912559)

CRSQ = .608 S.E.R. = .14 MEAN LHS = -3.793 RANGE: 1973 Q2 TO 1985 Qv

OW = 1.457 NOB

49 ESTIMATED: 9/87

72. MJE: (MED IMPORTS OF JAPAN FROM THE UK ~7REAL $

MJE = MJEV/(EEI * ( JAIME_EXUVI/100))

73. MJGV: (MJGV) IMPORTS OF JAPAN FROM GERMANY --NOMINAL $

LOG(MJGV/(GEI * GPXGUV )) = -2.14599 + .19837 % LOG( JGNPPOT ) (1.9651) (2.19285) + 1.89118 * (LOG(JGNP) - LOG( JGNPPOT )) (2.40883) + LAGCOEF1 * (LOG(GEI(I) * GPXGUV(T )/(JPGNP(I) * JEI(I)))) + .649263 * LOG(MUGV{-1)/(GEI(-1) * GPXGUV(-1))) - .038706 * Ql - 036911 * Q2 (7.05225) (1.38154) (1.32486 ) - .089864 * Q3 + MJG_LERR * BETA (3.22673) CRSQ = .88 S.E.R. = .067 MEAN LHS = -.081 RANGE: 1973 Q2 TO 1985 Qi:

DW = 2.003 NOB = 49 ESTIMATED: 9/87

LAG LAGCOEF1 T-STAT

(-0) -0.066 2.968

(-1) -0.106 2.968

(-2) -0.119 2.968

(-3) -0.106 2.968

(-4) -0.066 2.968 SUM: -0.464¢

74. MJG: (MJG) IMPORTS OF JAPAN FROM GERMANY --REAL $

MJG = MJGV/(GEI * GPXGUV)

- 15 -

75. MJCV: (MJCV) IMPORTS OF JAPAN FROM CANADA --NOMINAL $

LOS(MJCV/(CEI * CPXGUV)) = -2.19721 + .228746 * LOG( JGNPPOT ) (2.23104) * (2.62815) + 1.7275 *% (LOG(JGNP) - LOG(JGNPPOT )) (1.97329) + LAGCOEF1 * (LOG(CEI(I) % CPXGUVII)/( UPGNPII) *% JEI(I})}) + .325243 * LOG(MJCV(-1)/(CEI(-1) * CPXGUV(-1))) - .005472 * Ql + .132925 * Q2 (2.5323) (.188195) (4.19208) + .0¢4¢850 * Q3 + MUC_LERR * BETA (. 335654) RSQ = .699 S.E.R. = .069999 MEAN 5a = .607 RANGE: 1973 Q2 TO 1985 Q@2

i tH = 1.982 NOB = 49 ESTIMATED: 9/8

LAG LAGCOEF1 1-STAT

(-0} -0.035 2.545 C-15 -0.056 2.545 {-2) -0.063 2.545 (-3) . -0.056 2.545 (-¢) -0.035 2.545 ‘SUM - -0.243

76. MJC: (MJC} IMPORTS OF JAPAN FROM CANADA --REAL $

MJD = MJCV/(CEI * CPXGUV)

77. MJUV: (MJUV) IMPORTS OF JAPAN FROM THE US --NOMINAL $

LOI3( MJUV/UPXGUV } = -1.66279 + .194961 % LOG( JGNPPOT) + .969227 * {LOGI JGNP) - LOG( JGNPPOT)) (1.93876) i 2.35053) (1.42002) + LAGCOEF1] * (LOG(UPXGUV(T)1/( JPGNP(I) * JEICID))) + 92523 , % LOGI MJUV( -1)/UPXGUVI -1))} (9.0395 - .011751 * Ql + .014920 * Q2 - .041722 * Q3 + MJU_ERR * BETA t .522819) (.659253) (1.84746) CRSQ = .879 S.E.R. = .055 MEAN LHS = 2.179 RANGE: 1973 Q@2 TO 1985 Q2

DW = 2.422 NOB = 49 ESTIMATED: 9/87

AG LAGCOEF1 T-STAT

(--0) -0.026 2.247 (-1) -0.041 2.247 (--2) -0.047 2.247 (-3) -0.041 2.247 (-4) -0.026 2.247

uM: 4-0.181

78. MJU: imu RTS OF JAPAN FROM THE US --REAL $

- 16 -

79. MJLV: (MJLV) IMPORTS OF JAPAN FROM LDCS --NOMINAL $

LOGUMJLV/LPXGUV ) = -5.15718 + .531167 * LOG(JGNPPOT) + .496952 * (LOGIUGNP) - LOG( JGNPPOT }) (3.47703) (3.70138) ( .861566)

+ LAGCOEF1 * (LOG(LPXGUV(I)/( UPGNP(I) ® JEI(I)))) + LAGCOEF2 * (LOG( JFPXFTW(I)/( JPGNP(I) * JEI(I)})) + .562773 * LOG(MJLV( -1/LPXGUV( -1))

(5.65233) - .104462 * Ql + .028120 * Q2 - .032288 * Q3 + MUL ERR * BETA (5.57018) (1.38978) (1.74015) ‘us@ = ,989 S.E, oR. = .045 tAn LHS = 2.242 RANGE: 1973 Q2 TO 1985 Q2 fas = 2.472 NOB = 49 ESTIMAIED: 9/87 i AG LAGCOEF1 1-STAT LAGCOQEF2 T-STAT (-C> -0.077 2.869 0.95] 1.847 (-1)5 -0.122 2.869 0.082 1.847 (-2) -0.139 2.869 0.092 1.847 (-23 -0.123 2.869 & O62 1.847 (-4) -0.077 2.869 U.051 1.847 SUM: -0.539 0.357 80. MJL: (MJL) IMPORTS OF J#PAN FROM LDCS --REAL $ MJL = MJLV/LPXGUV 81. MJ1V: (MJIV) IMPORTS OF JAPAN FROM CTHER OECD--NOMINAL $ LUG( MJIV/IPXGUY ) = 1.12533 - .0:459. © LOG(JGNPPOT) + 1.11123 *® (LOG(UGNP! - LOG( JGNPPCT 3) (.769003) (.1+4049) (1.6485) - 1.35703 * LOG(IPXGUV/( JPGNP * JEI1) ¢ 1.09435 x LOG( JFPXFTW/( JPGNP * JEI}} (3.9367) (3.21041) + =.526664 * LOG(MJI].!-1)/IPXGUv(-i)) - .083946 * Q1 + .003028 * Q2 (5.40256) (3.731853 (.128699) - .024250 * Q@3 + MJI_ERR * BETA (1.08585) CRSQ = .905 S.E.R. = .055 MEAN LHS = 1.67 RANGE: 1973 Q2 TO 1985 Q2 DOW = 1.9% NOB = 49 ESTIMATED: 9/87

82. MJI: (MJI} IMPORTS OF JAPAN FROM OTHER OECD --REAL $

- i7 -

83. MJOV: (MJOV) IMPORTS OF JAPAN FROM OPEC --NOMINAL $

LOGIMJOV/OPOIL72) = 2.41979 - .136844 * LOG(JGNPPOT) + .406278 * (LOG(JGNP) - LOGI JGNP (1.47289) (1.0589) (5402535) LOGI JGNPPOT }) + LAGCOEF1 * (LOG(OPOIL72(I)/( JPGNP(I) ® JEI(I)))) # .607906 ® LOG(MJOV(-1)/0POIL72(-1)) (7.12001) - .043208 * Ql - .13764 % @2 - .086689 * Q3 + MJO_ERR * BETA (1.303) (5.8749) (2.6974) CRSQ = .812 S.E.R. = .059 MEAN LHS = 1.412 RANGE: 1974 Q@1 To 1985 @2 RHO = -.367 (3.2) NOB = G5 ESTIMATED: 9/87 LAG - LAGCOEF1 = T-STAT (-0) -0.013 3.759 (-1) -0.022 3.841 (-2) -0.027 3.949 (-3) -0.027 4.072 (-«) -0.022 4.067 (-5) -0.013 2.778 (-6) -0.000 0.018 (-7) 0.018 1.466 SUM: -0.108 84. MJO: (MJO) IMPORTS OF JAPAN FROM OPEC --REAL $ MJO = MJOV/OPOIL72 85. MJTV: (MJTV) TOTAL IMPORTS FROM JAPAN --NOMINAL $ MJTV = MJCV + MJEV + MUGV + MUUV + MUIV + MULV + MUOV + MUZV 86. XJTV: (XJTV) TOTAL EXPORTS OF JAPAN -- NOMINAL $ . XJTV = XUCV + XJEV + XUGV + XUUV # XJIV + XULV + XUOV + KUZy 87. JTRADE: (JTRADE) JAPANESE TRADE BALANCE JTRADE = XJTV - MUTV 88. XJCV: (XJCV) EXPORTS OF JAPAN TO CANADA -- NOMINAL $ SF RR aD eS a em 9a XJCV = .911927 % MCUJV + XJUC_ERR * BETA (120.091) Biv

CrsQ = .993 S.E.R. = .177 MEAN LHS = 1.497 RANGE: 1960 Q1 TO 1986 Q2 OW = 2.016 NOB = 106 ESTIMATED: 6/87

89. XJEV: (XJEV) EXPORTS OF JAPAN TO THE UK -- NOMINAL $

LOG XJEV) = .150297 + .958288 * LOG(MEJV) - .139785 * Ql - .155209 * Qz (3.90476) (69.3745) (2.59877) sy (288538) - - .067166 * Q3 + XJE_ERR * BETA ~~ (1.23708) ont CRSQ = .979 S.E.R. = .196 MEAN LHS = -.123 RANGE: 1960 Q1 TO 1986 Q2

DW = 1.744 NOB = 106 ESTIMATED: 6/87

- 18 -

90. XJGV: (XJGV) EXPORTS OF JAPAN TO GERMANY -- NOMINAL $

LOG(XJGV) = - .021660 + 1.02411 * LOGIMGJV) - .016214 * Ql - .071905 » Q2 (1.19732) (174.631) (639458 ) (2.83707) + .032698 * Q3 + XJG_ERR * BETA (1.2778) CRSQ = .997 S.E.R. = .091999 MEAN LHS = -.071 RANGE: 1960 Q1 TO 1986 Qi:

OW = 2.303 NOB = 106 ESTIMATED: 6/87

91. XJUV: (XJUV) EXPORTS OF JAPAN TO THE US -- NOMINAL $

LOG( XJUV ) = -.012868 + 1.00321 * LOG(MUJV) - .103151 * Ql - .036702 * Q2 (.926522) (249.519) (6.99702) (2.49016) - - .037370 * Q3 + XJU_ERR * BETA (2.51188)

CRSQ = .998 S.E.R. = .054 MEAN LHS = 2.192 RANGE: 1960 Ql TO 1986 Q2 DW = 1.989 NOB = 106 ESTIMATED: 6/87

92. JFPXFTW: (JFPXFTW) JAPANESE IMPORT PRICES FROM MCM COUNTRIES JFPXFTW == UPXGUV##0.353866 * (CPXGUV % CEI )¥#0.138461 * (EPXGUV * EEI) ##0.206897 % (GPXGUV * GEI )¥*#0.323077

93. MUEV: (MUEV) IMPORTS OF THE US FROM THE UK --NOMINAL $

LOG(MUEV/(EEI * JAIME_EXUVI)) = -8.55781 +¢ 1.13412 * LOG(UGNPPOT) (3.94959) (3.75187)

+ 1.59694 % (LOG(UGNP) - LOG(UGNPPOT ))

(2.30187)

+ LAGCOEF] * (LOG(EEI(I) * JAIME_EXUVI(I) * UTARIFF(I )/UPGNP(I)))

+ .575253 * LOG(MUEV( -1)/(EEI(-1) % JAIME_EXUVI(-1))) - .000026 * Ql + .099526 * Q2 (5.06864) (000662) (2.46605)

+ .059237 * Q3 + MUE_ERR * BETA (1.50014)

CRSQ = .904 S.E.R. = .091 MEAN LHS = -2.065 RANGE: 1973 Ql TO 1985 Q2 DW = 2.365 NOB = 50 ESTIMATED: 9/87

LAG LAGCOEF1 T-STAT

(-0) -0.991 3.438 (-1) 0.592 1.394 (-2) -0.035 0.078 (-3) 0.355 0.792 (-4) -0.406 0.903 (-5) -0.399 0.884 (-6) 0.735 2.176 SUM: -0.148

94. MUE: (MUE) IMPORTS GF THE US FROM THE UK --REAL $

MUE = MUEV/(EEI * (JAIME_EXUVI/100))

-19-

95. MUGV: (MUGV) IMPORTS OF THE US FROM GERMANY --NOMINAL $

LOG( MUGV/(GEI * GPXGUV)) = -12.6817 + 1.93026 * LOG(UGNPPOT ) (4.81247) (4.86119)

+ 2.75425 % (LOG(UGNP) - LOG(UGNPPOT )) (3.51151) :

+ LAGCOEF] * (LOG(GEI(I) * GPXGUV(I) % UTARIFF(I)/UPGNP(I))) + LAGCOEF2 * (LOG(UGPCOMP(I) * UTARIFF(I )/UPGNP(I))) ¢ .330933 * LOG(MUGV( -1)/(GEI(-1) * GPXGUV(-1))) + .004¢717 * Ql + .029175 * Q2

(2.20151) (.144067) (.893636) - 047988 * Q3 + MUG_ERR * BETA

(1.40746) CRSQ = .932 S.E.R. = .076 MEAN LHS = 1.57 RANGE: 1973 Ql TO 1985 Q2 DW = 2.139 NOB = 50 ESTIMATED: 9/87

LAG LAGCOEF1° T-STAT LAGCOEF2 T-STAT

(-0) -0.096 2.640 -0.018 0.037 (-1) -0.164¢ | 2.640 0.219 0.344 (-2) -0.205 2.640 0.483 0.743 (-3) -0.219 2.640 0.806 1.217 (-4) -0.205 2.640 -0.868 1.325 (-5) -0.164 2.640 -0.363 - 0.555 (-6) -0.096 2.640 0.260 0.536 SUM: -1.150 0.519

96. MUG: (19UG) IMPORTS OF THE US FROM GERMANY--REAL $

MUG = MUGV/IGEI * GPXGUV)

97. MUJV: (MUJV) IMPORTS OF THE US FROM JAPAN --NOMINAL $

mw em meee ee een Cee ee wees ee wwe oe we eee wer ww eoooowe

LOG(MIJV/(JEI % JPXGUV))" = -10.3444 + 1.59426 % LOG(UGNPPOT ) (3.94826) (4.00106)

+ 1.16886 * (LOGCUGNP) - LOG(UGNPPOT )) (2.12963)

+ LAGCOEF] ® (LOG(JEI(I) ® JPXGUV(I) % UTARIFF{I )/UPGNP(I))) - .038725 % LOG(UJPCOMP % UTARIFF/UPGNP) + .549535 * LOG(MUJV(-1)/( JEI(-1) % JPXGUVI-1)))

(.195528) (4.81053) 4+ .027926 ® Ql + .049231 * Q2 + .034968 * Q3 + MUJ_ERR * BETA ( . 958706) (1.6624) (1.17831)

CRSQ = .973 S.E.R. = .071 MEAN LHS = 2.62 RANGE: 1973 Q1 TO 1985 Q2 CW = 1.519 NOB = 50 ESTIMATED: 9/87

LAG LAGCOEF1 T-STAT

(-0) -1.0346 2.982 (-1) 1.092 2.281 (-2) -0.574 1.748 SUM: -0.516

98. MUJ: (MU) IMPORTS OF THE US FROM JAPAN ~-REAL $.

MUJ = MUJV/(JEI * JPXGUV)

- 20 -

99. MUCV: (MUCV) IMPORTS OF THE US FROM CANADA. --NOMINAL $

LOG( MUCV/(CEI % CPXGUV)) = -10.2932 + 1.86032 % LOG(UGNPPOT ) (4.81929) (6.36058)

+ 2.7251 % (LOG(UGNP) - LOG(UGNPPOT ))

(3.88447) + LAGCOEF] * (LOG(CEI(I) * CPXGUV(I) # UTARIFFCI)/UPGNP(I))) - .039549 * (a1 (2.65507) + .022661 * Q2 - .098199 * Q@3 + MUC_ERR * BETA (1.31712) (6.45849) CRSQ = .948 S.E.R. = .05 MEAN LHS = 2.945 RANGE: 1972 Q@4 TO 1985 Q2 RHO = .608 (4.8) NOB = 50 ESTIMATED: 9/87

LAG LAGCOEF1 T-STAT

(-0) -0.999 4.137 (-1) 0.053 0.310 (-2) 0.325 1.849 (-3) -0.185 0.767

:

100. MUC: (MUC) IMPORTS OF THE US FROM CANADA --REAL $

muc = MUCV/(CEI * CPXGUV)

101. MULV: (MULV) IMPORTS OF THE US FROM LDCS --NOMINAL $

mw we www wen mew eww mee ewe eee ore e we eee eer eeoeeooesooes

. LOG( MULV/LPXG6UV ) = -9.06081 + 1.43891 % LOGCUGNPPOT) ¢* .086317 % (LOG(UGNP) - LOGI UGNPPOT)) (3.73126) (3.7317) " (.234983)

+ LAGCOEF] * (LOG(LPXGUV(I) * UTARIFF(T)/UPGNP(I))) + 782883 |” LOG( MULV( -1 )/LPXGUV(-1))

+ .012781 * Ql + .005020 * Q@2 + .025215 * Q3 + MUL_ERR * BETA (607366) (. 240889) (1.17741)

CRSQ = .977 S.E.R. = .052 MEAN LHS = 3.01 RANGE: 1973 Q1 TO 1985 Q2 DW = 1.869 NOB = 50 ESTIMATED: 9/87

LAG LAGCOEF1 T-STAT

(-0) -0.018 2.380 (-1) -0.032 2.380 (-2) -0.039 2.380 (-3) -0.042 2.380 (-4@) -0.039 2.380 (-5) -0.032 2.380 (-6) -0.018 2.380 SUM: -0.221

102. MUL: (MUL) IMPORTS OF THE US FROM LDCS --REAL $

MUL = MULV/LPXGUV

- 21 -

103. MUIV: (MUIV) IMPORTS OF THE US FROM OTHER OECD--NOMINAL $

1.0G( MUIV/IPXGUV ) = -11.0066 + 1.78039 * LOG(UGNPPOT) + 1.47671 * (LOG(UGNP) - LOG(UGNPPOT)) (5.91649) (6.07863) (3.05965) + LAGCOEF1 * (LOGCIPXGUV(I) * UTARIFF(I)}/UPGNP(I))) + .288866 * LOG(MUIV(-1)/IPXGUV(-1)) (2.37048) - .010306 * Ql + .007628 * Q2 - .02224¢3 * Q3 + MUI_ERR * BETA (.409347) (.302318) (.864078) CRSQ = .954@ S.E.R. = .063 MEAN LHS = 2.431 RANGE: 1973 Ql TO 1985 Q2

OW = 2.222 NOB = 50 ESTIMATED: 9/87

LAG LAGCOEF1 T-STAT

(-0) -0.355 3.608 (-1) -0.179 4.490 (-2) -0.051 1.668 (-3) 0.030 0.596 (-4) 0.064 1.093 (-5) 0.051 0.995 (-6) -0.010 0.278 (-7) -0.117 2.364 (-8) ~0.272 2.519 SUM: -0.839

104. MUI: (MUI) IMPORTS OF THE US FROM OTHER OECD--REAL $

muI = MUIV/IPXGUV

105. MJOV: (MUOV) IMPORTS OF THE US FROM OPEC --NOMINAL $

|.0G( MUOV/OPOIL72 ) = 1.47359 + 2.26283 *% (LOG(UGNP) - LOG(UGNPPOT )) (3.36447) (1.50846 )

+ LAGCOEF1 * (LOG(OPOIL72(1) * UTARIFF(I)/UPGNP(I))) + .633566 % LOG(MUOV( -1)/0POIL72( -1))

(6.16408) - .024352 * Ql - .029716 * Q2 + .042953 * Q3 + MUO_ERR * BETA (.381529) (.471041) (.667227) CRSQ = .928 S.E.R. = .14¢3 MEAN LHS = 1.256 RANGE: 1976 Q2 TO 1986 Q2

OW = 2.063 NOB = 41 ESTIMATED: 9/87

LAG LAGCOEF1 T-STAT

(-0) -0.008 2.330 (-1) -0.016 2.330 (-2) ~0.022 2.330 (-3) -0.028 2.330 (-4) -0.032 2.330 (-5) -0.036 2.330 (-6) -0.038 2.330 (-7) -0.040 2.330 (-8) -0.040 2.330 (-9) -0.040 2.330 (-10} ~0.038 2.330 (-11) -0.036 2.330 (-12) -0.032 2.330 (-13) -0.028 2.330 (-14¢) -0.022 2.330 (-15) -0.016 2.330 (-16) -0.008 2.330 SUM: -0.482

106. MJO: (MUO) IMPORTS OF THE US FROM OPEC --REAL $

UO = MUOV/OPOIL72

- 22 -

107. MUTV: (MUTV) TOTAL IMPORTS OF THE US --NOMINAL $

MUTV = MUCV + MUEV + MUGV + MUJV + MUIV + MULV + MUOV + MUZV

108. XUTV: (XUTV) TOTAL EXPORTS OF THE US --NOMINAL $

XUTV = XUCV + XUEV + XUGV + XUUV + XUIV + XULV + XUOV + XUZY

109. UTRADE: (UTRADE) US TRADE BALANCE

UTRADE = XUTV - MUTV

110. XUCV: (XUCV) EXPORTS OF THE US TO CANADA --NOMINAL $

LOG( XUCV ) = -.026796 + .950034 * LOG(MCUV) - .013906 * Ql + .008312 * Q2 (2.00323 ) (240.165) (1.35367) ( .809043 ) + .000543 * Q3 + XUC_LERR * BETA (.05235) CRSQ = .998 S.E.R. = .037 MEAN LHS = 2.637 RANGE: 1960 Ql TO 1986 Q2

DH = 1.279 NOB = 106 ESTIMATED: 6/87

111. XUEV: (XUEV) EXPORTS OF THE US TO THE UK --NOMINAL $

LOG( XUEV ) = =.1049%4 + 1.03366 * LOG(MEUV) - .059005 * Ql - .084906 * Q2 (5.28696) (114.479) (2.84959) (4.10051) - .046155 * Q3 + XUE_ERR * BETA (2.20582) CRSQ = .992 S.E.R. = .075 MEAN LHS = 1.334 RANGE: 1960 Q1 TO 1986 G2

DH = 1.823 NOB = 106 ESTIMATED: 6/87

112. XUGV: (XUGV) EXPORTS OF THE US TO GERMANY --NOMINAL $

LOG( XUGV ) = =-.364822 + 1.07299 * LOG(MGUV) + .015915 * Ql - .011594 * @2 (10.9192) (69.8863) (484136) (352782 | + .016254 * Q3 + XUG_ERR * BETA ( .489067) CRSQ = .979 S.E.R. = .12 MEAN LHS = 1.257 RANGE: 1960 Ql TO 1986 Qi: OW = .992 NOB = 106 ESTIMATED: 6/87

113. XUJV: (XUJV) EXPORTS OF THE US TO JAPAN --NOMINAL $

LOG{ XUJV ) = -.143099 + 1.01282 * LOG(MJUV) - .041308 * Ql - .092451 * Q@2 (8.80429) (176.45) (2.57314) (5.75836) - .043979 * Q3 + XUJ_ERR * BETA (2.71351) CRSQ = .997 S.E. =R. = .058 MEAN LHS = 1.84 RANGE: 1960 Ql TO 1986 Q2

DW = 1.745 NOB = 106 ESTIMATED: 6/87

114. UGPCOMP: (UGPCOMP) THIRD COUNTRY PRICE FOR US-GERMAN TRADE

UGPCOMP = ((CPXGUV * CEI )#0.165 * (EPXGUV * EEI)**0.215 * (JPXGUV * JET) 380.245 ee 1/(1 - 0.375))

- 23 -

115. UJFCOMP: (UJPCOMP) THIRD COUNTRY PRICE FOR US-JAPAN TRADE

U.IPCOMP = ((CPXGUV *® CEI )#*0.165 % (EPXGUV * EEI )**0.215 *% (GPXGUV * GEI) #0375 eC I/01 - 0.245))

116. UCFCOMP: (UCPCOMP) THIRD COUNTRY PRICE FOR US-CANADA TRADE

UCPCOMP = ((EPXGUV * EEI )**0.215 % (GPXGUV * GEI )##0.375 % (JPXGUV * JEI) 380. 2465 Jee( 1/01 - 0.165))

117. UEFCOMP: (UEPCOMP) THIRD COUNTRY PRICE FOR US-UK TRADE

UEPCOMP = ((CPXGUV % CEI )¥#0.165 % (GPXGUV * GEI)#0.375 ™ (JPXGUV * JEI) 0.245 )e*(1/01 - 0.215))

118. UFPXFTW: (UFPXFTW) U.S. IMPORT PRICES FROM MCM COUNTRIES

UFPXFTW = (CPXGUV * CEI )%#0.165 * (EPXGUV * EEI )**0.215 % (GPXGUV * GEI) *#0.375 % (JPXGUV % JET 0.245

119. XEIV: (XEIV) EXPORTS OF THE UK TO OTHER OECD --NOMINAL $

LOG(XEIV/CEEI * JAIME_EXUVI )) = -6.2303 + 1.53892 * IGNPTRD (4.47206) (4.91071)

+ 1.04727 * (LOGCROWIPEEC) - IGNPTRD) - .320252 * LOG(EEI * JAIME_EXUVI/IPXGUV) (3.842) (2.34379)

- .432223 *% LOG(UEPCOMP/IPXGUV) + .418515 * LOG(XEIV(-1)/(EEI(-1) * JAIME_EXUVI(-1))) (1.39455) (3.45643)

- .082212 * Ql - .065332 * Q2 - .144812 * Q3 + XEI_ERR * BETA (3.57901) (3.1521) (6.76881)

CRSQ = .912 S$.E.R. = .046 MEAN LHS = -.666 RANGE: 1973 Q1 TO 1985 Q2

OW = 1.613 NOB = 50 ESTIMATED: 9/87

120. XE‘: (XEI) UK EXPORTS TO OTHER OECD -- REAL $

XEI = XEIV/(EEI * ( JAIME_EXUVI/100 ))

121. XGIV: (XGIV) EXPORTS OF GERMANY TO OTHER OECD --NOMINAL $

LIGG(XGIV/(GEI %* GPXGUV)) = -3.081 + 1.24233 % IGNPTRD + .952816 ® (LOG{ROWIPEEC) - IGNPTRD) (5.46947) (6.70395) (5.72815)

- .182959 * LOG(GEI * GPXGUV/IPXGUV) + .099950 * LOG{UGPCOMP/IPXGUV ) (1.64056) (.876437)

+ .307926 * LOG(XGIV(-1)/(GEI(-1) * GPXGUV(-1))) - .107172 * 07511 - .041585 * Ql (3.22427) (3.98517) (3.26878)

- .048786 * Q2 - .114721 * Q3 + XGI_LERR * BETA (4.18055) (9.96108)

CRSQ = .97 S.E.R. = .023 MEAN LHS = 3.631 RANGE: 1973 Ql TO 1985 Q2

OW = 1.531 NOB = 50 ESTIMATED: 9/87

- 24 -

122. XGI: (XGI) GERMAN EXPORTS TO OTHER OECD -- REAL $

XGI = XGIV/(GEI * GPXGUV)

123. XJIV: (XJIV) EXPORTS OF JAPAN TO OTHER OECD --NOMINAL $

LOG(XJIV/( JET * JPXGUV)) = -1.904¢8 +¢ .518775 * IGNPTRD + .403429 * (LOG(ROWIPEEC} - IGNPTRD) (1.08535) (1.2739) (1.24807) - .1914¢27 * LOG(JEI * JPXGUV/IPXGUV) + .673539 * LOG( UJPCOMP/IPXGUV ) (1.07154 } (.897134) + .761778 * LOG(XJIV(-1)/(JEI(-1) % JPXGUV(-1))) - .131784 * Ql - .025751 * Q2 (7.52895) (4.58949) (878861) + .013819 * Q3 + XJI_ERR * BETA ( .470808 ) CRSQ = .932 = .07 MEAN LHS = 1.763 RANGE: 1973 Q1 TO 1985 Q2

E

S.E.R. DW = 1.57 NOB = 50 ESTIMATED: 9/87

124. XJI: (XJI) JAPANESE EXPORTS TO OTHER OECD -- REAL $

XJI = XJIV/(JEI * JPXGUV)

125. XCIV: (XCIV) EXPORTS OF CANADA TO OTHER OECD --NOMINAL $

LOG(XCIV/(CEI * CPXGUV)) = -6.99262 + 1.60257 * IGNPTRD (4.47515) (4.60902)

+ 1.75712 * (LOG(ROWIPEEC) - IGNPTRD) - .4%28261 * LOG(CEI * CPXGUV/IPXGUV ) (4.22915) (2.43061)

- .82332 * LOG(UCPCOMP/IPXGUV) + .534988 * LOG(XCIV(-1)/(CEI(-1) * CPXGUV(-1))) (1.58756 ) (5.6131)

- .123899 * Ql - .030853 * Q2 - .118089 * Q@3 + XCI_ERR * BETA (3.64719) (. 943586) (3.4564)

CRSQ = .747 S.E.R. = .082 MEAN LHS = .565 RANGE: 1973 Q1 TO 1985 Q2

DW = 2.341 NOB = 50 ESTIMATED: 9/87

126. XCI: (XCI) CANADIAN EXPORTS TO OTHER OECD -- REAL $

xCI = XCIV/(CEI * CPXGUV)

127. XUIV: (XUIV) EXPORTS OF THE UNITED STATES TO OTHER OECD --NOMINAL $

LOG( XUIV/UPXGUV ) = -3.69356 + 1.11382 * IGNPTRO + .976349 * (LOG(ROWIPEEC:) - IGNPTRD) (3.27099) (3.84355) (3.38676) - .348051 * LOG(UPXGUV/IPXGUV) + .133837 * LOG( UFPXFTW/IPXGUV ) (2.46866 ) (.324734) + .518745 * LOG(XUIV(-1)/UPXGUV(-1)) - .042157 * Ql - .077695 *® Q2 (4.75385) (1.70143) (3.03437) - .172988 * Q3 + XUI_LERR * BETA (6.84305 ) CRSQ = .874 S.E.R. = .054 MEAN LHS = 2.713 RANGE: 1973 Q1 TO 1985 Q2°

DW = 2.252 NOB = 50 ESTIMATED: 9/87

128. XUI: (XUI) US EXPORTS TO OTHER OECD -- REAL $

XUI = = =XUIV/UPXGUV

- 25 -

129. MILV: (MILV) IMPORTS OF OTHER OECD FROM LDCS --NOMINAL $

LOG(MILV/LPXGUV) = -6.93328 + 2.064996 * IGNPTRD + 1.26825 * (LOG(ROWIPEEC) - IGNPTRD) (6.61021) (6.90546) (5.75356) - .479371 * LOG(LPXGUV/IPXGUV) + .210052 * LOG(MILV(-1)/LPXGUV(-1)) - .030655 * Ql (4.54299) (1.84452) (2.11164) - 001593 * @2 - .051996 # QS + MIL_ERR © BETA >

(1116783 (3.8709) : =: CRSQ = .967 S.E.R. = 083 LMS = 2.935 RANGE: Se a wie DN = 1.959" NOB 2°S0 ESTIMATED. “5787 130. MIL: (MIL) IMPORTS OF OTHER OECD FROM LDCS --REAL $ MIL = MILV/LPXGUV 131. MIOV: (MIOV) IMPORTS OF OTHER OECD FROM OPEC --NOMINAL $ LOG(MIOV/OPOIL72) = 1.41341 - .087714 % IGNPTRD + 1.31013 * (LOG(ROWIPEEC) - IGNPTRD) (1129999) (1375051) (4103431) + LAGCOEFL * (LOGLOPOTL72(T)/TPMGUV(T))) + 853212 % LOGIMEOV( -1)/0POIL72(-1)) (7.14 - .423281 * D7341 - .508747 * 07411 - .092216 * Ql - 095554 * Q2z (5.99002) (7706979) (3.48076) (3.74225) - .072872 * Q3 + MIO_ERR * BETA (2.80877)

CRSQ = .961 S.E.R. = .062 MEAN LHS = 1.798 RANGE: 1973 Q1 TO 1985 Q2 OW = 1.842 NOB = 50 ESTIMATED: 9/87 :

LAG LAGCOEF1 T-STAT

(-0) -0.018 2.093

(-1) -0.027 2.093

(-2) -0.027 2.093

(-3) -0.018 2.093 SUM: ~0.091

132. MIO: (MIO) IMPORTS OF OTHER OECD FROM OPEC --REAL $

MIO = MIOV/OPOIL72

133. XITV: (XITV) TOTAL EXPORTS OF OTHER OECD --NOMINAL $

MCTV + METV + MGTV + MJTV + MUTV + MLTV + MITV + MOTV + MZTV + XWTV_ERR - (XCTV + XETV + XGTV + XJTV + XUTV + XLTV + XOTV + XZTV)

XxITV

134. MITV: (MITV) TOTAL IMPORTS OF OTHER OECD --NOMINAL $

MITV = XCIV + XEIV + XGIV + XJIV + XUIV + (MILV + MIOV) * 0.7 + MIZV

135. ITRADE: (ITRADE) OTHER OECD TRADE BALANCE

ITRADE = XITV - MITV

- 26 -

136. MZTV: (MZTV) TOTAL IMPORTS OF RESIDUAL REGION ~- NOMINAL $

MZTV = XCZV + XEZV + XGZV + XJZV + XUZV * XIZV + XLZV + XOZV

137. XZTV: (XZTV) TOTAL EXPORTS OF RESIDUAL REGION -- NOMINAL $

XZTV = MCZV + MEZV + MGZV + MJZV + MUZV + MIZV + MLZV + MOZV

138. ZTRADE: (ZTRADE) REDIUAL REGION TRADE BALANCE

ZTRADE = X2TV - MZTV

139. XELV: (XELV) EXPORTS OF THE UK TO LDCS -- NOMINAL $

LOG(XELV/(EEI * JAIME_EXUVI)) = -.912372 + .053858 * LGNPTRD (2.89744) (1.66331)

“+ .203426 * (LOG(ROWIPLDC ) - LGNPTRD)

(2.00185)

+ LAGCOEF1] * (LOG(EEI(I-1) * JAIME_EXUVI(I-1)/LPXGUV(I-1)))

+ =.340447 * LOG(XELV(-1)/(EEI(-1) * JAIME_EXUVI(-1))) - .062280 * Ql - .016251 * Q2 (2.44527) (3.29063) (. 953048 )

- .066606 * Q3 + XEL_ERR * BETA (3.65228)

CRSQ = .298 S.E.R. = .04¢3 MEAN LHS = -1.588 RANGE: 1973 Q1 TO 1985 Q2

OW = 1.85 NOB = 50 ESTIMATED: 9/87

LAG LAGCOEF1 T-STAT

(-0) -0.021 1.359

(-1) -0.031 1.359

(-2) -0.031 1.359

(-3) -0.021 1.359 SUM: -0.103

140. XEL: (XEL) BRITISH EXPORTS TO LDCS ~--REAL $

XEL = XELV/(EEI * ( JAIME_EXUVI/100))

141. XGLV: (XGLV) EXPORTS OF GERMANY TO LDCS -- NOMINAL $

LOG( XGLV/(GEI * GPXGUV)) = .353214 + .093021 * LGNPTRD + .325177 * (LOG(ROWIPLDC) - LGNPTRD) (2.33933) (3.44421) (3.5108) - .450963 * LOGIGEI * GPXGUV/LPXGUV) + .133004 * LOG(UGPCOMP/LPXGUV ) (4.47377) (1.16576) + .686506 * LOG(XGLV(-1)/(GEI(-1) * GPXGUV(-1))) - .145604 * Ql - .061898 * Q2 (9.26714) (9.7978) (4.38306 ) - .100343 * Q@3 + XGL_ERR * BETA (6.99272) CRSQ = .902 S.E.R. = .035 MEAN LHS = 2.274 RANGE: 1973 Q1 TO 1985 Q2

DW = 2.066 NOB = 50 ESTIMATED: 9/87

142. XGL: (XGL) GERMAN EXPORTS TO LDCS --REAL $

XGL = XGLV/(GEI * GPXGUV)

- 27 -

143. XJLV: (XJLV) EXPORTS OF JAPAN TO LDCS -- NOMINAL $

LOG(XJLV/(JEI % JPXGUV)) = .064777 + .215446 ® LGNPTRD + .012468 * (LOG(ROWIPLDC) - LGNPTRD)

(.428661) (2.90715) (.120586 ) - .289614 *% LOG(JEI % JPXGUV/LPXGUV) - .218195 * LOG(UJPCOMP/LPXGUV } (2.48597) (1.28553) + Rett st ‘a LOG(XJLV( -1)/( JEI(-1) % JPXGUV(-1))) = Pet $4 # Ql - .033861 * Q2 (6.31528 8.49358) (1.80634) - .062862 ® Q3 + XJL_ERR * BETA (3.4825) CRSQ = .975 S.E.R. = .044 MEAN LHS = 2.759 RANGE: 1973 Ql TO 1985 Q2

OW = 2.079 NOB = 50 ESTIMATED: 9/87

144. XJL: (XJL) JAPANESE EXPORTS TO LDCS --REAL $

XJL = XJLV/(JEI * JPXGUV)

145. XCLV: (XCLV) EXPORTS OF CANADA TO LDCS -- NOMINAL $

LOG(XCLV/(CEI * CPXGUV)) = -.557672 + .211437 % LGNPTRD - .24507 * (LOG(ROWIPLDC) - LGNPTRD) (1.91874) (2.20551) ( .685626) - .5705 * LOG(CEI * CPXGUV/LPXGUV) + .2¢5006 * LOG(UCPCOMP/LPXGUV ) (1.91528) (834649) + .40726 * LOG(XCLV(-1)/(CEI(-1) % CPXGUV(-1))) - .168379 * Ql +* .073396 * Q2 (3.04021) (4.33395) (1.78804 ) - - 068665 * Q3 + XCL_ERR * BETA (1.69313) CRSQ@ = .826 S.E.R. = .095 MEAN LHS = .657 RANGE: 1973 @1 TO 1985 Q2

OW = 2.24 NOB = 50 ESTIMATED: 9/87

146. XCL: (XCL) CANADIAN EXPORTS TO LDCS --REAL $

xCL = XCLV/ICEI * CPXGUV) 147. XULV: (XULV) EXPORTS OF THE US TO LDCS -- NOMINAL $

LOG(XULV/UPXGUV) = .108578 + .094242 % LGNPTRD + .111764@ # (LOG(ROWIPLOC) - LGNPTRO) (.581961) (2.25373) (653252)

- .265489 ® LOG(UPXGUV/LPXGUV) + .065942 % LOG( UFPXFTW/LPXGUV ) (1.52394) (.364921)

+ .826457 * LOG(XULV(-1)/UPXGUV(-1)) - .025748 * Ql + .036043 * Q2 (13.1891) (1.35526) (1.8917)

- .029191 * Q3 + XUL_ERR * BETA 41.5111)

CRSQ = .995 S.E.R. = .047 MEAN LHS = 3.099 RANGE: 1973 Q1 TO 1985 Q2 OW = 1.65 NOB = 50 ESTIMATED: 9/87

148. XUL: (XUL) US EXPORTS To = ~TREAL $

- 2 -

149. MLIV: (MLIV) IMPORTS OF LDCS FROM OTHER OECD --NOMINAL $

we wwe eee eee ee ee wn ew ee eee we ee eee ere eer e ror ooece

LOG( MLIV/IPXGUV ) = .526459 ¢ .181271 % LGNPTRD +* .239779 ® (LOG(ROWIPLD) - LGNPTRD) (3.20928) (3.49623) (2.76362) - .26218 * LOG(IPXGUV/LPXGUV) ¢ .61077 * LOG(MLIV( -1)/IPXGUVI -1)) - | .172366 * Ql (2.39845) (5.77809) (9.22855) - .065035 * Q@2 - .14¢9828 * Q3 + MLILERR * BETA (4.48147) (9.46812)

CRSQ = .956 S.E.R. = .036 MEAN LHS = 3.224 RANGE: 1973 Ql TO 1985 Q2 DW = 2.362 NOB = 50 ESTIMATED: 9/87

150. MLI: (MLI) IMPORTS OF LDCS FROM OTHER OECD -- REAL $

MLI = MLIV/IPXGUV

151. MLOV: (MLOV) IMPORTS OF LDCS FROM OPEC --NOMINAL $

LOG( MLOV/OPOIL72 ) = .581092 + .045322 % LENPTRD + .44051 * (LOG(ROWIPLDC) - LGNPTRD) (2.22493) (596514) (1.95766): - .055587 * LOG(OPOIL72/LPXGUV) + .527734 * LOG(MLOV(-1)/0POIL72(-1)) - .580985 * 073461 (1.04201) (5.08365) (6.70051) + .083587 ® Ql + .005453 * Q@2 - .001622 * Q3 + MLO_LERR * BETA (2.44314) (.162164) (04848) crse@e = .757 S.E.R. = .08 AN LHS = 1.572 RANGE: 1973 Q1 TO 1984 @&

OW = 2.022 NOB = 48 ESTIMATED: 9/87

152. MLO: (MLO) IMPORTS OF LDCS FROM OPEC -- REAL $

MLO = MLOV/OPOIL72

153. MLTV: (MLTV) TOTAL IMPORTS OF LOCS -- NOMINAL $

MLTV = XCLV + XELV + XGLV + XJLV ¢-XULV + MLOV ¢ MLIV + MLZV

154. XLTV: (XLTV) TOTAL EXPORTS OF LOCS -- NOMINAL $

XLTV = MCLV + MELV + MGLV + MJLV + MULV + MILV ¢ MOLV ¢ XLZY, |

155. LTRADE: (LTRADE ) LDCS TRADE BALANCE

LTRADE = XLTV - MLTV

- 29 -

156. XEOQV: (XEOV) EXPORTS OF THE UK TO OPEC -- NOMINAL $

|.OGCXEOV/(EEI * JAIME_EXUVI )) = -1.01452 + LAGCOEF1 * (LOG(XOTV(I )/OPOIL72(1I))) (4.41635) + LAGCOEF2 * (LOG(EEI(I) * JAIME_EXUVI(T )/OPOIL72(1))) + .734025 * LOG(XEOV(-1)/(EEI(-1) % JAIME_EXUVI(-1))) - .354002 * 07911 - .065990 * Ql (14.7932) (5.21432) (2.40111) + .009074¢ * Q2 - .087810 * Q3 + XEQO_ERR * BETA (.347076) (3.36736) CRSQ = .903 S.E.R. = .061 MEAN LHS = -2.187 RANGE: 1974 Q2 TO 1984 Q4

OW = 2.168 NOB = 43 ESTIMATED:, 9/87

LAG LAGCOEF1 T-STAT LAGCOEF2 T-STAT

(-0) 0.061 2.721 -0.044 2.143 (-1) 0.091 2.721 -0.066 2.143 (-2) 0.091 2.721 -0.066 2.143 (-3) 0.061 2.721 -0.044 2.143 SUM: 0.304 -0.219

157. XEQO: (XEO} BRITISH EXPORTS TO OPEC --REAL $

XEO = XEOV/(EEI * ( JAIME_EXUVI/100))

158. XGOV: {(XGOV) EXPORTS OF GERMNAY TO OPEC -- NOMINAL $

|.OG(XGOV/(GEI * GPXGUV)) = (arin? + LAGCOEF1] * (LOG(XOTV(TI )/OPOIL72(1))) + LAGCOEF2 * (LOG(GEI(I) * GPXGUV(T )/OPOIL72(1))) + .741196 *% LOG(XGOV(-1)/(GEI(-1) *% GPXGUV(-1))) - .150017 * Ql - .034825 * Q2 (13.3009) (4.92903) (1.05602) - .030280 * Q3 + XGO_ERR * BETA ( . 924343)

cRsSQ = .899 S.E.R. = '.077 MEAN LHS = 1.565 RANGE: 1974 Q@2 TO 1984 Q4 DW = 1.709 NOB = 43 ESTIMATED: 9/87

LAG LAGCOEF1l T-STAT LAGCOEF2 T-STAT

(-0) 0.081 3.155 -0.061 3.409

(-1) 0.121 3.155 -0.092 3.409

(-2) 0.121 3.155 -0.092 3.409

(-3) 0.081 3.155 -0.061 3.409 SUM: 0.403 -0.307

159. X30: (X60) GERMAN EXPORTS TO OPEC --REAL $

“GO = XGOV/(GEI * GPXGUV)

- 30 -

160. XJOV: (XJOV) EXPORTS OF JAPAN TO OPEC -- NOMINAL $

LOG(CXJOV/(JEI * JPXGUV)) = -.431843 +¢ LAGCOEF] * (LOG(XOTV(I )/OPOIL72(1))) . (1.20389) + LAGCOEF2 * (LOG(JEI(I) * JPXGUVIT )/OPOIL72(1I))) + 631049 ® LOG(XJOV(-1)/( JEI(-1) * JPXGUV(-1))) - .172144 * Ql + .000018 * Q2 (10.2873) (5.73285) (.000636) - .054669 * Q3 + XJO_ERR * BETA (1.8942) CRSQ = .% S.E.R. = .068 MEAN LHS = 1.9% RANGE: 1974 Q2 TO 198 Q4 OW = 1.642 NOB = 43 ESTIMATED: 9/87

LAG LAGCOEF1 T-STAT LAGCOEF2 T-STAT

(-0) 0.048 2.517 -0.067 3.850 (-1) 0.072 2.517 -0.101 3.850

(-2) 0.072 2.517 -0.101 3.850

(-3) 0.048 2.517 -0.067 3.850 SUM: 0.240 -0.337

161. XJO: (XJO)} JAPANESE EXPORTS TO OPEC --REAL $

XJO = XJOV/(JEI * JPXGUV)

162. xXCOV: (XCOV) EXPORTS OF CANADA TO OPEC -- NOMINAL $

LOG(XCOV/(CEI * CPXGUV)) = Ppa + LAGCOEF] *® (LOG(XOTV(I )/OPOIL72(1))) -

+ LAGCOEF2 *® (LOG(CEI(I) %* CPXGUV(T )/OPOIL72(I))) ; + .516627 * LOG(XCOV( -1)/(CEI(-1) * CPXGUV(-1))) - .160765 * Ql - .020222 * Q2

(4.37939) (2.4056) (.310185 ), - .060261 * Q3 + XCO_ERR * BETA

(. 934627) CRSQ = .739 S.E.R. = .151 MEAN LHS = -.472 RANGE: 1974 Q2 TO 1984 Q@&

OW = 2.167 NOB = 43 ESTIMATED: 9/87

LAG LAGCOEF1 T-STAT LAGCOEF2 T-STAT

(-0) 0.089 2.233 -0.100 2.624 (-1) 0.134 2.233 -0.151 2.624 (-2) 0.134 2.233 -0.151 2.624 (-3) 0.089 2.233 -0.100 2.624 SUM:

0.445 -0.502

163. XCO: (XCO) CANADIAN EXPORTS TO OPEC --REAL $

xcO = XCOV/(CEI * CPXGUV)

-31-

164. XUOV: (XUOV) EXPORTS OF THE US TO OPEC -- NOMINAL $

LOG( XUOV/UPXGUV) == eet tt ey + LAGCOEF1 * (LOG(XOTV(TI)/OPOIL72(1)))

+ LAGCOEF2 * (LOG(UPXGUV(I)/OPOIL72(1))}) + in eee % LOG( XUOV( -1 )/UPXGUV( -1)) - 0686 )

- .104742 ® Ql + .018150 * Q2 - .062178 * Q3 + XUO_ERR * BETA (3.80917) (.678492) (2.33542)

CRSQ = .888 S.E.R. = .062 MEAN LHS = 1.944 RANGE: 1974 Q2 TO 198 @4 DW = 1.61 NOB = 43 ESTIMATED: 9/87

LAG LAGCOEF1 T-STAT LAGCOEF2 T-STAT

(-0) 0.050 3.187 -0.027 2.169 (-1) 0.075 3.187 -0.041 2.169 (-2) 0.075 3.187 -0.041 2.169 (-3) 0.050 3.187 ~0.027 2.169 SUM: 0.251 -0.137

165. XUO: (XU0) US EXPORTS TO OPEC --REAL $

166. MOIV: (MOIV) EXPORTS OF OTHER OECD TO OPEC -- NOMINAL $

LOG(MOIV/IPXGUV} = -.530978 + LAGCOEF] * (LOG(XOTVI(I )/OPOIL72(I)))

(1.48289) + LAGCOEF2 *® (LOG(IPXGUV(I)/OPOIL72(I})} + .730433 *% LOG(MOIV( -1)/IPXGUVI-1)) (13.2055) - .184372 # Ql - .05329%% * Q@2 - .15796 * Q3 + MOI_ERR * BETA (6.88797) (2.10902) (6.22609)

CRSQ = .958 S.E.R. = .059 MEAN LHS = 2.495 RANGE: 1974 Q2 TO 1964 Q4 DN = 2.612 NOB = 43 ESTIMATED: 9/87

LAG LAGCOEF1 T-STAT LAGCOEF2 T-STAT

(-0) oO. 057 2.883 0.064 3.511 (-1) 0.086 2.883 -0.0% 3.511 (-2) 0.086 2.883 -0.0% 3.511 (-3) 0.057 2.883 -0.064 3.511 SUM: 0.287 -0.318

167. MOI: (MOI) IMPORTS OF OPEC FROM OTHER OECD --REAL $

- 32 -

168. MOLV: (MOLV) EXPORTS OF LDCS TO OPEC -- NOMINAL $

LOG( MOLV/LPXGUV ) = -1.21074 + .464276 * LOG(XOTV/OPOIL72) (2.11486) (3.6311)

+ LAGCOEF] * (LOG(LPXGUV(I)/OPOIL72(1})) + .000799 * LOGI MOLV( -1)/LPXGUV(-1))

(005077)

+ .067447 ® Ql + .068482 * Q@2 + .028572 ™ Q3 + MOL_ERR * BETA (3.37502) (2.12902) (1.46659)

CRSQ = .968 S.E.R. = .06 MEAN LHS = 1.762 RANGE: 1974 Q1 TO 1984 @% RHO = .786 (6.7) NOB = 43 ESTIMATED: 9/87

LAG _ LAGCOEF1 T-STAT

(-0) -0.783 4.326

(-1) 0.049 0.303

(-2) -0.064 0.524

(-3) 0.014 0.121

(-4) -0.189 1.681

(-5) -0.125 1.130

(-6) -0.015 0.145

(-7) -0.137 1.463

SUM: -1.250

169. MOL: (MOL) IMPORTS OF OPEC FROM LDCS --REAL $

MOL = MOLV/LPXGUV

170. XOTV: (XOTV) TOTAL EXPORTS OF OPEC -- NOMINAL $

xOTV = (MCOV + MEOV + MGOV + MJOV + MUOV + MLOV + MIOV) # 0.8 + XOZV

171. MOTV: (MOTV) TOTAL IMPORTS OF OPEC -- NOMINAL $

MOTV = XCOV + XEOV + XGOV + XJOV + XUOV + (MOLV + MOIV)/CIFFOB + MOZV

172. OTRADE: (OTRADE) OPEC TRADE BALANCE

OTRADE = XOTV - MOTV

173. UNETX: (UNETX) US REAL TRADE BALANCE -- $

UNETX = MCU + MEU + MGU + MUU + XUI + XUL + XUO - MUC - MUE - MUG - MU - MUI - MUL

174. CNETX: (CNETX) CANADIAN REAL TRADE BALANCE -- $

CNETX = MUC + MEC + MGC + MJC + XCI + XCL + XCO

;

MCE - MCG - MCJ - MCI - MCL

175. ENETX: (CENETX) BRITISH REAL TRADE BALANCE -- $

ENETX = MUE + MCE + MGE + MJE + XEI + XEL + XEO

MEU - MEC - MEG - MEJ - MEI - MEL

176. GNETX: (GNETX) GERMAN REAL TRADE BALANCE -- $

GNETX = MUG + MCG + MEG + MUG + XGI + XGL + XGO - MGU - MGC - MGE - MGJ - MGI - MGL

- 33 - 177. JNETX: (JNETX) JAPANESE REAL TRADE BALANCE -- $

JNETX = MUS + MCJ + MEJ + MGJ + XJI + XJL + XJO - MUU - MUC - MUE - MUG - MUI - MUL - MUO

178. LNETX: (LNETX) LOCS REAL TRADE BALANCE -- $

LNETX = MCL + MEL + MGL + MJL + MUL + MIL + MOL - XCL - XEL ~ XGL - XJL - XUL - MLI - MLO

VARIABLE | EQUATION NUMBER

OT TC- s

BETA 1 3 5 40 40 41 75 75 77 2110111112 156 156 158

CEI 1 3 5 99 99 100

CER 13 15

CFPXFTW ll 14

CGNP 1 3 5

CGNPPOT 1 3 5

CIFFOB 171

CNETX 174

CPGNP 1 3 5

CPMGOL 13 15

CPXGOL 13

CPXGUV 23 24 «45 125 125 126

CTRADE 17

D7341 131 «151

D7411 50 54 56

D7511 121

D7731 4 52

07911 156

ECPCOMP 23 44

EEI 1 2 14 92 92 93

EGNP 23 25 27

EGNPPOT 23 25 27

ENE TX 175

EPGNP 23 25 27

EPXGUV 14 67 69

ETRADE 39

EUPCOMP 29 «45

GEI 3 4 14 96 96 115

GEPCOMP 68

GFPXFTW 54 «56 67

GGNP 4 48 50

GGNPPOT 46 48 50

GJPCOMP 50 69

GNETX 176

GPGNP 46 48 50

GPXGUV 3 4 14 141 141 142

GTRADE 62

GUPCOMP 52 70

IGNPTRD 119 121 123

IPXGUV ll 12 = 3 150 150 166

ITRADE 135

JAIME_EXUVI 1 2 4%

JEI 5 6 14 83 83 97

JEPXFIW 79 = 81 92

JGNP 71 73 75

JGNPPOT 71 73 75

JNETX 177

JPGNP 71 73 75

JPXGUV 5 6 14 123. 123 124

JTRADE 87

LGNPTRD 139 141 143

LNETX 178

LPxXGuUV 9 10 31 151 151 168

LTRADE 155

MCE 2 174 #175

MCE_ERR 1

MCEV 1 2 16

MCG 4 174 176

MCG_ERR 3

MCGV 3 4 16

MCI 12 174

MCI_ERR ll

MCIV ll 12 16

MCJ 6 174 177

MCJ_ERR 5

MCUV 5 6 16

MCL 10 174 178

MCL_ERR 9

MCLV 9 10 16

McO 15 174

MCO_ERR 13

MCOV 13 15 16

MCTV 16 17° «(133

CROSS REFERENCE LIST OF

48 145

131

23

29

29 70

25 116

52 52 52 158

125 167 47

77 77

77 27 143 145

32 169

63

154

170

49 146

25 114

31

31 92

26 117

114 79 79 79

144

147

27

33 33

33 114

4 118

FR gy

160 149 55

13 123 166

13 118

13 13

68 163

29 116 5 35 35 115

45 121

45

131 81 93

117

83 83

8&3 161 151

79

- 34 -

19 20 50 852 89 90 125 127 168

23 24 125 126

69 70 31 = 33 118 #119

116 (118

%6 48 122 141

73 74°

82 103

% #119 50 118 #123

50 51

80 101

21 54

91

129

48 145

75

35 120

50 142

92 104 120

67 124

67

102

22

93 131

49

146

76

45 139

52 158

95

119

139 143

129

23

95 139

67 162

92

% 140

96 let

140 70 144

70

130

VARIABLES AND EQUATIONS

25

63 141

68

163

99

47 156

115

123

156 160

97

139

27 64 99 143

69

100

67 157

116

125

157 161

141

29 65 101 145

70

114

69

73

117

127

73

114.

143

31 66 103 147

75

115

70

74

118

129

77

116

1465

33

105 149

76

117

71

92

121

131

79

117

147

35

110 151

92

118

72

95

122

149

81

118

149

VARIABLE | EQUATION NUMBER

U 8 173 174 MCU_ERR 7 McuV 7 8 16 110 MCZV 16 137 MEC 24 «174 «#2175 MEC_ERR 23 . MECV 19 23 24 37 MEG 26 175 176 MEG_ERR 25 MEGV 25 26 37 64 MET 34 «(175 MET_ERR 33 ' 33 34 37 MEJ 28 #175 177 MEJ_ERR 27 MEJV 27 28 37 89 MEL 32 #175 178 MEL_ERR 31 MELV 31 32 37 154 MEO 36 175 MEO_ERR 35 Ov 35 36 37 170 METV 37 39 133 MEU 30 173 175 MEU_ERR 29 29 30 37 #111 MEZV 37 137 : 49 174 176 MGC_ERR 48 20 48 49 60 47 175 176 MGE_ERR 46 MGEV 41 46 47 60 MG; 57 176 MGI_ERR 56 56 57 60 51 176 177 MGJ_ERR 50 Vv 50 51 60 90 55 176 178 MGL_iERR 5 5 5! 60 154 59 176 MGO_IERR 58 MGOV 58 59 60 170 MGTV 60 62 133 MGU 53 173 176 MGU_IERR 52 MGUV 52 53 60 112 MGZV 60 137 MIL 130 178 MIL_IERR 129 MILV 129 130 134 154 I 132 MIO_IERR 131 MIOV 131 132 134 170 MITV 133 134 135 MIZV 134 137 76 #174 #177 MJC_IERR 75 21 75 76 8s 72 #175 #177 MJE_IERR 71 42 71 72 85 74 =#+4176 #177 MJG_IERR 73 65 73 74 85 MJI 82 177 MJI_IRR 81 MJIV 81 82 85 MUL 80 177 178 MJL_ERR 79 79 80 8S 154 . & 177 MJO_ERR 83 MJOV 83 ro 85 170 MJTV 85 87 133 MUU 78 #4173 «#4177 MJU_ERR 77 MJUV 77 (78 & 113 MJZV 8S 137 MLI 150 178 MLI_ERR 149 Mi 149 150 153 MLO 152 178 MLO_ERR 151 151 152 153 170 MLTV 133 153 155 MLZV 137 153

VARIABLE

MOI_ERR MOIV MOL

MOL_ERR MOLV

M2T OPOIL72 OTRADE Ql

Q2 Q3

ROWIPEEC ROWI PLDC UCPCOMP UEPCOMP UFPXFTW UGNP

UGNPPOT UGPCOMP UJPCOMP UNETX UPGNP UPXGUV

UTARIFF UTRADE XCE_ERR XCEV XCG_ERR XCGV

| EQUATION NUMBER

167 178

19

20 174

125

21 174

145 174

162 18 22

136 40

41 175

119

171

169 172

174 100

175

94 176 107 177

98 178 107 107

133 138 58

126

178 146

163 133

120

171

107

107

107

107

154

170

99

141 143

99 29

99

134

153

171

134

101 101

101 30

101

103 %

103

105 105

105 52

105

- 36 -

106

53

67

132 21

131 21

131 21

93 131

68

22

139 22

139 22

139

69

152 23

141 23

141 23 58

141

77

78

158

27

101 145

27

101 145

27

101 145

160

29

103 147

29

103 147

29

103 147

162

31

105 149

31

105 149

31

105 149

164

33 110

151 .

33

110 151

33

110 151

166

35

111 156

35

111 156

35

111 156

92 127 128 147 148

VARIABLE | EQUATION NUMBER

XEJ_ERR XEJV XEL XEL_ERR XELV

XUL_ERR

xUO XUO_ERR xUOV

XUZV XWTV_ERR XZTV ZTRADE

42 175

139 175

156 39

43 136

63

64 176

121

65 176

141 176

158 62

66

136 135

89 90 177 123 177

143 177

160 87

91 136

170 110 111

112 173

127

113 173

147 173

164 109 136

‘137

178

140.

157 133

122 178

142

159 133

124 178 144

161 133

128

178 148

165 133

138

153

171

134

153

171

134

153

171

160

13¢

153

171

- 37 -

162 164 166 168 170 172

MNEMONIC {| EQUATION |

GNP EGNPPOT ENETX EPGNP EPXGUV ETRADE EUPCOMP GEI GEPCOMP GFPXFTW

GGNP GGNPPOT GJPCOMP

JAIME _EXUVI JEI JFPXFTW

xex ‘N +

8588

(EXOG )

- 38 - ALPHABETICAL LIST OF VARIABLES FOR MODEL

DEFINITION

US-CANADA EXCHANGE RATE (US$/C$)--INDEX 1972 US-CANADA EXCHANGE RATE (US$/C$)

CANADIAN FOREIGN IMPORT PRICES

CANADIAN REAL GNP--DOMESTIC CURRENCY--1972 CANADAIA POTENTIAL OUTPUT--DOMESTIC CURRENCY--1972 CIFFOB FACTOR

CANADIAN REAL TRADE BALANCE -- $

CANADIAN GNP DEFLATOR--DOMESTIC CURRENCY--1972 CANADA OIL IMPORT PRICE-DOMESTIC CURRENCY--1972 CANADA OIL EXPORT PRICE-DOMESTIC CURRENCY--1972 CANADIAN EXPORT UNIT VALUE--DOMESTIC CURRENCY--1972 CONAD TAN TRADE BALANCE

JUMMY

= 1 IN 1974Q1 DUMMY = 1 IN 1974Q1 DUMMY = 1 IN 1975Q1 OUMMY = 1 IN 1977Q3 OUMMY = 1 IN 19791

THIRD COUNTRY PRICE FOR UK-CANADA TRADE

US-UK EXCHANGE RATE (US$/POUND )--INDEX 1972 BRITISH REAL GNP--DOMESTIC CURRENCY--1972 BRITISH POTENTIAL OUTPUT--DOMESTIC CURRENCY--1972 BRITISH REAL TRADE BALANCE -- $

BRITISH GNP DEFLATOR--DOMESTIC CURRENCY--1972 BRITISH EXPORT UNIT VALUE--DOMESTIC CURRENCY--1972 UK TRADE BALANCE :

THIRD COUNTRY PRICE FOR UK-US TRADE

US-GERMANY EXCHANGE RATE (US$/DM)--INDEX 1972 THIRD COUNTRY PRICE FOR UK-GERMANY TRADE

GERMAN FOREIGN IMPORT PRICES

GERMAN REAL GNP--DOMESTIC CURRENCY--1972

GERMAN POTENTIAL OUTPUT--DOMESTIC CURRENCY--1972 THIRD COUNTRY PRICE FOR GERMAN-JAPAN TRADE

GERMAN REAL TRADE BALANCE -- $

GERMAN GNP DEFLATOR--OOMESTIC CURRENCY--1972 GERMAN EXPORT UNIT VALUE--DOMESTIC CURRENCY--1972 GERMAN TRADE BALANCE

THIRD COUNTRY PRICE FOR GERMAN-US TRADE

LOG OF OTHER OECD TREND OUTPUT--DOMESTIC CURRENCY--1972 OTHER OECD EXPORT UNIT VALUE--DOMESTIC CURRENCY--1972 OTHER O€CD TRADE BALANCE

NON-OIL UNIT VALUE FOR THE UK--1972=1

US-JAPAN EXCHANGE RATE (USS/YEN)--INDEX 1972 JAPANESE FOREIGN IMPORT PRICES

JAPANESE REAL GNP--DOMESTIC CURRENCY--1972 JAPANESE POTENTIAL OUTPUT--DOMESTIC CURRENCY--1972 JAPANESE REAL TRADE BALANCE -- $

JAPANESE GNP DEFLATOR--DOMESTIC CURRENCY--1972 JAPANESE EXPORT UNIT VALUE--DOMESTIC CURRENCY--1972 JAPANESE TRADE BALANCE

LOG OF LDCS TREND OUTPUT--DOMESTIC CURRENCY~-1972 LDCS REAL TRADE BALANCE -- $

LOCS EXPORT UNIT VALUE--DOMESTIC CURRENCY--1972 LOCS TRADE BALANCE

CANADIAN IMPORTS FROM UK 1972 PRICES

RESIDUAL IN MCEV EQUATION

CANADIAN IMPORTS FROM UK ($)

CANADIAN IMPORTS FROM GERMANY 1972 PRICES RESIDUAL IN MCGV EQUATION

CANADIAN IMPORTS FROM GER($)

CANADIAN IMPORTS FROM OECD 1972 PRICES

RESIDUAL IN MCIV EQUATION

CANADIAN IMPORTS FROM OECD ($)

CANADIAN IMPORTS FROM JAPAN 1972 PRICES

RESIDUAL IN MCJV EQUATION

CANADIAN IMPORTS FROM JAP($)

CANADIAN IMPORTS FROM LDC 1972 PRICES

RESIDUAL IN MCLV EQUATION

CANADIAN IMPORTS FROM LDC ($)

CANADIAN IMPORTS FROM OPEC 1972 PRICES

RESIDUAL IN MCOV EQUATION

CANADIAN IMPORTS FROM OPEC ($)

TOTAL CANADIAN IMPORTS ($)

CANADIAN IMPORTS FROM US 1972 PRICES

RESIDUAL IN MCUV EQUATION

CANADIAN IMPORTS FROM US ($)

CANADIAN IMPORTS FROM THE RESIDUAL REGION--NOMINAL $ UK _ IMPORTS FROM CANADA --REAL $

RESIDUAL IN MECV EQUATION

UK IMPORTS FROM CANADA--NOMINAL $

UK IMPORTS FROM GERMANY --REAL $

RESIDUAL IN MEGV EQUATION

UK IMPORTS FROM GERMANY--NOMINAL $

UK IMPORTS FROM OTHER OECD —-REAL $

RESIDUAL _IN MEIV EQUATION

UK IMPORTS FROM OTHER OECD--NOMINAL $

- 39 -

MNEMONIC | EQUATION | DEFINITION

MEJ 28 UK IMPORTS FROM JAPAN --REAL $ MEJ_ERR (EXOG) RESIDUAL IN MEJV EQUATION

ME JV 27 UK IMPORTS FROM JAPAN--NOMINAL $

MEL 32 UK IMPORTS FROM LDCS --REAL $ MEL_ERR’ (EXOG) RESIDUAL IN MELV EQUATION

MELV 31 UK IMPORTS FROM LOCS --NOMINAL $

MEO 36 UK IMPORTS FROM OPEC --REAL $ MEO_ERR (EXOG) RESIDUAL IN MEOV EQUATION

MEOV 35 UK IMPORTS FROM OPEC --NOMINAL $

METV 37 TOTAL IMPORTS OF THE UK --$

MEU 30 UK IMPORTS FROM THE US --REAL $ MEU_ERR (EXOG) RESIDUAL IN MEUV EQUATION

MEUV 29 UK IMPORTS FROM US--NOMINAL $

MEZV (EXOG) BRITISH IMPORTS FROM THE RESIDUAL REGION--NOMINAL $

MGC 49 GERMAN IMPORTS FROM CANADA --REAL $ MsC_ERR (EXOG) RESIDUAL IN MGCV EQUATION

MGCV 48 GERMAN IMPORTS FROM CANADA --NOMINAL $

MGE 47 GERMAN IMPORTS FROM THE UK --REAL $ MSE_ERR (EXOG) RESIDUAL IN MGEV EQUATION MGEV 46 GERMAN IMPORTS FROM THE UK --NOMINAL $ GERMAN IMPORTS FROM OTHER OECD e REAL $

MGI 57 MSI_ERR (EXOG) RESIDUAL IN MGIV EQUATION MGIV 56 GERMAN IMPORTS FROM OTHER OECD ~-NOMINAL $ MGJ 51 GERMAN IMPORTS FROM JAPAN --REAL $ MSJ_ERR (EXOG) RESIDUAL IN MGJV_ EQUATION MGJV 50 GERMAN IMPORTS FROM JAPAN --NOMINAL $ MGL 55 GERMAN IMPORTS FROM LOCS --REAL $ MSL_ERR (EXOG) RESIDUAL IN MGLV EQUATION MGLV 54 GERMAN IMPORTS FROM LDCS --NOMINAL $ MGO 59 GERMAN IMPORTS FROM OPEC --REAL $° MSO0_ERR (EXOG) RESIDUAL IN MGOV EQUATION MGOV 58 GERMAN IMPORTS FROM OPEC --NOMINAL $ MGTV 60 TOTAL IMPORTS OF GERMNAY --NOMINAL $ MGU 53 GERMAN IMPORTS FROM THE US --REAL $ MSU_ERR (EXOG) RESIDUAL IN MGUV EQUATION MGUV 52 GERMAN IMPORTS FROM THE US --NOMINAL $ MGZV_ (EXOG) GERMAN IMPORTS FROM THE RESIDUAL REGION--NOMINAL $ MIL 130 IMPORTS OF OTHER OECD FROM LDCS --REAL $ MIL_ERR (EXOG) RESIOUAL IN MILV EQUATION MILV 129 IMPORTS OF OTHER OECD FROM LDCS --NOMINAL $ MIO 132 IMPORTS OF OTHER OECD FROM OPEC --REAL $ MIO_ERR (EXOG) RESIDUAL IN MIOV EQUATION . MIOV 131 IMPORTS OF OTHER OECD FROM OPEC --NOMINAL $ MITV 134 TOTAL IMPORTS OF OTHER OECD --NOMINAL $ MIZV (EXOG} OTHER OECD IMPORTS FROM THE RESIDUAL REGION--NOMINAL $ MJC 76 IMPORTS OF JAPAN FROM CANADA --REAL $ MJC_ERR (EXOG) RESIDUAL IN MJCV EQUATION MJCV 75 IMPORTS OF JAPAN FROM CANADA --NOMINAL $ MJE 72 IMPORTS OF JAPAN FROM THE UK --REAL $ MJE_ERR (EXOG} RESIDUAL IN MJEV EQUATION MJEV 71 IMPORTS OF JAPAN FROM THE UK --NOMINAL $ MJG 74 IMPORTS OF JAPAN FROM GERMANY --REAL $ MJG_ERR (EX0G) RESIDUAL IN MJGV EQUATION KJGV 73 IMPORTS OF JAPAN FROM GERMANY --NOMINAL $ MJI 82 IMPORTS OF JAPAN FROM OTHER OECD --REAL $ MJI_ERR (EXOG) RESIDUAL IN MJIV EQUATION MJIV 81 IMPORTS OF JAPAN FROM OTHER OECD--NOMINAL $ MJL 80 IMPORTS OF JAPAN FROM LOCS --REAL $ MJL_ERR (EXOG) RESIDUAL IN MJLV EQUATION MJLV 79 IMPORTS OF JAPAN FROM LDCS --NOMINAL $ MJO & IMPORTS OF JAPAN FROM OPEC --REAL $ MJO_ERR (EXOG) RESIDUAL IN MJOV EQUATION MJOV 83 IMPORTS OF JAPAN FROM OPEC --NOMINAL $ MJITV 85 TOTAL IMPORTS FROM JAPAN --NOMINAL $ MJU 78 IMPORTS OF JAPAN FROM THE US --REAL $ MJU_ERR (EX0G) RESIDUAL IN MJUV EQUATION MJUV 77 IMPORTS OF JAPAN FROM THE US --NOMINAL MJZV (EXOG) JAPANESE IMPORTS FROM THE RESIDUAL REGION--NOMINAL $ MLI 150 IMPORTS OF LDCS FROM OTHER OECD -- REAL $ MLI_ERR (EXOG) RESIOUAL IN MLIV EQUATION MLIV 149 IMPORTS OF LOCS FROM OTHER OECD --NOMINAL $¢ MLO 152 IMPORTS OF LOCS FROM OPEC -- REAL $ MLO_ERR (EX0G) RESIDUAL IN MLOV EQUATION MLOV 151 IMPORTS OF LDCS FROM OPEC --NOMINAL $ MLTV 153 — TOTAL IMPORTS OF LDCS -- NOMINAL $ MLZV (EX0G) LOCS IMPORTS FROM THE RESIDUAL REGION--NOMINAL $ MoI 167 IMPORTS OF OPEC FROM OTHER OECD --REAL $ MOI_ERR (EX0G) RESIDUAL IN MOIV EQUATION MOIV 166 EXPORTS OF INDSUST TO OPEC -- NOMINAL $ MOL 169 IMPORTS OF OPEC FROM LDCS --REAL $ MOL_ERR (EX0G) RESIDUAL IN MOLV EQUATION MOLV 168 EXPORTS OF LDCS TO OPEC -- NOMINAL $ MOTV 171 TOTAL IMPORTS OF OPEC -~ NOMINAL $ MOZV (EXOG) OPEC IMPORTS FROM THE RESIOUAL REGION--NOMINAL $ MUC 100 IMPORTS OF THE US FROM CANADA --REAL $ MUC_LERR (EX0G) RESIDUAL IN MUCV EQUATION MUCV 99 IMPORTS OF THE US FROM CANADA --NOMINAL $

MUE 94 IMPORTS OF THE US FROM THE UK --REAL $

- 40 -

MNEMONIC | EQUATION | DEFINITION MUE_ERR (EXOG) RESIDUAL IN MUEV EQUATION MUEV 93 IMPORTS OF THE US FROM THE UK --NOMINAL $ MUG 96 IMPORTS OF THE US FROM GERMANY--REAL $ MUG_ERR (EXOG) RESIDUAL IN MUGV EQUATION MUGV 95 IMPORTS OF THE US FROM GERMANY - INAL MUI 104 IMPORTS OF THE US FROM OTHER O€CD--REAL 3 MUI_ERR (EX0G) RESIDUAL IN MUIV EQUATION MUIV 103 IMPORTS OF THE US FROM OTHER OECD--NOMINAL $ MUS 98 IMPORTS OF THE US FROM JAPAN --REAL $ MUJ_ERR = =( EXO0G} RESIDUAL IN MUJV EQUATION MUJV 97 IMPORTS OF THE US FROM JAPAN --NOMINAL $ MUL 102 IMPORTS OF THE US FROM LDCS --REAL $ MUL_ERR (EXOG) RESIDUAL IN MULV EQUATION MULV 101 IMPORTS OF THE US FROM LDCS --NOMINAL $ MUO 106 IMPORTS OF THE US FROM OPEC --REAL $ MUO_ERR (EXOG) RESIDUAL IN MUOV EQUATION MUOV 105 IMPORTS OF THE US FROM OPEC --NOMINAL $ MUTV 107 TOTAL IMPORTS OF THE US --NOMINAL $ MuUZV (EXOG) US IMPORTS FROM THE RESIDUAL REGION--NOMINAL $ MZTV 136 TOTAL IMPORTS OF RESIDUAL REGION -- NOMINAL $ OPOIL72 (EXOG) NOMINAL OIL PRICE INDEX FOR OPEC--1972=1 O7TRADE 172 OPEC TRADE SALANCE ; ; Ql (EXOG) DUMMY VARIABLE = 1 IN FIRST QUARTER - ZERO OTHERWISE Q2.(EXOG) DUMMY VARIABLE = 1 IN SECOND QUARTER - ZERO OTHERWISE Q3 ( EXO0G) DUMMY VARIABLE = 1 IN THIRD QUARTER - ZERO OTHERWISE ROWIPEEC (EXOG) OTHER OECD REAL GNP ROWIPLDC (EXOG) LOCS REAL GNP UCPCOMP 116 THIRD COUNTRY PRICE FOR US-CANADA TRADE UEPCOMP 117 THIRD COUNTRY PRICE FOR US-UK TRADE UFPXFTW 118 US FOREIGN IMPORT PRICES UGNP (EXOG) US REAL GNP--DOMESTIC CURRENCY--1972 UGNPPOT (EXOG) US POTENTIAL OUTPUT--DOMESTIC CURRENCY--1972 UGPCOMP 114 THIRD COUNTRY PRICE FOR US-GERMAN TRADE UJPCOMP 115 THIRD COUNTRY PRICE FOR US-JAPAN TRADE UNETX 173 US REAL TRADE BALANCE -- $ UPGNP (EXOG) US GNP DEFLATOR--DOMESTIC CURRENCY--1972 UPXGUV (EX0G) US EXPORT UNIT VALUE--DOMESTIC CURRENCY--1972 UTARIFF (EXOG) MULTILATERAL US TARIFF--INDEX 197221 UTRADE 109 US TRADE BALANCE XCE_ERR (EXOG) RESIDUAL IN XCEV EQUATION XCEV 19 CANADIAN EXPORTS TO THE UK ($) ACG_ERR (EXOG) RESIDUAL IN XCGV EQUATION XCGV 20 CANADIAN EXPORTS TO GERMANY ($) xCI 126 CANADIAN EXPORTS TO OTHER OECD -- REAL $ XCI_ERR (EXOG) RESIDUAL IN XCIV EQUATION XCIV 125 EXPORTS OF CANADA TO OTHER OECD --NOMINAL $ XCJ_ERR (EXOG) RESIDUAL IN XCJV EQUATION xCJV 21 CANADIAN EXPORTS TO JAPAN ($) xCL 146 CANADIAN EXPORTS TO LDCS --REAL $ XCL_ERR (EXOG) RESIDUAL IN XCLV EQUATION XCLV 145 EXPORTS OF CANADA TO LDCS -- NOMINAL xco 163 CANADIAN EXPORTS TO OPEC --REAL $ XCO_ERR (EXOG) RESIOUAL IN XCOV EQUATION xCOV 162 EXPCRTS OF CANADA TO OPEC -- NOMINAL $ xXCTV 18 CANADIAN EXPORTS TOTAL ($) XCU_LERR (EXOG) RESIDUAL IN XCUV EQUATION XCUV 22 CANADIAN EXPORTS TO THE US ($) xCZV (EXOG) CANADIAN EXPORTS TO THE RESIDUAL REGION--NOMINAL $ XEC_ERR (EXOG) RESIDUAL IN XECV EQUATION XECV 40 UK EXPORTS TO CANADA --NOMINAL $ XEG_ERR (EXOG) RESIDUAL _IN XEGV EQUATION XEGV 41 UK EXPORTS TO GERMANY --NOMINAL $ xEI 120 UK EXPORTS TO OTHER OECD -- REAL $ XEI_ERR (EXOG) RESIOQUAL IN XEIV EQUATION XEIV 119 EXPORTS OF THE UNITED KINGDOM TO OTHER OECD --NOMINAL $ XEJ_ERR = (EXOG) RESIDUAL _IN XEJV EQUATION XEJV 42 UK EXPORTS TO JAPAN --NOMINAL $_. XEL 140 BRITISH EXPORTS TO LDCS --REAL $ XEL_ERR (EXOG) RESIDUAL IN XELV EQUATION XELV 139 EXPORTS OF THE UK TO LOCS -- NOMINAL $ xEO 157 BRITISH EXPORTS TO OPEC --REAL $ XEO_ERR (EXOG) RESIDUAL IN XEOV EQUATION XEOV 156 EXPORTS OF THE UK TO OPEC -- NOMINAL $ XETV 38 TOTAL EXPORTS OF THE UK -~ NOMINAL $ XEU_ERR (EXOG) RESIDUAL _IN XEUV EQUATION XEUV 43 UK EXPORTS TO THE US --NOMINAL $ XEZV (EX0G) BRITISH EXPORTS TO THE RESIDUAL REGION--NOMINAL $ XGC_ERR (EXOG) RESIDUAL IN XGCV EQUATION XGCV * 63 GERMAN EXPORTS TO CANADA --NOMINAL $ XGE_ERR (EXOG) RESIDUAL IN XGEV EQUATION XGEV 64 GERMAN EXPORTS TO THE UK --NOMINAL $ xGI 122 GERMAN EXPORTS TO OTHER OECD -- REAL $ XGI_ERR (EXOG) RESIDUAL IN XGIV EQUATION XGIV 121 EXPORTS OF GERMANY TO OTHER OECD --NOMINAL $ XGJ_ERR (EXOG) RESIDUAL IN XGJV EQUATION XGJV 65 GERMAN EXPORTS TO JAPAN --NOMINAL $ XGL 142 GERMAN EXPORTS TO LOCS --REAL $

XGL_ERR (EXOG) RESIDUAL IN XGLV EQUATION

MNEMONIC | EQUATION |

xuI XUI_ERR XIV *XUJ_ERR muy XUL_ERR XULV

xuo XU0_ERR

YNTV_ERR XZTV ZTRADE

141

- 41 -

DEFINITION

EXPORTS OF GERMANY TO LDCS -- NOMINAL $ GERMAN EXPORTS TO OPEC --REAL $

RESIOUAL IN xXGOV

EQUATION

EXPORTS OF GERMNAY TO OPEC -- NOMINAL $ TOTAL GERMAN EXPORTS -- NOMINAL $

RESIDUAL IN XGUV

GERMAN EXPORTS TO THE US

EQUATION --NOMINAL

$ GERMAN EXPORTS TO THE RESIDUAL REGION--NOMINAL $

TOTAL EXPORTS OF

OTHER OECD --NOMINAL

$ OTHER OECD EXPORTS TO THE RESIDUAL REGION--NOMINAL $

RESIDUAL IN XJCV EXPORTS OF JAPAN RESIDUAL IN XJEV EXPORTS OF JAPAN RESIDUAL IN XJGV EXPORTS OF JAPAN JAPANESE EXPORTS RESIOUAL IN XJIV EXPORTS OF JAPAN JAPANESE EXPORTS RESIDUAL IN XJLV EXPORTS OF JAPAN JAPANESE EXPORTS RESIDUAL IN XJOV EXPORTS OF JAPAN TOTAL EXPORTS OF RESIDUAL IN XJUV EXPORTS OF JAPAN JAPANESE EXPORTS TOTAL EXPORTS OF

EQUATION

TO CANADA -- NOMINAL $ EQUATION

TO THE UK -- NOMINAL $ EQUATION

TO GERMANY -- NOMINAL $ TO OTHER OECD -- REAL $ EQUATION

TO OTHER OECD --NOMINAL $ TO LOCS --REAL $

TO THE US -- NOMINAL $ TO THE RESIDUAL REGION--NOMINAL $

Locs NAL $ LOCS EXPORTS TO. THE RESIOUAL RESTON--NOMINAL $

TOTAL EXPORTS O!

OPEC -- NOMINAL

OPEC EXPORTS TO THE RESTOUAL REGION--NOMINAL $

RESIDUAL IN XUCV EXPORTS

EQUATI

OF THE US TO CANADA --NOMINAL $

US EXPORTS TO OTHER OECD -- REAL $

RESIDUAL IN XUIV

EQUATION EXPORTS OF THE ra ore STATES TO OTHER OECD --NOMINAL $

RESIDUAL IN XUUV

EXPORTS OF THE US TO JAPAN --NOMINAL $ EXPORTS TO LOCS --REAL $

us RESIDUAL IN XULV

EQUATION

EXPORTS OF THE US TO LDCS -- NOMINAL $ US EXPORTS TO OPEC --REAL $

RESIDUAL IN XUOV

EQUATION

EXPORTS OF THE US TO opec -- NOTNAL 8

TOTAL EXPORTS OF

THE US --NOMINAL

US EXPORTS TO THE RESIDUAL REGION--NOMINAL $

RESIDUAL IN XITV TOTAL EXPORTS OF

EQUATION RESIDUAL RESTON -- NOMINAL $

RESIDUAL REGION TRADE BALANC

apaenbiew °9 Aarzzzer

zenbaey-yoeusS aoTuer zenbien ewyec

Auems ‘@'A'W'd Tseutyos ‘cf Azrey AeUuSTT °N,O 9FIg Aipuey ‘gq ppaeq uOoSSOTIY “Y [Ten

sodwey ey[tne

azadooy 1990g STATSH “IT WeTTTIM

ueuyTestds “4q ptaeqg

OTALSH “T WPTTTIM s[neg suuetg ‘g

zeuyog *§ 41aqoy Tleazazel *s Aazuay

UOSTTM ‘WY S@TIBYD JOUSTI “N,O OT39

uueW “JT euTrseyyze)

uosuyot “Fy 3z1EqQoy oyxsedoo] “gq atuuog

CS) aOHIAV

syueg “S$'n jo seyourig usqperz0g Ae sitsodaq uo sjueussessy soueinsul 3ysodeq

ase9 ueTenzeuen ay jo stsATeuy [eoTatduy uy :queudoj[eseq pue uot eIQUaDUND TeyTouPUTY

aZueyD 02 peMo{ty 21e SQUSTOFJJI0D usym STepow eqey aeZueyoxy jo aouPUAOJIeg ZutAseoei1oy aetdwes-jo-3ng ey]

yZnorzyl-sseg ejey eZueyoxy Jo [epop Vv seinpeso0ig ZuyZersay-eseug Jo [epoy enZoyeuy uy

stsA[Teuy Teotatdug uy “SQO86T 942 UT APOTFEM TPusIIAXY “S°N PuL

peuotzey st 3}peaD yueg uaym Auouooy ArquN0) padoTzeaeq ssatT ey Jo AIF }Fqeasul opTWOUODe0IOBYW

ONTBA aZuPyoXY saz IoeZ FW S,ZB[[°0q aya Jo seinseeW Jo Ausewssesseoy y

ZuTpue] yueg TeuoTAeua_AUL FO [epow uoTAe{TNUTS astduts Vv

SJjtazael yatm AjTodong [euopzeureqUuy:

awe aya JO satny ayl ZuyZueyD ‘:uoTQeT[e_Iey pue uot IIeR0I1g SdATISeYy [PUOCTIeUISQUI

1oj puewsag ay uo sasnoysuTiPeTy)

SperzAL TezAVAeTTATNW FO AoFFT PUL

ueder ut

sseo0ig quewqsn({py ey jo sqoadsy :aqey aZueyoxy AST [Od-uez eyQ Fo quoeuwustlesy

SHILIL

Saiodeg WoTssnostq Soueurq Teuotjeuiojuy

66¢

OO0€

TOE

co0E

COE

VOE

SOE

90€

LOE

80€

60€

OTE

TTe

da AWN dda

“TSSOZ

‘o'q ‘uoqZutysem ‘mweqsks aeartasey [erepe4

ay JO s1IoOUIeAOD Jo prze0og ‘HZ dojsg ‘aoueUTY TeUOTAJeUAeQU] Jo UOTSTATIG ‘siredeg

uoTSsnostq eoueUuTy TeuoTJeUraqUI 02 seTdod 10j sqasenber sseippe eseetg

eaUuTAe] ssoy

Azewg “J Aleqoy

piemoy “H praed LapuexeTV *S STMeyT

uuPW “J autirseyqe9 Jedooy 1030g

uossoTzg ‘WY [Ten

s[neg suuetg -g

12{A2IS STOT @TATH WETTTTM

aUTAS] Sssoy uopios) ‘g ptaeg

uou8ey “gq ydesor

uOoWIEH preyoTy

II] uewetog utor anq{TmM

zonbiew oawter

(S) ¥OHINV

seqey esueyoxg premi0og jo Butotag eu eutyg ‘uemtey ut Aotjtog Azeqauoy

Azjunog Zutdoteaeg e ut uozAezyT{Tqeas OTWoUC.e0IDBW pue seuTZay aqey eSuryoxy

setaqunog Zutdojeaeq peaqepuy ATyavey 1OJ Sanss] oTWou0Dg :suotsreauoy AqGeq

avousASTsSisg pue sesneg SI :Apyosseq Teuzeaxg “sn oul

seseyjoddky pe jsen-uonN pue pajsey Zutqjsay AoZ sotastzaezas Jo setjrzedorzg aetdues ejaTUuTY ey. pue AZoTopoyjep oT1eD equoW

seqeUTIsy seTes [}eIeYy 02 AseATTYy ueuTey aya jo uot jeot{ddy uy :eqeq [euoTsTaorg yo Adeanooy AsBoeIO™ ay. Butaordwy squnoooy suofqoersural

[euoTIeuISeAUI “S'n eYyQ UT saedTAras AayI0 puke sMOoUT JuaewASeAU]I But [epoy

Z86T

wM9TQOIg, IYZTTI Teatdeg ayy

sotmeukg epPAL TeuOCTIeUAeQUT pure sjsoD QueNqsn[py

T®POW HOUVO-WES eUL :AITOTASepeyso1zejey

TBUCTATpUuoD eATSSeIZe10Iny pezt{e1reuey YITm Tepoy suotjenby snosueqtnuts eul

Awouooq sourapy-ut-yse9 B UT [TeItTdep pue ‘Aserzaquy ‘Kauoy

ATOT Jaq epeAL sn eya jo stskTeuy pue uoTjIeWTISsY DTAAeWOU0D| usTei0gy JO SaTATOTASeTY eotAag pue aemoouy

Be6T SaTllL 7

siedeg UOTSSNoSTq SoueuTy TeuoTjeurejUuyT

:SMOTZ eperzl -

ete

eTe

VTE

STE

OTE

LTe

8Te

6Tt

O¢E

Ice

(ZAS

ECE

7CE

YAWN ddd