PREDICTING DEBT RESCHEDULING: A QUANTITATIVE APPROACH

• Document Type: CREST
• Collection: General CIA Records
• Document Number (FOIA) /ESDN (CREST): CIA-RDP86T00586R000300310007-7
• Release Decision: RIPPUB
• Original Classification: C
• Document Page Count: 33
• Document Creation Date: December 22, 2016
• Document Release Date: November 13, 2009
• Sequence Number: 7
• Publication Date: April 1, 1985
• Content Type: REPORT

Sanitized Copy Approved for Release 2009/11/13: CIA-RDP86T00586R000300310007-7
f Confidential
o
s n "R) Intelligence
Predicting Debt Rescheduling:
A Quantitative Approach
A Technical Intelligence Report
Confidential
GI 85-10114
DI 85-10020
April 1985
Copy 4 4 7
Sanitized Copy Approved for Release 2009/11/13: CIA-RDP86T00586R000300310007-7
 Sanitized Copy Approved for Release 2009/11/13: CIA-RDP86T00586R000300310007-7
Sanitized Copy Approved for Release 2009/11/13: CIA-RDP86T00586R000300310007-7
 Sanitized Copy Approved for Release 2009/11/13: CIA-RDP86T00586R000300310007-7
Directorate of
Intelligence
A Quantitative Approach
Predicting Debt Rescheduling:
A Technical Intelligence Report
Analytical Support 25X1
Group. Comments and queries are welcome and
may be directed to the Chief, Economics Division,
OGI, 25X1
This paper was prepared b=Office of 25X1
Confidential
GI 85-10114
DI 85-10020
Anil 1985
Sanitized Copy Approved for Release 2009/11/13: CIA-RDP86T00586R000300310007-7
 Sanitized Copy Approved for Release 2009/11/13: CIA-RDP86T00586R000300310007-7
Confidential
Summary
Information available
as of 17 April 1985
was used in this report.
iii Confidential
GI 85-10114
DI 85-10020
April 1985
Sanitized Copy Approved for Release 2009/11/13: CIA-RDP86T00586R000300310007-7
A Quantitative Approach
Predicting Debt Rescheduling:
intensive analysis of their specific situations.
In the last four years, about 40 countries have rescheduled their official or
private debts. Almost all the reschedulings occurred during economic or
political crises. In many instances, moreover, warnings preceded the actual
reschedulings. Although not supplanting well-reasoned, sound analysis of
debt situations, sophisticated quantitative examinations of linkages be-
tween economic conditions and debt reschedulings can be useful. In
particular, the multicountry nature of this approach allows a comprehen-
sive survey of a large number of countries and selection of those indicated
as potential trouble spots; those can then be followed up with more
sufficient data on suspected preindicators.
We have systematically explored the links between international and
domestic economic conditions and rescheduling for some 75 less developed
countries (LDCs), using logistic regression. This technique statistically
estimates the probability of a discrete event given trends in underlying
quantitative variables. In addition to its use in rescheduling analysis, it has
potential for predicting political events, such as coups or elections, using
appropriate preindicators. This technique is limited by the availability of
within a three-year period.
Our analysis indicates this technique does well in forecasting reschedul-
ings. In the 1977-83 period, for example, we found that applying logistic
regression to economic indicators, such as consumer price inflation, debt
and debt service, exports, imports, and reserves, correctly predicted in more
than three-fourths of the cases whether or not a country would reschedule
they may reschedule later in 1985.
Our analysis of 1983 economic trend data indicated that 50 countries
would reschedule in 1983, 1984, or 1985. Twenty-eight of these countries
have already rescheduled in 1983 or 1984. The remainder-Bangladesh,
Bolivia, Burma, Cameroon, Congo, Colombia, Egypt, Gabon, The Gambia,
Guyana, Ghana, India, Israel, Jordan, Kenya, Mauritania, Panama, South
Korea, Syria, Tanzania, Thailand, and Zimbabwe-were predicted to
reschedule by the technique in 1983-85 but have not done so as of yet. This
result could be due to an incorrect conclusion of the model or to the fact
 Sanitized Copy Approved for Release 2009/11/13: CIA-RDP86T00586R000300310007-7
Confidential
To attempt to delineate which of the two reasons seems more likely, we ob-
tained 1984 economic trend data for nine of the countries and applied the
technique again. As a result, we were able to refine somewhat the
rescheduling expectations for 1985 for these countries. We found that the
likelihood of rescheduling has gone up for Burma, and down for India,
Kenya, and Egypt. There was no real change for South Korea, Thailand,
Jordan, Israel, and Bolivia.
Confidential iv
Sanitized Copy Approved for Release 2009/11/13: CIA-RDP86T00586R000300310007-7
 Sanitized Copy Approved for Release 2009/11/13: CIA-RDP86T00586R000300310007-7
Confidential
Contents
Summary
iii
Choosing a Methodology
1
Evaluating the Models
2
Applying the Model
3
B. Data on Economic Indicators for Selected Foreign Countries,
13
1977-84
Sanitized Copy Approved for Release 2009/11/13: CIA-RDP86T00586R000300310007-7
 Sanitized Copy Approved for Release 2009/11/13: CIA-RDP86T00586R000300310007-7
Confidential
Predicting Debt Rescheduling:
A Quantitative Approach
Economic and financial theory suggest that a number
of preindicators exist for a country having to resched-
ule its debt. For example, low international reserves
relative to imports might indicate a need for resched-
uling. Similarly, rampant inflation often leads to
deteriorations in the domestic economy and in inter-
national payments balances requiring debt restructur-
ing
This study quantifies these relationships. Specifically,
we have tested three alternative statistical procedures
against a number of potential determinants. In two of
these-logistic regression and discriminant analysis-
we examined the linkage between economic indicator
trends in one year and whether the country resched-
uled in that or the succeeding two years. Thus, for
example, economic trends in 1982 were examined for
their linkages with rescheduling in 1982, 1983, or
1984.
A three-year period was chosen largely for practical
reasons. The lags in availability of international eco-
nomic indicator data prevent predicting for the next
year. Therefore, a three-year period is necessary to
have information soon enough to predict rescheduling.
For the third method-catastrophe theory-data re-
quirements prevented the use of the three-year period.
Hence, the results of this method are more interesting
than practical.
Choosing a Methodology
We developed our quantitative methodology for pre-
dicting rescheduling by examining three trial method-
ologies: logistic regression; linear discriminant analy-
sis; and a method based on mathematical catastrophe
theory (see appendix A):
the consumer price index. The technique fits an
equation to the observed cumulative probability of
the event. The fitted curve can then be used to
estimate probabilities.
? Discriminant analysis seeks to draw a divider be-
tween events. If the consumer price index were the
only predictor variable, discriminant analysis would
identify a value for the index. Countries with a price
index higher than this value would be predicted to
reschedule; countries with a lower index would not.
When more than one predictor variable is used,
discriminant analysis generates a dividing line,
plane, or higher dimensional linear shape. The
dividing shape is called the discriminant.
? Catastrophe theory is based on the notion of a
graphical relationship between one variable and
several other variables-in this case, a graph of
reschedulings versus the predictor variables. Be-
cause a country either reschedules or does not, the
graph will have a break between the two events,_
To determine the best methodology for predicting
reschedulings, we examined each model against a set
of indicator and rescheduling data for some 75 LDCs
for 1977 through 1983. For the indicator data, we
used economic trends commonly thought to influence
reschedulings: consumer price inflation, ratio of ex-
ports to imports, ratio of total international reserves to
imports, ratio of debt service to exports, ratio of
interest payments to exports, share of official debt in
total debt and debt service, and ratio of debt to
exports. The indicator data were obtained from IMF
publications and a CIA data base on debt and debt
service. Information on whether or not a country had
? Logistic regression uses statistics to estimate the
probability of an event, such as a rescheduling,
based on one or more predictor variables, such as
Sanitized Copy Approved for Release 2009/11/13: CIA-RDP86T00586R000300310007-7
 Sanitized Copy Approved for Release 2009/11/13: CIA-RDP86T00586R000300310007-7
Confidential
For testing the three models, the data were arranged
as follows:
? The information on rescheduling was transformed
into a binary variable equal to 0 if a country did not
reschedule in a given year or the two succeeding
years, and equal to 1 if it did.
? The economic indicator data were transformed into
ratios and percent changes as appropriate.
? All data on individual countries were pooled into a
single series for each variable.
As a result, the actual estimation procedures took
place against a series of 525 observations of pooled
data for 75 countries over seven years.
Each model was applied to the data and the percent-
age of successful classifications tallied. The results
were evaluated on two criteria:
? The percentage of correct classifications.
? The degree to which incorrect classifications were
evenly distributed between rescheduling and non-
rescheduling countries.
In evaluating the models it is necessary to establish a
trade-off between the two criteria. The problem can
be understood by a simple example. In the aggregate,
countries used roughly one-quarter of their opportuni-
ties to reschedule. Thus, a prediction that countries
never reschedule would be right nearly 75 percent of
the time. But none of the reschedulings would be
successfully predicted. This defeats the purpose of the
model
Ideally, the model should predict rescheduling and
nonrescheduling with equal accuracy. But achieving a
balance may reduce the model's overall performance.
Two of the methods-discriminant analysis and ca-
tastrophe theory-automatically set the trade-off be-
tween overall correctness and balance. Logistic re-
gression analysis requires the analyst to determine the
trade-off. Since the purpose is to discriminate between
countries that will reschedule and those that will not,
by about 2 percentage points.
an excess of errors in either direction is undesirable.
Therefore, a close balance was sought even though the
total number of correct classifications fell, typically
The differences in percentage of total correct classifi-
cations among the models are not large (table 1). In all
of the models not based on logistic regression, nonres-
chedulings are classified more accurately than resche-
dulings. This arises from the near certainty that a
country with favorable economic conditions will not
reschedule.
The discriminant analysis approach proved the most
successful in terms of overall correctness-but the
balance was quite poor. The catastrophe theory re-
sults also show an imbalance.
Logistic regression provided an acceptable level of
overall correctness (76.3 percent) and a good balance
(76.7 percent correct for reschedulings, 76.1 percent
for nonreschedulings). Consequently, we concluded
that the logistic regression model offered the best
approach for forecasting rescheduling.
Using the Model
In the logistic regression model chosen, the variables
included as predictors of rescheduling were: consumer
price inflation, the ratio of export earnings to imports,
the ratio of total reserves to imports, and the change
in the ratio of debt service to exports. This model was
used to calculate the probability of rescheduling for
the 75 countries examined in the study for 1977-85,
using economic indicator data for 1977-83.
To turn the probabilities of rescheduling into a predic-
tion as to whether or not a country will reschedule
within some period, a threshold probability had to be
established. Initially, one might guess that the cut-
point should be 50 percent; that is, if a country has a
Sanitized Copy Approved for Release 2009/11/13: CIA-RDP86T00586R000300310007-7
 Sanitized Copy Approved for Release 2009/11/13: CIA-RDP86T00586R000300310007-7
Confidential
Table 1
Comparison of Estimation Methods
75.9
76.7
75.0
74.1
76.0
76.0
greater chance of rescheduling than not, a reschedul-
ing should be predicted. However, a graph of the
percentage of correct classifications shows that the
optimal cutpoint is lower than one-half (figure 1 on
page 6). Ideally, the cutoff point should maximize the
percentage of correctly classified events, and make
equal the percentage of correctly classified reschedul-
ings and nonreschedulings. Figure 1 shows that the
cutoff point maximizing overall correct classifications
lies a little to the right of the point where the
"rescheduling" and "nonrescheduling" lines intersect.
Since the goal of the model is to predict whether
countries reschedule or not, a balance of correct
classifications between the two categories is impor-
tant. Assuming a country not to reschedule, for
example, would provide an accuracy of 76 percent,
but would not predict any reschedulings. As figure 1
demonstrates, a balance of correct classifications can
be obtained without a drastic decrease in the overall
percentage of correct classifications by setting the
cutoff point at 0.342. This choice sets the percentage
of correctly classified reschedulings to 76.7 percent;
nonreschedulings to 76.1 percent; and overall classifi-
cations to 76.3 percent
Although historical data must be used to estimate the
model, we used the model for predictions by:
1. Applying the 1983 values of the economic indicator
data for the four chosen variables to the model's
coefficient structure.
2. Choosing those countries predicted to reschedule in
1983-85 on the basis of that economic indicator data.
3. Comparing those predictions with those of the
countries that actually rescheduled in 1983-84.
4. Taking a closer look at some of the countries that
had not rescheduled, using estimates of 1984 data.
Using this procedure, the model predicted that 50 of
the 75 countries would reschedule in 1983-85, given
their 1983 economic indicators (table 2). Press and
government reports showed that only 28 of these 50
Sanitized Copy Approved for Release 2009/11/13: CIA-RDP86T00586R000300310007-7
 Sanitized Copy Approved for Release 2009/11/13: CIA-RDP86T00586R000300310007-7
Confidential
Table 2
Rescheduling Predictions for
1983, 1984, and 1985
Predicted to
Reschedule? a
Has
Rescheduled? b
Trinidad and Tobago
0.05
No
No
Central African Republic
0.06
No
Yes
Singapore
0.09
No
No
Hong Kong
0.10
No
No
Botswana
0.11
No
No
Malaysia
0.22
No
No
Papua New Guinea
0.22
No
No
Swaziland
0.23
No
No
Ethiopia
0.24
No
No
Sanitized Copy Approved for Release 2009/11/13: CIA-RDP86T00586R000300310007-7
 Sanitized Copy Approved for Release 2009/11/13: CIA-RDP86T00586R000300310007-7
Confidential
Table 2 (continued)
Predicted to Has
Reschedule? a Rescheduled? b
a In 1983, 1984, or 1985 from 1983 indicator data.
b In 1983 or 1984.
Sanitized Copy Approved for Release 2009/11/13: CIA-RDP86T00586R000300310007-7
 Confidentimi, Sanitized Copy Approved for Release 2009/11/13: CIA-RDP86T00586R000300310007-7
Figure 1
Effect of Cut-Off Point on Classifications
Non-
rescheduling
20 / \\
0 0 0.2 0.4 0.6 0.8 1
All
classifications
Rescheduling
countries had rescheduled by the end of 1984; the
remaining 22-Bangladesh, Bolivia, Burma, Camer-
oon, Congo, Colombia, Egypt, Gabon, The Gambia,
Guyana, Ghana, India, Israel, Jordan, Kenya, Mauri-
tania, Panama, South Korea, Syria, Tanzania, Thai-
land, and Zimbabwe-had note
The missed prediction for these countries could be just
that: a missed prediction. On the other hand, because
the model is looking only at the current year and two
years ahead, the prediction may be simply uncon-
firmed. In this case, the usefulness of the model is to
indicate which countries to watch in the last year of
the prediction period. We attempted to take a closer
look at nine of these countries using 1984 data and
found that the probability of rescheduling:
? Fell for India, Kenya, and Egypt.
? Remained about the same for South Korea, Thai-
land, Jordan, Israel, and Bolivia.
? Rose for Burma (table 3).
Table 3
Rescheduling Predictions for 1984,
1985, and 1986 Using 1984
Indicator Data
Bolivia
0.99
Yes
Burma
0.95
Yes
Egypt
0.85
Yes
India
0.16
No
Israel
0.99
Yes
Jordan
0.82
Yes
Kenya
0.60
Yes
South Korea
0.42
Yes
Thailand
0.49
Yes
This table is Confidential.
Logistic regression, or any statistical tool, cannot
supplant well-reasoned, sound analysis of the potential
occurrence of an event. In some instances it even may
provide potentially misleading results. In the case of
South Korea, for example, it predicts, albeit barely, a
1983-85 rescheduling, but most other evidence and
sources indicate such an outcome is unlikely. Indeed,
we believe South Korea is in a strong international
financial position. These tools can, nevertheless, pro-
vide useful complementary support. This model car-
ries out such a modest function, providing a way to
use leading economic indicators to predict changes in
the odds for or against a country formally asking its
creditors for a rescheduling of its debt.
Sanitized Copy Approved for Release 2009/11/13: CIA-RDP86T00586R000300310007-7
 Sanitized Copy Approved for Release 2009/11/13: CIA-RDP86T00586R000300310007-7
Confidential
Appendix A
Methods of Estimation
Models of rescheduling were developed using three
methodologies-logistic regression, linear discrimi-
nant analysis, and a method based on mathematical
catastrophe theory. Logistic regression provided the
model that performed best according to the evaluation
criteria. That model was used to develop the probabil-
ities and predictions in this report.
sharply. If their contribution to the model's perfor-
mance becomes sufficiently small, they will be
removed.2
Since the model must predict rescheduling on the
basis of current conditions, rescheduling was lagged
two years and "current" was assumed to be 1983. The
model was estimated for pre-1983 data. The predicted
rescheduling probabilities were calculated for 1983-
85.3
Logistic regression is a standard statistical technique
for estimating probabilities based on a set of continu-
ous predictor variables.' The model used to derive the
results on the debt rescheduling problem was obtained
from a stepwise logistic regression. This procedure
identifies the independent variable with the greatest
classificatory power and creates an initial model from
that variable. Then, the variable that can make the
greatest marginal contribution to the model is incor-
porated. The process repeats until none of the unin-
corporated variables can make a significant
contribution.
Independent variables that are highly correlated may
be removed from the model. The correlated variables
are not only predictors of the dependent variable-
they are also predictors of each other. Thus, when a
correlated variable enters the model, the marginal
contribution of other correlated variables will drop
Logistic regression is a variation on linear regression, the most
commonly used regression technique. In a linear model, the depen-
dent variable can be made arbitrarily large or small by selecting
appropriate values for the independent variables. But, if the
variable of interest is a probability (as in logistic regression), it must
never exceed unity or be less than zero. Logistic regression meets
this restriction by taking a linear combination of the independent
variables, then subjecting it to a transformation called the logistic
transform, or logit. In geometric terms, the logit bends the line into
an S-shaped curve that ranges from zero to one.
Logistic regressions were performed on three sets of.
independent variables:'
? The basic set.
? The basic set with the gross annual change' in each
variable.
? The basic set with the percentage annual change in
each variable.
'Correlations among the independent variables in a regression, if
severe, may warrant the construction of an artificial set of variables
with the correlations removed (as in factor analysis). Because the
independent variables are ratios, many of which have the same
numerator or denominator, there are correlations. The effects of
these on the regression were explored in some detail. Construction
of an artificial set of independent variables was not warranted.
' There are two undesirable aspects associated with this procedure.
First, the model must predict this year's decision based on previous
years' conditions, whereas the decisionmaker may use current
information, if available. This objection, of course, can never
entirely be overcome when predicting events. Second, the lag
reduces the amount of data available to the estimation procedure.
When generating the model, economic conditions in 1981 were
paired with rescheduling in 1981, 1982, or 1983. Data later than
1981 could not be used, because rescheduling information was not
available past 1983.
' When deciding whether or not to reschedule, decisionmakers may
consider not only current economic conditions, but also where the
economy is headed. Poor but rapidly improving economies may
yield a repayment. Good but rapidly deteriorating conditions could
trigger a rescheduling. Assessing the direction of an economy is
complex. For the present study, simple methods were used to
incorporate such information. The annual percentage change in
each variable was used as an indicator of economic direction. The
magnitude of the annual change was used as an alternative.
' The changes were calculated using the previous year as a base. For
example, the change in the consumer price index in 1978 was
obtained by subtracting the index in 1977 from the index in 1978.
Regressions using gross or percentage changes were not able to use
1977 data because no earlier data were available.
Sanitized Copy Approved for Release 2009/11/13: CIA-RDP86T00586R000300310007-7
 Sanitized Copy Approved for Release 2009/11/13: CIA-RDP86T00586R000300310007-7
Confidential
For the basic indicators with their gross annual
changes (the best performing set of independent vari-
ables), the variables with nonzero coefficients were:
? Consumer price index (CPID).
? Ratio of exports earnings to imports (EXPIMP).
? Ratio of total reserves to imports (RSIMP).
? Ratio of debt interest to export earnings (INTEXP).
? Change in the ratio of debt service to exports
(DDSEXP).
For readability, the linear part of the model and its
logistic transform can be written separately. Letting
BETA represent the linear part of the model,'
BETA = 2.2324 - 0.00138*CPID -
0.6983*EXPIMP - 17.053*INTEXP +
2.387*RSIMP + 4.823*CDSEXP
The probability of not rescheduling in a given three
years is estimated by the logit transformation of
BETA:'
P = EXP(BETA) / [1 + EXP(BETA)].
The probability of rescheduling is estimated by:
PROB=1-P
' When interpreting the model, little significance should be at-
tached to the list of variables that were included. Because correla-
tions are present among the independent variables, the exclusion of
a particular variable may not mean that it lacked predictive power,
but rather that some other variable presented the same information
in a marginally better form. Also, little significance should be
attached to the signs of the coefficients in the model equation.
When correlations are present, the signs require very careful
interpretation. Without such interpretation, some of the signs may
seem paradoxical. In the present model, for example, a high ratio of
export earnings to imports would seem to promote rescheduling.
This, of course, is not a real effect, but an artifact of correlations
among the variables. The presence of paradoxical signs does not
invalidate the model's overall performance. Rather, it means that
the component parts of the model (that is, terms in the regression
equation) cannot stand on their own as models of the effects of
individual variables.
'The S-shaped curve produced by the logit transformation is a
generic form used to estimate probability functions. It is the
standard form for preparing such estimates when the exact nature
of the probability form is not know. C. C. Brown's chi-square test
was used to assess any lack of fit between the shape of the logistic
curve and the shape of the data. The test gives the probability that
the differences between the ideal curve and the data are due to
sampling error, assuming that the errors are normally distributed.
For the model used to derive the key findings, the probability is
25.2 percent.
A debt rescheduling is predicted if the probability of
rescheduling is sufficiently large-for this problem,
the cutoff point was chosen to be 0.342. Accordingly,
76.7 percent of the reschedulings were correctly clas-
sifed; 76.1 percent of the nonreschedulings were cor-
rectly classified; and 76.3 percent of the overall
classifications were correct!
The optimal cutoff point is less than one-half (figure
1). Ideally, the cutoff point should maximize the
percentage of correctly classified events, and make
equal the percentage of correctly classified reschedul-
ings and nonreschedulings. The cutoff point maximiz-
ing overall correct classifications lies a little to the
right of the point where the "rescheduling" and
"nonrescheduling" lines intersect (figure 1).
Since the goal of the model is to discriminate resche-
dulings from nonreschedulings, a balance of correct
classifications between the two categories is impor-
tan.' As the figure demonstrates, this can be obtained
without a drastic decrease in the overall percentage of
correct classifications.
The foregoing analysis shows that the logistic curve
does not closely model the form of the econometric
data. Still, the overall performance of the regression
model is acceptable-and better than the perfor-
mance of two competing methodologies.
Logistic regression gives two kinds of results-a prob-
ability of rescheduling, and a prediction of whether or
not rescheduling will occur. The predictions are much
simpler to interpret than the probabilities. Although
the probability figures are useful in identifying close
calls, and in assessing the degree to which a country's
economic status has changed, interpreting the proba-
bilities involves subtleties and cannot be done
intuitively.
' The maximum percentage of total correct classifications is 77.0,
corresponding to a cutoff point of 0.358. The percentage of
correctly classified nonreschedulings is then 78.8, with 74.1 of the
reschedulings correctly classified. The cost of improving the bal-
ance between the categories is a decline of 0.7 in the percentage of
total correct classifications. This corresponds to an expected loss of
less than one correct prediction among the 75 countries.
' Some policymakers may prefer to err on the side of caution-that
is, to increase the percentage of correctly predicted reschedulings at
the cost of predicting fewer nonreschedulings correctly. Since that
is a judgmental matter, this study aims for equal predictive power
in both categories.
Sanitized Copy Approved for Release 2009/11/13: CIA-RDP86T00586R000300310007-7
 Sanitized Copy Approved for Release 2009/11/13: CIA-RDP86T00586R000300310007-7
conneential
Two methods of estimation were used as alternatives
to logistic regression-linear discriminant analysis,
and a method based on mathematical catastrophe
theory. Neither of the alternatives performed as well
as logistic regression. Discriminant analysis yielded a
higher percentage of correct classifications (77.7 per-
cent versus 76.3 percent for logistic regression), but
the classifications were unbalanced (74.1 percent for
reschedulings, 79.1 percent for nonreschedulings). Ca-
tastrophe theory scored 76.7 percent correct predic-
tions (76.0 percent for rescheduling, 75.1 percent for
nonrescheduling).
Linear Discriminant Analysis
Discriminant analysis is an outgrowth of the theory of
Gaussian distributions and assumes that the data are
samples drawn from two or more Gaussian distribu-
tions with the same variances, but different means 10
(figure 2). The distributions overlap. The goal of
discriminant analysis is to draw lines separating the
distributions. The lines serve a similar function to cut-
off point in logistic regression.
For predicting reschedulings, the data are not single
values as illustrated in figure 2, but are vectors of
values comprising the independent variables. The
corresponding discriminant is not a line, but a higher
dimensional analogue of a line. To view the distribu-
tions and the discriminant, it is possible to reduce a
multidimensional set of independent variables to a
single artificial variable, called the canonical variable.
The distributions of reschedulings and nonreschedul-
ings across values of the canonical variable illustrate
the fundamental source of difficulty in discriminating
reschedulings-the economic indicators characterize
countries that do not reschedule far better than
countries that do (figure 3). Nonreschedulings are
grouped at the higher values of the canonical variable.
Reschedulings occur at all but the highest values.
The ideal histograms shown in figure 2 can, in large
10 If the within-group variances in the raw data are not equal, the
data can be transformed to better meet the assumptions underlying
the method. The variances of the independent variables differ
markedly between rescheduling and nonrescheduling. These differ-
ences were reduced by the use of the cube root transform.
Figure 2
Example of a Linear Discriminant
part, be separated by a discriminant line. The actual
distributions in figure 3 substantially overlap. A clean
separation cannot be effected.
Catastrophe theory is a branch of topology concerned
with surfaces having a fold, cusp, or other discontinu-
ity. Its application to mathematical modeling lies in
the use of such surfaces as modeled response surfaces.
The kinds of regression discussed in the previous
section permit only one dependent variable. Other
methods of estimation allow multiple dependent vari-
ables. In that case, the dependent variables describe a
Sanitized Copy Approved for Release 2009/11/13: CIA-RDP86T00586R000300310007-7
 Confideti.9Sanitized Copy Approved for Release 2009/11/13: CIA-RDP86T00586R000300310007-7
Figure 3
Histogram of Canonical Variable
0 Yes 0 No
EIIII
H W
3LLLLLfHf
"HF
multidimensional surface. The points along the sur-
face predict how the dependent variables will respond
to sets of values of the independent variables; hence,
they are called response surfaces.
Most shapes used as models for response surfaces are
continuous-without breaks, folds, and so forth.
These will produce models in which the dependent
variable changes smoothly over time, provided that
the independent variables do likewise. This kind of
surface can only approximately model behaviors like
rescheduling, where economic conditions may change
smoothly, yet the response shifts from nonreschedul-
ing to rescheduling without any intermediaries.
Castastrophe theory offers a variety of surfaces hav-
ing "catastrophes" as candidate response surface
models. Some kinds of catastrophes involve discontin-
uities. When the dependent variable moves across a
discontinuity, an instantaneous change of value will
occur.
Catastrophe theoretic models also permit the response
surface to have a property called hysteresis. Hystere-
_~s means that the value of the dependent variables
-may depend not only on the independent variables'
current values, but also on their history.
Sanitized Copy Approved for Release 2009/11/13: CIA-RDP86T00586R000300310007-7
 Sanitized Copy Approved for Release 2009/11/13: CIA-RDP86T00586R000300310007-7
Confidential
Rescheduling behavior may exhibit hysteresis. Con-
sider a country with a problem economy. The country
reschedules. The decisionmakers may wish to avoid a
second rescheduling, for a variety of reasons-such as
avoiding further damage to the country's perceived
creditworthiness. If so, the country's rescheduling
history should be included as an independent variable.
Another possible source of hysteresis involves the
decisionmaker's expectations. Consider two countries
with similar problem economies. One's economy is
improving; the other's is deteriorating. Perhaps the
decisionmaker with the improving economy would opt
to delay rescheduling, while the other decisionmaker
might reschedule in the hope of ameliorating the
problem. In this case, the dependent variable, re-
scheduling, would depend in part on the independent
variables' history.
Since most applications of catastrophe theory are
qualitative, this study used a hybrid catastrophe
theory/discriminant analysis approach. An attempt
was made to use some of the insights provided by
catastrophe theory, by incorporating possible hystere-
sis effects into the model-that is, perhaps countries
that reschedule should be considered separately from
countries that did not reschedule. Accordingly, the
countries were divided into three groups-preresched-
uling (including those that did not reschedule at all),
rescheduling, and postrescheduling. When a discrimi-
nant analysis was run against the basic independent
variables with their annual percentage changes, this
method produced the highest overall percentage of
correct classifications (81.7 percent). However, only
76.0 percent of reschedulings were correctly classi-
fied-as opposed to 76.7 percent for logistic regres-
sion. The prerescheduling and postrescheduling coun-
tries were classified correctly 78.3 percent and 46.2
percent of the time, respectively." The three-year lag
that was applied to the independent variable in the
other two models was not applicable here because of
data constraints. Consequently, because this model
predicted rescheduling only for the same year as the
indicator data, the percentage accuracy is not directly
comparable to the other models.
" These figures were combined via a weighted average to find the
percentage of correctly classified nonreschedulings (75.1 percent)
given in table 1.
For each method of estimation, three models were
generated. These modeled the actual rescheduling
behavior of the 75 countries based on one of the
following sets of indicators:
? The basic indicators.
? The basic indicators and their annual percentage
changes.
? The basic indicators and their gross annual changes.
The basic indicators together with their gross annual
changes produced the most successful models.
The models were applied to each country separately
and the percentage of successful classifications re-
corded; these results were presented in table 1. The
results were evaluated according to two criteria:
? The percentage of classifications that were correct.
? The degree to which incorrect classifications were
evenly distributed between rescheduling and nonre-
scheduling countries.
In evaluating the models, it is necessary to establish a
trade-off between the two criteria. The problem can
be understood by way of a simple example. In the
aggregate, countries used roughly one-quarter of their
opportunities to reschedule. Thus, a prediction that
countries never reschedule would be right nearly 75
percent of the time. But none of the reschedulings
would be successfully predicted. This defeats the
purpose of the model.
Ideally, the model should predict rescheduling and
nonrescheduling with equal accuracy. But achieving a
balance may reduce the model's overall performance.
Two of the methods-discriminant analysis and ca-
tastrophe theory-automatically set the trade-off be-
tween overall correctness and balance. Logistic re-
gression analysis requires the analyst to determine the
trade-off. Since the purpose is to identify countries
that will reschedule from those that will not, an excess
of errors in either direction is undesirable. Therefore,
a close balance was sought even though the total
number of correct classifications declined slightly."
" For the most successful regression model, the decline was from
77.0 percent to 76.3 percent correct classifications.
Sanitized Copy Approved for Release 2009/11/13: CIA-RDP86T00586R000300310007-7
 Sanitized Copy Approved for Release 2009/11/13: CIA-RDP86T00586R000300310007-7
Confidennai
As was shown in table 1, the differences in percentage
of total correct classifications among the models is not
large. However, the degree to which reschedulings are
accurately classified varies considerably. In all of the
models not based on logistic regression, nonreschedul-
ings generally are classified more accurately than
reschedulings. It is much more certain that a country
with a favorable economy will not reschedule, than
that a country with an unfavorable economy will.
The catastrophe theory proved the most successful in
terms of overall correctness. Although catastrophe
theory classified 76.7 percent of the total cases cor-
rectly, only 76 percent of the reschedulings were
correctly classified. In other words, the model is
highly successful at telling when a country will not
reschedule, but does worse than logistic regression in
classifying reschedulings. The discriminant analysis
results show a similar imbalance.
In summary, logistic regression provided an accept-
able level of overall correctness (76.3 percent) and a
good balance (76.7 percent correct for reschedulings,
76.1 percent for nonreschedulings). Catastrophe the-
ory and discriminant analysis had marginally higher
levels of overall correctness, but for discriminant
analysis this fact was outweighed by the greater
imbalance between the percentage of correct predic-
tion to reschedulings and nonreschedulings. For catas-
trophe theory, the absence of the three-year lag
applied to the other methods limits the usefulness of
the results or forecasts. This method is of mainly
technical interest. Consequently, the estimates given
in this paper derive from logistic regression on the
basic variables and their annual changes.
Sanitized Copy Approved for Release 2009/11/13: CIA-RDP86T00586R000300310007-7
 Sanitized Copy Approved for Release 2009/11/13: CIA-RDP86T00586R000300310007-7
Confidential
Appendix B
Data on Economic Indicators
for Selected Foreign Countries,
1977-84
Sanitized Copy Approved for Release 2009/11/13: CIA-RDP86T00586R000300310007-7
 Sanitized Copy Approved for Release 2009/11/13: CIA-RDP86T00586R000300310007-7
Confidential
Key to Economic Indicator Abbreviations
Abbreviation Economic Indicator
CPID Rate of inflation based on consumer price index
DBEXP Ratio of debt burden to export earnings
DSEXP Ratio of debt service to export earning
EXPIMP Ratio of export earnings to imports
IMFIMP Ratio of International Monetary Fund reserves to imports
INTDS Ratio of debt interest to debt service
INTEXP Ratio of debt interest to export earnings
OFTLDB Ratio of official to total debt burden
OFTLDS Ratio of official to total debt service
RSIMP Ratio of total reserves to imports
RSCL Country rescheduled debt = 1
Country did not reschedule debt = 0
Sanitized Copy Approved for Release 2009/11/13: CIA-RDP86T00586R000300310007-7
 Sanitized Copy Approved for Release 2009/11/13: CIA-RDP86T00586R000300310007-7
COUNTRY
YEAR
CPID
OBEXP
DSEXP
EXPIMP
IMFIMP
INTDS
INTEXP
OFTLDB
OFTLDS
RSCL
RSIMP
--------------------
----
-------
-------
-------
-------
-------
-------
-------
-------
-------
-------
-------
ARGENTINA
77
176.587
1.335
0.266
1.358
0.000
0.312
0.083
0.168
0.163
0.000
0.659
78
175.323
1.538
0.418
1.669
0.034
0.292
0.122
0.168
0.083
0.000
1.033
79
159.562
1.797
0.325
1.166
0.023
0.480
0.156
0.133
0.100
0.000
1.087
80
100.763
2.103
0.421
0.761
0.025
0.526
0.222
0.113
0.108
0.000
0.514
81
104.500
2.857
0.614
0.970
0.025
0.660
0.405
0.073
0.059
1.000
0.314
82
164.743
3.773
0.723
1.429
0.017
0.674
0.487
0.069
0.065
1.000
0.454
83
343.812
3.672
0.677
1.744
0.000
0.605
0.409
0.071
0.077
1.000
0.276
77
3.213
0.045
0.016
0.839
0.002
0.157
0.003
0.124
0.632
0.000
0.019
78
6.096
0.045
0.009
0.853
0.002
0.322
0.003
0.116
0.164
0.000
0.018
79
9.046
0.026
0.008
0.949
0.001
0.323
0.003
0.086
0.135
0.000
0.015
80
12.108
0.021
0.006
0.891
0.002
0.406
0.002
0.116
0.128
0.000
0.013
81
11.100
0.046
0.011
0.996
0.002
0.557
0.006
0.104
0.095
0.000
0.024
82
6.031
0.080
0.020
0.890
0.002
0.629
0.012
0.082
0.039
0.000
0.032
83
4.075
0.064
0.013
1.685
0.005
0.717
0.010
0.083
0.067
0.000
0.052
77
10.300
4.850
0.147
0.409
0.000
0.373
0.055
0.943
0.716
0.000
0.166
78
13.200
5.091
0.180
0.363
0.000
0.441
0.080
0.939
0.728
0.000
0.161
79
8.200
5.084
0.279
0.345
0.000
0.245
0.068
0.804
0.346
0.000
0.155
80
18.500
4.765
0.125
0.292
0.000
0.385
0.048
0.933
0.540
0.000
0.091
81
13.200
4.982
0.148
0.293
0.000
0.422
0.062
0.941
0.606
0.000
0.045
82
9.276
5.820
0.184
0.334
0.003
0.733
0.135
0.940
0.631
0.000
0.073
83
8.084
5.606
0.220
0.445
0.011
0.570
0.125
0.944
0.709
0.000
0.255
77
8.154
0.659
0.114
0.354
0.012
0.227
0.026
0.566
0.136
0.000
0.112
78
9.531
0.604
0.082
0.415
0.010
0.430
0.035
0.559
0.318
0.000
0.147
79
16.867
0.598
0.110
0.357
0.007
0.452
0.050
0.589
0.349
0.000
0.119
80
14.547
0.513
0.085
0.434
0.010
0.474
0.040
0.620
0.224
0.000
0.119
81
14.600
1.051
0.153
0.340
0.009
0.648
0.099
0.476
0.275
0.000
0.150
82
10.297
0.949
0.150
0.478
0.000
0.628
0.094
0.469
0.422
0.000
0.200
83
5.221
0.904
0.139
0.589
0.003
0.548
0.076
0.409
0.300
0.000
0.190
77
7.983
2.339
0.272
1.073
0.012
0.360
0.098
0.483
0.276
0.000
0.329
78
10.311
2.876
0.600
0.816
0.012
0.249
0.149
0.486
0.153
0.000
0.199
79
19.753
2.616
0.383
0.852
0.000
0.445
0.170
0.495
0.263
1.000
0.178
80
47.275
2.433
0.327
1.416
0.000
0.554
0.181
0.509
0.284
1.000
0.165
81
32.100
2.904
0.329
0.991
0.000
0.671
0.221
0.497
0.310
1.000
0.125
82
123.618
3.276
0.362
1.677
0.000
0.845
0.306
0.509
0.328
1.000
0.347
83
275.558
2.944
0.681
2.723
0.000
0.340
0.231
0.491
0.251
1.000
0.424
84
600.000
5.076
1.676
0.913
0.000
0.217
0.363
0.491
0.251
1.000
0.342
77
13.166
1.688
0.158
0.654
0.005
0.358
0.057
0.588
0.196
0.000
0.299
78
9.003
1.289
0.172
0.627
0.003
0.476
0.082
0.412
0.211
0.000
0.327
79
11.563
0.677
0.094
0.838
0.004
0.563
0.053"
0.439
0.211
0.000
0.390
80
13.895
0.562
0.079
0.728
0.007
0.589
0.047
0.519
0.293
0.000
0.390
81
16.200
0.715
0.089
0.500
0.011
0.761
0.068
0.560
0.228
0.000
0.274
82
11.532
0.857
0.126
0.664
0.013
0.480
0.061
0.451
0.193
0.000
0.390
83
10.262
0.713
0.129
0.860
0.015
0.328
0.042
0.548
0.319
0.000
0.514
77
43.330
2.885
0.434
0.914
0.012
0.386
0.168
0.152
0.122
0.000
0.451
78
38.750
3.649
0.585
0.841
0.009
0.423
0.247
0.132
0.094
0.000
0.607
79
52.790
3.370
0.634
0.770
0.009
0.492
0.312
0.125
0.084
0.000
0.347
80
82.810
2.788
0.571
0.807
0.011
0.551
0.314
0.125
0.082
0.000
0.184
81
105.600
2.761
0.576
0.967
0.009
0.580
0.334
0.117
0.078
1.000
0.239
Sanitized Copy Approved for Release 2009/11/13: CIA-RDP86T00586R000300310007-7
 Sanitized Copy Approved for Release 2009/11/13: CIA-RDP86T00586R000300310007-7
BRAZIL
82
97.957
3.505
0.771
0.958
0.012
0.646
0.498
0.113
0.099
1.000
0.169
83
141.990
0.357
0.730
1.303
0.000
0.563
0.'411
1.140
0.123
1.000
0.249
77
-1.185
3.163
0.295
0.756
0.000
0.210
0.062
0.633
0.203
0.000
0.339
78
-5.994
4.546
0.356
0.716
0.000
0.247
0.088
0.643
0.225
0.000
0.252
79
5.632
3.720
0.376
1.203
0.000
0.256
0.096
0.640
0.233
0.000
0.512
80
0.604
3.827
0.392
1.336
0.000
0.298
0.117
0.647
0.216
0.000
0.604
81
0.300
4.417
0.502
1.279
0.024
0.263
0.132
0.617
0.174
0.000
0.555
82
4.985
6.292
0.591
0.930
0.032
0.381
0.225
0.662
0.201
0.000
0.253
83
8.428
6.206
0.722
1.062
0.020
0.330
0.238
0.642
0.269
0.000
0.266
84
5.999
5.259
0.612
2.088
0.020
0.421
0.257
0.642
0.269
0.000
0.449
77
6.863
0.457
0.034
1.200
0.000
0.200
0.007
0.838
0.333
0.000
1.052
78
23.857
1.003
0.042
0.706
0.048
0.345
0.014
0.848
0.345
0.000
0.638
79
36.566
1.061
0.040
0.682
0.029
0.310
0.012
0.889
0.452
0.000
0.451
80
9.421
2.190
0.101
0.390
0.044
0.424
0.043
0.928
0.333
0.000
0.445
81
7.800
2.496
0.132
0.442
0.046
0.351
0.046
0.873
0.266
0.000
0.329
82
9.926
2.545
0.120
0.409
0.034
0.448
0.054
0.881
0.286
0.000
0.126
83
8.861
2.548
0.197
0.515
0.049
0.264
0.052
0.882
0.518
0.000
0.134
77
14.737
1.198
0.086
0.899
0.000
0.466
0.040
0.603
0.444
0.000
0.045
78
12.451
1.555
0.154
0.760
0.003
0.423
0.065
0.581
0.364
0.000
0.039
79
6.643
1.652
0.157
0.880
0.005
0.491
0.077
0.549
0.270
0.000
0.076
80
9.290
1.609
0.166
0.864
0.008
0.587
0.098
0.534
0.299
0.000
0.093
81
10.600
2.068
0.234
0.774
0.008
0.620
0.145
0.566
0.304
0.000
0.052
82
12.297
2.077
0.317
0.880
0.012
0.559
0.177
0.592
0.285
0.000
0.051
83
15.157
1.895
0.302
1.096
0.006
0.477
0.144
0.609
0.322
0.000
0.092
77
11.053
1.325
0.051
1.272
0.000
0.341
0.017
0.462
0.707
0.000
0.333
78
11.483
1.765
0.064
1.308
0.030
0.348
0.022
0.448
0.261
0.000
0.343
79
9.300
1.763
0.008
1.143
0.027
2.167
0.016
0.496
0.500
1.000
0.486
80
17.357
1.825
0.024
1.426
0.000
0.143
0.003
0.624
0.679
1.000
0.536
81
12.000
2.779
0.086
0.830
0.000
0.676
0.058
0.695
0.706
1.000
0.633
82
12.000
2.960
0.060
0.789
0.011
0.689
0.041
0.723
0.800
1.000
0.442
83
12.000
3.267
0.285
0.789
0.021
0.201
0.057
0.755
0.762
1.000
0.426
CHILE
77
92.233
2.120
0.487
0.970
0.000
0.244
0.119
0.442
0.331
0.000
0.177
78
40.151
2.391
0.605
0.825
0.013
0.287
0.174
0.322
0.293
0.000
0.295
79
33.333
1.938
0.462
0.923
0.009
0.380
0.176
0.221
0.266
0.000
0.362
80
35.135
2.065
0.514
0.912
0.013
0.489
0.251
0.159
0.134
0.000
0.490
81
19.700
3.234
0.889
0.614
0.010
0.539
0.479
0.112
0.072
1.000
0.443
82
9.941
3.764
0.853
1.052
0.020
0.545
0.465
0.093
0.073
1.000
0.483 0
83
27.204
3.884
0.795
1.393
0.000
0.488
0.388
0.097
0.077
1.000
0.726
COLOMBIA
77
33.036
1.264
0.159
1.205
0.038
0.415
0.066
0.605
0.466
0.000
0.739
78
17.788
1.097
0.158
1.059
0.025
0.446
0.071
0.606
0.420
0.000
0.665 po
79
24.712
1.216
0.231
1.021
0.023
0.399
0.092
0.527
0.299
0.000
0.928
80
26.534
1.297
0.189
0.846
0.025
0.575
0.108
0.461
0.334
0.000
0.833
81
27.500
2.086
0.304
0.569
0.029
0.663
0.202
0.421
0.304
0.000
0.816
82
24.549
2.400
0.390
0.565
0.032
0.620
0.242
0.403
0.281
1.000
0.663
83
19.773
2.832
0.449
0.605
0.053
0.556
0.250
0.388
0.341
1.000
0.395
77
14.474
2.896
0.233
0.900
0.000
0.344
0.080
0.662
0.353
0.000
0.055
78
10.089
5.418
0.254
0.581
0.000
0.426
0.108
0.618
0.418
0.000
0.029
Sanitized Copy Approved for Release 2009/11/13: CIA-RDP86T00586R000300310007-7
 Sanitized Copy Approved for Release 2009/11/13: CIA-RDP86T00586R000300310007-7
C~
COUNTRY
YEAR
CPID
DBEXP
DSEXP
EXPIMP
IMFIMP
INTDS
INTEXP
OFTLOB
OFTLDS
RSCL
RSIMP
--------------------
----
-------
-------
-------
-------
-------
-------
-------
-------
-------
-------
-------
CONGO
79
8.121
1.812
0.242
1.751
0.000
0.354
0.086
0.621
0.276
0.000
0.111
80
7.296
1.054
0.113
2.222
0.000
0.402
0.045
0.581
0.362
0.000
0.157
81
17.100
1.126
0.119
1.079
0.002
0.356
0.042
0.480
0.387
0.000
0.107
82
12.724
1.426
0.285
1.210
0.004
0.504
0.144
0.444
0.258
0.000
0.042
83
7.803
1.748
0.453
1.580
0.005
0.331
0.150
0.438
0.251
0.000
0.014
77
4.131
0.940
0.114
0.811
0.000
0.405
0.046
0.502
0.303
0.000
0.156
78
6.019
1.178
0.292
0.712
0.007
0.278
0.081
0.519
0.182
0.000
0.128
79
9.290
1.490
0.293
0.669
0.005
0.338
0.099
0.442
0.216
0.000
0.067
80
18.064
1.658
0.201
0.675
0.000
0.626
0.126
0.464
0.314
0.000
0.078
81
37.100
2.337
0.202
0.796
0.000
0.545
0.110
0.418
0.354
1.000
0.094
82
90.080
2.872
0.158
0.994
0.000
1.097
0.174
0.445
0.637
1.000
0.239
83
32.617
3.473
0.548
1.298
0.000
0.520
0.285
0.463
0.380
1.000
0.501
77
12.966
1.169
0.160
0.800
0.000
0.326
0.052
0.430
0.329
0.000
0.156
78
3.562
1.526
0.209
0.684
0.000
0.460
0.096
0.426
0.283
0.000
0.124
79
9.172
1.331
0.351
0.716
0.000
0.284
0.100
0.429
0.150
0.000
0.153
80
16.686
1.496
0.215
0.587
0.000
0.634
0.136
0.538
0.269
0.00.0
0.099
81
7.500
1.383
0.245
0.712
0.000
0.560
0.137
0.600
0.428
1.000
0.119
82
7.721
2.605
0.447
0.532
0.000
0.498
0.222
0.647
0.371
1.000
0.083
83
4.374
2.896
0.516
0.534
0.005
0.368
0.190
0.706
0.542
1.000
0.113
ECUADOR
77
13.136
1.109
0.128
0.806
0.000
0.369
0.047
0.295
0.232
0.000
0.349
78
15.530
1.842
0.201
0.921
0.005
0.478
0.096
0.416
0.141
0.000
0.308
79
10.302
1.582
0.521
1.041
0.005
0.218
0.113
0.341
0.257
0.000
0.283
80
13.895
1.581
0.275
1.114
0.010
0.464
0.127
0.338
0.306
0.000
0.360
81
13.000
1.947
0.427
1.132
0.011
0.433
0.185
0.348
0.439
1.000
0.248
82
16.106
2.343
0.640
1.075
0.000
0.450
0.288
0.311
0.348
1.000
0.146
83
45.122
2.497
0.518
1.504
0.008
0.493
0.255
0.291
0.231
1.000
0.430
EGYPT
77
12.791
4.836
0.667
0.355
0.000
0.308
0.206
0.822
0.434
0.000
0.091
78
11.046
6.012
0.751
0.258
0.000
0.330
0.248
0.815
0.382
0.000
0.069
79
9.947
6.651
0.659
0.479
0.000
0.275
0.181
0.781
0.281
0.000
0.127
80
20.627
4.542
0.539
0.627
0.000
0.252
0.136
0.773
0.263
0.000
0.186
81
10.400
4.713
0.657
0.368
0.003
0.308
0.202
0.752
0.348
0.000
0.080
82
14.855
5.329
0.842
0.344
0.000
0.407
0.342
0.722
0.383
1.000
0.079
83
16.088
5.059
0.708
0.356
0.003
0.390
0.276
0.753
0.459
1.000
0.083
84
18.000
4.956
0.709
0.367
0.003
0.302
0.214
0.753
0.459
1.000
0.077
77
11.897
0.312
0.078
1.047
0.005
0.256
0.020
0.792
0.221
0.000
0.206
78
13.251
0.553
0.062
0.780
0.009
0.591
0.037
0.706
0.450
0.000
0.217
79
15.918
0.467
0.051
1.089
0.008
0.602
0.031
0.734
0.443
0.000
0.121
80
17.371
0.570
0.058
1.116
0.000
0.651
0.038
0.813
0.472
0.000
0.082
81
14.800
1.005
0.095
0.809
0.000
0.609
0.058
0.814
0.546
0.000
0.081
82
11.760
1.294
0.111
0.797
0.000
0.579
0.064
0.820
0.613
0.000
0.131
83
13.094
1.412
0.155
0.789
0.000
0.393
0.061
0.833
0.721
0.000
0.186
77
16.478
1.424
0.090
0.848
0.019
0.378
0.034
0.921
0.709
0.000
0.474
78
14.424
1.825
0.110
0.575
0.000
0.378
0.042
0.910
0.643
0.000
0.246
79
16.000
1.475
0.068
0.736
0.000
0.461
0.031
0.963
0.789
0.000
0.249
80
4.493
1.655
0.082
0.589
0.006
0.490
0.040
0.943
0.726
0.000
0.102
81
6.100
2.129
0.125
0.526
0.000
0.428
0.054
0.875
0.582
0.000
0.322
82
5.938
2.244
0.155
0.514
0.000
0.530
0.082
0.874
0.552
0.000
0.219
83
-0.712
2.469
0.216
0.466
0.005
0.383
0.083
0.886
0.620
0.000
0.147
Sanitized Copy Approved for Release 2009/11/13: CIA-RDP86T00586R000300310007-7
 Sanitized Copy Approved for Release 2009/11/13: CIA-RDP86T00586R000300310007-7
COUNTRY
YEAR
CPID
DBEXP
DSEXP
EXPIMP
IMFIMP
INTDS
INTEXP
OFTLDB
OFTLDS
RSCL
RSIMP
--------------------
----
-------
-------
-------
-------
-------
-------
-------
-------
-------
-------
-------
FIJI
77
7.563
0.492
0.049
0.588
0.011
0.477
0.023
0.680
0.511
0.000
0.396
78
5.990
0.490
0.098
0.559
0.009
0.359
0.035
0.672
0.446
0.000
0.292
79
7.248
0.465
0.059
0.547
0.007
0.480
0.028
0.690
0.539
0.000
0.221
80
14.547
0.537
0.066
0.670
0.009
0.364
0.024
0.617
0.413
0.000
0.234
81
11.200
0.758
0.099
0.493
0.008
0.577
0.057
0.648
0.479
0.000
0.184
82
7.014
0.859
0.094
0.559
0.010
0.840
0.079
0.778
0.701
0.000
0.226
83
6.723
1.278
0.169
0.587
0.020
0.638
0.108
0.760
0.676
0.000
0.274
77
13.840
1.162
0.133
1.874
0.004
0.269
0.036
0.166
0.111
1.000
0.012
78
10.774
1.453
0.290
1.795
0.000
0.387
0.112
0.205
0.105
1.000
0.029
79
7.955
0.847
0.182
3.476
0.000
0.383
0.070
0.259
0.137
0.000
0.030
80
12.334
0.646
0.193
3.226
0.000
0.294
0.057
0.281
0.118
0.000
0.126
81
8.700
0.510
0.152
2.611
0.000
0.346
0.053
0.315
0.137
0.000
0.203
82
16.651
0.500
0.157
2.985
0.000
0.337
0.053
0.298
0.120
0.000
0.391
83
11.719
0.417
0.148
2.880
0.009
0.305
0.045
0.332
0.235
0.000
0.339
77
12.465
0.615
0.027
0.612
0.000
0.231
0.006
0.829
0.308
0.000
0.259
78
8.744
1.186
0.105
0.390
0.000
0.122
0.013
0.570
0.122
0.000
0.199
79
6.115
1.339
0.100
0.413
0.000
0.259
0.026
0.641
0.052
0.000
0.010
80
6.724
3.577
0.143
0.192
0.000
0.178
0.025
0.691
0.089
0.000
0.027
81
6.100
4.890
0.316
0.219
0.000
0.407
0.128
0.720
0.151
0.000
0.024
82
10.839
3.428
0.254
0.456
0.000
0.625
0.159
0.781
0.295
0.000
0.083
83
10.629
3.153
0.232
0.426
0.000
0.614
0.142
0.871
0.596
0.000
0.026
77
116.406
0.803
0.041
0.887
0.000
0.396
0.016
0.681
0.697
0.000
0.118
78
73.085
0.845
0.071
1.106
0.000
0.356
0.025
0.631
0.598
0.000
0.225
79
54.415
1.021
0.083
1.165
0.000
0.359
0.030
0.660
0.667
0.000
0.256
80
50.083
0.979
0.095
1.190
0.000
0.243
0.023
0.717
0.350
0.000
0.154
81
116.490
1.186
0.084
0.961
0.007
0.382
0.032
0.730
0.525
0.000
0.125
82
22.297
1.413
0.108
1.238
0.000
0.450
0.048
0.770
0.634
1.000
0.203
83
122.537
2.640
0.229
0.634
0.011
0.386
0.088
0.782
0.633
1.000
0.205
77
12.444
0.334
0.032
1.164
0.012
0.563
0.018
0.682
0.346
0.000
0.540
78
8.000
0.555
0.059
0.866
0.010
0.593
0.035
0.633
0.352
0.000
0.457
79
11.482
0.609
0.066
0.845
0.009
0.709
0.047
0.693
0.393
0.000
0.364
80
10.742
0.586
0.060
0.974
0.014
0.729
0.044
0.728
0.448
0.000
0.230
81
11.400
0.970
0.088
0.746
0.005
0.644
0.057
0.761
0.472
0.000
0.088
82
0.449
1.303
0.116
0.842
0.000
0.572
0.067
0.798
0.569
0.000
0.086
83
1.950
1.407
0.159
1.100
0.007
0.376
0.060
0.764
0.584
0.000
0.193
77
8.247
1.559
0.123
0.822
0.000
0.476
0.059
0.514
0.295
0.000
0.060
78
15.252
1.484
0.168
1.060
0.000
0.343
0.058
0.565
0.286
0.000
0.160
79
17.762
1.696
0.312
0.921
0.000
0.271
0.084
0.585
0.254
0.000
0.042
80
14.090
1.425
0.179
0.982
0.000
0.365
0.065
0.612
0.420
0.000
0.025
81
24.700
1.838
0.228
0.805
0.000
0.448
0.102
0.711
0.340
0.000
0.014 ?
82
18.524
2.739
0.203
0.861
0.000
0.832
0.169
0.731
0.321
1.000
0.036
83
13.498
3.972
0.692
0.671
0.000
0.327
0.226
0.746
0.416
1.000
0.024
77
6.492
0.968
0.133
0.702
0.000
0.156
0.021
0.856
0.553
0.000
0.132
78
-2.594
1.229
0.127
0.666
0.010
0.198
0.025
0.901
0.497
0.000
0.129
79
13.049
1.234
0.067
0.681
0.016
0.240
0.016
0.917
0.512
0.000
0.156
80
17.786
1.437
0.122
0.518
0.000
0.227
0.028
0.867
0.458
0.000
0.035
81
5.700
2.411
0.176
0.335
0.000
0.290
0.051
0.735
0.474
0.000
0.046
82
9.177
2.545
0.128
0.308
0.000
0.444
0.057
0.775
0.338
0.000
0.008
Sanitized Copy Approved for Release 2009/11/13: CIA-RDP86T00586R000300310007-7
 Sanitized Copy Approved for Release 2009/11/13: CIA-RDP86T00586R000300310007-7
O
COUNTRY
YEAR
CPID
DBEXP
DSEXP
EXPIMP
IMFIMP
INTDS
INTEXP
OFTLDB
OFTLDS
RSCL
RSIMP
--------------------
----
-------
-------
-------
-------
-------
-------
-------
-------
-------
-------
-------
0.057
0.836
0.293
0.000
0.015
77
8.383
1.154
0.130
0.895
0.000
0.450
0.059
0.621
0.456
0.000
0.256
78
6.215
1.282
0.153
0.876
0.009
0.494
0.076
0.613
0.430
0.000
0.203
79
12.484
1.288
0.197
0.888
0.007
0.449
0.088
0.614
0.335
0.000
0.193
80
15.607
1.425
0.166
0.822
0.000
0.622
0.103
0.603
0.391
0.000
0.117
81
10.200
1.850
0.202
0.801
0.000
0.694
0.140
0.614
0.402
0.000
0.092
82
9.982
2.362
0.285
0.918
0.000
0.582
0.166
0.650
0.412
1.000
0.143
83
9.488
2.778
0.370
0.759
0.005
0.582
0.216
0.688
0.442
1.000
0.132
77
4.000
0.081
0.013
0.920
0.000
0.049
0.001
0.022
0.035
0.000
0.303
78
5.400
0.109
0.025
0.855
0.000
0.051
0.001
0.014
0.016
0.000
0.312
79
14.000
0.124
0.022
0.884
0.000
0.031
0.001
0.015
0.012
0.000
0.250
80
13.500
0.127
0.025
0.881
0.000
0.090
0.002
0.017
0.013
0.000
0.215
81
9.700
0.136
0.036
0.881
0.000
0.047
0.002
0.016
0.020
0.000
0.235
82
11.900
0.182
0.044
0.891
0.000
0.024
0.001
0.014
0.007
0.000
0.273
83
9.800
0.194
0.043
0.914
0.000
0.027
0.001
0.013
0.012
0.000
0.314
77
8.400
2.347
0.161
0.960
0.000
0.299
0.048
1.009
0.835
1.000
0.642
78
2.500
2.376
0.177
0.848
0.009
0.321
0.057
0.970
0.796
0.000
0.664
79
6.400
2.120
0.151
0.794
0.016
0.378
0.057
1.030
0.816
0.000
0.605
80
11.400
2.118
0.142
0.578
0.022
0.335
0.048
0.948
0.770
0.000
0.386
81
13.000
2.275
0.171
0.538
0.021
0.297
0.051
0.973
0.685
0.000
0.281
82
7.876
2.211
0.174
0.633
0.025
0.439
0.077
0.001
0.668
0.000
0.285
83
11.813
1.254
0.070
0.590
0.036
1.347
0.094
0.056
0.590
0.000
0.366
84
9.999
2.371
0.208
0.627
0.036
0.473
0.098
0.923
0.195
0.000
0.396
INDONESIA
77
11.107
1.137
0.128
1.742
0.011
0.318
0.041
0.573
0.199
0.000
0.332
78
8.107
1.247
0.199
1.740
0.010
0.222
0.044
0.578
0.185
0.000
0.303
79
21.904
0.968
0.158
2.165
0.010
0.312
0.049
0.559
0.221
0.000
0.429
80
18.525
0.759
0.098
2.022
0.015
0.383
0.037
0.565
0.287
0.000
0.398
81
12.200
0.787
0.107
1.677
0.015
0.411
0.044
0.568
0.289
0.000
0.333
82
9.500
0.905
0.123
1.322
0.013
0.424
0.052
0.545
0.311
0.000
0.176
83
11.300
1.168
0.171
1.288
0.004
0.359
0.061
0.506
0.330
0.000
0.223
84
10.999
1.166
0.188
1.668
0.004
0.478
0.090
0.506
0.330
0.000
0.338
77
p4.557
2.622
0.205
0.534
0.000
0.456
0.093
0.587
0.413
0.000
0.224
78
50.620
2.352
0.137
0.522
0.000
0.599
0.082
0.619
0.506
0.000
0.274
79
78.295
2.274
0.303
0.524
0.004
0.277
0.084
0.693
0.318
0.000
0.273
80
131.000
2.281
0.212
0.570
0.003
0.582
0.123
0.697
0.591
0.000
0.275
81
116.800
2.515
0.364
0.555
0.000
0.392
0.143
0.683
0.395
0.000
0.298
82
120.387
2.836
0.403
0.544
0.000
0.641
0.258
0.693
0.497
0.000
0.365
83
145.605
3.160
0.592
0.537
0.004
0.467
0.276
0.688
0.367
0.000
0.373
84
340.000
3.094
0.549
0.608
0.004
0.435
0.239
0.688
0.367
0.000
0.281
77
27.553
0.920
0.136
1.228
0.000
0.379
0.051
0.299
0.170
0.000
0.087
78
12.991
1.335
0.190
0.998
0.004
0.451
0.086
0.258
0.157
0.000
0.148
79
16.577
1.524
0.251
1.009
0.005
0.397
0.100
0.275
0.154
0.000
0.045
80
14.679
1.470
0.280
1.048
0.003
0.375
0.105
0.257
0.139
0.000
0.006
81
8.800
1.879
0.330
1.063
0.000
0.516
0.170
0.238
0.165
0.000
0.007
82
7.353
2.267
0.505
1.055
0.000
0.530
0.268
0.258
0.131
1.000
0.002
83
5.907
2.943
0.644
0.969
0.000
0.522
0.336
0.271
0.180
1.000
0.011
Sanitized Copy Approved for Release 2009/11/13: CIA-RDP86T00586R000300310007-7
 Sanitized Copy Approved for Release 2009/11/13: CIA-RDP86T00586R000300310007-7
JAMAICA
78
34.893
1.263
0.237
0.920
0.000
0.365
0.086
0.460
0.239
1.000
0.051
79
29.088
1.482
0.268
0.824
0.000
0.435
0.117
0.495
0.307
1.000
0.049
80
27.307
1.572
0.253
0.822
0.000
0.455
0.115
0.527
0.308
0.000
0.070
81
12.740
1.739
0.421
0.661
0.001
0.317
0.133
0.551
0.205
0.000
0.050
82
6.546
2.481
0.432
0.529
0.003
0.647
0.280
0.585
0.370
1.000
0.072
83
11.588
2.763
0.664
0.482
0.000
0.388
0.258
0.604
0.346
1.000
0.041
77
14.441
2.997
0.245
0.180
0.004
0.362
0.089
0.628
0.300
0.000
0.404
78
6.920
3.737
0.358
0.198
0.004
0.456
0.163
0.549
0.234
0.000
0.473
79
14.340
3.194
0.355
0.205
0.005
0.475
0.169
0.595
0.256
0.000
0.466
80
10.988
3.042
0.407
0.239
0.007
0.539
0.219
0.570
0.205
0.000
0.389
81
7.700
2.738
0.431
0.232
0.005
0.524
0.226
0.611
0.214
0.000
0.307
82
7.428
2.933
0.425
0.237
0.005
0.593
0.252
0.664
0.285
0.000
0.264
83
5.013
4.086
0.580
0.191
0.002
0.527
0.306
0.684
0.489
0.000
0.272
84
3.000
3.192
0.453
0.318
0.002
0.453
0.205
0.684
0.489
0.000
0.258
77
14.852
0.988
0.074
0.930
0.000
0.490
0.036
0.617
0.475
0.000
0.335
78
16.954
1.348
0.139
0.598
0.000
0.425
0.059
0.612
0.372
0.000
0.160
79
7.985
1.654
0.164
0.667
0.000
0.508
0.083
0.562
0.382
0.000
0.289
80
13.766
1.573
0.167
0.595
0.000
0.505
0.084
0.563
0.355
0.000
0.166
81
11.800
1.989
0.244
0.554
0.000
0.448
0.109
0.573
0.293
0.000
0.094
82
20.483
2.449
0.397
0.600
0.001
0.498
0.198
0.613
0.230
0.000
0.112
83
11.507
2.603
0.372
0.752
0.007
0.505
0.188
0.645
0.358
0.000
0.258
84
9.999
2.187
0.312
0.699
0.007
0.429
0.134
0.646
0.358
0.000
0.219
77
6.223
0.592
0.059
0.965
0.000
0.279
0.017
0.773
0.649
0.000
0.049
78
7.357
0.688
0.052
1.048
0.000
0.500
0.026
0.703
0.621
1.000
0.029
79
11.548
0.872
0.139
1.059
0.000
0.298
0.042
0.669
0.217
1.000
0.082
80
13.766
0.953
0.066
1.103
0.000
0.578
0.038
0.712
0.293
1.000
0.008
81
7.600
1.196
0.051
1.188
0.000
0.868
0.045
0.758
0.407
1.000
0.016
82
5.948
0.956
0.049
1.422
0.000
1.346
0.067
0.777
0.654
1.000
0.013
83
2.763
1.197
0.116
1.454
0.000
0.675
0.078
0.785
0.651
1.000
0.051
77
3.111
0.722
0.042
0.975
0.000
0.385
0.016
0.888
0.888
0.000
0.163
78
6.609
0.902
0.069
0.875
0.000
0.270
0.019
0.744
0.588
0.000
0.103
79
14.016
1.813
0.132
0.616
0.000
0.312
0.041
0.584
0.362
1.000
0.006
80
18.203
2.619
0.146
0.670
0.000
0.438
0.064
0.611
0.359
1.000
0.012
81
30.500
6.533
0.284
0.493
0.001
0.854
0.243
0.626
0.335
1.000
0.054
82
31.418
7.487
0.537
0.469
0.002
0.740
0.397
0.641
0.232
1.000
0.040
83
19.288
6.987
1.295
0.519
0.000
0.287
0.372
0.657
0.385
1.000
0.061
77
4.167
2.036
0.142
0.859
0.000
0.332
0.047
0.684
0.276
0.000
0.311
78
8.428
2.721
0.191
0.548
0.000
0.470
0.090
0.690
0.331
0.000
0.171
79
11.331
2.541
0.245
0.560
0.000
0.570
0.140
0.554
0.211
0.000
0.134
80
18.343
2.610
0.309
0.648
0.000
0.551
0.170
0.559
0.141
0.000
0.123
81
9.500
2.680
0.374
0.791
0.011
0.632
0.236
0.600
0.206
1.000
0.119
82
9.315
3.102
0.341
0.825
0.000
0.825
0.281
0.634
0.213
1.000
0.069
83
15.372
3.776
0.611
0.692
0.007
0.469
0.287
0.644
0.212
1.000
0.049
77
4.739
0.430
0.091
1.339
0.012
0.300
0.027
0.364
0.142
0.000
0.518
78
4.988
0.412
0.127
1.255
0.009
0.212
0.027
0.358
0.149
0.000
0.433
79
3.536
0.312
0.060
1.412
0.009
0.384
0.023
0.344
0.208
0.000
0.388
80
6.724
0.341
0.042
1.201
0.011
0.599
0.025
0.306
0.285
0.000
0.327
81
9.700
0.565
0.067
1.019
0.010
0.548
0.037
0.245
0.227
0.000
0.312
Sanitized Copy Approved for Release 2009/11/13: CIA-RDP86T00586R000300310007-7
 Sanitized Copy Approved for Release 2009/11/13: CIA-RDP86T00586R000300310007-7
COUNTRY
YEAR
CPID
--------------------
----
-------
MALAYSIA
82
5.834
0.849
0.111
0.971
0.009
0.522
0.058
0.174
0.137
0.000
0.282
83
3.704
0.962
0.148
1.070
0.012
0.409
0.061
0.146
0.120
0.000
0.280
77
4.167
3.478
0.074
0.784
0.000
0.337
0.025
0.954
0.761
0.000
0.030
78
8.428
4.627
0.091
0.392
0.000
0.324
0.029
0.949
0.814
0.000
0.024
79
11.331
3.595
0.086
0.411
0.000
0.344
0.030
0.923
0.633
0.000
0.014
80
18.343
3.375
0.068
0.468
0.012
0.417
0.028
0.937
0.446
0.000
0.028
81
9.600
4.938
0.103
0.424
0.022
0.535
0.055
0.941
0.472
0.000
0.044
82
9.763
6.641
0.300
0.439
0.027
0.318
0.095
0.835
0.172
0.000
0.048
83
9.726
6.529
0.421
0.489
0.026
0.214
0.090
0.830
0.421
0.000
0.047
77
10.300
2.935
0.262
0.758
0.000
0.219
0.057
0.679
0.431
0.000
0.200
78
7.100
4.792
0.212
0.681
0.000
0.366
0.078
0.694
0.450
0.000
0.338
79
9.000
4.215
0.450
0.567
0.000
0.234
0.106
0.773
0.182
0.000
0.335
80
11.100
3.721
0.154
0.680
0.000
0.430
0.066
0.816
0.430
0.000
0.385
81
19.090
3.194
0.209
0.975
0.000
0.339
0.071
0.880
0.383
0.000
0.524
82
12.629
4.316
0.171
0.849
0.000
1.166
0.200
0.890
0.751
1.000
0.462
83
0.887
4.945
0.574
0.939
0.000
0.392
0.225
0.898
0.917
1.000
0.439
77
9.195
0.268
0.031
0.693
0.000
0.289
0.009
0.851
0.381
0.000
0.123
78
8.491
0.505
0.042
0.650
0.000
0.616
0.026
0.667
0.471
0.000
0.072
79
14.526
0.650
0.064
0.665
0.000
0.635
0.041
0.548
0.419
0.000
0.041
80
42.674
0.722
0.088
0.708
0.000
0.567
0.050
0.516
0.289
0.000
0.118
81
14.500
1.038
0.164
0.590
0.000
0.654
0.107
0.589
0.248
0.000
0.055
82
11.354
1.104
0.195
0.787
0.000
0.552
0.108
0.624
0.329
0.000
0.077
83
5.647
1.226
0.259
0.832
0.000
0.399
0.103
0.000
0.313
0.000
0.041
MEXICO
77
28.959
5.917
1.029
0.768
0.000
0.330
0.339
0.133
0.079
0.000
0.241
78
17.544
5.460
1.255
0.789
0.000
0.285
0.358
0.116
0.062
0.000
0.196
79
18.060
4.194
1.296
0.743
0.000
0.294
0.381
0.101
0.062
0.000
0.136
80
26.422
2.794
0.638
0.800
0.005
0.472
0.301
0.103
0.061
0.000
0.123
81
27.900
2.760
0.561
0.805
0.007
0.585
0.328
0.100
0.065
1.000
0.149
82
58.952
2.798
0.528
1.412
0.000
0.697
0.368
0.118
0.073
1.000
0.054
83
101.869
3.002
0.761
2.573
0.011
0.429
0.326
0.148
0.156
1.000
0.465
77
12.600
3.270
0.230
0.407
0.000
0.532
0.122
0.464
0.413
0.000
0.137
78
9.800
3.707
0.419
0.508
0.000
0.465
0.195
0.439
0.260
0.000
0.168
79
12.000
3.421
0.460
0.532
0.000
0.525
0.241
0.423
0.268
0.000
0.122
80
15.000
3.092
0.528
0.574
0.000
0.507
0.267
0.459
0.219
0.000
0.079
81
12.500
3.514
0.577
0.542
0.000
0.505
0.291
0.519
0.188
1.000
0.050
82
10.578
4.677
0.749
0.478
0.000
0.551
0.413
0.508
0.173
1.000
0.051
83
5.611
5.815
1.171
0.492
0.000
0.459
0.537
0.529
0.294
1.000
0.036
77
9.804
0.895
0.035
0.479
0.000
0.500
0.017
1.000
0.786
0.000
0.710
78
7.398
0.958
0.030
0.411
0.011
0.407
0.012
1.000
1.000
0.000
0.527
79
3.563
1.132
0.030
0.428
0.009
0.455
0.014
1.000
1.000
0.000
0.495
80
14.679
2.164
0.050
0.235
0.015
0.475
0.024
1.000
1.000
0.000
0.434
81
11.100
1.653
0.035
0.381
0.016
0.531
0.019
1.000
1.000
0.000
0.486
82
11.701
3.385
0.068
0.222
0.015
0.683
0.047
1.000
1.000
0.000
0.470
83
12.832
3.682
0.094
0.202
0.014
0.583
0.055
1.000
0.988
0.000
0.286
77
11.449
1.493
0.184
0.836
0.000
0.483
0.089
0.474
0.257
0.000
0.161
78
4.612
1.658
0.182
1.084
0.000
0.494
0.090
0.511
0.274
1.000
0.067
79
48.096
2.082
0.122
1.573
0.000
0.867
0.106
0.587
0.481
1.000
0.083
Sanitized Copy Approved for Release 2009/11/13: CIA-RDP86T00586R000300310007-7
 Sanitized Copy Approved for Release 2009/11/13: CIA-RDP86T00586R000300310007-7
NICARAGUA
80
35.318
3.767
0.170
0.508
0.000
0.495
0.084
0.550
0.737
1.000
0.034
81
23.900
4.421
0.376
0.509
0.000
0.487
0.183
0.561
0.538
1.000
0.030
82
24.778
6.929
0.709
0.523
0.000
0.466
0.331
0.711
0.683
1.000
0.039
83
31.048
7.334
0.741
0.564
0.000
0.420
0.312
0.715
0.670
1.000
0.000
77
23.274
1.290
0.159
0.816
0.016
0.325
0.052
0.547
0.243
0.000
0.424
78
10.026
1.099
0.106
0.925
0.016
0.470
0.050
0.553
0.225
0.000
0.324
79.
7.337
0.906
0.093
0.971
0.011
0.713
0.066
0.492
0.169
0.000
0.218
80
10.254
1.076
0.146
0.953
0.010
0.691
0.101
0.454
0.300
0.000
0.167
81
22.900
1.545
0.187
0.892
0.012
0.638
0.119
0.543
0.416
1.000
0.179
82
11.635
2.080
0.397
0.752
0.014
0.629
0.250
0.564
0.329
1.000
0.061
83
-2.478
2.484
0.358
0.969
0.033
0.795
0.285
0.657
0.344
1.000
0.189
77
19.474
0.147
0.022
1.062
0.031
0.368
0.008
0.488
0.307
0.000
0.316
78
18.649
0.260
0.017
0.822
0.029
0.425
0.007
0.332
0.492
0.000
0.115
79
11.139
0.241
0.025
1.719
0.029
0.688
0.017
0.221
0.235
0.000
0.414
80
11.359
0.211
0.031
1.595
0.022
0.793
0.024
0.176
0.141
0.000
0.484
81
20.900
0.342
0.065
0.943
0.021
0.627
0.041
0.147
0.092
1.000
0.161
82
7.527
0.602
0.110
1.121
0.000
0.622
0.069
0.115
0.073
1.000
0.101
83
20.308
1.024
0.222
1.460
0.000
0.484
0.108
0.109
0.091
1.000
0.122
84
34.999
1.051
0.254
2.921
0.000
0.328
0.083
0.109
0.091
1.000
0.196
77
9.000
5.802
0.282
0.486
0.000
0.439
0.124
0.940
0.779
0.000
0.174
78
7.300
5.259
0.282
0.449
0.000
0.485
0.137
0.935
0.716
0.000
0.114
79
9.300
4.060
0.282
0.507
0.000
0.435
0.123
0.915
0.696
1.000
0.055
80
12.600
3.515
0.260
0.489
0.000
0.411
0.107
0.892
0.674
1.000
0.084
81
11.880
3.189
0.214
0.512
0.000
0.373
0.080
0.903
0.599
1.000
0.121
82
5.899
4.127
0.297
0.439
0.011
0.449
0.133
0.884
0.548
0.000
0.173
83
7.444
3.204
0.280
0.577
0.017
0.390
0.109
0.900
0.705
0.000
0.366
PANAMA
77
4.552
5.591
0.694
0.291
0.000
0.444
0.308
0.272
0.144
0.000
0.068
78
4.225
7.669
2.265
0.272
0.004
0.223
0.505
0.231
0.055
0.000
0.123
79
7.985
7.120
1.327
0.256
0.002
0.511
0.679
0.231
0.105
0.000
0.078
80
13.766
6.513
1.338
0.249
0.006
0.545
0.729
0.239
0.101
0.000
0.066
81
7.300
7.826
1.631
0.213
0.000
0.578
0.943
0.240
0.165
1.000
0.067
82
4.287
8.362
1.840
0.237
0.000
0.541
0.995
0.233
0.100
1.000
0.059
83
2.055
8.099
1.431
0.340
0.007
0.559
0.799
0.383
0.212
1.000
0.158
PAPUA NEW GUINEA
77
4.319
0.500
0.040
1.064
0.000
0.697
0.028
0.355
0.369
0.000
0.548
78
5.897
0.530
0.043
0.927
0.000
0.739
0.032
0.385
0.363
0.000
0.405
79
5.687
0.458
0.056
0.978
0.003
0.545
0.030
0.394
0.249
0.000
0.425
80
12.108
0.452
0.054
0.959
0.003
0.485
0.026
0.418
0.222
0.000
0.284
81
8.100
0.749
0.081
0.661
0.000
0.671
0.054
0.382
0.220
0.000
0.309
82
5.458
0.981
0.124
0.635
0.000
0.737
0.091
0.359
0.172
0.000
0.262
83
7.895
1.022
0.138
0.667
0.004
0.582
0.080
0.366
0.205
0.000
0.370
77
9.298
1.419
0.135
0.905
0.021
0.375
0.051
0.684
0.356
0.000
0.716
78
10.590
2.068
0.199
0.671
0.017
0.454
0.090
0.637
0.329
0.000
0.900
79
28.257
2.168
0.263
0.586
0.016
0.474
0.125
0.558
0.291
0.000
0.890
80
22.399
2.527
0.359
0.505
0.024
0.609
0.219
0.518
0.300
0.000
0.974
81
14.000
3.174
0.412
0.493
0.042
0.596
0.246
0.485
0.258
0.000
1.156
82
6.754
3.366
0.370
0.491
0.042
0.834
0.308
0.501
0.289
1.000
0.923
83
13.394
2.855
0.446
0.877
0.061
0.519
0.231
0.568
0.363
1.000
1.234
Sanitized Copy Approved for Release 2009/11/13: CIA-RDP86T00586R000300310007-7
 Sanitized Copy Approved for Release 2009/11/13: CIA-RDP86T00586R000300310007-7
COUNTRY
YEAR
CPID
DBEXP
DSEXP
EXPIMP
IMFIMP
INTDS
INTEXP
OFTLDB
OFTLDS
RSCL
RSIMP
--------------------
----
-------
-------
-------
-------
-------
-------
-------
-------
-------
-------
-------
PERU
77
38.057
3.628
0.544
0.903
0.000
0.364
0.198
0.328
0.183
1.000
0.172
78
57.855
3.723
0.556
0.991
0.000
0.448
0.249
0.352
0.226
1.000
0.171
79
66.693
2.275
0.377
1.918
0.000
0.517
0.195
0.344
0.233
1.000
0.657
80
59.210
2.084
0.489
1.560
0.000
0.430
0.211
0.389
0.231
1.000
0.640
81
75.390
2.511
0.732
0.944
0.000
0.379
0.277
0.383
0.225
1.000
0.315
82
64.445
2.812
0.655
0.914
0.000
0.451
0.296
0.364
0.274
1.000
0.353
83
111.150
3.625
0.797
1.201
0.000
0.365
0.291
0.335
0.270
1.000
0.691
77
7.967
1.606
0.274
0.732
0.000
0.300
0.082
0.285
0.186
0.000
.0.294
78
7.530
1.846
0.346
0.661
0.000
0.303
0.105
0.303
0.167
0.000
0.273
79
18.908
1.593
0.311
0.695
0.000
0.394
0.123
0.314
0.213
0.000
0.269
80
17.786
1.523
0.226
0.692
0.000
0.532
0.120
0.312
0.192
0.000
0.277
81
13.300
1.805
0.339
0.668
0.000
0.514
0.174
0.329
0.159
0.000
0.230
82
10.944
2.415
0.530
0.601
0.000
0.488
0.258
0.333
0.171
1.000
0.196
83
10.899
2.761
0.561
0.614
0.000
0.463
0.260
0.000
0.190
1.000
0.096
77
14.505
0.802
0.014
0.749
0.017
0.769
0.011
0.977
0.615
0.000
0.556
78
12.598
1.376
0.025
0.401
0.016
0.944
0.024
0.987
0.611
0.000
0.376
79
15.691
1.115
0.016
0.592
0.028
0.889
0.014
0.970
0.389
0.000
0.601
80
7.216
2.125
0.040
0.311
0.034
0.733
0.029
0.966
0.500
0.000
0.602
81
6.600
2.189
0.057
0.290
0.025
0.596
0.034
0.950
0.511
0.000
0.527
82
12.195
2.158
0.074
0.316
0.024
0.612
0.045
0.969
0.776
0.000
0.405
83
6.773
2.488
0.089
0.290
0.035
0.603
0.053
0.976
0.836
0.000
0.373
77
11.401
0.800
0.113
0.816
0.000
0.311
0.035
0.485
0.218
0.000
0.037
78
3.329
1.680
0.278
0.559
0.003
0.270
0.075
0.487
0.167
0.000
0.020
79
9.785
1.704
0.273
0.575
0.000
0.308
0.084
0.504
0.191
1.000
0.017
80
8.696
2.125
0.437
0.453
0.000
0.276
0.120
0.566
0.188
1.000
0.007
81
5.900
2.540
0.298
0.514
0.000
0.347
0.103
0.642
0.215
1.000
0.009
82
17.280
3.488
0.354
0.459
0.001
0.477
0.169
0.693
0.386
1.000
0.011
83
11.675
3.188
0.498
0.577
0.001
0.358
0.179
0.727
0.441
1.000
0.015
77
11.651
1.728
0.164
0.680
0.000
0.243
0.040
0.514
0.188
1.000
0.152
78
7.536
1.795
0.296
0.599
0.000
0.185
0.055
0.482
0.272
1.000
0.096
79
21.294
1.860
0.311
0.612
0.000
0.176
0.055
0.470
0.118
1.000
0.112
80
11.111
1.874
0.239
0.486
0.000
0.175
0.042
0.613
0.073
1.000
0.056
81
23.300
2.902
0.450
0.431
0.003
0.165
0.074
0.594
0.068
0.000
0.045
82
31.062
3.404
0.531
0.372
0.000
0.272
0.144
0.608
0.209
1.000
0.027
83
63.335
3.644
0.412
0.311
0.000
0.292
0.120
0.652
0.318
1.000
0.038
77
3.297
0.139
0.013
0.787
0.001
0.494
0.007
0.357
0.362
0.000
0.303
78
4.728
0.137
0.034
0.777
0.001
0.287
0.010
0.344
0.146
0.000
0.312
79
4.063
0.121
0.021
0.807
0.001
0.449
0.009
0.307
0.326
0.000
0.250
80
8.460
0.088
0.018
0.807
0.002
0.478
0.008
0.331
0.341
0.000
0.215
81
8.200
0.087
0.016
0.760
0.002
0.583
0.009
0.282
0.222
0.000
0.235
82
3.882
0.098
0.019
0.738
0.002
0.510
0.009
0.232
0.198
0.000
0.273
83
1.157
0.098
0.022
0.775
0.002
0.363
0.008
0.233
0.150
0.000
0.314
77
10.559
6.499
0.124
0.277
0.000
0.321
0.040
0.938
0.372
0.000
0.435
78
9.963
5.156
0.085
0.442
0.000
0.385
0.033
0.951
0.429
0.000
0.404
79
24.279
6.126
0.184
0.387
0.000
0.441
0.081
0.878
0.206
0.000
0.118
80
58.831
5.312
0.115
0.512
0.000
0.117
0.013
0.953
0.549
0.000
0.044
81
44.400
4.535
0.124
1.004
0.000
0.360
0.045
0.929
0.652
0.000
0.136
82
23.615
5.229
0.150
0.708
0.000
0.633
0.09.5
0.904
0.641
1.000
0.026
Sanitized Copy Approved for Release 2009/11/13: CIA-RDP86T00586R000300310007-7
 Sanitized Copy Approved for Release 2009/11/13: CIA-RDP86T00586R000300310007-7
COUNTRY
YEAR
CPID
DBEXP
DSEXP
EXPIMP
IMFIMP
INTDS
INTEXP
OFTLDB
OFTLDS
RSCL
RSIMP
--------------------
----
-------
-------
-------
-------
-------
-------
-------
-------
-------
-------
-------
SOMALIA
83
27.469
9.439
0.896
1.045
0.000
0.270
0.242
0.918
0.848
1.000
0.094
77
10.173
0.945
0.136
0.929
0.000
0.351
0.048
0.437
0.218
0.000
0.226
78
14.460
1.001
0.164
0.849
0.001
0.329
0.054
0.397
0.195
0.000
0.142
79
18.265
1.029
0.194
0.740
0.001
0.318
0.062
0.368
0.169
0.000
0.111
80
28.700
1.004
0.176
0.785
0.000
0.446
0.078
0.371
0.205
0.000
0.103
81
21.300
0.988
0.188
0.813
0.000
0.439
0.083
0.352
0.186
0.000
0.089
82
7.255
1.072
0.208
0.901
0.000
0.506.
0.105
0.356
0.186
0.000
0.105
83
3.38
1.094
0.200
0.933
0.002
0.411
0.082
0.355
0.235
0.000
0.086
84
1.99
0.914
0.179
1.053
0.002
0.415
0.074
0.355
0.235
0.000
0.087
77
1.253
1.059
0.168
1.083
0.000
0.178
0.030
0.886
0.492
0.000
0.345
78
12.108
1.229
0.110
0.874
0.000
0.289
0.032
0.930
0.580
0.000
0.318
79
10.745
1.218
0.099
0.676
0.000
0.416
0.041
0.878
0.593
0.000
0.272
BO
26.167
1.333
0.094
0.516
0.000
0.448
0.042
0.850
0.649
0.000
0.095
81
17.900
1.540
0.100
0.571
0.001
0.588
0.059
0.770
0.527
0.000
0.149
82
10.857
1.891
0.158
0.512
0.003
0.638
0.101
0.728
0.356
0.000
0.159
83
14.002
2.027
0.180
0.584
0.009
0.561
0.101
0.719
0.409
0.000
0.157
77
16.755
3.599
0.209
0.611
0.000
0.285
0.060
0.588
0.430
1.000
0.018
78
19.874
5.560
0.347
0.433
0.000
0.270
0.094
0.547
0.323
1.000
0.018
79
30.813
6.545
0.193
0.482
0.000
0.419
0.081
0.733
0.564
1.000
0.046
80
25.360
7.483
0.228
0.344
0.000
0.377
0.086
0.761
0.620
1.000
0.024
81
24.600
8.107
0.301
0.437
0.000
0.147
0.044
0.639
0.389
1.000
0.010
82
25.682
13.658
0.997
0.388
0.000
0.255
0.255
0?.555
0.055
1.000
0.015
83
29.349
11.263
1.741
0.461
0.000
0.281
0.489
0.569
0.393
1.000
0.012
77
9.703
0.018
0.003
0.779
0.000
0.333
0.001
1.000
0.889
0.000
0.208
78
10.556
0.083
0.005
0.909
0.012
0.667
0.003
0.170
0.500
0.000
0.255
79
13.936
0.066
0.007
1.081
0.012
0.656
0.005
0.139
0.438
0.000
0.318
80
13.250
0.055
0.006
1.020
0.016
0.645
0.004
0.106
0.419
0.000
0.298
81
8.700
0.058
0.006
0.834
0.014
0.655
0.004
0.109
0.379
0.000
0.317
82
7.268
0.063
0.006
0.838
0.016
0.792
0.004
0.111
0.250
0.000
0.315
83
4.460
0.058
0.009
1.041
0.007
0.487
0.004
0.040
0.308
0.000
0.139
77
20.863
0.306
0.016
0.805
0.009
0.536
0.008
0.935
0.750
0.000
0.348
78
7.589
0.617
0.035
0.631
0.009
0.522
0.018
0.615
0.507
0.000
0.280
79
16.598
0.743
0.054
0.538
0.006
0.543
0.030
0.680
0.378
0.000
0.199
80
18.624
0.568
0.056
0.596
0.007
0.617
0.035
0.695
0.378
0.000
0.207
81
20.000
0.508
0.059
0.635
0.007
0.655
0.039
0.747
0.345
0.000
0.140
82
10.833
0.668
0.079
0.590
0.000
0.656
0.052
0.812
0.465
0.000
0.133
83
16.842
0.902
0.111
0.406
0.003
0.547
0.061
0.811
0.544
0.000
0.148
77
11.797
1.434
0.104
0.400
0.000
0.304
0.032
0.867
0.539
0.000
0.160
78
5.041
1.941
0.257
0.431
0.000
0.352
0.090
0.875
0.717
0.000
0.130
79
4.811
1.430
0.221
0.494
0.000
0.315
0.070
0.903
0.802
0.000
0.141
80
18.920
1.175
0.184
0.511
0.002
0.346
0.064
0.914
0.837
0.000
0.071
81
18.390
1.239
0.182
0.417
0.001
0.378
0.069
0.918
0.837
0.000
0.056
82
14.309
1.369
0.203
0.505
0.000
0.388
0.079
0.912
0.817
0.000
0.052
83
7.536
1.632
0.250
0.421
0.000
0.370
0.093
0.936
0.870
0.000
0.039
77
11.602
2.224
0.116
0.725
0.000
0.351
0.041
0.877
0.415
0.000
0.311
78
11.386
2.860
0.200
0.416
0.000
0.403
0.081
0.804
0.284
0.000
0.067
79
13.778
2.927
0.277
0.494
0.000
0.461
0.128
0.721
0.228
0.000
0.047
Sanitized Copy Approved for Release 2009/11/13: CIA-RDP86T00586R000300310007-7
 Sanitized Copy Approved for Release 2009/11/13: CIA-RDP86T00586R000300310007-7
TANZANIA
80
30.208
3.455
0.340
0.404
0.000
0.501
0.170
0.731
0.255
0.000
0.013 a
81
25.600
3.235
0.440
0.478
0.002
0.401
0.176
0.750
0.276
0.000
0.013
82
28.981
4.588
0.526
0.395
0.000
0.425
0.224
0.777
0.372
0.000
0.004
83
27.037
5.133
0.554
0.539
0.000
0.391
0.216
0.000
0.441
0.000
0.024
77
7.634
0.573
0.081
0.756
0.007
0.395
0.032
0.448
0.335
0.000
0.341
78
7.801
0.675
0.088
0.763
0.000
0.496
0.043
0.454
0.337
0.000
0.304
79
10.000
0.768
0.106
0.740
0.000
0.552
0.058
0.409
0.268
0.000
0.207
80
19.617
0.896
0.121
0.706
0.000
0.643
0.078
0.396
0.242
0.000
0.142
81
12.700
1.036
0.155
0.706
0.000
0.691
0.107
0.393
0.207
0.000
0.158
82
5.235
1.227
0.195
0.812
0.000
0.660
0.129
0.410
0.218
0.000
0.173
83
3.710
1.556
0.248
0.622
0.003
0.541
0.134
0.420
0.294
0.000
0.159
84
1.499
1.589
0.289
0.617
0.003
0.465
0.135
0.420
0.294
0.000
0.147
TOGO
77
22.437
2.014
0.356
0.561
0.007
0.173
0.061
0.341
0.060
0.000
0.134
78
0.485
2.596
0.199
0.538
0.004
0.234
0.047
0.348
0.140
0.000
0.121
79
7.246
3.920
0.170
0.421
0.006
0.259
0.044
0.448
0.562
0.000
0.097
80
12.613
2.727
0.180
0.608
0.000
0.502
0.090
0.543
0.581
0.000
0.111
81
20.100
4.093
0.195
0.477
0.000
0.475
0.093
0.641
0.645
1.000
0.301
82
10.741
4.095
0.169
0.444
0.000
0.896
0.151
0.667
0.567
1.000
0.338
83
9.023
4.100
0.610
0.444
0.000
0.403
0.246
0.671
0.391
1.000
0.369
TRINIDAD & TOBAGO
77
11.794
0.125
0.008
1.205
0.015
0.489
0.004
0.309
0.522
0.000
0.675
78
10.253
0.214
0.017
1.037
0.015
0.675
0.012
0.193
0.362
0.000
0.705
79
14.690
0.203
0.025
1.240
0.018
0.778
0.020
0.215
0.219
0.000
0.773
80
17.509
0.196
0.064
1.283
0.020
0.281
0.018
0.254
0.083
0.000
0.687
81
14.300
0.268
0.047
1.204
0.025
0.607
0.028
0.252
0.158
0.000
0.921
82
11.461
0.367
0.070
0.831
0.026
0.471
0.033
0.222
0.189
0.000
0.756
83
16.719
0.559
0.132
0.950
0.048
0.300
0.040
0.195
0.138
0.000
0.804
URUGUAY
77
58.750
1.349
0.428
0.832
0.000
0.242
0.104
0.302
0.235
0.000
0.535
78
44.488
1.288
0.641
0.886
0.022
0.151
0.097
0.290
0.128
0.000
0.514
79
66.758
1.443
0.206
0.653
0.013
0.562
0.116
0.262
0.254
0.000
0.299
80
63.399
1.253
0.224
0.652
0.016
0.549
0.123
0.242
0.178
0.000
0.259
81
34.000
1.393
0.214
0.740
0.017
0.706
0.151
0.187
0.178
1.000
0.297
82
19.030
2.035
0.283
0.927
0.000
0.755
0.214
0.175
0.165
1.000
0.185
83
49.216
2.747
0.479
1.627
0.014
0.441
0.211
0.179
0.232
1.000
0.463
VENEZUELA
77
7.886
0.662
0.123
0.873
0.076
0.286
0.035
0.079
0.105
0.000
0.618
78
7.018
1.027
0.132
0.781
0.050
0.514
0.068
0.053
0.086
0.000
0.428
79
12.432
0.861
0.143
1.342
0.038
0.469
0.067
0.037
0.047
0.000
0.558
80
21.507
0.741
0.191
1.625
0.041
0.464
0.089
0.030
0.026
0.000
0.472
81
16.000
0.788
0.203
1.536
0.042
0.601
0.122
0.023
0.023
0.000
0.566
82
9.741
1.189
0.302
1.312
0.054
0.530
0.160
0.015
0.017
1.000
0.506
83
6.284
1.419
0.314
1.786
0.104
0.482
0.151
0.014
0.017
1.000
0.911
ZAIRE
77
69.048
2.940
0.123
1.622
0.000
0.584
0.072
0.377
0.325
1.000
0.196
78
48.503
3.907
0.158
1.571
0.000
0.638
0.101
0.414
0.325
1.000
0.182
79
108.595
3.203
0.225
2.215
0.000
0.504
0.114
0.382
0.260
1.000
0.277
80
42.086
2.590
0.272
1.954
0.000
0.508
0.138
0.539
0.416
1.000
0.204
81
34.910
6.510
0.516
0.991
0.034
0.531
0.274
0.596
0.358
1.000
0.214
82
37.210
7.808
0.654
1.185
0.000
0.560
0.366
0.657
0.387
1.000
0.267
83
53.176
10.022
1.795
0.969
0.000
0.379
0.681
0.000
0.499
1.000
0.470
Sanitized Copy Approved for Release 2009/11/13: CIA-RDP86T00586R000300310007-7
 Sanitized Copy Approved for Release 2009/11/13: CIA-RDP86TOO586ROO0300310007-7
COUNTRY
YEAR CPID
DBEXP
DSEXP
EXPIMP
IMFIMP
INTDS
INTEXP
OFTLDB
OFTLDS
RSCL
RSIMP
--------------------
---- -------
-------
-------
-------
-------
-------
-------
-------
-------
-------
-------
ZAMBIA
77 19.624
1.668
0.237
1.140
0.000
0.315
0.075
0.529
0.206
0.000
0.077
78 16.405
1.968
0.328
1.156
0.000
0.310
0.102
0.563
0.178
0.000
0.063
79 9.681
1.544
0.227
1.518
0.000
0.290
0.066
0.599
0.217
0.000
0.075
80 11.732
1.699
0.236
1.073
0.000
0.376
0.089
0.659
0.271
0.000
0.053
81 14.000
2.359
0.316
0.842
0.006
0.316
0.100
0.661
0.345
1.000
0.044
82 12.456
2.494
0.235
1.062
0.000
0.584
0.137
0.654
0.386
1.000
0.060
83 19.832
2.963
0.456
1.000
0.000
0.381
0.174
0.685
0.469
1.000
0.066
77 10.305
0.175
0.007
1.234
0.000
0.295
0.002
0.286
0.852
0.000
0.092
78 9.737
0.465
0.009
1.312
0.000
0.506
0.005
0.099
0.588
0.000
0.174
79 13.789
0.498
0.014
1.133
0.000
0.500
0.007
0.074
0.358
0.000
0.254
80 5.374
0.505
0.031
0.977
0.000
0.222
0.007
0.144
0.161
0.000
0.124
81 13.100
0.675
0.078
0.781
0.000
0.321
0.025
0.214
0.104
0.000
0.090
82 10.698
0.868
0.170
0.778
0.000
0.419
0.071
0.196
0.091
0.000
0.086
83 23.083
0.825
0.190
0.931
0.000
0.490
0.093
0.000
0.111
0.000
0.058
Sanitized Copy Approved for Release 2009/11/13: CIA-RDP86TOO586ROO0300310007-7
 Sanitized Copy Approved for Release 2009/11/13: CIA-RDP86T00586R000300310007-7
Confidential
Confidential
Sanitized Copy Approved for Release 2009/11/13: CIA-RDP86T00586R000300310007-7