The Foreign Exchange Gap, Growth and Industrial
Strategy in Turkey: 1973-1983                                                         SWP306
World Bank Staff Working Paper No. 306
November 1978
The views and interpretations in this document are those of the authors
and should not be attributed to the World Bank, to its affiliated
organizations or to any individu,al acting in their behalf.
Prepared by: Kemal Dervis
Sherman Robinson
Economics of Industry Division
Development Economics Department                                        O      ;
1978
UB        Bank
G          et, N.W.
881.5     DnCl 204.3- LJ-S ,                                                                  VtA
W57
167                                                                                               0
lo .306






The views and interpretations in this document are solely those of the
authors, and should not be attributed to the World Bank, to its affiliated
organizations, or to any individual acting in their behalf.
WORLD BANK
Staff Working Paper No. 306
November 1978
THE FOREIGN EXCHANGE GAP, GROWTH
AND INDUSTRIAL STRATEGY IN TURKEY: 1973-1983
T'his study is an examination of the interaction between. trade,
trade policy and growth in the Turkish economy. The analysis relies to
a great extent on a multi-sector general equilibrium growth model of the
economy. T'he model focuses on trade and industry and attempts to
capture the basic mechanisms that link economic performance and
structure to trade policy in the medium run. The time period covered
is 1973 to 1983 with first an evaluation of the past five years-
(1973-1977) and then an analysis of future prospects and alternatives
(1978-1983). The study was undertaken during the first half of 1978,
by the authors, who work in the World Bank's Economics of Industry
Division.
Prepared by: Kemal Dervi,s
Sherman Robinson
Economics of Industry Division
Development Economics Department
Copyright   Q   1978
The World Bank
1818 H Street, N.W.
Washington, D.C. 20433 U.S.A.






-ii-
This study has been prepared in the Economics of Industry
Division of the Development Economics Department during the Spring and
Summer of 1978. It concentrates on an analysis of the foreign-exchange
gap and the interrelationships between trade, growth and industrialization
in Turkey. The analysis is largely, but not exclusively, based on
results obtained by using a multi-sector general equilibrium growth
model of the economy covering the period from 1973 to 1983. All experi-
ments were completed in July 1978 and the forward looking projections
reflect the data and estimates as they were available at that time.
We have benefited from close collaboration with the regional
economists and would like to thank Ram Chopra, David Berk, Shakil Faruqi
and Adrian Wood for their help and advice. We have worked particularly
closely with Adrian Wood and many of his ideas are reflected in the
formulation, of the model and discussion of the experiments. Vinod Dubey,
Attila Karaosmanoglu, Don Keesing and Larry Westphal made important
suggestions which have helped a great deal in writing the first draft.
Jaime de Melo gave us extensive comments and much in this study reflects
ideas developed with him over the last two years of joint work. Finally,
we would like to thank Hollis Chenery for his encouragement and constant
interest.
We have had the expert research assistance of Murat K6prUi.1i
throughout this study and thank him for his help. We are also indebted
to Jeff Lewis who did the "sources of growth" computation and prepared
the figures in Part 5. Margot Clark typed the whole manuscript and
Rob Kish typed the first two drafts. We thank them for their paLience.






-iii-
Remaining errors and weaknesses are of course ours alone and
the views and policy conclusions expressed in this study are those of
the authors alone and should not be attributed to the World Bank, to its
affiliated organizations or to any individual acting in their behalf.






The Foreign Exchange Gap, Growth
and Industrial Strategy in Turkey: 1973-1983
Page
1.  Introduction                                                               1
2.  Distinctive Features of the TGT Model                                      9
2.1  Introduction                                                          9
2.2  Imporl: Demands and Relative Prices                                  11
2.3  Fixed Exchange Rates and Quantitative Restrictions                   14
2.4  Price Level Normalization, Inflation and the Exchange Rate           19
2.5  Production and Supply                                                22
2.6  The Treatment of Exports                                             25
2.7  Macroeconomic Aspects and the Flow-of-Funds                          27
2.8  Dynam:ic Linkages                                                    34
3.  Analysis oi. the Turkish Economy from 1973 to 1977:  The Making
of a Crisis                                                               38
3.1  Introduction                                                         38
3.2  Summary of Recent Events and Policy Reactions                        39
3.3  Exchange Rate Drift and Structural Imbalances:   Decomposing the
Change in the Equilibrium Exchange Rate                              47
3.4  Growth, Trade and Structure:   1973-77                               59
3.5  Conclusion                                                           69
4.  Prospects f'or the Future:  An Economy-Wide Perspective for 1978-1983     72
4.1  Introduction                                                         72
4.2  Constant Price Deflated Exchange Rate Policy                         72
4.3 Macroeconomic Consequences of Alternative Trade and Exchange
Rate ]'olicies                                                       89
5.  Microeconomic Analysis of the Impact of Trade Policy on Industrial
Structure and the Sources of Industrial Growth                           109
5.1  Introduction:  Trade Policy and Resource Allocation                 109
5.2  Sectoral Aggregation and Sectoral Trade Characteristics in the
TGT Model                                                            i1
5.3  Growth and Industrial Structure under Alternative Trade and
Exchange Rate Policies                                              116
5.4  Export Expansion, Import Substitution and the Sources of Growth     128
6.  Conclusion                                                               150
7.  Appendix A:  The Equations of the TGT Model
8.  Appendix B:  The Data






1. Introduct:Lon
The purpose of this study is to analyze Turkish industrialization
and growth in the 1970's and to evaluate prospects for the 1980's using a
mullti-sector general equilibrium growth model of the economy. The very
serious foreign exchange crisis that emerged in 1977 has again emphasized
the importance of trade and trade policies as major determinants of Turkey's
overall economic performance. It therefore seems appropriate that the
major focus of the discussion be on the interaction between foreign trade
and growth and the analysis of the foreign exchange constraint.
Turkey's growth rate has been impressive in the past, averaging
6.5 percent over three decades (1947-1977). This relatively high growth
rate was achieved without the availability of particularly valuable resources
such as oil, with only a moderate amount of foreign aid and within the
framework of basically democratic political institutions. Finally, while
income is quilte unequally distributed (with a very large rural-urban gap
and a Gini-coefficient above 0.500), basic needs are reasonably well met
and problems of malnutrition, basic health care, basic education and shelter
are less acute than in many countries with equal or even higher per capita
incomes.L/
- The initial conditions from which Turkey started after World War
I were not favorable. For example, both in terms of physical infrastruc-
ture and human resources, Egypt was significantly ahead of Turkey at
1/ For an evaluation of Basic Needs in Turkey, see Karaosmanoglu and
Durdag (1977). For an analysis of the distribution of income, see
Dervi, and Robinson (1977).



- 2 -
the beginning of the century.A/ Particularly in terms of human resources,
all the Southern European countries such as Bulgaria, Greece, Serbia, Croatia,
Spain and Portugal were far ahead of Turkey before and after World War I.
Furthermore, the rate of population growth in Turkey remained betweeen 2.5
and 3.0 percent throughout the century and, while on a declining trend,
it is still more than double that in the rest of Southern Europe. With a
per capita income of about $1000 in 1977, Turkey remains poorer than most
countries in the semi-industrial category.
Growth, while rapid on average, has not proceeded at a steady
pace. The foundations of Turkish industrialization were laid in the decade
before World War II and, in spite of the world depression, Turkey achieved
substantial growth in the 1930's with important investments in infra-
structure and the creation of State Economic Enterprises that successfully
led to the beginnings of industrial growth. The war and the diversion of
resources and change of priorities it created in spite of Turkey's neutrality
were probably the major causes of the complete economic standstill that
followed in the 1940's.i/
Since 1950, which marks the beginning of regular national
accounting as well as an important political turning point, Turkey seems
to have gone through three rather similar cycles. Each starts with a period
of quite rapid industrial growth and ends with a major foreign-exchange
1/ See C. Issawi (1978).
2/ See Bulutay and others (1975) for estimates of national income in the
1930's and 1940's.  See also Herschlag (1968) and Land (1970).



- 3-
crisis, a large devaluation and a transitory slowdown in industrial growth.L/
In Figure 1, two-year moving averages of industrial growth rates
have been plotted against time. The three cycles are quite apparent
from the graph. Each downswing is associated with an acute foreign-exchange
crisis and a major effective devaluation, close to 100% in 1958, about 50%
in 1970 and again about 50% in the period from September 1977 to March 1978.-/
Wh:Lle a clear cyclical pattern emerges from Figure 1, one has
to be careful in interpreting the cycles in too mechanistic a fashion.
Common factors and aspects exist but one crisis has not been a simple
repetition of its predecessor. Thus, while the 1958-1960 crisis followed
a period of a]Lmost hyperinflation and was followed by a period of remark-
able price stability, exactly the opposite is true of the 1970 crisis.
It followed a period of relative price stability but was followed by a
period of substantial inflatiosi. The impact on export performance has
also varied. The 1958 de facto devaluation was not followed by a major
upward surge of exports. Between 1957/58 and 1961/62, exports increased
by only 23% in value. In contrast export revenue increased by 153% between
1969/1970 and 1973/1974.
There was reason for much optimism in the early 1970's. The
foreign-exchiange constraint that had plagued the Turkish economy throughout
1/ Economy-wiLde growth has not always followed the movements of industrial
growth because of the extreme volatility of agricultural growth, heavily
dependent on weather conditions.
2/ See Krueger (1974) for the computation of changes in effective exchange
rates in 1958 and 1970.   Note that by industrial we are here referring
to the non-agricultural sectors including services.



I~~- . 
.  -'-- -.f-1'_- 4-          1'-4'---''. I-  -------I 
____ _ ______  _ _________  ______ ___ _  _________=  _  ___ ______=_  ;__ __
I_  _   _  __-_.------            <  -.        t      ,-
.   -j, j ____ _____.-.-.I...-___--.. ...-.. -.-..-. . ...F_
_ _   _ I .__ .       L   I.                         _  _ -_  __I_____1__ _ - ---- -  --- ---
_I    _i_  -         .  .   _.  __.___._ __        __<__---7'''---_.  _ _. __
o t- -    --     ;
K.~                I  __         I _---       I----- ------
..   o  _,_   .__   _ ___   _  .__   .................................... --- 1  = : .  .
.- ~3i- _     ______ ____ S __ ___ ___ - ____----t_  ___ ______
_. i~  3t   "    _ __      __            ___      - t___ - 
._ , _         .____ , ___'_  ............__  ._ ___  _ _____.......... ___.__ 
,L.      _ -____,_- -5<!--_i__                - -_
~~~4i-       ___ ___     _____ ___3 _ ____      i     ; - . . _._
42N~~~~~
--.---..-.- .-   I.-. -   1--                  .----- =  t  ii
r-i~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~f
-t:-o:---t-- -=  I  t  4 :=---~~~~~I -------.-F1---.--- ~--- =-1---t------=  f
___r , _ ___ __ _ _  _, ___  L-  -_  _ --__      ___--  ------  -___-  =
I'o
____h  C-'-__  0____  -  -  ____--__-__ -__  ________  ____  __-_--___  - D- -- -- 1 E- -;--
----f-S I  ---------i-  =-<~-- ;--  = i ------------_- T....- - . .t- --
- O I        L-.-  -.-  .            .        .
~~J  -~~~~  N  N~~~~  -. N  N  ~~~~~~~I  N  N  ~~~~~~  ~~~  ....N~~~~~ ~~~'
A- _.....4
I -  -  I             I.                  I
Ij   _ _ _ _ -   ..II        _  _ _ 1 ~ . _ _ _ _   _ _    _



-5-
the 1950's and 1960's had all of a sudden disappeared. For the first
time ever in post-war history, the current account deficit declined to
zero in 1972 and turned into a 484 million dollar surplus in 1973. Foreign
exchange reseirves reached 2 billion dollars at the end of 1973, a figure
equal to annual imports of merchandise. The major factor responsible for
this substantial accumulation of reserves was the sudden and massive in-
flow of workers' remittances from Western Europe: while only 141 million
dollars in 1969, remittances were 471 million dollars in 1971 and reached
a peak of 1426 million dollars in 1974. Tourism revenues and manufactured
exports also increased. Total export revenues increased by 15% in 1971, 31%
in 1972 and 49% in 1973. Manufactured exports grew at an annual rate of
40% in the same period. Value-added in industry grew at 10.6% in the 1970-73
period. Performance in agriculture was mixed, with a 10% decline in 1973
but a 13% growth-rate in 1971.   While the weather is still a major deter-
minant of agricultural output, increased use of fertilizer and mechanization
proceeded at a very rapid pace in the first half of the seventies and the
average growth rate of agricultural output between 1970 and 1975 was 4.5%.
The- situation seemed to point to an increase in the possible
trend growth rate of the Turkish economy from about 6.5 percent to about
8.0 percent. In fact, between 1970 and 1976, GDP grew at an annual average.
rate of 7.7 percent. It was even on an accelerating trend after 1973,
seemingly in spite of the oil crisis and the world depression that followed.
There was general agreement in Turkey that the Fourth Five Year Plan due
to start in 1978 should aim at an annual growth rate of at least 8.0 percent.



-6-
This kind of growth rate is perceived to be a necessary minimum for the
absorption of underemployment and for a significant narrowing of the
absolute income gap that separates Turkey from other Southern European
countries, an objective planners hope to achieve by the end of this
century.
With the full emergence of the current foreign exchange crisis,
these goals are seriously called into question. The cumulative current
account deficit since 1974 has reached 8 billion dollars, with the 1977
deficit alone about 3 billion dollars. While the current foreign-exchange
crisis is similar in many ways to the 1958 crisis, it is probably deeper,
with a much greater percentage resource gap. The debt-service ratio has
climbed from 11.4 percent of exports and workers' remittances in 1976 to
15.6 percent in 1977 and it will be above 20% for the coming years. Payments
for imports have been delayed on a wide scale, foreign exchange reserves
equal only one month's worth of imports, industry is lacking crucial imported
inputs as well as energy and it appears that growth has come to a halt in
the latter half of 1977.
Predictions and projections for the future vary widely. In
the short-run, that is to say in 1978 and 1979, it seems clear that only
a very substantial net new inflow of borrowed funds can allow increased
capacity use and prevent fixed investment from actually declining. Can
a 5 to 6 percent annual growth rate be achieved in 1978-1979? How much
new external finance would appear necessary for the achievement of such a
target? What are Turkey's growth prospects in the longer-run? Can an 8



-7-
percent growth rate be achieved over the Fourth Five Year Plan now covering
1979 to 1983? What kind of policy package complemented by what amount of
foreign resources could be expected to allow the realization of an 8 per-
cent growth target? What are the implications for the growth of private
consumption and public consumption? What are the implications for em-
ployment and sectoral structure? Is the exchange-rate of 25 TL to the
dollar arrived at after the March 1978 devaluation a realistic one that
should be preserved in real terms or are further devaluations needed?
These are the basic questions we want to address in the following
sections emphasizing and analyzing the impact of policy. But before
attempting to provide projections for the future and evaluating possible
alternative policy packages, it is necessary to analyze the nature of the
crisis that emerged in 1977. How did Turkey move from a 484 million current
account surplus to a 3.2 billion dollar deficit in 4 years? Is the crisis
one that is largely due to factors endogenous to Turkey's development policies
or can it be explained by exogenous shocks coming from the world economy?
Are present problems the necessary outcome of Turkey's inward-looking
import-substituting development strategy or would the crisis never have
occurred had it not been for the oil price increase and increased dis-
bursements for military hardware due to the American embargo? Without at
least finding tentative answers to these questions and assigning approximate
weights to the various factors that led to the 1977 collapse, one cannot
really appraise future prospects and policy alternatives. Part 3
will therefore attempt to provide a quantitative analysis of the 1977 crisis



with special emphasis on the role of the oil price increase. Part 4 will
turn to the future and offer an analysis of the impact of alternative
policy packages on overall growth and economy-wide performance in the
Fourth Five Year Plan period. Part 5 will turn to a microeconomic analysis
and will focus on the pattern of sectoral growth and the role of import
substitution and export expansion. Both the macroeconomic analysis in Part 4
and the microeconomic analysis in Part 5 will rely to a large extent,
although not exclusively, on a computable general equilibrium model of the
Turkish economy that allows conditional policy experiments to be conducted
and with which we attempt to quantify the general equilibrium mechanisms
that link trade, growth and employment to policy variables such as the
exchange rate, tariffs, import rationing, export subsidies, taxes and
government expenditure patterns. Part 2 below provides a detailed
description of the most important features of the model. A complete
statement of the model equations is available in Appendix A.



2. Distinctive Features of the TGT Model
2.1 Introduction: General Equilibrium Modelling
This section describes the most important features of the
general equilibrium growth model of the Turkish economy on which much
of the discussion will be based. While of an economy-wide nature, it
concentrates on the industrial sector and on issues of trade and in-
dustrialization.   We shall call it the TGT model (Turkey, Growth and Trade).
The model is in the tradition of the non-linear computable
general equilibrium (CGE) models that have been built over the past few
years for devielopment planning purposes.-/ The original inspiration for
this class of models can be found in Chenery and Uzawa (1958) and Leif
Johansen's 1960 study of the Norwegian economy, but computational diffi-
culties prevented, at that time, the full implementation of these ideas.
Since then, computational difficulties have greatly diminished and it has
become possible to implement very large and highly non-linear models.
Progress, in this field as in others, is not costless. The data re-
quirements of large models characterized by price endogenous feedback
mechanisms are larger than those of simpler linear models and often
the specification is ahead of the data. The availability of Social
Accounting Matrices of the sort recently built by Stone, Pyatt and others
1/  See for instance Dervis (1975), Adelman and Robinson (1978), Taylor
(1978), De Melo (1978) and Ahluwalia, Lysy and Pyatt (1977).
See also Chenery and Raduchel (1971) who, with a 4-sector illustrative
model, stressed the importance of modelling price-sensitive direct
substitut:Lon mechanisms. For an approach based on linearization rather
than expl:Lcit solution see Celasun (1975).



- 10 -
is therefore a development from which CGE modelling can greatly benefit.l/
A second area of concern is the difficulty of keeping track of the various
causal chains implicit in a complex non-linear model. Nevertheless, while
still in their first phase of implementation, the recent models that
endogenize prices and incorporate direct substitution constitute a notable
advance in the field of development planning because they make possible
an explicit analysis of policy packages that work through the price mech-
anism. They should therefore, when properly used, allow a much richer
dialogue among economic theorists, model builders and policy makers. It
is in this dialogue that one should see their most useful function.
The model of Turkey builds on recent work we have done and
aspects of it can already be found in De Melo and Dervis (1977), Robinson
and De Melo (1976) and Dervis (1977a & b)-. But, particularly in the treat-
ment of trade policy, the specification of adjustment to a fixed exchange
rate, the treatment of exports and the incorporation of macro-accounts,
the present model provides several new features.
In the next subsection, we describe a number of the distinctive
features of the model. The full set of model equations, both static and
dynamic, is given in an appendix. We do not discuss the solution
1/ See Pyatt and Thorbecke (1976), Stone (1970), and United Nations (1975).
An aggregate social accouting matrix for the model is presented in
Section 2.6 below.



- 11 -
algorithm but note only that we are able to solve the rather large non-
linear general equilibrium system quite economically.l/
2.2     Import: Demands and Relative Prices
Emp:Lrically, one of the most unrealistic assumptions of trade
theory is the treatment of foreign and domestic goods of the same
sectoral classification as identical. The assumption is essentially
harmless when it is used to obtain the many important qualitative results
and theorems of trade theory. But when it is incorporated into applied
planning or model-building exercises, it leads to extremely unrealistic
and misleading results. Under such a homogeneity assumption, the
domestic prices of tradables are fully tied to the price of imports and
the structure of production and consumption in a given country will be
extremely sensitive to slight relative-price variations between foreign
and domestic goods, leading to overestimation of the effects of exchange-
rate policies as well as a tendency to specialize in the production of a
few commodities.
An elegant formulation that allows one to keep aggregative
commodity categories across countries, but introduces product differentiation
by countries of origin into the structure of demand for commodities in
any given country, was proposed and implemented in a partial equilibrium
1/  Our solution strategy follows the basic approach described in Adelman
and Robinson (1978). We have, however, developed a new algorithm for
solving this type of model that seems more robust and easier to apply
than previous algorithms.



- .12 -
framework by Armington in 1969.1/ The crucial assumption is that "mar-
ginal rates of substitution between any two products of the same kind
(i.e., commodity category) must be independent of the quantities of the
products of all other kinds."
Within the framework of a single country model, the basic idea
is to define a "composite" commodity that is a C.E.S. aggregation of
commodities produced abroad or imports, Mi, and commodities produced and
consumed at home, Di. The aggregation takes the familiar C.E.S. form:
i= y[6 M-pi + (l -      )D-Pi]l/Pi        il, . . ., n       (1)
iii
where yi, %i and Pi are the parameters of the C.E.S. function in sector i,
with       = ° defining the elasticity of substitution.    M. and D   are
1+P i i                                              21i
like inputs "producing" the aggregate output. Their ratio to each other
is sensitive to relative prices and the degree of sensitivity varies with
the elasticity of substitution. Consumers and producers are assumed to
minimize the cost of obtaining the "composite" goods by choosing the cost-
minimizing ratio of imports to domestically produced goods. After solving
1/ See Armington (1969) for a more complete discussion of the assumption
underlying the treatment of product differentiation. Robinson and
De Melo (1976) used the idea in a single country CGE model formulation
and De Melo, Dervis and Robinson (1977) formulated a global trade model
using the same approach. The approach has also been used in a somewhat
different spirit by Pyatt (1978) and in a still different form by Petri
(l376).  See De Melo and Robinson (1978) for a partial equilibrium analysis
of the price relations implicit in Armington's formulation.



- 13 -
the first order conditions, the import demand functions are given by:
6 i     PDi
M      -     i (PM- ) i Di
i     ~i
where PDi - domestic good price and
PM, - imported good price
Government policy directly affects the domestic price of im-
ported commodities. Adopting the small country assumption, assume fixed
world prices Hm. Denoting ad valorem tariffs by tmi and the exchange
i 
rate by ER, the domestic price of imports is given by:
PM  = Tim(l + tm)ER                  l, . .., n
Since government policy determines tmi and, depending on the exchange-rate
regime, ER, then PMi is fixed. The prices of domestically produced
commodities (PDi) on the other hand, are free to vary so as to equate the
supply of domestically produced goods to the demand, which is also
sensitive to the PMi/PDi ratio. Note that in the models following the
assumptions of pure trade theory, there is no distinction between the
foreign and the-domestic components with a given sectoral aggregation
and hence PM =PD = i( +tm )ER. This leads to the complete determination
of all prices of tradable commodities by the world price and tariff
equations alone, with demand and supply conditions playing no role whatso-
ever in the determination of relative prices of tradable commodities.
The model becomes one in which quantities passively adjust to predetermined



- 14 -
relative prices. This of course greatly exaggerates the actual control
trade policy has over domestic relative prices. In the C.E.S. formulation
adopted here, not only the prices of non-tradable commodities but also
the prices of domestically produced tradables are free to vary and cannot
be tightly controlled through tariff policy, although they will of
course be influenced by changes in the prices of imported commodities
caused by tariff changes or exchange-rate adjustment. But the degree of
autonomy of the domestic price system will be much greater and the links
leading from exchange-rate and tariff policy to domestic prices are
weaker and more complex than in standard trade theory. Only if the
elasticities of substitution are very large will the model behave like
the pure models of trade theory.
2.3     Fixed Exchange Rates and Quantitative Restrictions
While for some purposes a flexible exchange rate is a desirable
model specification, it is often necessary to specify a fixed exchange
rate and allow only discrete, government-determined adjustments to it.
This certainly reflects the realities of the Turkish situation.
Model builders have usually dealt with a fixed exchange rate
by dropping the foreign exchange balance equation (i.e., the foreign-
exchange constraint) from the model while retaining the mechanism deter-
mining imports. The balance-of-payments deficit becomes endogenous and
whatever additional foreign borrowing is necessary is assumed to be forth-
coming. If much too high, the borrowing requirement may then be inter-
preted as a signal suggesting the unfeasibility of the proposed growth path.



- 15 -
Such a treatment of adjustment to a fixed exchange rate has
many undesirable features. It may yield empirically quite unrealistic
paths and also it is very difficult to evaluate the merits of alternative
policy package!s when net foreign capital flows are not held constant
across experimients. To give an example, suppose one wants to compare
the effects of raising income taxes to the effects of raising tariffs
in order to achieve a given increase in government revenue. With a fixed
exchange rate, these two policy changes will generate substantially
different net capital inflows and it will therefore be very difficult
to separate the direct effects pf the tax policy changes alone from
the effects of induced changes in capital flows.
In the Turkey model we have taken a different route. Total
imports (valued in dollars) are set equal to total foreign exchange
earnings composed of export earnings., factor income from abroad and net
foreign capital flows. Capital flows are exogenously given to the model
and will therefore remain constant across experiments. Given that total
imports are thus determined by total exports and that world import prices
are fixed, it can no longer be true that imports are also equal to
"desired" imports as determined by the customs clearance price (c.i.f. +
tariff) and the cost-minimization procedure outlined above. Something
has to give, and there are essentially two alternatives:
Case a.   There is a market for import licenses and/or imports can be
resold. The domestic user price of imports is then bid up over and above



- 16 -
the customs clearance price until actual imports equal desired imports.
Import users are thus on their demand curves and the premium separating
the market price from the customs clearance price accrues to the
recipients of import licenses.    The burden of adjustment thus falls on
the premia which will endogenously move so as to equate desired im-
ports to what is made possible by available foreign exchange earnings.
The user cost of imports will now be:
PMPi   PMI + PR TmER
i  i     i i
where PRi denotes the premium in sector i. It is PMP i/PDi not PM i/PDi
that will determine the import ratio and enter the import demand functions.
If quotas are fixed sector by sector, each individual PRi will have to
move so as to equate sectoral import demands to the quotas. When total
imports are restricted by available foreign exchange earnings, with the
sectoral structure not rigidly fixed by quotas, the average level of
import premia will adjust so as to clear the balance of payments. In
that case, the model behaves as if the exchange rate were flexible but
on the import side only: a systematic bias remains against exports.
Case b.   For many producer goods, the sale of licenses or the resale of
the goods themselves may not be possible. A domestic producer may
receive a license to import specific machinery necessary for the expansion
of his production facility and such a license will not be transferable.
Although the scarcity value of the imported machinery may be much higher,
in this case the user will only pay the c.i.f. + tariff price. Import



- 17 -
users will be "off" their demand curves in the sense that at pre-
vailing user prices the desired sectoral import ratios are higher than
the actual realized ratios. No explicit import premia appear in the
model, but they are implicit in the profits made by producers. A certain
part of these profits must be viewed as rents generated by import
restrictions and the fixed exchange rate.
The user cost of imports now remains what it was before the
introduction of quantitative restrictions, namely
PMi,=     <(  + tm )ER
and desired imports will therefore be:
a       PD
-I      )a      i
i   (1-= i  i       i)ai Di
But the sum in world prices of desired imports may exceed
total foreign exchange earnings available to be spent on imports.     Letting
RM denote the ratio of actual imports to total desired imports we have:
RM   TIM/EflmMD
ii
where TIM-is total foreign exchange available for imports.
A simple allocation rule to determine actual imports in proportion
to desired imports is to multiply desired imports by RM in each sector:
Mi = RM * MDi



- 18 -
In the case of a general foreign exchange shortage, the burden
of adjustment is now on RM which will endogenously move so as to satisfy
the foreign exchange constraint.
The quantity adjustment mechanism outlined above, although
clearly a stylized and simplified story, closely resembles what actually
took place in Turkey during the last few years. Import rationing has
been very important and the assumption that foreign exchange is being
rationed roughly in proportion to the levels of "desired" imports is an
acceptable representation of reality.-L/ The remaining question is
whether, in Turkey, the ultimate users of imported commodities pay the
customs clearance price or a higher price incorporating an importer's
premium.
It is difficult to give a clear-cut answer: the situation has
varied over time and varies across sectors. A first point to keep in
mind is that finished consumer goods constitute an insignificant pro-
portion of Turkish imports, less than 5% on average in the 1970's.
Secondly, while in the early 1960's probably more than half of import
licenses were granted to "importers," this proportion has been dramatically
reduced in the 1970's.!/ For quota list items, the share of importers
has steadily declined and was already down to 23% in 1970 (see Krueger,
page 151), with the remainder going directly to "industrialists."
1/ This has also been true during previous foreign exchange shortages.
See for instance the discussion in Krueger (1974), pages 144-171.
2/ "Importers" are intermediaries and are not the final users of the
imported goods.



- 19 -
Furthermore, the importance of user-specific quotas and also licenses
tied to use, even for items on the liberalized imports list, greatly
increased during the 1970's. Almost all capi'tal-goods imports and a
great deal of intermediate-goods imports took place linked to user-
specific "encouragement certificates" that grant the user various tax
rebates and import duty waivers. It is also illegal for industrialists
to resell their imports.
Thus it may seem that Turkey fits Case b more closely than
Case a and that it is the direct users of producer goods that reap most
of the rents implicit in the restrictive import regime. This conclusion
is probably substantially correct although it should not be overstated:
illegal or semi-legal means of resale exist and for certain intermediate
goods very important trader's premia are being realized.
We have therefore chosen Case b as the basic specification in
the model, leaving Case a and possible mixtures as variants to be ex-
plored. It is worth stressing that this way of modelling a fixed-exchange-
rate regime that allows capital flows to remain fixed across experiments
constitutes an important improvement over usual practice.
2.4     Price-:Level Normalization, Inflation and the Exchange Rate
ks a general equilibrium model, the present model only determines
relative prices. The absolute price level must be determined separately
by an additional normalization equation. A wide variety of price normal-
ization equations have been used in planning models. Johansen, in his



- 20 -
1960 study, fixed the wage of labor and thus expressed all prices in
terms of wages. One could alternatively fix the price of any one
commodity and express all prices in terms of this numeraire.
For a model to be used as a tool of analysis and policy
formulation, it seems best to use a price normalization that provides
a "no inflation" benchmark against which all price changes are relative
price changes. The equation used here will be of the form:
EPii 'PL
ii
where the Qi are weights defining the index FL that one wants to hold
constant. We shall use base year value-added weights. The normalization
equation provides a numeraire and interacts with the real structure of the
model only in the sense that a numeraire does. But it could be used to
become a link between the CGE model and a separate macro-monetary model
that would determine the value of the price index.
Note that it is composite commodity prices that enter the
determination of the price index. Given that the value of the composite
commodity in each sector must equal the sum of the value of imports and
the value of domestically produced goods, the prices of the composite
commodities are given by:
i =[      i   PM  Mi/Di /f i [i/Di, 1]
where i /Di is the ratio of imports to domestic goods.
In the absence of rationing, when cost-minimizing behavior is



- 21 -
unconstrained, the equation for the price of the composite commodities
simplifies to the cost function that corresponds to the C.E.S. aggregation
function:
P   ' 1 ('6T i   ai + tl - d  ai PD    i]l/l ai
Since PMi = TP( + tmi)ER, a devaluation will lead to a general
increase in the price of imported commodities. For the price-level
defined in terms of the P 's to remain constant, a devaluation must be
matched by a decline in at least some of the domestic good prices. The
downward pressure on a given PD will vary directly with (l-.ti), the
initial domestic share parameter and inversely with the substitution
elasticity ai in the aggregation function. But the response of the
domestic price system to a devaluationi will also depend on the various
demand and supply elasticities embodied in the demand system and the
production and factor market equations. Some domestic prices may in fact
go up after a devaluation. The story is further complicated when rationing
is constraining the sectoral import ratios.
What needs to be emphasized at this point is that by normalizing
around the overall price level, we are simply implying that the determination
of the price level is exogenous to the model. If an overall rate of in-
flation of 10% is given to a model, an exchange rate adjustment will
not automatically affect this rate. The fact that there must be
certain specific monetary mechanisms at work that lead to the 10 percent
inflation is undeniable but they are not .explicitly modelled and are
assumed to be separable from the rest of the model. It is of course



- 22 -
perfectly possible to estimate the effect of exchange-rate changes
or other variables on the price-level and then to incorporate these
estimates into the projections of price-level changes that the general
equilibrium model takes as given. We have not so far, however, at-
tempted to endogenize the price-level explicitly. It is projected
separately and while it affects the model, the feedback from the
model to the projections is kept informal.
2.5.    Production and Supply
Production technology is given by two-level C.E.S. production
functions in which "capital" is a fixed proportions aggregate-l and
labor is itself a C.E.S. aggregate of different sub-categories. Inter-
mediate goods are required according to fixed input-output coefficients.
Sectoral capital stocks are fixed in each period and can only be changed
by. depreciation or investment. Short-run sectoral supply curves are thus
always upward sloping. The short-run supply elasticities will vary
directly with the substitution elasticities specified in the production
functions and will also depend on the way labor markets have been modelled.
Given the use of neoclassical production functions and the
assumption of some substitutability between domestic and imported inter-
mediate inputs, supply will never be rigidly constrained by an upper
l/ Note that while no substitution is allowed between "machines" and
"buildings," there is substitution between "foreign" and "domestic"
machines as discussed above in the section on imports.



- 23 -
bound on capacity. Provided prices are bid up high enough, a supply
response will always be forthcoming.-/ While reasonable for medium-term
and long-term analysis, this treatment of supply does not make it possible
to capture the more short-term effects of specific shortages of inputs
and will tend to overestimate the ability of supply to adjust within
one or two year periods. Given the sudden and severe nature of the
crisis that Turkey is going through, it is necessary to modify the
treatment of supply outlined above for those experiments- focusing on
the anatomy of a foreign-exchange crisis. There is strong evidence in
Turkey that the recent severe rationing of intermediate imports has
forced many sectors to reduce significantly their degree of capacity
utilization.
Since the model is to be used to explore the impact of import
rationing on the economy, it seems important to capture its impact
on capacity utilization. The way this is done is to assume that the
productivity parameter in the production function depends on the overall
severity of import rationing and on the degree of dependence of a given
sector on imported intermediate goods. The function is given by:
Ui =: (RM) imi
where Ui is the utilization rate in sector i,
RM is the ratio of total desired to available imports discussed
above,
mi is the ratio of imported intermediates to total intermediate
1/ Of course, for CES production functions, if the elasticity of substitution
is greater than one, the isoquants intersect the axes and there will exist
an extreme wage-rental ratio that will set an upper bound on supply.



- 24 -
inputs in sector i, and
a   is a fixed parameter greater than zero.
Note that Ui is applied to the productivity parameter and
hence affects both labor and capital. Also, since O<RM<l, a>O, and
m >0, then O<Ui<l.
Agricultural labor is treated as separate and immobile within
any period, with endogenous rural-urban migration taking place between
periods. Within the urban industrial sector we distinguish between
modern-sector, largely unionized "organized" wage labor and traditional
"unorganized" labor (small scale enterprise employees, family workers,
self-employed). But rather than explicitly differentiating firms by
size or type, the two kinds of labor are treated as imperfect substitutes
entering the same sectoral production functions. The real wage of
organized labor is exogenously fixed and only parametrically varied. The
real wage of'unorganized labor has been considered, unless otherwise in-
dicated, as fully flexible. The unorganized sector thus absorbs any
surplus urban labor. Under these conditions no open unemployment will
normally appear in the model. The "surplus" labor problem will be reflected
in low wages of traditional labor and consequent urban poverty rather
than in an open unemployment rate that still does not have much meaning
1/
in the Turkish context.-
1/ Official census figures report as unemployed only an insignificantly
small proportion of the labor force. On the other hand estimates of
urban underemployment or labor surplus for the mid-1970's varied
between 12% and 15% of the urban labor force. (See S.P.O., Annual
Programs)



- 25 -
j.6.    The Treatment of Exports
Uncler the small-country assumption, a country's export prices,
ie, are fixed in the world market independently of the quantities exported.
With a constarLt exchangesrate and subsidy system, the per unit revenue of
domestic exports would be given by:
PE1   TrE(l + tei) ER
With PEi > PD   and the assumption of no supply constraints specifically
affecting exports, no domestic sales would take place and whatever is pro-
duced domestically would be exported. In fact, given the C.E.S. aggregation
function that regulates the share of domestically produced output in total
use of any commodity, Di can never fall to zero and we could therefore never
have PEi > PD     The constraints implied by the model on the export side could
thus be summarized as follows:
PE  < PD   and Ei = ° whenever PE    < PD
While faithful to the small-country assumption, this treatment of
exports is quite inconsistent with the "view of the world" implicit in the
specificatAon of product differentiation and imperfect substitution on the
import side. Once one adopts Armington's assumption that products are
differentiated by country of origin, one cannot assume that a world price
exists for the exports originating from an individual country. Rather,
the world price of a certain category of products, H e, is an aggregate
reflecting the C.E.S. aggregation of the various components distinguished by



- 26 -
country of origin while the world price facing the buyers of our country's
specific product can be written as follows:
PWE  = PEi/[ER(l + tei)]                  i-,         n ,
with PEi 3 PDi whenever Ei > °.
Contrary to the small-country case, the chain of causality is thus
reversed, with domestic prices determining export prices rather than the
opposite. The quantity of exports demanded will now be a function of the total
level of world demand for the aggregate commodity in question and the ratio
of our country's export price to the aggregate world price reflecting all
other countries' production costs, trade policies and export prices.     We get:
e
Ei   EBi        i                     i-l,     ., n
where EBi captures the effect of aggregate world demand and the country's
initial share in it, and n is the price elasticity of demand for exports.
The supply of exports is equal to total domestic production net
of domestic use and will therefore normally rise with increases in PD i. Ex-
ports are determined by the interaction of domestic supply and foreign
demand with the foreign demand elasticities and domestic supply elasticities
jointly determining the sensitivity of exports to changes in the structure
of relative prices.
For agriculture, mining and textiles, we have actually retained
the small-country assumption by specifying that the world price of Turkish
exports is exogenous for these sectors. Wheat, ginned cotton and minerals
dominate exports from these sectors and it is reasonable to assume that



- 27 -
Turkey is a price taker on world markets for these commodities. Further-
more, rather than insisting on equality between PDi and PEi, we allow devia-
tions and the share of domestic production going to exports is assumed to
be a symmetric logistic function (with lower asymptote of zero) of the
ratio of PDito PEi.
When analyzing trade policy, the most important effect of having
largely abandoned the small-country assumption on the export side, is, of
course, the now endogenous nature of the terms-of-trade. A devaluation
will worsen the terms of trade and, apart from its more indirect effects,
will lead to a terms-of-trade induced reduction in real income. Note
however that the small country assumption has only been dropped on the
export side: import supply remains completely elastic at given world
prices, a quite realistic assumption.
2.7.    Macroeconomic Aspects and the Flow-of-Funds
The TlT model is a trade and growth model of primarily micro-
economic nature. However, all computable general equilibrium models, by
their very nature, must provide a complete specification of the circular
flow in the economy. Table 1 presents a social accounting matrix (SAM)
which summarizes the aggregate flow-of-funds in the system and which
captures the important linkages in the model economy.11 It presents a
summary picture of the model whose equations are presented in Appendix A.
1/ For a discussion of the conventions of social accounting, see Pyatt
and Thorbiecke (1976).



- 28 -
The SAM distinguishes between "activities" and "commodities."
In the model, "activities" are identical to "sectors" in an input-output
table and produce domestic goods. These goods are either exported or
combined with imports to produce composite "commodities" (as discussed
above). The "activities" purchase inputs (intermediate goods, labor and
capital) and also pay indirect taxes. Factor income is distributed to
two of the "institutions": labor households and enterprises. Enterprise
income is in turn either retained, distributed to "capitalist" households,
or paid to the government in direct taxes.
Columns I and 2 of the SAM summarize the factor and product
markets in the model economy. Row and column 10 summagrize international
trade. The rest of the matrix summarizes the distribution of income and
its allocation. The capital account, in particular, summarizes the deter-
mination of aggregate saving and investment and involves a number of macro-
economic issues that must be discussed in some detail.
The model is essentially savings driven, with the savings
propensities of the different institutions determining the pace of capital
accumulation. Table 2 presents the capital accounts in more detail,
distinguishing among different types of financial flows. The two kinds of
households have been aggregated and a new institution, the "financial
system," has been included to collect savings and distribute them to
enterprises to purchase investment goods. The financial system has no
independent existence in the model economy - it is merely a conduit and
does not appear explicitly in the model equations.



Table 2.1
Aggregated Social Accounting Matrix
\  Expendi-      1       j     23                         4              5           6              7           8             9]1 10
s   Activities     Commo-             Factors:                               Institutions:                         Capital     Rest of
t   AurevitiesIJCaia                                                                                                           Rs
Receipts     \      __________ j dities          Labor       Capital      Lbr.HHshld.         shld. lEnterprises      Govt.      Account      the Ibrld
1   Activities                       Domestic.                                                                                                    Exports
commodity
supplies
2   Commodities         Interme-                                              Consump-       Consump-                   Consump-    Invest-
diate                                                   tion          tion                        tion        ment
i      inputs ____                                                           ____
Factors:
3     Labor            lWages
q   hCapital           Capital
rentals
Institutions:
5      Labor hshlds.                                 Labor                                                              Transfers                Worker's
income                                                                                      remittan-
-_ _.____-_. - . . _  _ .___________                                                                                  _ __I            ces
6      Capitalist                                                                                        Distributed    Transfers                Short-
households                                                                                           income                                term capi
tal in flc
7     Enterprises                                               Capital
.                     ~~~~~~income
8     Government        Indirect      Tariffs                                 Direct         Direct      Direct                     Seigno-      Long-
taxes                                                taxes          taxes       taxes                        rage       term capi
tal ir.fla-Y
9   Capital                                                                   Saving         Saving      Retained       Saving                   Reserve
account                                                                                             earnings                                accumula-
_____  __  _                                             _   ___    _ _ _ _ _       tion
10  Rest of the                       Imports
world
Total Expen-       Total         Commodity       Factor     Factor        Hshld.        Hshld        Enterprise      Govt.     Total         Imports
ditures           costs          supplies       income     income        income        income        income      Expendi-     invest-
ture         ment



- 30 -
Table 2.2
Capital Accounts
enditures                                                 Financial      Rest of
Receipts           Households      Enterprises    Government       System        Economy
Households                                        Seignorage                    Disposable
transfer                       income a/
Enterprises                                                      Total          Retained
investment     earningsb/
Government         Seignorage                                                   Net govt.
revenue C
Financial system   Voluntary      Retained        Government                    Reserve
saving        earnings        investment                    accumulation
Rest of economy    Private         Total          Government     Reserve
consumption    investment      consumption    accumulation
Notes:
a/ Labor income + distributed profits + government transfers + workers' reniittances +
short-term foreign capital inflow - direct taxes
b/  Undistributed capital income (net of indirect taxes)
c/ Direct taxes + indirect taxes + long-term foreign capital inflow - transfers



- 31 -
Households receive their disposable income and are assumed
first to set some of it aside as "idle balances" or "forced saving." This
forced saving is assumed to arise from money creation by the government
which thus receives the proceeds as "seignorage." Households are then
assumed to apply fixed "voluntary" savings rates to their remaining income,
to deposit the saving in the financial system, and to spend the rest on
consumption goods.
The government is treated as a distinct entity. Line 8 of Table 1
gives its revenue sources and column 8 its expenditure categories.
The government sector receives all taxes (direct and indirect), long-term
foreign capital inflow and a share of the seignorage arising from new
money creation. Tax income is determined by applying fixed rates to the
appropriate magnitudes. Long-term foreign capital inflow is set exogenously.
Seignorage income may be determined endogenously and depends on how
government consumption and savings behavior is specified.
The specification of government expenditure behavior represents
the crucial point where various different types of "closure" rules are
imposed on the model economy. We have included different alternatives that
can be used, with different implications for the macroeconomic behavior of
the model. The first alternative is very simple. Government simply divides
its income between saving and consumption by applying a fixed saving rate,
just as do labor and capitalist households.
There are a number of possible variants on the first alternative.
Instead of specifying a government savings rate, one can specify the level



- 32 -
of real or nominal government savings exogenously. In this case, government
consumption is determined residually as government income minus savings.
Another pair of variants would be instead to specify the real or nominal
level of government consumption exogenously and to determine government in-
vestment residually. In all these variants, new money creation -- seignorage --
is either zero or specified exogenously.
The second alternative is to specify the level of government
investment and consumption spending exogenously. Again, there are possible
variants depending on whether government spending is specified
in either real or nominal terms. The problems with this alternative
is that some other element in the flow of funds must now adjust
endogenously in order to balance the accounts. There are a number of different
ways to achieve the necessary balance. For example, one might simply adjust
all tax rates, or some subset of them, so that government revenue equalled
the exogenously specified expenditure. In the Turkey model, a different
approach is taken that incorporates a very simple monetary story and allows
the possibility of incorporating the rate of inflation endogenously into
the model.
The assumption made is that the government simply creates the new
money necessary to finance its target expenditures. The proceeds from the
creation of new money -- seignorage -- are assumed to be shared between
government and capitalists in fixed proportions. This view of how seignorage
is shared is very simplistic and really conceals a host of issues in monetary
theory. In this model, one might assume that some Schumpeterian process is
going on such that the economy will validate the expenditure decisions of



- 33 -
government ancl capitalists and that this process is summarized by the simple
division of the seignorage between government and capitalists.
For the flow of funds to balance, the nominal value of forced
savings by households must exactly equial the nominal value of seignorage given
to capitalist households and to government. Furthermore, under the second
alternative, the value of seignorage accruing to the government is determined
endogenously so that government resources exactly equal expenditures (whose
value was set exogenously). The idle balances held by households are forced
savings that t:hey are required to hold. In--a more:complete model in which
the rate of irLflation depends, in turn, on new money creation, these forced
savings can be seen as reflecting the incidence of inflationary finance on the
different institutions in the economy. The division of the holding of idle
balances between the two kinds of households is by fixed proportions or, in
a different variant, in proportion to their shares in aggregate disposable
income.
The seignorage account can be seen as a kind of transfer mechanism
whereby funds are transferred from some institutions to other institutions.
It is a reasonable fable to identify the transfer with a simple monetary
mechanism, but: one must be careful not to read too much into the fable. It
is certainly not anything approaching an adequate behavioral specification.
For example, t;nere is no consiaeration of the portfolio situation of the
different institutions, nor are there any explicit paper assets being intro-
duced. Indeed, we do not even choose to keep track of the stock of idle
balances accunmulated dynamically by the different institutions since they
have no effect: whatsoever on behavior in the model. Instead, the specification



- 34 -
should be seen as perhaps the simplest possible way to introduce the important
fact that in Turkey government deficits are largely financed by money creation
into the model without adding institutions such as banks or financial inter-
mediaries and without having to build an elaborate model of the financial
sector.
It is important to emphasize that while a flow-of-funds specification
has been added to the microeconomic core of the model and while we attempt
to capture the effects of money creation on government resources and aggregate
investment, we have not attempted so far to endogenize the inflation rate.
The price level as such will simply be projected separately although this
projection can be linked to the variables appearing in the model.
In both Tables 1 and 2, it is convenient to distinguish "enterprises"
and to treat their accounts separately since corporations are behaviorally
distinct and important entities. In the model, however, they do not have
any separate behavioral rules and serve only as a conduit linking the factor
and activity accounts with the household and government accounts. In the
equations presented in the appendix, the enterprise accounts are aggregated
with those of "capitalist households" into a single conglomerate institution.
Thus, for example, the savings of "capitalist households" include retained
earnings of enterprises.
2. 8. Dynamic Linkages
The overall dynamic model is partitioned into a static within-
period general equilibrium model and a separate between-period model which
provides the necessary intertemporal linkages. In general, the role'of



- 35 -
the intertemporal model is to update all the exogenous variables entering
the static model which will then be solved for the next period. In turn,
when various variables are updated for the following period, the inter-
temporal model will take past solutions of the static model as given.
Most of the variables to be updated are projected using simple
time trends, growth rates, exogenous projections, or accounting. The
variables relating to the sectoral allocation of investment and to rural-
urban labor supplies are behaviorally more interesting and important since
they determine the basic structure of the supply of factors in. the next
period.
Theoretically, probably the most satisfying way to model the
sectoral allocation of investment would be to specify both the supply of
and demand for investable funds. Unfortunately, given the data, it would be
impossible to implement a complete model of the loanable funds market in
Turkey. Instead, we have chosen a much simpler approach in which we assume
that there is a "normal" or historical set of sectoral allocation shares
which are modified over time as a function of the relative profitability
of different sectors. The normal shares are given by the last period's
sectoral shares in the aggregate capital stock.l/
The simple investment model has been formulated in a lagged
version as part of the intertemporal model.    However, once we give up the
1/ This approach to determining investment allocation is- identical to that
used by De Melo and Dervis (1977) except that they defined "normal" in-
vestment shares as equalling sectoral shares in total profits rather than
in total capital stock. Using capital stock shares fits the Turkish
historical data much better.



- 36 -
notion of an intertemporally efficient investment allocation procedure --
some kind of intertemporal tatonnement process -- we could in fact in-
corporate the determination of investment in the within-period model.
Simply use current instead of past profit rates and include the equation
in the within-period model. On the assumption that there are no serious
oscillations in the underlying technological and taste parameters, not much
should change. Experiments with the model indicate that it makes little
difference empirically which approach is used.
The rural-urban migration model treats migration as being a function
of the differential between the rural and urban wages. Migrants are
assumed to be attracted by the average urban wage compared to the rural
wage. Total rural and urban labor are assumed to have exogenously
specified natural rates of growth. The migration equations are given by:
MIGt a £     - 1]LA
w    =  (W2L2i + W3L3i)/LU
where  W     is the rural wage,
LA   is agricultural labor,
U
L    is urban labor,
W2   is the wage of   organized urban labor (L2i),
W3   is the wage of unorganized urban labor (L3 ), and
£    is the migration response parameter.



- .37 -
The average wage W can.be interpreted as the
expected wage a.new migrant would receive if his probability of employment
in the organized and unorganized labor markets were equal to their shares
in the total urban labor force. It is also consistent with the way Harris
and Todaro (1970) define the expected urban wage since we assume that any
excess supply of labor to the organized market (where the wage is fixed)
is simply absorbed into the unorganized market. There are, of course, a
number of alternative ways to define the probabilities of entering the
organized or unorganized markets, but this approach is both reasonable in
the Turkish context and empirically implementable.-/
This concludes the description of the distinctive features of
the TGT model. A full statement of the equations is available in
Appendix A.  The model has been applied to analyze the actual
developments in the Turkish economy and to provide policy conditional
projections for the medium-term future. Part 3 below turns to a model-
based discussion of the makings of the 1977 crisis. Parts 4 and 5 go on to
evaluate future prospects and analyze the impact of alternative policy
packages.
1/ See MundLak (1976) for an empirical analysis.



- 38 -
3. The Origins of a Crisis: An Analysis of the Turkish Economy from 1973 to 1977
3.1 Introduction
The 1977 crisis, like its predecessors in 1958 and 1970,
appears primarily as a foreign exchange crisis. By the end of 1977
Turkey, with an import bill of 5.8 billion dollars, had only 1.7
billion dollars of exports. Workers' remittances provided another billion
dollars, leaving a huge 3 billion dollar gap to be financed. The gap,
which had been of similar size in 1976, was financed by massive short-term
external borrowing and a complete running down of foreign exchange re-
serves. By the end of 1977, the situation had become one of acute crisis,
with foreign lenders declining to make further loans, commercial arrears
close to 2 billion dollars, and the economy unable to continue to grow
without imports that could no longer be financed.
After the crisis had reached these alarming proportions, a new
government under Prime Minister Ecevit was formed in January 1978.
The Ecevit government has embarked on a new program of readjustment and
stabilization that will have a large impact on Turkey's economic
performance in the next few years. The beginning of the Fourth Five
Year Plan has been delayed a year, until 1979, when it is hoped that the
worst part of the crisis will have been overcome.
In this section we will analyze the determinants of the 1977
crisis and its immediate impact. Part 4 will turn to prospects for
the future, focusing on the Fourth Five Year Plan period from 1979 to 1983.



- 39 -
3.2 Summary of Recent Events and Policy Reactions
The period from 1970 to 1977 marks the third cycle in Turkey's
post-war path of industrialization (see Figure 1, page 4). We shall briefly
describe events in this period that started with the 1970 devaluation and
led to the present crisis.
The latter half of the 1960's was characterized by severe
foreign exchange shortages and consequently increasingly severe import
rationing. While growth performance particularly in the industrial sector
was impressive, exports virtually stagnated. Between 1960/61 and 1969/70,
exports increased at an annual rate of only 5.9 percent in current dollar
value which reflects near stagnation in real terms. Over the same period
imports could ionly grow by 6.7 percent in current dollar value or about
3 percent in real terms compared to an average annual growth of real
GDP above 6 percent and an annual industrial growth rate of about 10 percent.
Whilea foreign exchange shortages were chronic and net incentives
had drifted more and more against exports, the situation in 1970 was
far less serious than it had been in 1958. The main reason was probably
a still small but significant flow of workers' remittances, averaging
about 100 million dollars a year since 1965, which compensated for about
40% of the trade deficit. Import substitution was also proceeding relatively
successfully, particularly in transport equipment and machinery, and the
foreign exchangye situation was not really deteriorating rapidly. The
timing of the devaluation that occurred in August of 1970 must be explained



- 40 -
as much by political as by purely economic factors. To quote Anne Krueger,
"the fact that a foreign exchange shortage had continued for so long
meant that it could continue longer."-l/
It is quite possible that it was the dismal performance of
exports in general and of manufactured exports in particular that con-
stituted the single most important factor leading to the 1970 devaluation.
At the State Planning Organization in particular, the failure of any
manufactured exports to materialize was perceived as a serious bottleneck
to further growth and as an indication that Turkey's industrial development
was lacking an important dimension. In fact, when the decision to devalue
was made in 1970, it was accompanied by a significant increase in the
average value of subsidies to manufactured exports, a fact indicating
that the need to start exporting manufactured products was strongly felt'
by policy makers. The 1970 policy adjustment was a substantial one, in-
creasing the effective exchange rate for imports by about 50%, for tra-
ditional agricultural exports by 28% (tobacco, hazelnuts, dried fruits,
raw cotton) and for manufactured exports by 57%.2/
The devaluation was followed by three years of extremely rapid
increase in foreign exchange receipts which made possible not only an
unprecedented increase in imports but also an even more dramatic ac-
cumulation of foreign exchange reserves.
1/ See Krueger (1974), page 312.
2/ See Krueger (1974).



- 41 -
Exports, in sharp contrast to what had happened after 1958,
responded vigorously to the 1970 exchange rate adjustment. Their total
value increased from 537 million dollars in 1969 to 588 million in 1970,
677 in 1971, E885 in 1972 and 1317 million dollars in 1973. This
represents an annual average growth of 25 percent. Turkish exports had
been close to 300 million dollars in 1950 and 1951. Thus in the two decades
from 1950 to 1969, total export value failed to double, increasing by only
80% over 19 years, a growth of only 3 percent per annum. In contrast
the near tripling of exports earnings between 1969 and 1973 constituted
an unprecedented achievement.
The overwhelming source of export expansion in the early 1970's
was in the food processing and textile sectors. Exports of processed
food products increased from 200 million dollars in 1969 to 390 million
dollars in 1973. Exports of textiles including ginned cotton increased
from 127 to 391 in the same period. But what is equally significant is
that from a base close to zero, significant exports appeared in the following
categories of manufactured products: clothing, footwear, inorganic
chemicals, cement, glass and glassware and metal products.
Thuis the 1970-73 period can be taken as an indication that the
potential fox export expansion in a wide range of manufactured products
exists in Turkey provided the structure of incentives is conducive to such
an expansion and provided foreign market conditions are appropriate.
Unfortunately neither the structure of incentives nor foreign market con-
ditions remained conducive to export expansion for more than a few years.



- 42 -
In addition to increased exports, Turkey aquired foreign exchange
from workers' remittances which increased from 141 million dollars in 1969
to 273 in 1970, 471 in 1971, 740 in 1972, and 1183 in 1973. This represents
an annual average growth rate of 70 percent. By 1973, the flow of remittances
was financing half of imports. The increase of remittances was only
partly a reaction to devaluation. Between 1969 and 1973, the number of
workers abroad itself grew at an average annual rate of about 35 percent.
On the rough assumption that remittances are proportional to total
income earned abroad and noting that nominal income per worker measured in
dollars grew at an annual rate of at least 10 percent one would estimbate
a 50 percent annual growth rate. To this was added the redirection into
official channels that no doubt followed the devaluation as well as some
repatriation of accumulated savings that constituted a direct response
to'the exchange rate adjustment. The great surge in remittances was
iramatic and largely unexpected.-/
Inflation was moderate in the 1960's, accelerating somewhat towards
the end of the decade, but averaging only about 5 percent per annum from
1960 to 1970.  The situation changed sharply after 1970.    The wholesale
price index rose by 16 percent in 1971, 18 percent in 1972 and 20.5 percent
in 1973, making for an average inflation rate of 18.2 percent per year in
the 1970-1973 period. During the same period worldwide inflation, expressed
in dollars, also increased substantially but did not average more than
1/ The Third Five Year Plan prepared in 1972 underestimated remittances
in 1973 and 1974, projecting them at one third of their actual value.



- 43 -
10 percent per year.l/ Relative incentives had thus significantly drifted
against exports by 1973 and the price deflated exchange rate had been
revalued by about 30 percent.-/
Between 1973 and 1976, the annual spread between Turkey's
domestic inflation rate and worldwide inflation remained between 8 and
10 percentage points. Turkey did start a series of minor exchange rate ad-
justments in this period, devaluing the Turkish Lira against the dollar
by an average of 5 percentage points a year, which did not fully compensate
for the inflation differential. The upward drift in the price-deflated
exchange rate thus continued, slowly but steadily, and by the end of 1976
the real exchange rate was therefore again close to what it had been before
the 1970 devaluation.
Since 1974, export revenues have grown only very slowly azid
erratically, officially recorded workers' remittances have declined by
about 40%, while imports almost tripled between 1973 and 1977. Figure
2 illustrates these divergent trends.
1/ See "Histo-rical Rates of Change in US$ GDP Deflators: 1961-1975,"
World Bank, Economic Analysis and Projections-Department, February 1978.
2/ The Turkish Lira was actually formally revalued by 6 percent against
the dollar between 1970 and 1973.



Figure 2: Exports. Imports and Workers' Remittances: 1970-1977
1 on Dllar3
pr s
rlces  iR ittances1
ol O"00 ~ ~ ~ ~  ~    ~   ~    I
II                                                        I~~~~~~~~~~~~~~~~~~0
pI 
-I--                                         I.      II~~~~~~~~~~~~~~~1
.1  I             I          .    .~~~~~~~~~~~~~0



- 45 -
As can be observed from Figure 2, exports and imports were
growing at about the same rate between 1970 and 1973 while workers' re-
mittances grew somewhat more rapidly with the sum of exports and remittances
actually overtaking the value of imports in 1972 and producing a sizeable
current account surplus in 1973. Imports jumped from 2.1 billion dollars
in 1973 to 3.8 billion in 1974, growing by 80% in one year. The foreign
exchange gap created by this sharp increase in imports in 1974 was
still however moderate, consisting of 819 million dollars or about 2.8
percent of GDP. It is in 1975 that the danger of a major foreign ex-
change crisis became apparent. In that year, both exports and remittances
declined while imports continued to grow rapidly, increasing by 25% over
1974. The foreign exchange gap (imports - exports - remittances) reached
2 billion dollars or about 5.4 percent of GDP.
It is thus quite clear that the crisis was already apparent in
1975 and that the level'and growth rate of imports experienced in 1974
and 1975 were inconsistent with the amount of export earnings and remittances
that materialized. The situation did not improve in 1976 or 1977. On the
contrary, in 1977 the foreign exchange gap reached 3 billion dollars
which at the March 1978 exchange rate of 25 TL to the dollar constituted
about 9 percent of GDP.
The gap was temporarily closed by massive international borrowing
and the running down of the substantial foreign exchange reserves that had
accumulated in 1972 and 1973. 'But by the end of 1977, there were no more



- 46 -
reserves to be run down and Turkey's borrowing capacity had reached its
limits. The situation was no longer tenable and a major readjustment
had become inevitable.
While the upward drift in the price deflated exchange rate and
the anti-export biased shift in incentives resulting from Turkey's high
inflation rate constitute one major element explaining the foreign exchange
crisis, there have been other important developments. A significant part
of the 1974 upward jump in imports that is apparent in Figure 2 can be
attributed to the oil price increase. Turkey imported about 70 percent
of its oil needs during the 1974-1977 period and there is no doubt that
the oil price rise has had a major adverse impact oh the balance of payments
and the economy. In fact the Turkish government tried to insulate the domestic
economy from the effects of the oil price increase by setting up a special
fund to subsidize the price of gasoline in the domestic market. While this
probably helped keep up real wages and profits domestically, at least for an
interim period, it probably heightened the impact of the oil crisis on the bal-
ance of payments since it weakened any possible substitution effect against
oil-intensive activities. By 1977, oil imports were almost equal in value
to total merchandise exports! But how important has the oil price increase
really been in explaining the present crisis? Has it been the major cause
of the great widening in the foreign exchange gap?
Another development in the middle 1970's that has been suggested
as one of the contributing causes of the current crisis is the major invest-



- 47 -
ment program undertaken by the government. The aggregate investment rate
(as a proportion of GDP) increased steadily during the period, from about
18 percent to 24 percent. Since investment is relatively import-intensive,
this increase has led to additional strain on the balance of payments.
It is important to attempt to explore the relative importance of
the different shocks that the Turkish economy has undergone in recent years
in order to understand the nature of the present crisis and to evaluate
future policy choices. We have used the TGT model to explore quantitatively
the relative impact of the different factors by means of' a number of counter-
factual or "as-if" experiments. These experiments are described below.
We first consider the impact of events on the equilibrium exchange rate
and then discuss their impact on the general performance of the economy.
3.3 Exchange Rate Drift and Structural Imbalances: Decomposing the Change
in the Evuilibrium Exchange Rate
The major question to be explored in this section is what would
have happened,if Turkey had pursued a flexible exchange rate policy during
the 1973-77 period. Would the "equilibrium" or "shadow" exchange rate
have had to depreciate at only a rate equal to the difference in inflation
rates between rurkey and its trading partners? How did the oil price
rise affect the equilibrium exchange rate? How different would economic
performance have been?
Instead of assuming that import rationing is the mechanism by
which the balance of payments is equilibrated, one can run the TGT model
with a flexibLe rather than fixed exchange rate. The exchange rate ad-



- 48 -
justs in each year so that total demand for foreign exchange equals the
total supply. However, the "equilibrium" exchange rate determined by equating
the annual demand and supply flows of foreign exchange will be sensitive
to the exogenous flows of foreign capital and reserve accumulation. For
example, in 1976 and 1977, Turkey borrowed massive amounts of foreign ex-
change which are given in the basic run but which in no way reflect some
it
normal" or "equilibrium" amount of borrowing. The result in the basic run
is that there is little import rationing until 1977. Running the model
with a flexible exchange rate but assuming the same massive borrowing
as in the basic run yields a market-clearing exchange rate in 1977 about
6-7 percent higher than-the fixed rate of 18.2 in the rationing-constrained
basic run..
To compute an "equilibrium" exchange rate, one thus has to
assume an amount of borrowing and a level of reserves that can be sus-
tained over time. It is not obvious what such a "normal" or "sustainable"
level of reserves and borrowing would have been, but one can make certain
rough assumptions. Regarding the stock of reserves, it is reasonable to
assume that it should equal about 30% of annual imports.-/ At the end
of 1972, Turkey's foreign exchange reserves stood at 1.2 billion dollars
compared to an import bill in 1972 of 1.6 billion dollars. The stock of
reserves was already too high. But another 700 million were allowed
1/ In 1973 this ratio has been 3-7.for Greece, 31% for Yugoslavia, 69% for
Spain, 61% for Germany, 23% for India and 50% for Mexico.
Source: IMF International Financial Statistics.



- 49-
to accumulate during 1973 leading, at the end of 1973, to one of the
highest reserves-to-imports ratios in the world.    Thereafter, starting in
1974, Turkey started running down its reserves, arriving in 1977 at a
reserves-to-imports ratio below 10%.
For the equilibrium exchange rate experiments, we assume that
instead of first accumulating and then decumulating reserves, Turkey keeps
its 1972 stock: of 1.2 billion until the end of 1975. By that time, the
reserve-to-imports ratio reaches 30% and, to preserve that ratio, reserves
have to be slowly accumulated from 1976 onwards.
With respect to borrowing, it is more difficult to define a normal
level. In Turkey's case, it is possible to argue that until 1975, the
total amount of debt and the levels of annual borrowing were too low
rather than too large given Turkey's size and total foreign exchange
earnings. But the situation was reversed quite dramatically in 1976 and
1977, with massive borrowing leading to a debt/GDP ratio close to 40%
by the spring of 1978. Taking the 1973-1977 period as a whole, Turkey
borrowed at levels that could not be sustained over time given the trend
increase in exports and the flow of workers' remittances. To compute an
"equilib_ium" exchange rate, we therefore assume that a "normal" level
of borrowing would have implied a total cumulative net flow of 3.5 billion
dollars instead of the 5.5 billion that were actually realized. Starting
from 500 million dollars in 1973, the net foreign capital inflow is
assumed to grow at 17% annually (about 7.0 percent in constant dollars)



- 50 -
reaching 940 million dollars in 1977 and summing to a total cumulative
flow of 3.5 billion dollars. Table 3.1 compares capital inflow and
reserve decumulation in the basic run, which reflects the actual events
over the 1973-1977 period, with the figures assumed for the experiments.
Table 3.1
Flexible Exchange Rate Experiments:
Reserve Decumulation and Net Capital Inflow
______________ __  1973  1974  1975    1976   1977   Cumulative
Basic Run
Reserve decumulation    -716    433    222    262    550      1,251
Net capital inflow       222    129  1,011  1,740  2,385      5,493
Sum                    -488     562  1,733  2,002  2,935      6,744
All Experiments
Reserve decumulation       0      0      0   -200   -250       -450
Net capital inflow       500    585    685    800    940      3,510
Sum                     500     585    685    600    690      3,060
Units: million dollars
Once the path of what is considered to be an equilibrium level
of reserves and foreign borrowing is specified, it is possible to use the
flexible exchange rate version of the TGT model to explore the values that
the market-clearing or "equilibrium" exchange rate would have taken if it
had been allowed to move to equilibrate the supply of and demand for



- 51 -
foreign exchange.  We are not here referring to a shadow rate in the sense
of a free trade exchange rate, but to the rate that would have been an
equilibrium rate given Turkey's structure of tariffs and in the absence
of quantity rationing.
In the basic run, which is used as a reference point for compari-
sons, the exchange rate is fixed at the official parity levels in each ,
year, as are also the levels of capital inflow, remittances, and reserve
decumulation. The intent of the basic run is to approximate as closely
as possible the actual path the economy followed in the 1973-77 period.
All the flexible-exchange-rate experiments were conducted with the assump-
tions about normal levels of foreign borrowing and reserve decumulation
discussed above. In the first experiment (A-1), there are no other
changes. It thus provides a reference path of equilibrium exchange
rates given the shocks that the economy actually underwent, but assuming
a flexible exchange-rate policy and more normal capital and reserve
behavior. The equilibrium exchange rates in each year from the experiment
are compared wilth the actual rates in Table 3.2.
Table 3.2
Experiment A-1: Equilibrium Exchange Rates, 1973-77
Experiment A-1:     Basic run:
equilibrium rates   official parity
1973                  10.4                 14.0
1974                  14.7                 13.5
1975                  17.8                 14.5
1976                  20.3                 16.0
1977                  28.2                 18.2



- 52 -
According to the TGT model the equilibrium exchange rate was
10.4 in 1973, substantially lower than the official parity of 14 TL
to the dollar. Thus, in terms of flow-equilibrium conditions, Turkey
seems to have had a significantly undervalued exchange rate in 1973.
While this may, at first, seem surprising, one should remember that
Turkey accumulated 1.5 billion dollars of reserves between 1971 and
1973. The reserves to imports ratio rose to above 100%, far in excess
of what can be considered normal or required. Furthermore, net
borrowing in that period was minimal. The explanation for this dramatic
reversal of the chronic foreign exchange shortage that had characterized
the post-war period is primarily to be found in the massive increase of
workers' remittances. To this must be added the very good export
performance in the early 1970's.
However, the situation did not last. Starting in 1974, the
downward trend in the equilibrium value of the Turkish Lira is very steep
and the degree of overvaluation steadily increased. Note that the single
biggest percentage change occurred between 1973 and 1974, reflecting the
impact of the oil price increase. By 1977, the TGT model results indicated
that the Turkish Lira was overvalued by more than 50%.
In experiments A-2 through A-5, we progressively add the major
causal factors discussed above in order to explore their separate contri-
butions to the decline of the equilibrium exchange rate. Table 3.3 gives
a summary description of the experiments and the equilibrium exchange rate
in 1977. Experiment A-2 is the same as A-1 except that the domestic
inflation rate is set equal to the world inflation rate throughout the



- 53 -
period. This experiment thus isolates the effect of differential
inflation rates in explaining the depreciation of the equilibrium
exchange rate. The third experiment (A-3) is the same as A-2 except
that there is no special rise in the world price of oil. These experi-
ments are designed to isolate the most important factors affecting the
equilibrium exchange rate discussed in the previous section.
Two additional experiments were run to explore the role of
other factors. Experiment A-4 is the same as A-3 except that the invest-
ment rate is held roughly constant throughout the period at about the 1973
level of 18 percent of GDP. Experiment A-5 is the same as A-4 except
that remittances continue to grow instead of stagnating after 1974.
Their growth rate is assumed equal in real terms to the economy-wide
growth rate.
Table 3.3
Flexible Exchange Rate Experiments:
Summary Description and 1977 Equilibrium Exchange Rate
Exchange
rate: 1977            Experiment                      Description
18.2                  Basic Run           Fixed exchange rate, historical run
28.2                     A-1              Basic run + flexible exchange rate
+ moderate borrowing
20.6                     A-2              A-1 + moderate inflation
18.3                     A-3              A-2 + no oil price increase
17.3                     A-4              A-3 + no investment rate increase
14.0                     A-5              A-4 + higher remittances



- 54 -
The experiments described above enable us to decompose the
change in the equilibrium exchange rate that occurred between 1973 and
1977. The diagram below plots the equilibrium exchange rates in 1977
for experiments A-1, A-2, and A-3 along a line whose origin is the
equilibrium rate in 1973. The change in exchange rate between adjacent
points is attributable to the single effect which is different between
the corresponding experiments.
10.4                  18.3     20.6                   28.2
A-3      A-2                    A-1
_____ Other             Oil   ___ Differential      4
factors          price        inflation
Reading from right to left, the devaluation due to:
differential inflation = 28.2/20.6 = 1.369
oil price rise          = 20.6/18.3 = 1.126
other factors           = 18.3/10.4 = 1.760



- 55 -
Or, to express the relationship anoth,er way, the equilibrium exchange
rate in 1977 is given by the expression:
ER1977 = (1.369) - (1.126) * (1.760) = 2.712    ER1973
The percent chiange in the exchange rate is given by taking the logarithm
of both sides of this expression and subtracting the log of ER1973 from
both sides. Ihe shares of the logarithms of the three terms in parentheses
as a percent cif the logarithm of their product (log of 2.712) provides
a decompositicon of their relative shares of the total change in the
equilibrium exchange rate. Thus, the shares of the three effects in the
total change in the equilibrium exchange rate are:
differential inflation:   32%
.oil price rise         : 12%
other factors          : 56%
total          : 100%
This decomposition indicates that differential inflation was more important
than the oil price rise in determining the change in the underlying value
of the equilibrium exchange rate. The oil price rise in itself led to a
20 percent depreciation of the equilibrium rate while differential inflation
caused more than a 50 percent decline. The residual in this decomposition
is the change in the equilibrium exchange rate due to all causes other than
the oil price increase and the differential inflation rate. It includes
some of the indirect effects of the oil crisis such as changes in world
market demand for Turkish exports, particularly textiles as well as lower
workers remittances, both in part, but not wholly, due to the deflationary
impact of the oil crisis.



- 56 -
Experiments A-4 and A-5 can be used to explore further the
"other factors" that seem to be significant in the 1973-77 period. The
1977 equilibrium exchange rates for experiments A-3, A-4, and A-5 are
plotted below:
10.4                         14.0             17.3        18.3
I                           I                I           j
A-5              A-4          A-3
<    -   Residual -          *-Remittances-2>   Invest- -
ment
The shares of these effects in the change in the exchange rate
due to "other factors", and in the total change, can be calculated exactly
as was done above. The results give the relative contribution of the
different effects to:
"other factors"            total
devaluation          devaluation
increased investment rate        10%                  5%
remittance deceleration          37                  21
residual factors                *53                  29
sum          100%                  56%
In experiment A-4, the investment rate is kept at roughly.its
1973 value throughout the period. The reduced levels of capital-goods
imports leads to some lessening of pressure on the balance of payments, but
this only accounts for 10 percent of the "other factors", or 5 percent of
the total 1973-77 change in the equilibrium exchange rate. It has half






- 58 -
required devaluation to differential inflation very seriously under-
estimate the degree of adjustment that is really required. The oil price
increase alone would have required a significant exchange rate adjustment.
A continuous upward trend in Turkey's investment rate and changes in
sectoral structure would also in themselves require some adjustment.
Finally, the role of workers' remittances is very important. If
everything in the economy, including exports and imports, were growing
with GDP at around 8% annually but workers' remittances remained
stationary, it is clear that the foreign exchange gap would steadily grow.
Thus, unless remittances can continue to grow at a rate close to that
of other variables in the economy, exports must grow more rapidly than
imports to preserve equilibrium. In turn, either the real exchange
rate or other policy variables will have to change continuously to
generate this more rapid growth of exports. This is a simple fact brought
out clearly by the experiments described above. Unfortunately, it has
not been stressed sufficiently in the policy discussions about Turkey.
We do not necessarily want to imply that remittances should be expected
to remain stationary in the future. But given their initially very high
level, close to total exports, their failure to grow in the 1973-1977 period
should have been compensated by growth of exports significantly in excess
of the growth of imports. The opposite occurred, leading to the present
crisis.
Concluding this section, it is worth emphasizing once more that
looking at differential inflation rates is not an adequate way to evaluate
the degree of overvaluation reached by a currency. In Turkey the oil
price increase, the increase in the investment rate and the behavior of



- 59 -
remittances were together much more important factors leading to the
dramatic foreign exchange gap causing the 1977 crisis. Finally, note
that an exchange rate policy aimed at continuous annual flow equilibrium
may not necessarily be the best policy to follow at all times. This
section should not be interpreted as advocating such a policy as always
optimal. But better knowledge of what is the equilibrium rate and a
better understanding of the factors that cause its change over time
would seem to be prerequisite for adequate planning and policy formulation.
The steep downward trend in the equilibrium value of the Turkish Lira has
been seriously underestimated by those using informal methods, a deficiency
which affects policy formulation and discussions about the future.
3.4 Growth, T'rade and Structure: 1973-77
The experiments described in the last section can also be used
to explore the impact of the various effects they embody on the grow'th and
structure of t:he Turkish economy as well as on the equilibrium exchange
rate and the balance of payments. In this section, we use the experiments
to explore the impact of the oil price rise, the investment boom, and the
foreign-exchange constraint on the economy.
In evaluating the impact of the individual effects introduced in
each experiment, there are two different kinds of comparisons to be made.
First, the results from each experiment can be compared with those from
the basic run which is designed to replicate as closely as possible the
actual path the Turkish economy followed during the period. Second, the
incremental impact of a given factor can be determined by comparing the
results with those from the adjacent experiment which differs by only the






-61 -
Table 3.4
Flexible Exchange Rate Experiments:
1973 Values and Ratios to Basic Run
Basic          All
1973:  Values                                    run        experiments
Export/import ratio Ma/                           62.3         36.6
Investment/GDP ratio (%)b/                       17.6          18.1
Exchange rate (TL/$)                             14.0          10.4
Balance of merchandise trade-'                    0.8           1.7-
C'
Capital inflow + reserve decumulation-          - 0.5           0.5
GDP-/                                            301.3        304.1
1973: Ratios to basic run values (%)
Export/import ratio                             100            59
Investment/GDP ratio                            100           103
Exchange rate                                   100            75
Balance of merchandise trade                    100           213
GDP                                             100           101
a/Merchandise exports/merchandise imports.
b/Gross fixed capital formation/GDP
c/Units are billion dollars
-/Units are billions of 1973 Turkish lira.



- 62 -
Tables 3.5 and 3.6 provide the basic macroeconomic results of
the experiments. Comparing the basic run with experiment A-1 (which
differs only in assuming more moderate borrowing and a flexible exchange
rate), it is interesting to compare the trade flows. Cumulative exports
are about a billion dollars higher over the period while cumulative
imports are almost 2 billion dollars less. The rate of growth of exports
in the basic run is 7.8 percent a year, slightly less than that of GDP.
In experiment A-1, the rate of growth of exports is much higher, 26 percent
a year, although starting from a much lower base year value. Starting from
the basic run value of exports in 1973, the rate of growth in experiment
A-1 is 17.4 percent a year. While high, such rates are comparable to
those for other countries and to recent Turkish experience. As noted
above, Turkey achieved a rate of growth of exports of 25 percent a year
in the 1969-1973 period.
Comparing experiment A-1 with the basic run indicates clearly
how dependent Turkish growth was on the massive inflow of foreign borrowing
during the period. The 1973-77 growth rate in experiment A-1 is 6.7
percent, compared to an actual rate in the basic run of 8.0 percent.
Aggregate consumption grows by only 3.4 percent a year, compared with
7.6 percent in the basic run. In 1977, aggregate consumption is 20.9
percent lower than the basic-run value.
Experiments A-2 through A-5 can be used to explore the impact
of various changes on the economy. In experiment A-2, the rate of
inflation was assumed to be lower (11.7 percent a year compared to 21.0
percent in experiment A-1).   This experiment had virtually no effect on



Table 3.5
Flexible Exchange Rate Experiments:      1973-77 Data
Basic Run        A-1           A-2          A-3          A-4           A-5
973-77: Cumulativc flows (billions $)                R r    I             I S  I        |     3 5    1     3-5          3.5
Foreign cavital inflow                                   3 _  -                           -_  _ ._  _-__.__          _
Reserve decumulaion.                        1.2          -.4          -.4    f      --            -.4          -.
Sum                                         6.7          3.1          3.1            3.1    J     3.1          3.1
Workers' remittances                  J     5.9          6.4          6.4     1      6.4          6.4          8.5
Pezrcleum ir?orts                     I     4.7          4.4     1     42.2                       2.2          2.3
Machinery & transport                       8.0          73           7-3            8.0          7.8    j      .3
ecui=2ent imoorts                          8                     , ,                 0                        _
|Totai r merchandise exports                7.9          8.8          8.8            8.1     !     .9    j     7.3
Total merchandise imoorts                  20.8         19.0         19.0     | -  18.3     1    18.1    !     9.3
Balance of merchandise trade               12.9         10.2     |   10.2           10.2         10.2         12.0
1973-77: Growth rates (Z)
CDP                                         8.0          676.7                       7.7          7 .4         7.9
Consumption                                 7.6          3.4          3.4            5.4          6.0    i     7.3
investment                                 14.5          6.7     !    6.8            8.3                 1 5_4_7.3    !
Domestic inflation        '                21.0         21.0         11.7           11.'7        11.7         11.7
Exchange rate                               6.8     |   28.3         18.7      |    15.2         13.5          7.8
Notes:   All experiments assure moderate borrowing (see text) and a flexible exchange rate.
The experiments are:    A-i:   Basic rum. except for moderate borrowing and flexible exchange rate.
A-2:   Same as A-1 but includes moderate inflation.
A-3:   Same as A-2 but includes no oil price increase.
A-4:   Sane as A-3 but includes no increase in investment rate.
A-5:   Same as A-4 but includes higher remittances.



Table 3.6
Flexible Exchange Rate Experiments:    1977 Values and Ratios
Basic Run      A-i           A-2         A-3          A-4          A-5
1977:  Val ues s                                                             II
E:.:-?ort!i:ort retio  (Z)                 1   37.8         52.5        52.0          50.0         48.5         36.1
Tnves   n.ent/GD? ratio  (Y)     -             23.7         23.5        23.4      1   22.2         18.7     1   18.5
Price index (1973 = 100)                      214.1        214.1       155.4         155.4        155.4     1 155.4
Exch2-.ge ra_e (-LI$)                      |   18.2         28.2         20.6         18.3     i   17.3         14.0
Bai2r.ce of nrcha ..d........... . .   d-se trade a /  -..... .  t 4.0  -  2.2  |  2.2  2.2  j  2.2  3.2
Ca?ital stock index (1973     100)            120.8        119.4       119.3      r  120.9        125
CGD? b/                                       410.0 i      394.1       393.5         409.7        405.2        412.1
1977:   RezioZ co bzasic run v2e (,)              i            i                                                   -
P?rice index                                  100          100           73           73           73           73
Exch.ange rate                            1100          | 155          113           101       |   95           77
Balance of merchandise trade                  100           55          55            55           55           80
Ca-ita1 stock                                 100           99          99         i10         1O
100       j   90       190               97         100       j105       *
Investmer.t                                   100           84          84            90           8_           86
CD?   -                                    |     _100    1  96          96         100             99          I 10
1977:   Ra.ies co adi2cant exneriment (%) c/      i           -
?rice index                                    -        | l00           73           100          100          100
Excaange raze                                  -       {   155          73        !   89           95           81
Balance of merchandise trade                   -       .1   55        100            150          100 i        145
Capital stock.<                                _            99      1|00             101           99          101
Consunlption                                   _            90       [100         I  108          102          105
invesz-en.                                                  84       |100            108           88          107
GD? P                                          -            96         100           104       |   99          102
a! Units are billion dollars
b/  Units are billions of 1973 Turkish Lira
c/  Column for A-1 is rat-o of A-1 values to basic run, column for A-2 is ratio
of A-2 values to A-1 values, and so forth.



- 65 -
any of the real variables In the economy since, with a flexible exchange
rate, there was no change in relative prices. This experiment emphasizes
the fact that: the TGT model does not behave like a standard macro-economic
model. Its equations are almost homogenous of degree zero in all prices
(including the exchange rate). As described in section 2 above, the
model is designed to be linked with a separate macroeconomic model through
variables such as the aggregate price level, but they have been treated
as exogenous variables in the TGT model. Note, however, in the fixed-
exchange-ratet experiments described in the next section, the real variables
in the model are sensitive to the aggregate price level since changing
the price level changes relative prices.
Experiment A-3 indicates the importance of the oil price rise
to the economiy. Eliminating the price rise leads to a 1 percentage point
increase in the growth of GDP (from 6.7 to 7.7 percent a year). The
effect on the cumulative dollar value of oil imports is also quite
dramatic -- they fall by 2.1 billion dollars over the period. The total
volume of trade also falls -- both exports and imports are 0.7 billion
dollars less -- indicating that oil is a necessity for which it is
difficult to substitute and impossible to expand domestic production
cheaply. When the price rose dramatically in the 1974-75 period, it was
therefore necessary either to expand exports or to borrow in a massive
way in order to pay for it.
The government sought to soften the impact of the oil price rise
by lowering the tariff on imported oil and by subsidizing the domestic
price of gasoline. These policy responses are reflected in the basic



- 66 -
run and are hence included in experiments A-1 and A-2. To see their
impact, we ran a separate experiment (not reported here) in which the
oil price rise was kept but in which the policy response was removed.
The major impact of the experiment is on the structure of the economy.
The infrastructure sector, which includes transportation, grows much
more slowly when the economy is forced to adjust fully to the higher
price of imported oil. Considering the political significance of the
infrastructure sector, it is not surprising that the government attempted
to maintain the domestic price of petroleum products by subsidies. The
major cost of the policy was felt in increased strain on the balance
of payments which, in the basic run, is reflected in increased import
rationing in the last few years.
The experiment in which the oil price increase is removed is
a relatively pure experiment in that it does not include any other
concomitant effects that would likely have occurred. For example,
without the oil crisis, there probably would have been higher levels
of workers' remittances and also a better world economic situation.
Even without these additional effects, experiment A-2 indicates that the
oil price increase cost Turkey over two billion dollars in foreign
exchange during the period. The response of policy makers was to delay
the necessary adjustment and to borrow in order to sustain growth.
However, one cannot blame the current foreign-exchange crisis on the
rise in the price of oil alone. As the discussion in the previous section
indicates, the oil price rise was not the major cause of the imbalance in



- 67 -
Turkey's equilibrium exchange rate. Removing the oil crisis does not
change the basic adverse trends in the real exchange rate, exports,
imports and remittances. Given the size of these trends, it appears
that the lack of an oil price rise would have delayed a major foreign-
exchange cr:Lsis by up to two years, but no longer.
Experiment A-4 shows the impact on the economy of keeping the
rate of investment roughly constant during the period. Instead of rising
from 18 percent of GDP in 1973 to 24 percent in 1977, the rate was kept
around 18-19 percent throughout the period. The cumulative effect on the
total capital stock over such a short time period is quite small, but
the effect on the structure of production is significant. Aggregate
consumption in 1977 is 4 percent higher and investment 12 percent lower
than in the adjacent experiment. The rate of growth of consumption over
the period is a full percentage point higher while that of investment is
3.4 percentage points lower. Thus the strain of the investment boom is
not felt so much on the balance of payments -- where, as discussed in
the previous section, its impact was relatively slight -- but rather on
aggregate consumption. As the foreign exchange crisis unfolded, the
continued steady increase in the investment rate resulted in a squeeze
on aggregate consumption and increased strain in the economy.
In, the last experiment (A-5), remittances-are assumed to grow
at the same rate as foreign capital inflow (17 percent a year), and are
thus assumed to be analogous to foreign capital. Over the period, the
cumulative difference amounts to 2.1 billion dollars. Comparing the



- 68 -
results to those from the adjacent experiment (A-4), the differences are
really quite dramatic. GDP grows at 0.5 percentage points more a year,
consumption by 1.4 percentage points, and investment by 1.9 percentage
points. In 1977, consumption is 6 percent higher and GDP 2 percent higher
than in experiment A-4.
From this experiment, it is interesting to note how dependent
on foreign trade is Turkish economic performance. Two billion dollars
represents about 10 percent of total cumulated foreign exchange receipts
during the period (in experiment A-4), or about 5 percent of 1977 GDP.
Yet it makes a half percentage point difference in the rate of growth of
GDP over the entire period! The reason for this dependence is that Turkey
is importing intermediate and capital goods which cannot be cheaply pro-
duced domestically and for which it is difficult to substitute. While
Turkey has a relatively low ratio of imports to GDP given its size and
per capita income, the goods it does import are important to the economy
and cannot be easily squeezed.
Indeed, an important lesson from this last experiment is that
trade dependence should not be measured by the ratio of imports to GDP.
One must consider what goods are being imported and their elasticity of
substitution in use with corresponding domestic goods. As will be seen in
the forward-running experiments discussed in the next section, import
substitution that is not selective but tries to proceed on all fronts is
expensive, and adjusting to foreign exchange shortages by severe import
rationing inflicts serious costs on the economy.



- 69 -
3.5 Conclusion
There are a number of lessons to be drawn from the analysis of
the 1973-1977 period. Given the variety and magnitude of shocks that
Turkey has undergone, and the speed with which they developed, it is
understandable why Turkish policy makers and other observers were caught
by surprise. The oil price rise, the world-wide recession, and the
swing in remnittances all contributed to confusing the picture. As the
analysis in section 3.3 indicates, under these circumstances it is
impossible to estimate the required devaluation from an examination of
differential inflation rates. Indeed, when so much is going on, it
would seem that any partial-equilibrium analysis is very likely to be
misleading and that a general-equilibrium framework such as the TGT
model is preferable.
The oil price rise and the swing in remittances both were very
important and can be considered as the proximate causes of the foreign
exchange crLsis that hit at the end of the period. Adding together the
cumulative Lncrease in net foreign-exchange inflow from experiments A-3
and A-5 yie:Lds over 4 billion dollars in additional foreign exchange
from lower oil prices and increased remittances. However, the basic
underlying l:rends in the real exchange rate, exports and imports were
still adverse and would have led eventually to a crisis. Indeed, one
might argue that while the oil price increase hastened the crisis by one
or two years, the completely unexpected increase in remittances in the
1970-1973 period postponed it by two to four years. The fact that the
oil price rise coincided with the leveling off of remittances caused the



- 70 -
denouement to occur extremely rapidly, but it would have occured in
any case.
A major goal of post-war Turkish economic policy has been to
promote industrialization in general, and import substitution in particular.
Turkish policy has strongly favored import substitution both through
direct protection and through import rationing and exchange rate policy.
The result has been severe discrimination against exports and a rather
autarkic development strategy. The fact that Turkey has had a relatively
low level of trade for a country of its size and per capita income may
appear consistent with a desire to be independent of other countries.
However, an important conclusion from our experiments is that the policy
regime that Turkey has followed may in fact have increased the country's
dependence on foreign resources. Discrimination against exports reduces
the ability of the economy to acquire foreign exchange and the resulting
restriction on all but essential imports reduces the ability of the economy
to adapt if necessary. The kind of exchange-rate policy followed inevitably
leads to a foreign-exchange crisis. As has happened twice in the past,
such a crisis cannot help but cause serious dislocation to the economy.
Turkey has been able, in the past, to overcome these crises
relatively rapidly. Growth did resume and even accelerated after the 1958-
1960 and 1970 crisis episodes. In both cases, the reasons for the rapid
improvement are clear. After 1960, government policy and economic
management greatly improved, economy-wide planning was introduced and the
public sector was reorganized. Turkey also received a substantial amount



- 71 -
of foreign aid. The 1970 crisis was,not followed by a similar period of
reform and reorganization although an early attempt was made. But an
exogenous factor, workers remittances, transformed the early 1970's into
a period of foreign exchange abundance, permitting a great increase in
imports and creating a false sense of security.
The present situation seems closer to 1960 than it is to 1970.
No great excigenous flow of foreign exchange seems likely to appear. It
is economic policy and internal reform that must pull Turkey out of the
crisis. Very important strategy decisions have to be made and priorities
determined. It is to a general-equilibrium analysis of these issues
that we turn in the next sections.



- 72 -
4. Prospects for the Future: An Economy-Wide Perspective for 1978-1983
4.1 Introduction
This section turns to an analysis of future prospects concentrating
on the 1979-1983 period which spans Turkey's Fourth Five Year Development
Plan (FFYP). We will first attempt to provide an economy wide perspective
focusing on the interaction between trade policy, trade performance
and overall economic growth. The macroeconomic perspective provided in
this section is of course based on developments in the 19 individual
sectors distinguished in the TGT model. Section 5 will turn to a
microeconomic analysis and a discussion of the sectoral growth and
trade prospects that underliethe alternative economy-wide scenarios
described below.
All projections generated by the TGT model are conditional pro-
jections based on assuming specific policy packages. Government policy
has a determining influence on the characteristics of the projected
growth paths and it is therefore not possible to take any one projected
path as the "predicted" path unless one is willing to predict that the
policy assumptions underlying that particular path are in fact the
policies that will be followed. It is very important always to keep in
mind the conditional nature of the projected paths.
4.2    Constant Price Deflated Exchange Rate Policy
The analysis presented in Part 3 suggests that the degree of
overvaluation of the Turkish Lira has been much greater in the 1975-1977



- 73 -
period than generally perceived or than one would estimate by simply com-
paring differential inflation rates. The March 1978 devaluation which
brought the parity of the dollar to 25TL, substantial though it was, did
not really alter the situation in real terms. Between mid-1977 and mid-
1978, the nominal exchange rate was devalued by over 40% against the
dollar. But in the same period Turkish price indices went up at rates
between 40% and 50%. Taking account of the depreciation of the dollar
against some major currencies, world inflation or the rate of change of
a dollar denominated price index was about 15%. So differential inflation
in Turkey was at least 25%, probably closer to 30%. Between 1977 and 1978
Turkey did not, therefore, devalue by more than about 10% in real terms.
Until the middle of 1977, the great widening of the underlying
foreign exchange gap that had been underway since 1975, did not have a
dramatic impact on the Turkish economy because of massive short-term
borrowing. Even the first half of 1977 was still characterized by a
very high leveal of imports, partly in anticipation of post-election deval-
*uation and import restrictions. It is only in the latter half of 1977
that the foreign exchange gap developed into an acute crisis with no more
short-term finance forthcoming, imports declining in absolute nominal value
and import raitioning reaching dramatic proportions. According to the TGT
model, the avearage degree of import rationing was only '4 percent in 1977.
This probably reflects the fact that during the first six to eight months
of the year, imports were allowed to be much higher than they would have
been without the speculative anticipation of events to come and the Demirel's
governments' pre-election willingness to encourage every possible source of



- 74 -
short-term finance. Thus cumulative January-August imports were still
21.2% higher in 1977 than in 1976. It is only in the Fall of 1977, when
possibilities for further financing had greatly diminished, that the painful
process of adjustment had to begin and it is only at that time that the
underlying crisis situation became fully apparent.
Since then the nominal exchange rate adjustment has been substan-
tial and the new government formed in January 1978 has undertaken or is
attempting to undertake important complementary measures ranging from
monetary tightening and proposals for comprehensive tax reform to even
longer-range plans for reorganizing State Economic Enterprises. But as
argued above the exchange rate adjustment came after inflation had increased
to an annual rate of almost 50 percent in the latter half of 1977. It is
quite clear that a 25TL = 1 dollar parity for 1978 implies a price-deflated
exchange rate that is not much higher than it had been at the beginning
of 1977. Thus even if the nominal exchange rate is henceforth adjusted
in step with differential inflation, the Turkish Lira will remain
substantially dvervalued and the necessity of severe import rationing will
persist.
Such a situation would of course not be a new one for Turkey.
With the exception of a brief period after the 1958-1960 devaluation-cum-
stabilization episode and of a somewhat longer period after the 1970
devaluation, there has always been upward pressure on the price of foreign
exchange in Turkey. Another manifestation of the same disequilibrium
situation can be found in the fact that desired imports have always been
above the actual imports that could be financed by available foreign



- 75 -
exchange earnings and that foreign exchange has always been more or less
severely rationed. It is only when the underlying disequilibrium became
too overwhelming and shortages too severe that exchange rate adjustments
were undertaken in the past. This has amounted to a development strategy
strongly orienited towards the domestic market and consisting of successive
phases of import substitution efforts, first concentrated on the light
consumer indusitries and then on consumer durables and basic intermediates.
Turkey now perceives itself to be at the threshhold of another phase of
intensive import substitution this time aimed at the capital goods
industries.
The first forward-looking experiment undertaken with the TGT
model assumes ithat the price deflated exchange rate will remain constant
throughout the FFYP period reflecting a continued strong inward orientation
of Turkish deveslopment strategy. We assume a small drift to 26TL = 1 dollar
in 1978 and devaluations every year thereafter to compensate for differential
inflation -- but no more. The inflation we project for 1979-1980 is 18
percent and we assume a 15 percent domestic inflation rate thereafter.
World inflation is projected at 9 percent per annum. A constant PLD exchange
rate therefore implies a 5.5 percent upward adjustment in the exchange
rate every year after 1979.* The projected scenario is thus one where no
attempt at a major exchange rate adjustment is made and where the 1978
devaluation is considered sufficient by Turkish policy makers. A series
of mini-devaluaitions is allowed, but they leave the real price deflated
exchange rate constant. It is not difficult to foresee that the resulting
growth path would be characterized by continued excess demand for foreign
*The term PLD exchange rate means "price level deflated" and is used by
Krueger and Bhagwati.   See Krueger (1978).



- 76 -
exchange and import rationing. The Turkish economy would retain its
domestic market orientation. But how severe would import rationing
have to be? What kind of overall growth performance could be expected
under these conditions? What level would be reached by the debt-service
ratio under alternative assumptions about capital inflow and how
dependent on foreign borrowing would the economy remain? These are the
major questions addressed in this section.
Important variables that must be projected exogenously are
workers' remittances and the net new flow of foreign borrowing. There
is much disagreement on the likely values of these flows and we have
experimented with many alternative values. For the experiment reported
here, a net capital inflow of 1,200 million dollars is assumed for 1978,
followed by a flow of about 650 million dollars annually thereafter.
Workers' remittances are projected at 1,250 million dollars for 1978,
increasing to 2,200 million dollars by 1983. There is little doubt that
these are optimistic projections but perhaps not unduly so. In the spring
of 1978, Turkey appears to have received new pledges of assistance from
various trading partners such as Germany, the USSR, Libya, Rumania and
Bulgaria. The IMF and the World Bank may jointly provide as much as 600
million dollars disbursement in 1978. Thus the figure of 1,200 million
net for 1978 should be feasible. It assumes a rollover and restructuring
of most of Turkey's short-term debt leading to a debt service and amortiza-
tion flow of about 0.8 billion dollars in 1978. Thus the projected gross
new borrowing figure is 2.0 billion dollars. But many of the credits
presently extended to Turkey have been made possible by special economic



- 77 -
and political circumstances. They are unlikely to be repeated for another
year. We project an average of 650 million dollars net for the 1979-1983
period, which would still probably constitute gross new borrowing of
about one and a half billion dollars a year. Again, this is probably on
the optimistic side.
The projections for workers' remittances reflect the rather
discouraging level of remittances realized in the first 5 months of 1978.
Despite the March devaluation remittances seem to be falling. Assuming
that it will be possible in the future to redirect the significant portion
of remittances that enters Turkey unofficially into official channels, we
project 1,250 million dollars for 1978 and a slowly rising trend thereafter
with a projection of 2,200 for 1983, the terminal year of the Fourth Five
Year Plan. This implies a 12% per annum average increase in the current
dollar value of remittances and therefore implicitly assumes that Turkish
workers in Western Europe remain constant in number and continue to send
home a constant proportion of their income. If this assumption does not
hold and workers abroad begin to feel more like permanent residents in
the countries they work in, living with their families and not necessarily
planning to return, our projections may be too high. When evaluating
these figures one should not forget that a small but significant amount
of outmigration still continues, mostly to the Middle East and that the
stock of Turkish workers abroad is unlikely to diminish. It may even
increase. Thus a 12% per annum average increase in current dollar value
may be a reasonable guess. How much of it will come through official
channels depends of course on the policies and regulations adopted.



- 78 -
Regarding government expenditures, we assume that current
expenditure will increase at a target rate of 6.5% per annum in real terms.
Tax rates, seignorage and transfers are exogenously projected so that
government investment expenditure is a residual arrived at after satisfying
the 6.5% current expenditure growth target. Private savings rates are
for the moment assumed constant at their 1977 levels. Total investment
is the sum of private and government investment. In real terms investment
will be sensitive in particular to overall GDP growth, to total government
revenue and to the relative price of capital goods. Since the savings
rates are fixed as nominal proportional rates, a decline in the relative
price of capital goods will lead to a rise in real capital accumulation.
Table 4.1 summarizes the macroeconomic characteristics of the
growth rate generated by the TGT model assuming the constant real exchange
rate policy and the capital flows described above.
The results underline the depth of the present crisis and the
difficulties that lie ahead. Under the fixed PLD exchange rate policy,
the TGT model predicts a growth rate of GDP around 2.5% in 1978 and 1979
and of about 6.2% thereafter until 1983. Consumption (in base year prices)
grows by less than 1% annually in 1978 and 1979, recovering and growing
at a rate close to 6.0% thereafter. This implies an actual decline in
per capita consumption for the 1978-1980 period.
Investment declines substantially in 1978 and recovers its real
1977 level only in 1982! The real wage also declines in 1978 and continues
to fall marginally in 1979. It then settles on a 2.0% per annum growth after
1980. Given that the real wage in the organized sector grows exogenously
the burden of the initial decline in the average real wage is assumed to
fall on the unorganized sector and will result in increased underemployment.



- 79 -
Table 4.1: Results of a Constant PLD-Exchange Rate
Policy, M!acroeconomic Indicators                    FFYP Average
Experimetnt Base                           Growth Races
1978    1979      1980      1981    1982    1983      1979-1983
GDP Growth Ratek**               2.52     2.6%     6.1%     6.4%      6.2%    6.2%       5.5%
Value Added Growth **i
Agriculture                  5.6%    2.3%      4.2%     4.3%     4.3%     4.3%       3.9%
Industry                     3.1%    4.2%      7.0%     7.8%     7.4%     7.4%       6.7%
Services                     4.5%    3.1%      5.9%     6.2%     6.0%     6.0%       5.4%
Growth of Consuaption            0.9%    0.3%      5.9%     6.1%      5.8%    5.9%       4.8%
Growth of Inves;ment           -14.6%    0.2%      6.3%     6.5%      5.8%    5.8%       4.9%
Tocal Urban Labor Force          7115    7435      7766     8093      8411    8724        __
(1000)
Growth Rate                      5.0%    4.5%      4.5%     4.2%      3.9%    3.7%       4.2%
Employment in Industry           2319    2426      2512     2609      2691    2768        __
(1000)
Growth Rate                      0.2%    4.5%      3.5%     3.9%      3.1Z    2.9Z       3.6%
Growth of Average Urban         -6.3%   -1.3%      1.7%     2.0%      2.1%    2.4%       1.4%*
Wage, Real
Average Economy-Wide            18.9%   17.2%     17.3%    17.5%    17.6%    17.7%      17.5%*
Profit Rate
Profit Rate in Manufacturing    32.2%   33.3%     34.3%    35.3%    36.0%    36.6%      35.1Z*
Investment Rate (of GDP)        21.3%   22.6%    22.9%     23.1%    23.1%    23.2%      23.0%*
ICOR                             9.1     7.2       3.0      2.9.     3.0      3.0        3.8*
Exports (million $, current)     2310    2679      3113     3601     4181     4847      16.0%
Imports (million S. current)     5064    5163     5958      6889     7916    9058       12.3%
Import-Elas.                    -9.2    -3.1      0.9       0.9      0.9     0.8          __
Net Invisibles ($)                407     628       755      907     1093     1322      26.6%
Remittances (S)                  1250    1400     1600      1800     2000     2200      12.0%
Net Capital Inflow (;)           1200     550       600      650      700      750        650
Debt Service Ratio**            21.7%   24.5Z    28.5%     33.17%   36.1%   35.7%       31.6%*
Degree of Import Rationing      40.0%   52.0%    54.0%     55.0%    57.0%    58.0%      52.7%*
Domestic Inflation              42.0%   18.0%    15.0%     15.0%    15.0%   15.0%       15.6%*
Exchange Rate                   26.0    28.2     29.7      31.3     33.1    34.9         6.0%
(to the dollar)
Devaluation (Z                  42.9%    8.3%     5.5%      5.5%     5.5%    5.5%        6.0%
*  Average over FFYM period
** Amortization + interest divided by exports + remittances + afs.
n    Constant market prices, 1973 base.
*n** Real value added, double deflated, 1973 base. Includes indirect taxes but excludes tariffs.
GDP equals total value added plus tariffs.



- 80 -
On the foreign trade side, TGT predicts 2.3 billion dollars
worth of merchandise exports for 1978 and 400 million dollars of net
revenue from invisibles including tourism. In conjunction with the
exogenously projected capital flows, these foreign exchange earnings
allow about 5 billion dollars worth of merchandise imports. In constant
dollars this represents a 20% decline over 1977. Desired imports are 8.3
billion dollars in 1978 and the degree of rationing is 40%. This constitutes
more severe import rationing than occurred in 1977. Capacity use there-
fore further diminishes and in spite of the substantial investment rate
in 1977, growth is projected to be slower in 1978 than in 1977. At this
point it should again be emphasized that 1977 was an awkward year, with
the first half quite different from the second half. The degree of
excess capacity and the severity of import rationing projected for 1978
is similar to that prevailing during the second half of 1977, but it is
worse than the 1977 average.
The foreign trade and balance-of-payments situation does not
improve in later years. On the contrary, the degree of rationing becomes
more severe every year, reaching 58% in 1983. The entire FFYP period
would thus be characterized by severe import rationing.
1978 constitutes the worst year of the crisis with GDP growth
only 2.5%, investment and real-wages falling and industrial employment
stagnating. The situation improves somewhat in 1979, but 1979 is still
very much a crisis year. A moderate recovery is finally achieved in 1980
and GDP growth settles at about 6.3% per year for the rest of the FFYP
period.



- 81 -
The growth path the econoxy reaches by 1980 is characterized
by chronic excess demand for imports, sluggish export growth and sub-
stantial import substitution (see section 4.3 below). Between 1978 and 1983,
imports grow by less than 2.8 percent per year in real terms, while GDP
grows at an average rate of 5.6 percent. The propensity to import would
thus be declining significantly, in sharp contrast to what happened
in the early L970's.
Whi:Le the path generated by a-fixed price deflated exchange rate
policy is certainly not an attractive one, and while the growth rates
generated fall far..below policy objectives.and expectations in Turkey, it
should be stressed that the TGT model suggests that the economy can
probably continue.to grow    at a moderate pace without a great increase in
imports and in spite of foreign exchange shortages. Indeed, the path
described abovre appears not unlike the path Turkey followed in the 1960's.
Between 1963 and 1968 real imports were growing at 2.9% while output
was growing at: 6.6% per year in real terms.-L/ Turkey has in the past been
able to grow moderately rapidly under conditions of foreign exchange
shortage and declining import-GDP ratios.    These periods have been periods
of substantial import substitution and it is quite possible that Turkey
is again preparing to go through a similar phase of intensified import
substitution, particularly in such sectors as steel, petrochemicals,
fertilizers and machinery.
1/ See the tables prepared for the "Sources" project of the Economics of
Industry I)ivision by Merih Celasun.



- 82 -
Without going more into the microeconomic details of the path
summarized above, one cannot really form a judgment on how likely a path
it is in terms of the possible pace of import substitution. But three
important observations can be made just looking at the macroeconomic
indicators:
a. All exogenous flows and variables have been projected at
optimistic levels. In spite of that, the overall growth
rate remains quite low.
b. The degree of import rationing increases throughout the
period reflecting continuous and very strong excess demand
for imports and foreign exchange.
c. At the end of the FFYP period, total merchandise exports are
still below 5 billion dollars, manufactured exports still
have not reached the one billion mark and the trade deficit
exceeds 4 billion dollars.
To test the sensitivity of the growth path to the exogenously
projected variables which we have projected at optimistic levels so far, we
undertook an experiment with the following "pessimistic" assumptions:
1978    1.979   1980    1981     1982    1983
Workers' Remittances          1100    1200    1300    1400    1500    1600
(M $)
Net Foreign Capital Inflow    1000     500     500     500     500     500
(M $)
Total Inflow                  2100    1700    1800    1900    2000    2100
(M $)



-83 -
This should be compared with the total i.nflow assumed so far which was:
1978    1979     1980    1981    1982     1983
Total Inflow                  2650    1950     2200    2450    2700     2950
(M $)
For this more "pessimistic" experiment, we also assumed that the
inflation rate would reach 47% between 1977 and 1978 instead of 42%
assumed previous.ly so that the 26 TL = 1 dollar initial exchange rate is
lower in real terms.
The results are not encouraging and underline Turkey's vulnerability
to shortfalls in foreign capital inflows and workers' remittances. Table 4.2
provides a condensed summary of the macroeconomic results.
Table 4.2: Results of Prolected
Low Levels of Borrowing and- Workers' Remittances             FFYM Average
Growth-Rates
1978    1979     1980    1981    1982     1983   1974-1983
GDP Growth Rate           0.8%    2.4%     5.4%    5.7%    5.9%     5.9%     5.0%
Growth of Consumption     0.6%    1.2%     5.4%    5.5%    5.6%     5.5%     4.8%
Growth of Investmnent   -18.0%    1.5%     4.0%    5.0%    5.3%     5.6%     4.3%
Degree of Import         45.0%   57.0%    59.4%   61.0%   62.1%    63.2%    60.6%*
Rationing
* Average over FFYP period
What iLs perhaps most important is that compared to a 2.5% growth
rate in 1978 when foreign borrowing and remittances are forthcoming at
optimistic levels, growth in 1978 is close to zero when foreign capital
flows are scaled down by 20% in 1978. The difference between 2.5% and 0.8%



- 84 -
on GDP growth for 1978 is a very important one. It may constitute the
crucial difference between a return to socioeconomic stability or further
serious disintegration. Given the crisis atmosphere currently prevailing
in Turkey, the total standstill that may well occur if foreign funds are
not forthcoming at the substantial levels we projected previously, may
lead to a collapse of expectations and a deepening of the crisis that
would vitiate the basic assumptions about savings rates, continued
investment and social stability that underlie our projections for a
recovery by 1980.
While our initial optimistic projections may be achieved with
enough foreign cooperation, it is quite clear that without a much more
substantial increase in exports, Turkey remains extremely vulnerable to
shortfalls in workers' remittances and foreign borrowing. While
optimistic assumptions about borrowing and remittances allow moderate
recovery and growth without a major devaluation, the situation appears to
be a very tenuous one. What is quite clear and needs to be strongly
emphasized is that the March 1978 devaluation is very far from being one
that corrects the fundamental imbalance behind the foreign exchange gap.
It simply compensates for the differential inflation rates in 1977 and
1978 but it does not constitute an adjustment to past differential
inflation rates, to the reality of a high oil price and to the fact that
workers' remittances can no longer be expected to grow significantly in
real terms. Quite apart from these factors, the equilibrium value of
foreign exchange may be following an upward trend in Turkey due to ongoing
shifts in the structure of output and national expenditure.



- 85 -
It is therefore quite likely that Turkish policy makers,
watching the inadequate performance of exports, feeling the pressure of
increasing excess demand for imports and realizing that GDP growth could
be increased if the foreign exchange constraint was not so tight, will
eventually consider important policy changes aimed at export expansion,
better growth performance, and less vulnerability to foreign capital flows.
The need for an expansion of manufactured exports as a basic condition
for rapid growt-h and continued industrialization is increasingly recognized
in Turkey.  Amlbitious targets are being set in this area.  But the funda-
mental fact is that incentives are very heavily biased against exports
and that the exchange rate is by far the most important variable determining
the structure of incentives and the degree of price competitiveness of
Turkish manufactured exports. It will not be possible to achieve sub-
stantial export growth without a major readjustment of the exchange rate.
All other measures are bound to remain marginal compared to what could be
achieved by a major readjustment of the real exchange rate.
It should be stressed that an adjustment in the real price deflated
exchange can only be achieved if monetary and fiscal policy is able to con-
trol the price level in the post-devaluation period (see Section 2 above).
A real devaluation is a package in wh:Lch the nominal rate of
devaluation is only one part that must be complemented by measures to
keep the price level from rising enough to dissipate the effects of the
devaluation. A real devaluation must be able to change relative prices
and incentives. While all prices may rise after a devaluation, some prices



- 86 -
must rise more than others to allow a change in the structure of incentives.
In the TGT model a devaluation of the nominal exchange rate, holding
everything else constant, implies a decline in at least some domestic
prices relative to a situation of no devaluation. While we keep the
overall inflation rate at a constant exogenously specified level, the
structure of price increases will be strongly affected by a change in
the exchange rate.
This paper does not attempt to predict the policy changes that
will actually take place or their timing. A constant price deflated
exchange rate from 1978 onwards may be quite a likely scenario. It is
even possible that the nominal exchange rate will be kept constant for a
certain period. On the other hand the exchange rate may be allowed to
drift somewhat faster than differential inflation, leading to a progressive
real devaluation. The constant PLD exchange rate projection captures the
essential characteristics of growth paths reflecting continued adherence
to an inward oriented growth and industrialization strategy. Whether the
nominal exchange rate is devalued a bit more or a bit less rapidly than
required by differential inflation does not change the fundamental nature
of the growth paths characterized by massive import substitution in the
basic intermediates and capital goods sectors.
What the TGT model suggests is that while such a path may well
be feasible it is not one that can generate very rapid growth in the
medium term. An annual growth rate of 5.5% constitutes an upper bound
for the FFYP period if extreme inward orientation persists and exports
fail to expand rapidly.



- 87 -
In terms of self-reliance and economic independence, the growth
path generated by our basic run is characterized by conflicting features.
On the one hand it can be interpreted as promoting a more independent
economy through the rapid building up of the capital goods industries.
But on the other hand the tightness of the foreign exchange constraint
will tend to further increase Turkey's dependence on foreign capital and
foreign creditors. The debt service ratio increases to levels above 30%
threatening a crisis of confidence that could lead to a collapse of the
flow of borrowed funds. Turkey would be walking on a tightrope with
another severe foreign exchange crisis always threatening. It is therefore
hard to see how a strategy that fails to promote dramatic export expansion
can really succeed in successfully strengthening economic independence
and self confidence.  Furthermore, as will become apparent when discussing
the microeconomic nature of the basic run, the extent and pace of import
substitution implied would have to constitute a record performance when
compared to performance in a sample of semi-industrialized economies in
the past two decades. While the trade substitution elasticities embodied
in the microeconomic structure of the TGT model are quite low and are
therefore unlikely to understate the difficulty of substituting domestic
production for imports, it is possible that we underestimate the effects
continuous severe import rationing might have on capacity utilization,
technical progress and x-efficiency in the economy. The basic run
projections provide an optimistic upper bound to what can be achieved with
a single minded strategy of import substitution. If everything goes right,
Turkey might overcome the crisis by the end of 1979 and settle on a steady



- 88 -
growth path of about 6% a year after 1980 without a major reorientation
of industrial strategy towards exports. But it is far from clear that
everything can go right and that foreign money will be forthcoming in
sufficient amounts under those circumstances. And even if everything
that is optimistically projected does materialize, a 6% growth rate
remains well below what is considered desirable or even acceptable in
Turkey. As the basic constraints limiting other growth performance become
more obvious and as it becomes clearer that a major export drive is an
indispensable condition for a more rapid trend growth rate in the 1980's,
it is quite possible, therefore, that a very serious reorientation of
industrial strategy will be considered by Turkish policy makers. The
most important signal for such a change would have to be a major real
devaluation and a shift of incentives and resource pulls towards exports.
Real savings would have to be embodied more in exports and less in
domestic import substitutes. It is the implications of such a shift
that we explore in section 4.3 below.



- 89 -
4.3 Macroeconomic Consequences of Alternative Trade and Exchange
Rate Policies
The real exchange rate is one of the most important relative
prices characterizing the price and incentive structure of a mixed market
economy. Its absolute nominal value does not carry much significance but
its relative real value has a determining influence on resource allocation,
resource pulls and the direction of growth. What determines resource pulls
and the relative structure of incentives is primarily the relative struc-
ture of net prices or "value added" prices. Trade policy influences net
prices through the exchange rate, import taxes, export subsidies and
import rationing. The causal links are complex and we shall return to a
more detailed discussion later. Note only at this stage that the mechanisms
linking trade policy instruments to relative prices and incentives in a
model that recognizes product differentiation and reflects strict exchange
control are different from those discussed in the simpler models of
pure trade theory assuming perfect substitutability and free converti-
bility.
In this section, we will report on two experiments with the TGT
model and compare the results with those obtained from the basic run which
assumed a constant price-deflated exchange rate. The focus of the analysis
will be on macroeconomic performance. However, both experiments involve
changes in the real exchange rate, and it is these changes which dominate
the results. Thus it is important to explore the incentive and resource-
pull effects, even at the macro level. Detailed sectoral analysis will be
done in Part .5 below.



- 90 -
Experiment B-1: A Higher Exchange Rate Policy
Experiment B-1 assumes a further major devaluation bringing
the value of the Turkish Lira to 36 TL = 1 dollar in 1979 and letting
the parity slide down thereafter by 10% (nominal) every year. Given a
differential inflation rate that is projected to be only 5.5% after
1979, this implies a real devaluation of about 30% in 1979 and of 4.3%
annually thereafter.
Experiment B-2: A Manufactured-Exports and
Investment-Oriented Policy Package
Experiment B-2. assumes the same path for the exchange rate as
experiment B-1, but the high exchange rate policy is complemented by the
following measures:
A once-and-for-all 50% increase in export subsidies for
all nontraditional manufactured exports in 1979
(Sectors 5 through 16).
An increase in public investment complemented by an
increase in the savings rates of labor households
designed to increase the average investment rate to
25.5% by 1983.
A substantial increase in the effective average tax
rate on capital income necessary to allow the non-
inflationary financing of the above policies.
Before starting to analyze and discuss the results obtained
from the two experiments it is important to stress again that the 30%
devaluation refers to a 30% change in the exchange rate, ER, relative
to the domestic price-level, PL. It should be clear that a 30% real
devaluation in 1979 can be achieved by various combinations of domestic



- 91 -
inflation rates and nominal exchange rate changes. The particular one
we are assuming is an 18% domestic inflation rate, a 9% world inflation
rate and a 313.5% nominal devaluation. But a lot of other combinations
could lead to the same amount of real devaluation that matters. The
real variables of the TGT model in its forward running form are not
affected by the inflation rate as long as the latter is accompanied
by an appropriate amount of nominal exchange rate adjustment. If,
however, the ratio of the nominal exchange rate to the price level
changes, in other words if there is a real devaluation, relative prices
and resource allocation will change and the real variables of the model
will be affected.
Another point that needs emphasis is that the TGT model does
not specify a lag structure in export demand. While there is an explicit
distinction between short run and long run responses on the supply side
via fixed capital stocks and profit rate responsive investment, the
demand response to exchange rate adjustment is specified to take place
within a single year. This may overstate the impact effect of devaluation.
The five year averages and terminal 1983 projections should therefore be
considered more reliable than the 1979 figures that may overstate the
speed of response to a major policy shift.
Tables 4.3 and 4.4 summarize some of the most important macro-
economic indications derived from experiments B-1 and B-2. These results
should be compared to those presented for the basic run in Table 4.1
above. Tables 4.7, 4.8 and 4.9 are comparative tables and summarize
results frorm all three experiments.
First consider the high exchange rate experiment whose results
are summarized in Table 4.3 and in various tables below. The re-



- 92 -
Table 4.3: Results of a High Exchange Rate
Policv, Macroeconomic indicarors                   FFYP Average
ExDeriment B-1                              Growth Rates
1978     1979     1980      1981    1982    1983     1979-1983
GDP Growth Rate***              2.5%     7.3%     7.6%     7.82     7.9%    8.0%       7.7%
Value Added Growth****
Agriculture                 5.6%    5.6%      5.4%     5.5%     5.7%    5.9%       5.6%
Industry                    2.6%    9.1%     8.8%     8.9%      8.9%    8.9Z       8.9%
Services                    4.5%    7.1%     7.2Z      7.6%     7.5%    7.7%       7.4%
Growth of Consumotion           0.9%    1.6%      6.5%     6.7%     6.8%    6.8%       5.7%
Growth of Investment          -14.6%    9.2%      8.6%     8.5%     8.6%    8.4%       8.7%
Total Urban Labor Force         7115     7435     7744     8043     8334    8616        --
(1000)
Growth Rate                     5.0%    4.5%     4.2%      3.9%     3.6%    3.4%       3.9%
Employment in Industry          2319    2400      2490     2569     2648    2716
(1000)
Growth Rate                     0.2%     3.5%     3.8%     3.2%     3.1%    2.6%       3.2%
Growth of Avera:e Urban        -6.3%    0.1%     2.3%      2.9%     3.1%    3.4%       2.3Z*
Wage, Real
Average Economy-Wide           18.9%   17.8%    18.0%     18.1%    18.3%   18.3%      18.12*
Profit Rate
Profit Rate in Manufacturing   32.2%   30.5%    30.8%     30.6%    30.2%   29.4%      30.3%*
Investment Rate (of GDP)       21.3%   22.9%    23.2%    23.3%     23.5%   23.6%      23.3%*
ICOR                            9.1     2.6      2.5      2.5       2.5     2.5        2.52*
Exports (million $)             2310    3350     4076     4943      5986    7294      25.9%
Imports (million S)             5064    5706     6736     7928      9315   10943      16.7%
Import-Elas.                   -9.2     0.7      1.1      1.1      1.1      1.1         --
Net Invisibles ($)               407'    496      563      640       723     805      14.6%
Remittances   )                 1250    1400     1600     1800      2000    2200      12.0%
Set Capital Inflow CS)          1200     550      600      650      700      750        650
Debt Service Ratio**           21.7%   21.8%    24.7%    27.9%    29.7%   28.5%      26.5%*
Degree of Import Rationing     40.0%   30.0%    27.0%    24.0%    20.0%   15.0%       23.2%*
Domestic Inflation             42.0%   18.0%    15.0%    15.0%    15.0%   15.0%      15.6%*
Exchange Rate                  26.0    36.0     39.6     43.6     47.9    52.7       15.2Z
(to the dollar)
Devaluation (%)                42.9%   38.5%    10.0%    10.0%    10.0%   10.0%      15.22
* Average over FFY? period
**  A,mortization + interest divided by exports + remittances - nrs
*** Constant market prices, 1973 base.
**** Real value added, double deflated. 1973 base. Includes indirect taxes but excludes tar'ffs.
GCP equals total value added plus tariffs.



- 93 -
Table 4.4: Results of a High Zxchanie Rat*/Dn14cv 'ULU zg
Policv, macroeconomic Indicators                     FY    vrg
~~ t  -  -  ~~~~~ I ~FFYP Average
Exreriment B-2                               Growth Rates
1978      1979     1980      1981    1982    1983      1979-1983
GD? Growth Rate***               2.5%    7.3%      7.7%     7.8%      8.0%    8.2%       7.8%
Value Added Growth****
Agriculture                  5.6%    5.7%      5.3%     5.4%      5.6%    5.8%       5.5%
Industry                     2.6%    9.3%      9.3%     9.4%      9.6%    9.7%       9.5%
Services                     4.5%    6.9%      7.2%     7.5%      7.5%    7.6%       7.3Z
Growth of Consumption            0.9%     1.3%     6.2%     6.3%      6.3%    6.3%        5.3%
Growth of Investment           -14.6%     9.2%    10.3%    10.02     10.3%   10.5%       10.1%
Total Urban Labor Force          7115     7435     7744     8047      8343    8633         -
(1000)
Growth Rate                      5.0%     4.5%     4.2%     3.9%      3.7%    3.5%        3.9%
Employment in :ndustry           2319     2423     2531     2632      2735    2842         --
(1000)
Growth Rate                      0.2%    4.5%      4.5%     4.0%      3.9%    3.9%       4.2%
Growth of Average Urban         -6.3%    0.0%      2.9%     3.3%      3.6%    3.9%       2.72*
ltage, Real
Average Economr-Wide            18.9%    17.8%    17.9%    18.0%     18.0%   17.9%      17.9%*
Profit Rate
Profit Rate in Manufacturing    32.2%   30.8%     31.1%    31.1%     30.8%   30.3%      30.8Z*
Investment Rate (of GDP)        21.3%   22.9%     23.6%    24.3%    24.9%    25.6Z      24.3%*
ICOR                             9.1     2.6       2.5      2.5      2.5      2.5        2.5Z*
Exports (million $)              2310    3388      4124     5010     6082     7409      26.2Z
Imports (million $)              5064    5729      6775     7980     9393    11058      16.9%
Import-Elas.                    -9.2     0.7       1.2      1.1      1.1      1.1         --
Net Invisibles (S)                407     490       553      630      716.     807      14.7%
Remittances ($)                  1250    1400      1600     1800     2000     2200      12.0%
Net Capital Inflow ($)           1200     550       600      650      700      750        650
Debt Service Ratio              21.7%   21.7%     24.6%    27.7%    29.5Z    28.2%      26.3Z*
Degree of Import Rationing      40.0%   30.0%    27.0%     25.0%    22.0%    19.0%      24.6%*
Domestic Inflation              42.0%   18.0%    15.0%     15.0%    15.0%   15.0%       15;6%*
Exchange Rate                   26.0    36.0     39.6      43.6     47.9    52.7        15.2%
(to the dollar)                                                                   I
Devaluation (%)                 42.9%   38.3%    10.0%     10.0%    10.0%   10.0%       15.2%
* Average over MYP period
** Amortizatiorn + interest divided by exports + remittances + nfs
*    Constant market prices, 1973 base.
**** Real value addied, double deflated, 1973 base. Includes indirect taxes but excludes tariffs.
GDP equals tot:al value added plus tariffs.



- 94 -
covery sets in with the devaluation in 1979, leading to a 7.3%
GDP growth rate instead of the 2.5% projected in the basic run. GDP
growth remains continuously higher and the FFYP average reaches 7.7%
compared to 5.5% in the basic run. Investment grows substantially more
rapidly with the higher exchange rates. Consumption also grows more
rapidly, although the difference is much smaller. The effect of deval-
uation on the average real urban wage is positive. It no longer
declines in 1973 and its growth throughout the FFYP period is one
percentage point higher than in the basic run, reaching 3.4% instead of
2.4% by 1983. Note finally that while the average economy-wide profit
rate is not affected by the higher exchange rate policy, the profit
rate in manufacturing does fall by about 5 percentage points. The
impact is every uneven with the export oriented subsectors gaining at
the expense of the import substituting subsectors. But on average the
static resource pull created by a devaluation adversely affects manu-
facturing. In the TGT model this does not mean that dynamically
devaluation will slow down growth in manufacturing output. The whole
economy, including manufacturing, grows more rapidly after an effective
devaluation that succeeds in generating a substantial increase in
exports.
There are essentially three reasons for the positive macro-
economic growth effect of a high exchange-rate policy: resource real-
location effects, capacity utilization increases and more rapid real
capital accumulation.



- 95 -
A successful real devaluation will lead to a reallocation of
resources both statically and dynamically. We shall discuss the resource
reallocation, mechanism in greater detail in Section 5 below. Essentially
what happens is that statically labor is reallocated to the more export-
oriented sectors away from the sectors producing very high-cost import
substitutes. This static reallocation alone is responsible for a once-
and-for-all 2.6 percentage point increase in GDP. Dynamically, the
structure of prices and profit rates is altered by the devaluation and
the growth of both employment and capital stock becomes more rapid in
the more export-oriented sectors.   These are characterized by lower
incremental capital-output ratios and therefore the dynamic reallocation
process leads to a lowering of the economy-wide ICOR. This adds about
0.75 percentage points to the annual growth rate. The static and
dynamic rescurce reallocation effects are therefore the most important
explanation of the increase in the growth rate that occurs after a
successful clevaluation.
Ch,anges in the degree of capacity utilization constitute a
second important factor. They lead to a 2.0 percentage point increase in
GDP in 1979 and to about 0.5 percentage point dynamically from 1980
onwards.
The specification adopted implies that it is the change in
the degree of rationing that leads to a contraction in effective
resources. Once the economy settles on a "isteady-state" degree of
import rationing, the growth rate of GDP ceases to be affected.



- 96 -
While there is ample documentation on the link between the degree of
capacity utilization and the degree of import rationing, 1/ there is
much less knowledge about possible dynamic links between the pace of
technical progress and trade policies. The TGT model does not contain
any such dynamic links. If one believes that export orientation and
the absence of quantity allocation processes has a positive effect on
the pace of technical progress and innovation, the growth rate differ-
entials generated by the TGT model should be considered as lower
bounds. On the other hand, it can be argued that technical progress
and learning is inherently more rapid in the import-substituting
"heavy" industries, and that therefore an inward-oriented strategy
emphasizing growth in those sectors will tend to increase the average
economy-wide rate of technical progress. 2/ By not specifying any
dynamic link between technical progress and trade policy we adopt a
cautious attitude that is probably justified, particularly given our
relatively short time horizon.
What is important to remember when evaluating the resource
allocation and capacity utilization effects discussed above is that the
economy is not operating on its transformation frontier when the exchange
1/ See Krueger (1978).
2/ See De Melo and Dervis (1977) and Findlay (1974) for models that
emphasize these dynamic issues.



- 97 -
rate is fixecl and imports are rationed. If the economy were operating
on its transformation frontier, any increase in exports could only be
obtained by a decrease in domestically used consumer and investment
goods. The situation is quite different in the experiments as empha-
sized by Table 4.5 below.
Consider first the impact effect of devaluation in 1973. The
economy produces 42% more exports in real terms and, in exchange, gets
only 15% more imports, reflecting a quite severe terms-of-trade effect.
Even so, investment increases by almost 9% and consumption also increases
marginally. All this reflects a movement of the economy towards its
transformation frontier. Thus, if export expansion is the outcome of
a more efficient allocation of resources and of better capacity
utilization,,it need not occur at the expense of investment or consump-
tion at home:. If, to the contrary, it were to reflect a simple diver-
sion of goods from the domestic market achieved by administrative
methods without an underlying change in resource allocation and
efficiency, ithen it would have to be paid for by quite severe cuts
in domestic consumption and/or investment. This is a point that needs
to be strongLy stressed, particularly when facing an unfortunate tendency
to deal with exports by administrative methods.
Dynamically, considering the cumulative sums in the second
part of Tables 4.5, the effective resource augmentation effect of a high
exchange-rate policy is even stronger. The cumulative stream of consump-
tion is higher with than without a major devaluation. The difference for
the stream of investment is a very substantial 14%. The static and dynamic



- 98 -
Table 4.5: Effects of Devaluation on Consumption, Investment
and the Trade Balance in Constant Domestic Prices
1979
Static Impact                Basic Run                Experiment B-1
Consumption a/                 346.9                      349.5
Investment                      85.4                       92.7
Exports                         35.8                       50.9
Imports                       -36.9                       -42.3
GDP                            431.2                      450.8
1979-1983
Cumulative Sum                Basic Run               Experiment B-1
Consumption a/                1,958.4                   2,000.6
Investment                      481.0                     547.7
Exports                         207.5                     333.6
Imports                        -206.2                    -250.7
GDP                           2,440.7                   2,631.2
Units: billions at 1973 TL
a/ Includes private plus government consumption.



resource reallocation and resource augmentation effects are thus sufficient
to more than outweigh the adverse terms-of-trade effect.
We have so far discussed resource allocation and capacity
utilization. But the total amount of real capital accumulation is also
affected by devaluation. Total real capital formation in any one year is
determined by the size of the nominal investment fund and the price of
investment goods. Both are affected by trade policy.
With a higher exchange rate, government tariff revenue increases
substantially. Given that we are exogenously fixing government consump-
tion expenditure in real terms, increased total revenue leads to greater
government investment. This is, however, partly offset by a decrease in
profit incomes induced by the devaluation. Table 4.6 below summarizes
these results.
Table 4.6: Effects of Devaluation on Total Nominal Investment
in 1979
Basic           Experiment
1979                      Run                 B-1
Total profit income                       719.0            692.3
Total government revenue                  370.4            384.4
Government investment                     112.8            122.8
Private & corporate sector investment     245.6            242.0
Total nominal investment                  358.4            364.3
Total nominal fixed investment            341.9            334.4
Units: billions of TL.



- 100 -
The increase in the investment fund depends crucially on the
type of government behavior specified. If the government spent its
additional revenue on consumption, the investment fund would decline.
Thus the assumption that the government keeps the real level of its
current expenditure constant is important.
Real capital accumulation is determined by dividing nominal
investment by the price of aggregate capital goods. In experiment B-i,
the latter falls by 7% after the 1979 devaluation, leading to a signifi-
cant increase in real capital accumulation. The opposite is often
assumed to occur. The relative price of capital goods is taken to
increase after a devaluation, reflecting the higher cost of imported
machinery. But this reasoning again assumes conditions of free converti-
bility. While the price of imported capital goods may be lower before
devaluation, the demand for imports cannot be satisfied and the domestic
user is forced to buy much more expensive domestic substitutes. The
same occurs for intermediates that enter the production of capital goods.
With the relaxation of the import constraint after a successful devalua-
tion, domestic production costs decrease as resources are utilized more
efficiently. While imports may be more expensive than before the exchange
rate adjustment, they are still less expensive than the domestic substi-
tutes produced under conditions of tight rationing. On balance, these
trends lead to a significant decline in the relative cost of capital
goods and the rate of growth of the economy-wide aggregate capital stock,
valued in constant base year prices, rises from 4.5% to 5.2% in experiment
B-1 (See Table 4.7 below).



- 101 -
The results of experiment B-1 reflect a growth path that is
much closer to the target path of Turkish policy makers, at least in
terms of overall growth rates. It does however have an important impact
on industrial structure and the pattern of industrialization. Sectors
such as chemicals, steel and machinery grow significantly less rapidly then
when protected by severe import rationing as in the basic run. Exports
remain heavily concentrated in food, textiles and clothing, and indus-
trial structure shifts significantly towards the lighter consumer goods
industries. This indicates that while overall growth is substantially
higher, the implied structure of growth is not consistent with the
desire of Turkish policy makers to achieve rapid development of heavy
industries.
Rapid growth in the capital goods and basic intermediate
industries has to proceed on a broad integrated front if there is to be
substantial net import substitution. This in turn does not release
enough resources for rapid export expansion. An inward-oriented
strategy based on massive import substitution in heavy industry would
constrain the economy to a rather moderate growth rate of 5 or 6 percent.
To achieve a growth rate in the vicinity of 8%, the emphasis would
have to be shifted, at least initially, to the export-oriented sectors
and to subsectors producing exportables in the basic intermediate and
capital goods industries. While the possible growth rate of exports
originating in the capital-intensive and technology-intensive sectors



- 102 -
may be very high, their base is so small that it is inevitable that the
bulk of exports will consist of primary products, food, textiles, clothing
and other relatively simple products. This is certainly true in the
short to medium run. In the long run, it can be argued that a prior
condition for the successful development of the technologically more
advanced industries is the end of the continuous crisis situation and a
resource-use and allocation mechanism that is more geared to world-
market norms and conditions. But from a shorter-run perspective, a trade-
off does exist between the objective of quickly expanding the capital-
goods and intermediate-goods sectors on a broad front, and the objective
of overcoming the foreign-exchange constraint by a massive export drive.
To test the feasibility of combining rapid overall growth
performance with an industrial strategy that continues to emphasize the
development of the capital goods and intermediate goods sectors, and to
analyze the extent of the resource mobilization effort necessary to
achieve success with such a strategy, we conducted a third experiment,
assuming the manufactured exports and growth-oriented policy package
described above.
The macroeconomic results of complementing the high exchange
rate policy with substantial export subsidies to nontraditional exports
and an investment drive financed by higher taxes on capital income and
higher savings from labor income are summarized in Table 4.7 and 4.8.
The aggregate results are not dramatically different from those reported



- 103 -
Table 4.7: Effects of Alternative Policies on Growth
Performance in the FFYP Period
Average Annual Growth Rates: 1979-1983
Basic Run              B-1                  B-2
Manufactured Exports
Constant Price Deflated   High Exchange    & Investment Oriented
Exchange Rate Policy       Rate Policy        Policy Package
GDP                                5.5%                 7.7%                7.8%
Agriculture                      3.9                  5.6                 5.5
Industry                         6.7                  8.9                 9.5
Services                         5.4                  7.4                 7.3
Economy-wide
Capital Stock                    4.5                  5.2                 5.4
Urban Labor Force                  4.2                  3.2                 3.9
Real Average Urban
Wage                             1.4                  2.3                 2.7
Exports,  current dollars         16.0                 25.5                26.2
constant dollars        6.4                  15.5               15.8
Imports,  current dollars         12.3                 16.7                16.9
constant dollars        3.0                   7.1                7.2



Table 4.8: Effects of Alternative Policies on Macroeconomic
Variables in 1983
Ratios (%) to Basic Run
Experiment Experiment         Experimnent  Experinment
1983                  Basic Run      B-1       B-2                B-1         B-2
GDP                                          548.5     608.9       611.5             110.0%       114.72
Total consumption                            438.9     454.0       446.5             103.4        101.7
Cross fixed investment                       100.9     120.3       128.4             119.2        127.3
Economy-wide capital stock                 1,153.8   1,196.3     1,205.7             103.7        104.5
Capital stock in manufacturing               225.8     235.0       237.1             104.0        105.0
Emuploymilent in manufacturing             2,128.1   1,997.1     2,053.4              93.8         96.5
Econoimy-wide profit rate                     17.5%     18.3%       17.9%            104.6        102.3
Profit rate in manufacturing                  35.1      29.4        30.3              83.8         86.3
Capital goods price index (1973  100)        756.6     673.9       693.1              89.1         91.6
Merchandise imports (billion $)                9.1      10.9        11.1             120.8         122.1
Herchandise exports (billton $)                4.8       7.3         7.4             150.5        152.9
Manufacturing exports (billion S)              0.9       1.5         1.6             162.6        170.4
Exchange rate (TL/$)                          34.9      52.7        52.7             151.0        151.0
Degree of imiport rationing a/                58.0      15.0        19.0              25.8         32.8
Debt service ratio b/                         35.7      28.5        28.2              79.8         79.0
Net capital inflow/merchandise exports        15.5      10.8        10.1              66.5         65.2
Units: CDP, consumption, investmuent, capital stock: billion 1973 TL.
Employmtient: thousand workers
a/ Defined as 1 - actual imports/desired imports.
b/ '(Interest + amortization)/(exports + nfs + workers' remittances)



- 105 -
for the high, exchange rate case (B-1). GDP grows only marginally faster.
Exports in 1983 are not much higher than before. The attempt at more
rapid real capital accumulation faces two difficulties: the relative
price of investment goods increases and the foreign exchange constraint
becomes relatively more binding, leading to an increase in rationing.
The subsidies to nontraditional exports do not have a major impact on
the total quantity of exports because of their very small initial share
(14% of total exports in 1973). Overall economic growth remains con-
strained by the growth of exports as well as the small base from which
they grow.
There are, however, significant differences between the policy
package experiment (B-2) and the high exchange rate scenario (B-1), and
the differences become more important towards the end of the planning
period when the gradually rising investment rate has significant effects.
Industrial growth is 0.6 percentage points higher over the plan period
(1979-83) and almost a full percentage point higher in the final year
than in the devaluation-only experiment (B-1). While consumption in
1983 is slightly lower in experiment B-2, the economy-wide capital
stock is about one percent higher than in experiment B-1, leading to
an increase in urban employment and the real urban wage, both of which
grow significantly faster in experiment B-2.
The experiments also differ substantially with respect to the
share of government in national expenditure. While current government
expenditure and transfers are always held constant in real terms across



- 106 -
all experiments, government investment increases substantially in
experiment B-2. Table 4.9 compares some of the results with respect to
the government's share and role in the economy. 1/ In addition, experi-
ment B-2 significantly reduces the after tax share of profits and the
composition of saving.
Table 4.9
Comparing the Effects of Alternative Policies on
Government Revenue, Profits, and the Financing of
Investment in 1983
(percent)                  Basic Run      Experiment B-1   Experiment B-2
Government Expenditure/GDP            26.1%            27.0%             28.6%
Direct Taxes/GDP                      13.0             11.5              13.1
After Tax Profit Income/GDP           40.5             39.3              35.3
Share of Investment Financed by:
Labor Saving                         5.6              5.7               9.2
Direct Government Saving            34.4             38.3              40.8
Capitalist Saving & SEE's           60.0             56.0             50.0
While important changes are brought about by the manufactured-
exports and investment-oriented policy package, overall growth performance
remains very close to what it was in experiment B-1. But the sectoral
pattern of growth is in fact quite different. The combination of export
subsidies to "infant" export sectors and increasing investment demand
1/  Note the total government expenditure refers to expenditure from the Con-
solidated Budget and does not include the SEE's which the TGT model in its
present form treats as consolidated with the private corporate sector.



- 107 -
creates important resource pulls favoring the heavy industries and the
machinery sector in particular. Thus the policy-package strategy is one
that achieves relatively rapid and rising overall growth without shifting
the pattern ofE growth too strongly towards the light consumer industries.
It is this kind of strategy that conforms best to Turkey's objectives.
It does however imply not only a major devaluation in 1979, but also a
very successful effort at resource mobilization and a relatively slow
growth of consumption.
The results described so far have provided a consistent view
of the macroeconomic variables over the FFYP period and their sensitivity,
in particular, to the exchange rate. They seem to provide a strong case
for a major exchange rate adjustment to take place as soon as possible.
Without being too optimistic about the response of exports to devaluation,
the combined effects of export expansion and lessening of import rationing
that would follow such a major adjustment allow the growth rate to increase
from 5.5% in the basic run to 7.7 or 7,8% in experiments B-1 and B-2.
Comparing the basic run with the two policy experiments indicates
that, during the fourth five-year plan period, Turkey probably cannot
achieve her historical 7-8% rate of growth without dealing with the
foreign-exchange problem through a major real adjustment in the exchange
rate and a shift to export-intensive growth. This conclusion is further
supported by the analysis of sectoral changes implied by the different
policy regimes that is presented in the next section. However, as
discussed above, there are trade-offs and there is no policy that



- 108 -
achieves all objectives without costs. It is thus especially important
to consider carefully the implications of the different policy regimes on
the structure of the economy,to identify the sectors which gain or lose,
to get a better sense of the timing and sequences of sectoral growth
involved, and to evaluate the reasonableness of the different regimes
from the point of view of their sectoral implications.



- 109 -
5.  Microeconomic Analysis of the Impact of Trade Policy on Industrial
Structure and the Sources of Industrial Growth
5.1  Introduction:  Trade Policy and Resource Allocation
The economy-wide perspective and the alternative scenarios
presented so far rely on developments in each of the 19 sectors distin-
guished by the TGT model. While a 19 sector disaggregation is not enough
to allow a real link between economy-wide analysis and analysis at the
microeconomic project level, consideration of sectoral developments
implicit in the macro-perspective is important in evaluating the macro
results. The reasonableness of the overall growth projections depends on
the reasonableness of the underlying sectoral growth rates, and 19 sectors
represent enough disaggregation to provide a useful framework for more
detailed sectoral analysis. Evaluation of sectoral growth must focus not
only on the rise in domestic production as such but also on the pace of
import substitution and/or export expansion that is implicit in the
sectoral growth projections. It is not only overall growth performance,
but also the sectoral pattern of growth that concerns most policy makers.
What is required is not only rapid growth but also a deepening of the
industrialization process through the development of the technologically
more advanced industries.
Relative prices and relative price effects play a very important
role in determining the resource allocation and resource growth that in
turn determine the structure of sectoral growth. The low-exchange-rate



- 110 -
experiment that keeps constant the price-deflated exchange rate leads to
quite different resource pulls than those of the high exchange rate
strategies of experiments B-1 and B-2. In the first case, severe import
rationing provides a great stimulus to import substitution and resources
are pulled into such sectors as chemicals, basic metals and machinery.
Exports on the other hand are not encouraged and resources are pulled away
from export-oriented sectors. On the other hand, a high exchange rate
changes the structure of relative prices and incentives in favor of
export-oriented activities. The rise in imports and relaxation of import
rationing allowed by export expansion diminishes the relative profitability
of import substitution and pulls resources away from import substituting
sectors. In the policy package experiment, this tendency is partly
counteracted by the higher investment rate and the subsidies to non-
traditional exports.
The mechanisms at work in the TGT model are quite different from
the simpler mechanisms stressed by pure trade theory and models that analyze
devaluation under conditions of free convertibility. As recently stressed
1/
by Anne Krueger in her comparative analysis of trade liberalization attempts
in several countries, including Turkey, a devaluation under conditions of
tight exchange control can have very different effects from a devaluation
under free convertibility. In the latter case a devaluation simply results
in an increase in the relative price of all tradeable commodities, be they
exportables or import substitutes. Resources will be pulled into all
tradeable sectors, away from the non-tradeable sectors. Within the
1/ See Krueger (1978)



framework of the TGT model, the situation is quite different. First of
all there is no clear-cut distinction between tradeables and non-tradeables.
Instead, there is a continuum: some sectors where import or export shares
are large and substitution elasticities are high are very "tradeable,"
while other sectors where import or export shares are small and the substi-
tution elasticities are low can be characterized as more "non-tradeable."
Furthermore, import rationing and exchange control imply that a successful
devaluation will lead not to a decline but to an increase in realized
imports, and therefore to a reduction in the relative profitability of
import substitution. Thus, contrary to traditional analysis, a devaluation,
while favoring export expansion, will not encourage import substitution.
Finally, the TGT model specifies an inverse link between capacity
utilization and the degree of import rationing.   This can also be inter-
preted to mean that rationing absorbs real resources that would be released
with a decrease in the degree of rationing. Real resources are used not
only by the government in the administration of the rationing system but
also by producers who have to use some of their resources to compete for
licenses and 'Lobby for their share of imports.
These fundamental points should be kept in mind when evaluating
the results presented below and when relating them to the microeconomics
of neoclassical trade theory.
5.2 Sectoral Aggregation and Sectoral Trade Characteristics in the TGT
Model
The TGT model distinguishes 19 sectors which are aggregations
of the sectors distinguished in the 1973, State Institute of Statistics



- 112 -
Input-Output Table. Details of the sectoral aggregation are provided in
Appendix B, but some important characteristics of the aggregation should
be kept in mind when evaluating the microeconomic results presented in
this section.
First, lightly processed agriculture commodities appear in the
manufacturing sectors, not in agriculture. Thus, preserved fruits and
vegetables, olive oil and tobacco are included in the food sector and
cotton ginning is included in the textiles sector. Second, crude oil and
natural gas are aggregated with refining in an integrated petroleum and
petroleum products sector. These sectoral definitions are especially
important to keep in mind when analyzing the trade projections.
Table 5.1 below summarized some important data that help
characterize the individual sectors' role in trade. A complete description
of sectoral production function parameters is given in Appendix B. The
first column presents the ratio of imports to domestic goods in domestic
use of each category of commodities. This RMDi ratio is measured in constant
base year (1973) prices. The second column presents the proportion, IMDi,
of imported intermediates in total intermediate input use for each sector,
measured in base year prices. The third column tabulates the Armington
trade substitution elasticities, ai, assumed in the TGT model. Column four
provides the ratio REX. of exports to total domestic production, again in
constant prices. Finally, column five contains the assumed export demand
elasticities.
Several things should be noted in Table 5.1. Turkey does not
import significant amounts of finished consumer goods or agricultural



Table 5.1: SECTORAL IMPORT DEPENDENCE, EXPORT SHARES, AND TRADE ELASTICITIES
Ratio of         Share
Imports to       of Imported                      Share
Domestically     Intermediates   .Armington       of Exports     Sectoral
Produced Goods   in Total         Trade           in Total       Export
in Total         Intermediate     Substitution    Domestic       Demand
Domestic Use     Tnptttq          Rlasticities    Production     Elasticities
RMDi   IMDi      -i             REX1
1.  Agriculture                 1.0            8.6             2.00            1.5          Fixed World Price
2. Mining                       6.9           18.5             0.50           15.2          Fixed World Price
3.  Food                        0.9            2.1             0.66           11.0            2.00
4.  Textiles                    1.7            8.2             0.66           14.3          Fixed World Price
5. Clothing                     2.0            4.0             0.66            5.7            2.00
6. Wood & Wood Products         0.4            3.6             0.66            0.4            2.00
7. Paper & Printing             8.1           10.8             0.66            0.4            2.00
8.  Chemicals                  59.1           31.0             0.33            1.5            2.00
9.  Rubber & Plastics          15.5           22.6             0.33            1.1            2.00
10.  Petroleum & Pet. Prod.     21.5           20.8             1.50            1.2            2.00
11.  Non-Met. Min. Prod.         5.9           10.5             0.66            2.7            2.00
12. Basic Metals                29.7           20.6             0.50            1.2            2.00
13.  Metal Products             15.4           22.8             0.50            2.0            2.00
14.  Non-Electric Machinery     86.1           44.3             0.33            0.5            2.00
15. Electric Machinery          50.3           27.3.            0.33            0.2            2.00
16. Transp. Equipment           29.1           22.9             0.75            0.1            2.00
17. Construction                 --            17.0             --               --             --
18. Infrastructure               1.7           16.3             0.20            5.8            1.25
19.  Services                    1.4            4.4             0.20            3.4            1.25
Note: Following the 1973 S.I.S. Input-Output Table trade and transportation margins are not distributed
to the individual sectors but appear in the last two sectors.



products and the proportion of imports in total domestic demand for "light"
manufactured goods such as processed food, textiles, clothing and wood
products is minimal. But Turkey imports almost half of its non-electrical
machinery, and is heavily dependent on imports of chemicals, electrical
machinery, basic metals and transport equipment. Turkey is also very
dependent on imports of crude petroleum. Almost all refining is done
domestically, so that the import ratio in sector 10 which aggregates crude
petroleum with petroleum products is only 21.5%. But in this case it would
be more meaningful to look at the share in current prices that partly
1/
reflects the much higher world price of oil.   The current price ratio is
35% for 1977.
The RMDi ratios' tell us something about which sectors will face
strong excess demand when import rationing increases. But the degree of
import dependence and the degree to which capacity utilization will be
effected by rationing is reflected in the IMDi ratios that give the pro-
portion of imported intermediates in total intermediate inputs. Note that
the IMDi ratios provide only "first-round" direct estimates of import
dependence since they do not take into account the indirect dependence
created through input-output linkages, capital requirements and general
equilibrium price effects. Food, clothing, wood products and services
are the sectors that appear least dependent on imports, followed by textiles
and agriculture. The average IMDi ratio in manufacturing is 15.2% and the
economy-wide ratio is 11.7%. The sectors most dependent on imported inter-
mediate inputs are non-electrical machinery with an IMDi ratio of 44%,
followed by chemicals, 31% and electrical machinery, 27%.
1/ Turkey had not, as of July 1978, fully adjusted the domestic price to
the world price.



- 115 -
Columnn three provides the assumed substitution elasticities.
These are guessed parameters attempting to capture intrasectoral product
mix characteristics. They are lower than the kind of elasticity estimates
one gets from the econometric studies of trade between industrialized
1/
countries and reflect the much more important quality and product mix
differences that exist between a developing economy and its developed
trading partners. They constitute an intermediate specification between
the pure complementarity assumption of fixed-coefficients models and the
perfect substilutability assumption made by pure trade theory.
Column four in Table 5.1 gives the share of exports in total
domestic output. There are really only 3 sectors that have a substantial
2/
export ratio: mining, textiles and food. Other sectors with an export
ratio above 2% are clothing, non-metallic mineral products and metal products.
Trade and transportation margins, as well as tourism revenues and revenues
from air-traveL,lead to significant REXi ratios also in infrastructure and
services. It is striking how low the export ratio is in most manufacturing
sectors. If one excludes food, clothing and textiles, the export ratio in
manufacturing averages to only 1%1 The bulk of Turkish exports still
consists of only lightly processed agricultural products: tobacco, hazelnuts,
dried fruit, ginned cotton, cotton cloth and, in recent years, wheat. To
this must be added a small amount of mineral products; principally raw
borates and chromium.
1/ See, for instance
2/ Excludes crude petroleum.



- 116 -
While manufactured exports remain insignificant, they showed the
first signs of life in the 1970-1973 period with the appearance of exports
in clothing, leatherware, cement, building materials, glass and glass
products, fabricated metal products and light electrical machinery. With
appropriate policy support, these exports could probably have grown rapidly
in the 1973-1977 period, taking advantage of the vicinity of the booming
Middle Eastern market. Their very small base should also have allowed
rapid expansion in the markets of industrialized countries. But in real
terms these exports declined over the 1973-1977 period reflecting the
increasing overvaluation of the Turkish Lira and the extreme domestic-
market orientation of economic policies, relative prices and incentives.
The last column in Table 5.1 presents the assumed export demand
elasticities for all sectors except agriculture, mining and textiles where
it was preferable to make the small country assumption, Turkey essentially
being a price taker for such homogenous products as cotton, wheat and
mining products that dominate trade in these sectors. All other elasti-
cities were set equal to 2, which reflects an intermediate position between
the extreme elasticity pessimism common in Turkey and the more optimistic
view of many outside researchers such as Krueger. It should also be
emphasized that these elasticities are short to medium-term elasticities.
Truly long-run elasticities should probably be set higher.
5.3 Growth afid Industrial Structure Under Alternative Trade Exchange
Rate, and Investment Policies
Table 5.2 presents changes in sectoral prices, net prices (value-
added prices), wages and profit rates that generate the important resource pulls



Table 5.2:  RATIOS TO THE BASIC-RUN VALUES OF DOMESTIC PRICES, NET PRICES,
WAGES, AND PROFIT RATES IN 1983
Experiment B-1                             Experiment B-2
Devaluation Only                            Policy Package
Domestic     Net                Profit     Domestic     Net                Profit
Price     Price      Wage      Rate        Price     Price      Wage      Rate
PD.       PN.                  Ri          PD.       PN.                  Ri
- - - - - - - - - - - - - - - - -  (Percent) - - - - - - - - - - - - - - - - - -
1. Agriculture                 110.1     115.9      120.0    135.4        107.8     112.9     116.8     127.5
2. Mining                       82.6      94.2      100.2    106.6         83.8      94.9      99.8     108.2
3. Food                        107.4     118.7      106.0    147.5        106.0     116.9     106.6     140.1
4. Textiles                     84.3     110.0      105.5    159.2         83.3     110.5     106.1     151.3
5. Clothing                    102.8     112.9      113.9    141.5        103.6     115.5     116.5     142.6
6. Wood & wood products        105.9     112.5      114.1    135.1        105.7     115.3     116.5     136.4
7. Paper and printing           93.9      89.8      106.1    100.7         94.1      89.7     106.8      97.0
8. Chemicals                    64.7      37.2      101.6     33.2         66.4      40.1     101.5      35.6
9. Rubber and plastics          84.2      76.9      107.9     76.3         85.2      78.4     109.0      76.3
10. Petroleum & pet. prod.       88.1      76.8      100.2     94.5         90.6      79.7      99.8      95.1
11. Non-met. min. prod.          98.8     104.8      104.6    133.7        100.5     108.3     105.1     138.4
12. Basic metals                 70.2      54.5      100.2     53.1         74.6      61.4      99.8      61.1
13. Metal products               83.3      99.1      112.3    118.0         85.9     100.3     114.4     115.8
14. Non-electric machinery       68.2      41.9      106.8     41.5         73.0      50.4     107.6      51.5
15. Electric machinery           75.6      63.9      108.4     63.4         78.9      68.9     109.6      68.8
16. Transp. equipment            86.1      81.1      109.6     87.7         88.9      85.2     111.2      91.8
17. Construction                 92.4      98.5      100.2    125.8         95.3     102.7      99.8     134.5
18. Infrastructure               99.9     101.2      100.2    119.6         99.5      98.0      99.8     110.7
19. Services                    101.9     102.1      109.8    114.1        101.7     102.0     113.3     109.6



- 118 -
in the TGT model. Taking the basic run values as 100.0, the Table presents
changes in these values due to experiments B-1 (devaluation only) and the
policy package experiment B-2 (devaluation complemented by export subsidies
and an investment drive financed by taxes on capital income).
Compare first the basic-run constant price-deflated exchange-rate
scenario to experiment B-1 characterized by a substantial real devaluation
and a continued high exchange rate policy. Both scenarios are characterized
by a moderately rising nominal investment rate that increases from about
21.5% in 1978 to 23.5% in 1983. Note, however, that real capital accumulation
is significantly higher in experiment B-2 due largely to a decline in the
average cost of capital goods allowed by the relaxation of import constraints.
The sectors that clearly gain from a "pure" devaluation policy are
the export oriented light manufacturing sectors as well as agriculture and
mining. On the other hand, basic intermediates and capital goods sectors are
adversely affected by the higher exchange rate policies--particularly chemi-
cals, basic metals and machinery. The greater availability of imports
diminishes the need for these three sectors to produce high-cost import
substitutes and the resources attracted to the basic industries under tight
import rationing are released into the more export-oriented light manufac-
turing activities. Note that the policy package (B-2) to some extent counter-
acts this tendency and raises the prices and relative profitabilities of
basic industries.
The mechanism at work in these experiments is very different from
the mechanism stressed by pure trade-theory models that assume free convert-
ibility. As discussed above, the analysis of devaluation under conditions



- 119 -
of exchange control is quite different from the usual analysis that argues
that all sectors producing tradeables will be positively affected by a
devaluation. En the TGT model the mechanism at work correctly reflects the
dilemma perceived in Turkey: a high exchange rate policy that stimulates
exports and overcomes the foreign exchange constraint does lead to negative
resource pulls affecting basic intermediates and capital-goods sectors and
also to a pattern of growth that does not reflect the Turkish desire to
deepen the industrialization process.
Table 5.3 describes sectoral growth and economic structure under the
three alternative scenarios. Total gross output grows at 5.9% in the basic
run (reflecting an inward-looking, import-substitution-biased growth
strategy), at 7.2% with the high exchange rate policy and at 7.5% when the
high exchange rate policy is complemented by an investment and manufactured
exports oriented policy package. Note that while the differences in growth
rates are subst:antial, 5 years is too short a period for them to really have
an impact on the structure of output. Table 5.4 summarizes the growth of
sectoral export: earnings in current dollars and emphasizes why a major
devaluation is considered necessary in the first place.
Let Us again start by comparing the basic run with pure devaluation
experiment. The sectors that gain and lose from a high exchange rate policy
are grouped and ranked by the absolute value of average output growth
differential from the Basic Run in Table 5.5 below.
There is not a one-to-one correspondence between Table 5.2
on the one hand and Tables 5.3 and 5.5 on the other. Several other
factors, notably changes in capacity utilization are at work besides
prices and profit rates. The sectors that grow faster include not only



Table 5.3
Sectoral Growth and Structure
Annual Growth Rates 1978-83                       Structure of Output, 1983
Constant    I High Exchange     Policy     1    Constant      High Exchange     Policy
PLD-ER            Rate         Package          PLD-ER           Rate          Package
Base Run           B-1            B-2           Base Run          B-1             B-2
1.  Agricuture                4.6%            5.5%           5.4%            19.1%          18.7%            18.5%
2.  Mining                    8.2             9.3           10.0              1.0            1.0              1.0
3.  Food                      5.7             9.0            9.0              8.4            9.2              9.1
4.  Textiles                  4.7            10.4           10.2              4.3            5.2              5.2
5.  Clothing                  7.0            10.1           10.5              2.1            2.2              2.3
6.  Wood & Wood Products      5.6             7.0            7.7              1.3            1.3              1.3
7.  Paper & Printing          6.7             6.5            6.7              1.1            1.0              1.0
8.  Chemicals                10.3             5.7            6.5              2.9            2.2              2.3
9.  Rubber & Plastics         7.5             5.9            6.3              1.2            1.0              1.0
10.  Petroleum & Pet. Prod.   10.0            13.2           13.4              3.9            4.2              4.2
11.  Non-Met. Min. Prod.       5.7             8.1            9.0              1.2            1.3              1.4     0
12.  Basic Metals              9.8             7.4            8.8              4.2            3.6              3.8
13.  Metal Products            5.1             6.7            6.8              1.1            1.2              1.1
14.  Non-Elec. Machinery      10.7             5.7            7.8              3.1            2.4              2.5
15.  Elec. Machinery           9.0             5.7            7.0              1.3            1.1              1.1
16.  Transp. Equipment         8.3             7.6            8.7              3.0            2.8              2.9
17.  Construction              5.3             8.6           10.3              5.4            6.0              6.4
18.  Infrastructure            7.4             9.4            9.5             12.8           13.2             13.0
19.  Services                  6.8             7.4            7.3             22.6           22.4             22.0
A,erage              5.9             7.2            7.5           100.0           100.0           100.0



Table 5.3
Sectoral Growth and Structure
Annual Growth Rates 1978-83                       Structure of Output, 1983
Constant    I High Exchange     Policy     1    Constant      High Exchange     Policy
PLD-ER            Rate         Package          PLD-ER           Rate          Package
Base Run           B-1            B-2           Base Run          B-1             B-2
1.  Agricuture                4.6%            5.5%           5.4%            19.1%          18.7%            18.5%
2.  Mining                    8.2             9.3           10.0              1.0            1.0              1.0
3.  Food                      5.7             9.0            9.0              8.4            9.2              9.1
4.  Textiles                  4.7            10.4           10.2              4.3            5.2              5.2
5.  Clothing                  7.0            10.1           10.5              2.1            2.2              2.3
6.  Wood & Wood Products      5.6             7.0            7.7              1.3            1.3              1.3
7.  Paper & Printing          6.7             6.5            6.7              1.1            1.0              1.0
8.  Chemicals                10.3             5.7            6.5              2.9            2.2              2.3
9.  Rubber & Plastics         7.5             5.9            6.3              1.2            1.0              1.0
10.  Petroleum & Pet. Prod.   10.0            13.2           13.4              3.9            4.2              4.2
11.  Non-Met. Min. Prod.       5.7             8.1            9.0              1.2            1.3              1.4     0
12.  Basic Metals              9.8             7.4            8.8              4.2            3.6              3.8
13.  Metal Products            5.1             6.7            6.8              1.1            1.2              1.1
14.  Non-Elec. Machinery      10.7             5.7            7.8              3.1            2.4              2.5
15.  Elec. Machinery           9.0             5.7            7.0              1.3            1.1              1.1
16.  Transp. Equipment         8.3             7.6            8.7              3.0            2.8              2.9
17.  Construction              5.3             8.6           10.3              5.4            6.0              6.4
18.  Infrastructure            7.4             9.4            9.5             12.8           13.2             13.0
19.  Services                  6.8             7.4            7.3             22.6           22.4             22.0
A,erage              5.9             7.2            7.5           100.0           100.0           100.0



- 122 -
Table 5.5: THE SECTORAL IMPACT OF EXCHANGE RATE
POLICY (B-1): GAINERS AND LOSERS
(in decreasing order)
Sectors that grow faster than in      '    Sectors that grow more slowly
Basic Run                               than in Basic Run
Textiles                                   Non-electrical machinery
Food products                              Chemicals
Construction                               Electrical machinery
Petroleum & petroleum products             Rubber and plastics
Clothing                                   Basic metals
Non-metallic mineral products              Transport equipment
Wood and wood products                     Paper and printing
Infrastructure
Metal products
Services
Mining
Agriculture
all the export-oriented sectors but also construction, petroleum and
petroleum products and, less importantly, infrastructure and services.
The sectors that grow more slowly constitute a more homogenous group: they
are the import-substituting sectors for which rationing creates the greatest
excess demand. But they are also the technologically more advanced manu-
facturing sectors and it is their rapid development that is taken to reflect
deepening of the industrialization process and constitutes a major objective
of Turkish development policy.



- 123 -
It is therefore clear that a successful devaluation alone cannot
be regarded as an optimal policy. It does lead to export expansion, it
does overcome the foreign exchange gap and it does lead to rapid economy-
wide growth but: at the expense of a slowdown in the development of the
basic intermediates and capital goods industries. While in the short-run,
Turkey may not have much choice given the absolutely overriding need to
expand exports, it appears that a high-exchange-rate policy, undertaken on
its own, does rLOt constitute an acceptable solution to Turkey's problems.
The third scenario (experiment B-2) reflects the search for a
better policy package that reconciles the need for expanding export earnings
with the desire to emphasize development of the basic intermediates and
capital-goods siectors. While experiment B-l only assumes a high-exchange-
rate policy, the "policy package" experiment B-2, complements higher
exchange rates with a 50% increase in the subsidies accorded to non-tradi-
tional exports and a significant investment drive that raises the investment
rate from a terminal year value of about 23.5% in the basic run and
experiment B-1, to 25.6% in experiment B-2.
As can be observed in Table 5.3, relative to the high-exchange-
rate policy alone, output grows faster with the policy package in all sectors
except agricult.ure, food, textiles and services. The growth differential is
substantial in mining, chemicals, non-metallic mineral products, basic metals,
machinery, transport equipment and construction. Thus in basic metals the
average annual growth rate goes up from 7.4% to 8.8%. In non-electrical
machinery it goes up from 5.7% to 7.8%. These are not average rates that are
as high as the ones projected by the basic run, but they are significant



- 123 -
It is therefore clear that a successful devaluation alone cannot
be regarded as an optimal policy. It does lead to export expansion, it
does overcome the foreign exchange gap and it does lead to rapid economy-
wide growth but: at the expense of a slowdown in the development of the
basic intermediates and capital goods industries. While in the short-run,
Turkey may not have much choice given the absolutely overriding need to
expand exports, it appears that a high-exchange-rate policy, undertaken on
its own, does rLOt constitute an acceptable solution to Turkey's problems.
The third scenario (experiment B-2) reflects the search for a
better policy package that reconciles the need for expanding export earnings
with the desire to emphasize development of the basic intermediates and
capital-goods siectors. While experiment B-l only assumes a high-exchange-
rate policy, the "policy package" experiment B-2, complements higher
exchange rates with a 50% increase in the subsidies accorded to non-tradi-
tional exports and a significant investment drive that raises the investment
rate from a terminal year value of about 23.5% in the basic run and
experiment B-1, to 25.6% in experiment B-2.
As can be observed in Table 5.3, relative to the high-exchange-
rate policy alone, output grows faster with the policy package in all sectors
except agricult.ure, food, textiles and services. The growth differential is
substantial in mining, chemicals, non-metallic mineral products, basic metals,
machinery, transport equipment and construction. Thus in basic metals the
average annual growth rate goes up from 7.4% to 8.8%. In non-electrical
machinery it goes up from 5.7% to 7.8%. These are not average rates that are
as high as the ones projected by the basic run, but they are significant



- 125 -
Table 5.6:  COMPARISON OF SECTORAL OUTPUT GROWTH RATES
IN 1983
Base run    B-1       B-2       Base    B-1      B-2
Output    Output    Output     Rank    Rank     Rank
1. Agriculture                 4.9%      5.8%       5.8%       19      17      19
2. Mining                      8.6       9.4      10.5          5       4       3
3. Food                        5.9       9.0        7.9        16       7      12
4. Textiles                    5.1      10.1        9.2        18       2       9
5. Clothing                    7.1       9.0       10.0        10       7       5
6. Wood & wood products        6.3       6.9        7.7        15      14      13
7. Paper and printing          7.1       7.1       7.2         10      13      17
8. Chemicals                  10.1        5.6       6.8         2      18      18
9. Rubber and plastics         8.1       6.7        7.6         9      15      15
10. Petroleum & pet. prod.     11.8      14.4       14.7         1       1       1
11. Non-met. miii. prod.        6.6       9.3        9.8        13       5       6
12. Basic metals                9.5       7.6        9.5         4      10       7
13. Metal products              5.2       7.3        7.6        17      12      15
14. Non-electric machinery      9.7       5.4       8.2          3      19      10
15. Electric machinery          8.5       6.2       8.0          6      16      11
16. Transp. equ:Lpment          8.4       8.1        9.4         7       9       8
17. Construction                6.7       9.2      11.1         12       6       2
18. Infrastructure              8.2      10.1      10.1          8       2       4
19. Services                    6.5       7.6        7.6        14      10      15
Total                  7.0       8.1       8.5



- 126 -
that merchandise exports in 1983 would have reached 7.4 billion dollars
instead of 4.8, the debt-service ratio would be 24% instead of 32% and GDP
would be 11.4% higher in real terms! With the policy package (B-2), total
consumption would have to be 1.7% lower than with pure devaluation policy
(B-1), but it would still be 1.6% higher than in the basic run. Investment
would be 7% higher than in experiment B-1 and as much as 27% higher than
in the basic run.
The message that emerges from these experiments is relatively
clear. If Turkey were to devalue and resolutely start pursuing export-
oriented industrial strategy without significantly raising the investment
rate and without giving special support to "infant" manufacturing exports;
the result would be a growth path biased towards the light manufacturing
sectors, agriculture and services. If, on the other hand, Turkey were to
raise its investment rate and attempt to achieve rapid growth centered on
the advanced industrial sectors without first achieving a real devaluation
and a massive increase in exports, the attempt would fail because of the
binding foreign-exchange constraint. What is best, therefore, is a mixed
strategy. A real devaluation and substantial export-oriented production
increases in the light manufacturing sectors where Turkey can be immediately
competitive and has a market, are necessary conditions for overcoming the
crisis. But, complementing the real devaluation, it is necessary to provide
additional subsidies to infant manufactured-exports sectors and, even more
importantly, to raise the economy-wide investment rate. A target of 25.5%
for 1983 should not be too difficult to achieve. If it can be realized,
the initial bias towards light manufacturing activity, that is necessary



- 127 -
at the beginning of the FFYP period, can quickly be overcome and Turkey
could enter the Fifth Five-Year Plan period with a healthy base of
manufactured exports, ready for rapid and more export-oriented growth
in the basic intermediate and more advanced engineering industries.



- 128 -
5.4 Export Expansion, Import Substitution and the Sources of Growth
In this section we continue the analysis of the underlying
sectoral developments implied by the various scenario experiments under-
taken with the TGT model by decomposing sectoral growth into three
components: domestic demand expansion, export growth and import sub-
stitution. The decomposition will complement the preceding analysis of
sectoral production growth rates and lead to a much more complete view
of what is implied by different policies and growth paths.
The general approach was used by Chenery, Shishido and Watanabe
(1962) in their analysis of Japanese growth and the decomposition measures
have been further refined by Syrquin (1976). The decomposition measures
start from the material-balance equations of the input-output system and
then derive the expression for changes in sectoral output as a function
of changes in the various components of demand.
In the TGT model, the material-balance equations for the supply
of and demand for domestically produced goods can be written as:
Xi= Pi(Fi + wi) + Ei
where     Xi= domestic production in sector i,
i= ratio of domestic goods to composite goods
Fi = final domestic use demand for composite goods
W= intermediate domestic use demand for the composite goods
Ei = export demand for the domestically produced commodity



- 129 -
Since, in the model, the composite goods (Fi and Wi) are CES aggregations
of domestic and imported goods, the ratios p1i are functions of the
parameters of the aggregation function and of the ratio of the import
price to the domestic-good price in each sector.
Intermediate demands are determined by fixed input-output
coefficients, aii.   In matrix notation, the material-balance equation
can thus be written:
X = (I - uA)    (uF + E)
where u is a diagonal matrix of the pi ratios, A is the matrix of input-
output coefficients, and X, D, and E are vectors. The matrix uA is the
matrix of domestic-goods input-output coefficients.
Denoting the change in a variable by A [AX = X(t+l) - X(t)],
the change in total domestic demand can be written (after some algebraic
manipulation) as:
AX =     Ru(AF)           domestic demand expansion
+ R(AE)            export expansion
+ R(Au)(F + W)     import substitution
+ Ru(AA)X          change in input-output coefficients
A-l
where R    (I - uA)
This equation gives the basic decomposition of the change in sectoral
output into different sources (i.e., AF, AE, Au, and AA). In the TGT
model, input-output coefficients are assumed to remain constant, so the



- 130 -
last term will always be zero. Sectoral growth is thus allocated among
three sources: domestic demand expansion, export expansion, and import
substitution.
A few points are worth noting about the decomposition
equation. Import substitution is defined sectorally as arising from
changes in the ratio of imports to total composite-good demand. The
aggregate contribution of import substitution to growth is thus sensitive
to the level of sectoral disaggregation. For example, it is possible
to have positive import substitution in every sector, but have the ratio
of total imports to total composite-goods demand increase because of
changes in the sectoral composition of demand. Second, each term in the
decomposition equation is pre-multiplied by the inverse of the matrix
of domestic input-output coefficients. It therefore measures both the
direct and indirect impact on total output of each effect, taking account
of the indirect linkages through induced intermediate demands. Note
also that there is an index-number problem implicit in the decomposition
equation because the decomposition can be defined using the combinations
of initial and terminal year weights which are analogous to Paasche and
Laspeyres price indices. In the tables below, both indices have been
separately computed for the decomposition in each period and the averages
of the two results are presented.
When evaluating the decomposition measures derived from the
TGT model for Turkey, it is important to be able to place them in a
wider context and to compare our projections to the experience of other
countries as well as to the performance of Turkey in the past.



- 131 -
Table 5.7 presents the percentage decomposition by different
sources of aggregate growth (obtained by summing algebraically over the
sectoral changes) for five countries: Japan, Korea, Norway, Taiwan,
and Turkey.  The countries are part of a group being studied in a
World Bank research project and represent all of the countries for which
the decomposition has been done so far. Unfortunately, they are not a very
representative collection of countries.-   With the exception of Turkey,
they have all followed an open development strategy with exports being
very important. In spite of the rather special nature of the sample it is
useful to take T'able 5.7 as a point of reference. It is particularly
interesting to compare our projections for the FFYP period to Turkey's
performance in the past.
Unfortunately the Turkish data start from 1958. Indications are that
import substitution was very important during the years of tight rationing
between 1953 and 1958.2/ It seems clear that there has been a definite cyclical
sequence in the post-war period in Turkey. There was substantial import sub-
stitution in 1953-58. In the 1958-63 period, the contribution of import
substitution to growth was negative, indicating that import coefficients
actually rose af'ter the 1958 devaluation. In the 1963-68 period, when
there was signif'icant rationing of imports, the contribution of import
substitution to growth was positive and much larger than that of export
1/ The other countries in the project (Mexico, Colombia, and
tentativelv, YugoslaviA) would nrovide a T.7ider b'asis for comparison.
Unfortunately, the data construction for these countries is not yet
completed.
2/ See Krueger (1974).



- 132 -
Table 5.7
Selected Countries,
Decomposition of Aggregate Growth,
Percentage Composition By Source
Domestic                                   Input-Output
demand        Export        Import         coefficient
Country       Period       expansion     expansion    substitution   change
Japan:       1955-70        85.4%         13.9%         -3.1%           3.8%
Korea:       1955-63        74.5          10.0          21.4           -5.9
1963-73       67.8           37.7          -2.5           -3.0
Norway:      1953-69        60.0          46.0         -12.3             6.1
Taiwan:      1956-61        54.3          23.9          15.1             6.7
1961-66       61.3           37.6          -1.1            2.2
Turkey:      1958-63        96.3           6.4          -7.5             4.8
1963-68       83.6            4.9           8.3            3.2
1968-73       81.8           16.3          -1.4            3.3
1958-73       85.3           10.5           0.6            3.6
Notes:
Data are from World Bank research project, "A Comparative Study of
the Sources of Industrial Growth and Structural Change", (RPO 671-32)
the results tabulated are preliminary.
Aggregate percentage contributions are computed by algebraically summing
changes in sectoral output and sources of growth over all sectors and
then dividing by total change in aggregate output between the two
benchmark years.
The 1958-73 results for Turkey are obtained by chaining the results for
the three subperiods. A more detailed analysis of the historical trends
in Turkey is given by Celasun (1977).



- 133 -
expansion. In the 1968-73 period, the role of exports increased and
again the contribution of import substitution was negative (but small).
While the benchmark years in our data are 1968 and 1973, the actual
turning points are closer to 1970 and 1977.
It is somewhat surprising, given Turkey's consistent and
strongly inwarcl-oriented development strategy, that the relative contri-
bution of import substitution has not been very great in any of the
subperiods. Over the whole 1958-1973 period, it turns out to be negligible.
But the end-poaints for this period are very special years with 1958
representing the peak of a foreign exchange crisis and a very low level
of imports and 1973, on the contrary, being the only year in the last
two decades in which Turkey accumulated a massive amount of foreign
exchange. It iLs therefore more interesting to focus on the 1963-1968
period characterized by chronic foreign exchange shortage. The contribution
of import substitution was 8.3 percent during those years, which is
substantial but: not enormous. It is a smaller contribution than that
experienced by Korea and Taiwan in the late fifties and early sixties.
Historically, the role of exports in growth has also varied.
From the end of the war to about 1970, export expansion contributed little
to growth. Howqever, the substantial expansion of exports after the exchange
rate adjustmenit in 1970 contributed significantly to aggregate growth.
From Table 5.7, export expansion constituted 16.3 percent of total growth
in the 1968-73 period, much greater than in any previous subperiod. The
rapid expansion of exports in the 1968-73 period seems to indicate that,



- 134 -
with proper policies, exports can play an increasingly important role
in generating growth. After 1973, with the drift in incentives against
exports discussed earlier, their role diminished. In the 1973-77 basic
run of the TGT model (see Table 5.8), their contribution to aggregate
growth was even slightly negative.
Table 5.8 presents the decomposition of aggregate growth for
the various experiments with the TGT model. In analyzing the forward
runs, it is important to be aware of the problems created by choosing
different starting points. Since the start of the fourth five-year
plan is 1979 and since the 1977-78 results do not vary across the
experiments (the alternate policy simulations all start in 1979),
we have so far taken 1978 as our base.   However, 1978 is far from being
a "normal" year. It constitutes the depth of the crisis, with extreme
import rationing. Measuring import substitution starting from the
extremely low import ratios of 1978 may not be reasonable. The contri-
bution of import substitution to growth in 1977-78 is enormous; while
for the 1978-83 period it is very small. This result is to be expected
since, given the extreme situation reached in 1978, further import
substitution would be very difficult to achieve. For the sources of
growth decompositions we have therefore chosen 1977 as our base. One
can reasonably argue that the import ratios realized in 1977 reflect
"normal" levels. The period from 1974 to 1976 was characterized by
abnormally high import levels. In 1977, imports stabilized in real terms.
While severe import rationing started in the fall of 1977, the yearly



- 135 -
Table 5.8
TGT Model Experiments,
Decomposition of Aggregate Growth,
Percentage Composition By Source
Domestic
demand        Export        Import
TGT Model Runs     Period       expansion      expansion     substitution
Historical Run:    1973-77       100.4           -1.0           0.6
Basic Run:         1977-78       -32.6           36.0          96.6
(Constant          1978-83        69.8           9.9          20.3
PLD-ER)            1977-83         57.4          13.1          29.5
Experiment B*-l:   1977-78       -32.6           36.0          96.6
(Devaluation)      1978-83        71.5          29.2          -0.7
1977-83        61.2          29.9           8.9
Experiment B--2:   1977-78       -32.6           36.0          96.6
(Policy Package)   1978-83        70.1          29.7           0.2
1977-83        60.2          30.3            9.5
Notes:
The contribution of "input-output coefficient change" is zero by
assumption. 1977-83 results are chained from 1977-78 and 1978-83
results.



- 136 -
average import ratios are not extremely low when compared to the 1970-76
averages. Thus 1977 provides a more reasonable base than 1978 from
which to measure relative growth contributions.
Table 5.8 gives the decompositions for both the 1977-83
and 1978-83 periods separately. The results for 1977-83 are calculated
by chaining the sources-of-growth measures for the two subperiods. We
can thus separately consider the "crash" of 1978.
The depth of the crisis is clearly evident in 1978. Domestic
demand collapses and, with severe rationing, the squeeze on imports is
very severe. The negative contribution of domestic demand expansion to
growth in 1977-78 is mostly due to the large decline in aggregate invest-
ment (which falls 13 percent in real terms). The direct and indirect
effects of this decline are felt strongly in the producer goods sectors.
Indeed, construction and non-metallic minerals (cement) decline absolutely.
Consumer goods sectors such as food and clothing are much less affected.
After 1978, all the forward runs move onto relatively smooth
(but different) paths. In the basic run, the 1978-83 period is
characterized by massive import substitution (contributing 20.3 percent
to aggregate domestic output change). Considering the entire 1977-83
period, the contribution of import substitution to growth is even higher
(29.5 percent). This projected contribution in the basic run is much
higher than has previously occurred in Turkey, and is also higher than
in any of the other countries in Table 5.8. Only the post-war reconstruc-
tion period in Korea exhibited a comparable contribution (21.4 percent).



- 137 -
The high contribution of import substitution to growth which is projected
by the basic rtn underlines the magnitude of implied effort and brings
into question the feasibility of achieving the targets and of holding
to the modelled policy regime.
On the other hand, both the high-exchange-rate and policy-
package experiments (B-1 and B-2) yield reasonable projections of growth
contributions. In both experiments, the contribution of import substitution
to growth is high (8.9 and 9.5 percent), but is comparable to the contri-
bution in the :1963-68 period. The projected contribution of export
expansion (around 30%) is greater than in previous periods in Turkish
history, but not unreasonably so and still smaller than the 38 percent
achieved by Korea between 1963 and 1973. It does assume a serious
re-orientation of policy in favor of exports and a corresponding response
on the part of the domestic economy that is greater and much more
sustained than that which occurred after the 1970 devaluation.
The discussion in section 5.3 above of changes in the structure
of the economy implied by the various forward-running experiments provided
some evidence of the "reasonableness" of the various scenarios. To
complement the earlier analysis of structural change, we now turn to a
discussion of the decomposition of growth contributions at the sectoral
level.  Table 5.9 presents the percentage composition by source of growth
in each of the 19 sectors for the basic run and for experiments B-1 and
B-2. In general, the sectoral results from the two experiments are
broadly similar to one another and quite different from the basic run.



Table 5.9
TCT Model Experiments, 1977-83
Decomposition at Sectoral Growth
Percentage Composition By Source
Domestic demand expansion       Export expansion          Import substitution
Sector          Basic Run  B-1       B-2    Basic Run   B-1      B-2     Basic Run  B-1       B-2
I                            I
1. Agriculture             76.8     62.2     60.4 1    15.1     33.5      34.8 1   8.1       4.3      4.8
2. Mininig                 27.2     44.9     44.7      20.2     42.2      41.7 1 52.6       12.9     13.7
3. Food                    60.3     37.4     34.4 1    33.5     61.4      64.1 1   6.2       1.2      1.5
4. Textiles                51.2     51.1     49.2 1    33.4     45:9      47.3 1 15.4        3.0      3.5
5. Clothing                71.6     56.1     50.0 1    20.5     41.8      47.8  1  7.9       2.1      2.2
6. Wuod 6 Wood Products    83.3     88.2     87.4 1     4.2      8.7       9.4  1 12.5       3.1      3.2
7. Paper & Printing        36.9     56.8     52.1 1     8.2     25.6      27.5  1 54.9      17.6     20.4
8. Chelimicals              4.1     37.8     32.3       4.5     32.0      32.6    91.4      30.2     35.0
9. Rubber & Plastics       14.7     59.3     54.7  1    6.7     26.3      27.2  1 78.6      14.4     18.1
I0. Petroleum 6 Pet Prod    45.8     43.1     42.2 1     8.8     20.4      21.0  1 45.4      36.5     36.8
11. Non-Met. Mtn. Prod.     49.2  i  63.6     62.6  1   12.8     28.1      28.9  1 38.0       8.3      8.5
12. Basic Metals            18.8     55.3     56.4  1    1.3     16.4      16.1  1 79.9',    28.3     27.5
13. Metal Products           7.6     55.4     48.6  1    6.9     26.4      30.6  1 85.5      18.2     20.8
14. Noni-Elec. Machinery     5.1     52.2     56.8 1     0.6      8.4       7.5 1 94.3       39.4     35.7
15. Elec. Machinery          6.9     69.9     69.2  1    0.6      3.8       3.6  1 92.5      26.3     27.2
16. Transp. Equiipipment    28.5     57.1     58.3 :     3.2      8.7       8.3    68.3      34.3     33.4
17. Constructiont          100.0    100.0    100.0  I    0.0      0.0       0.0  I  0.0       0.0      0.0
18. Infrastructure          74.8     68.2     67.3      16.0     29.6      30.2     9.2       2.2      2.5
19. Services                77.6     73.0     72.0  I   11.5     24.1      24.8  I 10.9       2.8      3.2
Total             57.4      61.2     60.2  1   13.1     29.9     30.3  1 29.5       8.9       9.5
Notes; The contribution of "input-output coefficient chiange" is zero by assumption. Results are chained from
1977-78 and 1978-83 figures.



- 139 -
As noted above, the increasing investment rate implicit in the policy
package shows its real effects only at the end of the planning period.
Over the 1977-1983 period as a whole, experiments B-1 and B-2 give
quite similar results.
Compared to the basic run, the higher exchange rate in both
experiments leads to an increase in the relative contribution of export
expansion to growth and a decrease in the relative contribution of import
substitution in every sector. The changes in the relative role of domestic
demand expansion vary across sectors. The sectors for which the relative
contribution of domestic demand expansion decreases the most compared to
the basic run are the major export sectors: agriculture, food products,
clothing and textiles. On the other hand, those sectors for which the
relative contribution of domestic demand expansion increase the most are
the large import-substitution sectors: machinery (both electric and non-
electric), metal products, basic metals, chemicals, and rubber & plastics.
Note that this implies that while Turkey produces less basic industrial
products after devaluation it uses more of them. This point can easily
be overlooked when looking at production growth rates alone.
Since Table 5.9 gives only the relative contributions of the
separate effects in each sector, it cannot tell much about the intersectoral
linkages at work and the actual magnitudes of the different effects in
each sector. It is thus useful to consider the contributions of the
different effects on the change in physical output by sectors. We have
selected eight particularly interesting sectors to analyze: processed



- 140 -
food, textiles, chemicals, basic metals, metal products, non-electrical
machinery, electrical machinery, and transport equipment.-/ Figures
5.1 to 5.8 present the total output change in each of these sectors for
the basic run and for the experiments, and its decomposition into changes
due to domestic demand expansion, export expansion and import substitution.-/
Processed food and textiles include the bulk of Turkish exports
and, no matter what happens in other sectors, they will continue to dominate
exports for the coming five years simply by virtue of their large initial
base. Together, they account for about two thirds of total merchandise
exports in 1977.3/ From Figures 5.1 and 5.2 it is clear that there is a
substantial impact of the real devaluation on both sectors. In the case
of food processing, the change is due entirely to export expansion
while, for textiles, the change is due both to export expansion and
domestic demand expansion. Import substitution is very small in both
sectors.
The reason for the different impact of domestic demand expansion
in the two sectors can be deduced from an examination of the behavior of
prices in the two sectors (given in Table 5.3 above). While net prices
rise in both sectors (compared to the basic run), the gross domestic price
actually falls in the textiles sector. The fall in domestic price induces
an increase in domestic demand and hence a significant contribution of
1/ We do not here include petroleum and petroleum products because so
much depends on the growth potential for local production of crude and
we do not feel qualified to discuss this.
2/ When looking at the figures it is important to note that the scale is
specific to each sector.
3/ Note that the sectors include ginned cotton and tobacco products.



... .... .... ....
.   .   .   .   .   .  .   .   .   .  .   .   .   .   .   .   .
z
.. . . . . .  . . . . .                . . .      . . . . . .
.   .   .  .   .   .   .   .   .   .   .   .   .   .   .   .     r   T'
... .......
w
M      x      x
.. . . . . . .. .                    .        . . . . .
a
w
-.4  co                                                                                                                        U     0                    A .
m
17                                                                  - - ----
---- --------

TT:FOUTT9
7 i
4-1 j
il-i J.
'7 r T                         ...........
-7F-7L
T-i
A-i
CT
I' i     E             --  -------
-4,                                                                                                                            A---
7. .

2 P4   ...
:-44
co.
IL
Po     W      W              ......
77
2.:
w.:
l..; 2
z
iJ
L-1  i,!.  i-
Cd                                                                                    KII, 77 7/
r
M X
Ul
H  P, 
14
0     3!
H 0 u
m
.  .  .  .  .  .  .  .  .  .                                                                                                        jI
I:            1                    vi-11-1 uollTie          . . . .  . .



- 142 -
domestic demand expansion to growth in textiles. One of the reasons
that the textile sector exhibits both a fall in domestic price and a
rise in net price is that the net price is a weighted average of net
receipts from both domestic and export sales. In the case of textiles,
the world price is assumed not to be affected by the volume of Turkish
exports. Thus, after the devaluation, the export price remains fixed
at the higher level and the net price rises even though the domestic
price falls. In the case of processed food, the export price is
sensitive to the volume of Turkish exports and so the domestic and
export prices are more closely linked. Another factor explaining the
behavior of textiles is the price of chemicals which provide important
intermediate inputs into textiles. The cost of chemicals declines when
import rationing is relaxed, positively affecting the net price of
textiles.
It is also important to note that in both textiles and food
processing, exports expand through increases in production and not through
squeezing domestic demand. In food processing, domestic demand does
decline, but only marginally. In textiles, as already noted, it actually
expands. It is therefore clear that the resources that make this
possible must either have been "created" (increases in efficiency,
capacity utilization and investment) or they must have been attracted
from other sectors.
Figure 5.3 presents the growth decomposition for the chemicals
sector, one of the most interesting import-dependent sectors in the Turkish
economy. First consider the basic run with a constant price-deflated



4
-Y-
IJ_
f-4,
FIGU M 5                                                                                                         W
.3
7
-CHE
-4-1 i.
-BASICIMETALS
4A                                   7
19                                                                                 L '_F: f
J
_-F- 1977-8                           :Ann; il
3                               u
r                                    
4
Groirth -
F
L  _L                                    I  4,
4-                                                                      Rates-
7 1, i -
-:Basic Run     10.1,%
i 7i
tj LL        intial"                                                                 T:
I iT                                    B-1
'Growth                                                                                                          '8.61
Rates                                                                                                  er.
4
J
-:Exp    B-i
-.Basic-Run    11. 2%          J1-- L!
j
4                                               
J
Exper. B-L.
T!  i                    W
a                                                                                                                J.'!
'7::
]'-
_
Exper,   8  0%.;
s-.0                                                                                                        -7-t
jT                       J
T-     :J,
T
V.
Lj
L:          t

;_l  L;             _7                                                                         L  F.-
2.
77'                                                                -7.
i Al-H4 4-
q
rt
A                                                        8.4
4
LJ
J,  J:
A+
 :i
L.L.                      :A.
Tr                              !-Aj
J
+ IJ
I-AA -.4-..
J                                                                                                                      a
Lx'
TF
q
H.
:T               DOMEST              ORT         I4ORT                                                             DMIES IC         EXPORT
         I
TORT
DE'KklTD'               '_F SUBSTITUTIqN                            i                 T          DDWiD              A73ION    SUBSTITUrl t    i m. .
L,    XPANSION. j                                                       OUTFU.
A.                                                                             CHANGE
NS                                                                                       I 1- .EXPAN: ;I
m .       EXP   Im                                                                                                 ON:
t:i
m
7 .
-J7
i-4
A      j
d            J
T
j I



- 144 -
exchange rate scenario. Growth is very rapid, but it is almost entirely
due to import substitution. Historically, chemicals has been a sector
in which the import ratio has steadily increased. To assume that it
could decline dramatically over the next five years may well be over-
stating the adaptability of the sector. Fertilizers constitute an
important part of chemicals and the kind of import substitution pro-
jected by the basic run would imply dramatic increases in fertilizer
production. It is true that a lot of capacity is presently idle but
whether such increases are feasible must be judged by a more detailed
examination of the sector than is possible here. The basic-run numbers
do seem unrealistically high.
With higher real exchange rates, the situation is quite different
in experiments B-1 and B-2. Total growth is roughly evenly distributed
among domestic demand expansion, export expansion, and import substitution.
The total projected change in sectoral output is also much smaller. The
two high-exchange rate scenarios thus seem much more realistic. Note
that export expansion can be significant for chemicals with the appro-
priate trade policies. Chemicals is a large and diversified sector
dominated by high-cost import substituting activities. However, it also
contains subsectors processing low-cost domestic minerals that may provide
a base for future export expansion.l/
1/ Raw borates, for example, are in plentiful supply in Turkey and could,
with increasing degrees of domestic processing, lead to significant
exports from the chemicals sector.



- 145 -
With chemicals, basic metals is the most important import-
substituting, intermediate-goods-producing sector. Figure 5.4 shows
the impact of the experiments on the sector. The story is quite similar
to that of chemicals. With a constant price-deflated exchange rate,
import substitution is responsible for almost all of total output growth.
With high exchange rates, its role diminishes. Note, however, that
contrary to chemicals, overall growth does not much decline with higher
exchange rates. One of the reasons is that the basic metals sector
has very strong forward linkages through intermediate deliveries to a
number of sectors: metal products, non-electric machinery, electrical
machinery, transport equipment, and construction. The input-output
coefficients from basic metals to these sectors are .42, .13, .26, .13,
.14 respectively. Chemicals has significant forward linkages only to
textiles and rubber & plastics (coefficients of .08 and .20 respectively).
Thus, the induced increases in intermediate demand reduce the decline in
the price of basic metals.
Figures 5.5 to 5.8 show the impact of the experiments on a number
of sectors "downstream" from basic metals. In the case of metal products,
the forward linkage from basic metals is so strong that the high exchange
rate experiments actually lead to an increase in output, overcoming the
negative gross-price effect of the devaluation. In all these "downstream"
sectors, import substitution is the most important source of growth in the
basic run and is replaced by domestic demand expansion in the high exchange
rate experimenits.



7
I
t
TTr -
r
..... .. ...
-1 +                             L       J           i  i.kl. i
t
NOISKVdXa          MNYHO                                                                                  WM           .39NVHD
NOIgN
Kolmlilms                                            Indlao                                                ISULIS    riolskv
I   f                 d
l   .1-   GONM             InalnO
isoaKi                                                                : .  ,  ;  .           I  
  , II     .. t
ontawou             wloj,                                            ;Ldo
LdUd2Ld                                               F:
'BiT     11 :uogl            oilsaKoa          .,;viol
G-G-
A
.........
-A
L                                                                       Jf.
T,
I L
A                                                         I-p
t
- I -  I A                       I7
a                                         N                                                                                                                     v
A I             
41               4 I
A
T
+ -
L:                                   m
'.44- 4                                             q .  .
Z
7
7
'
 +
%
t

i J i
A
T I                                                                                             ....    ..
J                             O  1
7
. . . ... . . . .
%6'1                                                                                 %Q. 9,
'%0.9     T-a: %led
;L,g.     -a '2
'Pci
3 -
xa -
.... ......
i-i  %            OTS 9
... ...        7U_     -r-
ES                                                                        H
w .                                                                                                                             ....   ...  ES-    L61
r ...    aNIHOVIA          -mo
slor"UMid 'IVIIR
rl .
9,19
_: : rs' 5. MMI-1
J
a
......... ... ..
L.
r                         VT  ;-I -r- _1



* - -I              -~ !-. -.- -   - --  :I                ;    |.r-      .     I - 
|  .  .   F  t  IR  5  i  -   a~.1   -.  : 3 z  I  t                      I , ,tj  i  -r !1 .. 3 I 
;                 --   ' ls~~~~~~~~~~i77F77-aT7']1..                     [     .::t;:.:: Gr 
-h          IHTI
- 1     1 i''' t;r , -   ,  * lc Run.L {[ ' .8.  1                1   -  -         _ 7  R
F--I--- - - |E;5 7                               1DFICURES
. ~ ~ ~ ~ ~ _______ ' ! . . . .. - * - i . :   I !   ;  --  ,      I          ::_____:::_
_ n 2 (~~~~~1               …                                      _ ______ _____ _____  g: 1 : : :   0!j 
ELECTRICALI MACHINERY              ~-TANPR                        EQUIPMEN 
83   '       t~~~~~~~~~ n, , , a I,1I 1 i        1   1 9 7 7 8           A n n u a l
< f W V XFX I S tAt~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~i  1 ;W: 
____             I          ~~~~~~ ~~~~~ ~~~~~~~~Rate's  m' , lI                       -Rates
.. . .   ...,, .' . j. .                                                       xpe , ,j ._. ,
Expe ~ ~ ~ ~ ~  ~    ~   ~   ~   ~   ~    1 B--.--6-:.
* ~ ~ ~ ~ j  i:                  -, I~I               n . :Z~3                         B CR f :.  8.;4 %!t'E' .i:i'  . -j  -  -L  -  : i:!,  :! ; j, 
6~~~~~~~~~~~~~~ 0%
______  (~~~~~E~B2                        l0-O----_6_......       .~ Expier~  B-2  8.8%
*   j   :    S  i t: !;  !! .. 1::  I  :  j  -::  .  !.1 ,   '; :':[,,1  :: ti 1:i.  I  1  --  _ -_i  _. ; 
:_____    -.: -                                                                .  .                -.....--. ' ,  - - 
I-,'IN
a            H~~~~~ 
I               iI                     I TOLAL A
m~~
CHANGE  EXPA3~~~~~                                   I         EXP'IT_O         I
..ION  Fl                             CHANCE    AASN~~~...   ....
iIA



- 148 -
In the metal products and the electrical and non-electrical
machinery sectors, domestic demand expansion is virtually nil in the
basic run. This result reflects the substantial decline in real invest-
ment that takes place from 1977 to 1979. In the basic run, total fixed
investment does not reach its 1977 level before 1982, with obvious
consequences for these sectors.
The story that emerges from the analysis of these sectors
emphasizes that looking at domestic production growth alone may be mis-
leading. For example, in the chemicals and machinery sectors, the growth
of output in the basic run appears very impressive. But one must realize
that it is all due to the import squeeze and represents the high-cost
production of domestic substitutes. It does not reflect an increased
availability of the products to the domestic market -- indeed, quite
the contrary. With higher exchange rates and increased foreign exchange
earnings, the necessity for import substitution diminishes and total
availability of the goods to domestic users increases substantially, at
lower cost.
Regarding exports, the analysis of the composition of the
sources of growth indicates that, in the next few years, the expansion
of non-traditional exports must come primarily from the currently important
export sectors: food processing and textiles. The initial base in all
other sectors, with the possible exception of clothing and mining, is
just too small to allow them to have a substantial impact within the next
five years. But the TGT model does project very significant export



- 149 -
expansion in a large number of manufacturing sectors if relative prices
and incentives are strongly altered ia favor of export activities. As
can be seen by referring back to Table 5.4, experiment B-2 projects the
highest export growth rates in non-electrical machinery (38.1%),
chemicals (37.8%) and electrical machinery (34.1%). While exports from
these sectors will still be very small in 1983, their growth rate might
be very high. And it is this kind of performance that is needed to
create strong confidence both inside and outside the country in Turkey's
ability to overcome the crisis, speed up growth and deepen the indus-
trialization process.



- 150 -
6. Conclusion
After two decades of remarkably rapid growth and development,
Turkey faces in 1978 a very serious crisis and a difficult challenge.
Economic and social change has been very important over the last decades.
From an overwhelmingly rural and traditional society, Turkey has become
a predominantly urban and industrial economy in a very short period of
time. Old values and restraints have crumbled and the demand for a
better life style, once restricted to a small elite, has become universal.
Given the essentially free and democratic political environment and the
extremely competitive nature of the political process, a constant strain
on existing resources and capabilities is natural and healthy.
In the last few years, however, the tension has become too
great and the gap between expectations and reality, from being a source
of dynamism, has become a threat to continued development, social
stability and growth. This gap must be narrowed both by scaling down
and controlling the extreme form that consumerism and excess demand
have taken in the early 1970s and by rapidly expanding the productive
capacity of the Turkish economy. That much is clear to everyone and
however difficult it is politically, serious attempts are and will be
made in the right direction.
But there are many ways to attempt to expand productive capacity.
Everything cannot be achieved at once, priorities must be set and certain
sequences must be chosen as better than others. To try and achieve
everything at once is a dangerous strategy that courts defeat. Human



- 151 -
and material resources are scarce and will lose all effectiveness if
spread too thin.
Turkey is entering the Fourth Five Year Plan period with a huge
foreign exchange gap and the effective constraint on growth over the next
few years will be foreign exchange. Assuming a high exchange rate policy
that would achieve a real devaluation of about 50% between 1978 and 1983,
starting with a 30% devaluation in 1979, the TGT model projects the
feasibility of 25 billion dollars worth of merchandise exports during
the F.F.Y.P.   Merchandise imports wouild sum to 40 billion dollars.  The
15 billion dollar deficit in merchandise trade and a 500 million dollar
reserve accumulation target would be financed by 3.2 billion dollars of
net invisible earnings and 12.3 billion dollars of remittances and net
foreign capital inflow.l/ This scenario allows for a GDP growth rate
between 7.5 and 8.0 percent. But to achieve 25 billion dollars worth of
exports requires a major shift in priorities and industrial strategy
that will require great determination to succeed.
The constant price deflated exchange rate scenario, on the
contrary, assumes continued inward orientation of the economy. Exports
would only sum to 18 billion dollars, constraining imports to 33 billion
dollars. The ratio of imports to domestic output would be continuously
falling requiring rapid import substitution on a broad front. The
average annual growth rate would fall to 5.5 percent.
1/ These figures are in current dollars and assume a 9.0% "dollar"
world inflation rate.



-. 152 -
It will be very difficult to achieve massive import substitution
in all basic industries simultaneously. The strain that this would put
on Turkey's administrative, managerial and technical capabilities would
be extreme. By assuming the continuation of historical rates of technical
progress into a period of massive import substitution aimed at all basic
industries, the basic run of the TGT model is probably overstating the
growth potential implicit in a continued inward-oriented industrialization
strategy. But even if the 5.5 percent annual growth rate projected by
the basic run were achieved, it is far from the 7 or 8 percent growth
rates that constitute Turkey's objective and that are necessary for the
rapid improvement in the standard of living of a population that is still
growing at 2.5 percent every year.
It is clear that a dramatic expansion of exports is the necessary
precondition for realizing annual GDP growth rates in the vicinity of 8
percent. This need for export expansion is widely recognized in Turkey.
But what is less clearly understood is that for such a massive expansion
in exports to take place resources must be diverted from other activities
into export-oriented production. It will not be possible to achieve
dramatic growth in the production of chemicals, basic metals and machinery
for the domestic market and simultaneously succeed in a massive export
expansion program. Resources are not unlimited and priorities must be
established. If massive export expansion is to take place, the required
resources must be released to export-oriented activities.



- 153 -
It must also be realized that in a mixed economy where the
market and prof'itability calculations play an important role, resources
will not be attracted to export-oriented production if domestic market
oriented activity remains much more profitable.
Finally, it must be recognized that Turkey will be able to
expand her share in the world market rapidly enough only if the dollar
price of Turkish exports is sufficiently competitive.
All this emphasizes that exports can only be expected to expand
if the price and incentive relationships are such as to attract suffi-
cient resources into export production and allow the sale of those exports
on the world market. -It is these price and resource pull relationships
that have been emphasized throughout this study and that have led us to
conclude that a major real devaluation is a necessary condition for
substantial export expansion and rapid growth over the FFYP period.
Administrative measures and subsidies cannot succeed in an atmosphere
of extreme excess demand and severe import rationing. While a realistic
exchange rate is certainly not the only policy variable that should be used
to promote exports, it is by far the single most important variable.
To try and embark on a major export drive with a grossly overvalued
exchange rate cannot lead to success. And Part 3 above has attemped
to demonistrate just how overvalued the exchange rate has become. Further-
more a once and for all devaluation is not enough. What is crucial is a
clear commitment to keep exports profitable and to keep the real exchange
rate from drifting back down again.



- 154 -
From the point of view of microeconomic structure and sectoral
growth, it must be recognized that export expansion for the coming few
years, will require more rapid growth in sectors where Turkey can be
immediately competitive and does not face important bottlenecks of tech-
nology and scale. Production must be increased rapidly and at moderate
cost, and an important exportable surplus must be generated quickly. Once
the peak of the crisis is overcome and confidence is restored, a more
ambitious program of selective import substitution can start simultaneously
with continued and broad based export expansion. Our microeconomic results
suggest that it may be efficient to concentrate import substitution efforts
on finished products in the metal working and machinery sectors rather than
pressing for rapid import substitution in basic intermediates and forcing
domestic producers to use the high cost domestic inputs. When choosing
priority sectors for import substitution it is'iiiportant to ensure that
export potential also exists in the near future. By this method the
structure of production and the structure of exports could be gradually
changed with more diverse and technologically more advanced products steadily
gaining in weight. Metal products, machinery and domestic resource based
chemicals may gain an increasing share in manufactured exports. The speed
of this transformation process will primarily depend on the overall degree
of resource mobilization as reflected in the investment rate. In guiding
and stimulating this process, export subsidies and preferential treatment
have a major role to play. While export subsidies cannot substitute for a
realistic exchange rate, they can and should be used to stimulate "infant"
exports from the capital goods and intermediate goods sectors. In an economy



- 155 -
characterized by a history of price distortions and rigidities in factor
allocation, a sudden move towards fully unified effective exchange rates is
neither possible nor desirable.
It is worth repeating that the TGT model attempts to capture
medium term mechanisms and does not include an analysis of macroeconomic
monetary mechanisms and very short term adjustment processes. Nevertheless,
to reach the longer run Turkey must bridge the short run, and do it in
such a way that an effective export-oriented increase in production can
materialize quickly. While our model has nothing to say about the inter-
action between monetary policy and real magnitudes, the relationship is
important. A monetary squeeze which is too severe may have strong negative
effects on production and capacity utilization. It is important to weigh
the gains from lower inflation rates against the losses in potential output,
employment, and perhaps even exports. A strongly deflationary approach to
crisis management may in fact make the beginning of an export-led growth
effort more difficult to achieve. In our opinion, a gentler approach
including adjustments and indexation in exchange rate and interest rates,
coupled with a substantial injection of new foreign borrowing made possible
by the introduction of resolute export-oriented policies, may be a much
more successful policy package for short term crisis management than
severe deflation policies.
Concluding, it is important to stress again that an attempt at
rapid inward-oriented growth with a greatly overvalued exchange rate and
without the support of an adequate export performance is unlikely to
succeed and could lead to an even deeper crisis. On the other hand,
devaluation and' export expansion alone do not constitute a solution to



- 156 -
Turkey's long run industrialization problems. What is required is the
combination of a realistic trade and exchange rate policy that recognizes
the crucial importance of the exchange rate for trade performance, with
an adequate domestic resource mobilization policy that succeeds in
generating the real domestic savings necessary for the achievement of the
growth and industrial deepening targets widely embraced in Turkey. A
lot must be achieved and everything cannot be decided and allocated
bureaucratically from the center. It is therefore important that the
market mechanism be able to play an important and constructive role,
which can only happen if relative prices in general and the relative price
of foreign exchange in particular reflect the real scarcities constraining
Turkish growth.
To advocate a more rational price structure and a greater
role for the price mechanism is not enough. The market mechanism can
only become effective and socially acceptable if the distribution of
ownership and wealth is significantly altered in a more egalitarian direction.
Prices do not only allocate resources between sectors but also economic
welfare between people. A policy designed to get "prices right" must be
combined with a policy to get "incomes right."  This is no small task,
but it will not be possible to reach economic equilibrium in Turkey without
at the same time making great progress towards social equilibrium. Convesely,
for Turkey's social problems to find a solution, an efficient economy and
rapid growth are indispensable preconditions.



- 157 -
REFERENCES
Adelman, I. ancl S. Robinson,.1978, Income Distribution Policy in Developing
Countries, Stanford, Stanford University Press.
Ahluwalia, M., F. Lysy and G. Pyatt, "A Price Endogenous Model of Malaysia:
Some Static Experiments," World Bank, mimeo.
Armington P., 1969, "A Theory of Demand for Products Distinguished by Place
of Production," Staff Papers, Vol. 16, pp. 159-178.
Bhagwati, J. and T.N. Srinivasan, 1978, "The Evaluation of Projects at World
Prices under Trade Distortions: Quantitative Restrictions,
Monopoly Power in Trade and Nontraded Goods," World Bank, mimeo,
May.
Bruno, M., 1976, "The Two Sector Open Economy and the Real Exchange Rate,"
The lmerican Economic Review, September.
Bulutay, H., et:. al., 1974, TUrkiyinin Milli Geliri, 1923-1948, S.B.F. Yayinlari,
Ankara.
Celasun, M., 1978, "Sources of Growth and Industrialization in Turkey,
1953--73, " World Bank, mimeo.
Celasun, M. and 0. Caglarcan, 1975, "Policy Analysis with a Price-Responsive
Inter-Industry Model, Prototype Country Models," Working Note
No. 6, World Bank, mimeo.
Chenery, H. and W.J. Raduchel, 1971, "Substitution in Planning Models" in
H. Chenery, ed., Studies in Development Planning, Cambridge,
Harvard University Press.
Chenery, H., arid H. Uzawa, 1958, "Non-linear Programming in Economic
Development," in K.J. Arrow, L. Hurwicz and H. Uzawa, eds.,
Studi.es in Linear, and Non-linear Programming, Stanford University
Press.
Chenery, H., S. Shishido, and T. Watanabe, 1962, "The Pattern of Japanese
Growth, 1914-1964," Econometrica, Vol. 30, No. 1.
Ciller, T.<, 1977, Tulk Sanayiinde Ithal Ikamesi ve Koruma Politikasi,
1963-1974, I.K.V. Yayini, Istanbul
De Melo, J., 19I78, "Estimating the Costs of Protection: A General equilibrium
Approach," Quarterly Journal of Economics, May,
De Melo, J. and K. Dervis, 1977, "Modelling the Effects in a Dynamic
Framework," Journal of Development Economics, Vol. 4, pp. 149-72.
De Melo, J. and S. Robinson, 1978, "Tradability in Trade Theory," World
Bank, mimeo, July.
De Melo, J., K. Dervis, and S. Robinson, 1977, "A Global Model of North
South Trade and Growth," RPDS Discussion Paper No. 72, Princeton
University, March.



- 158 -
Dervis, K., 1975, "Substitution, Employment and Intertemporal Equilibrium in
a Non-Linear Multi-Sector Planning Model for Turkey," European
Economic Review, Vol. 6, pp. 77-79.
Dervis, K., 1977a., "Exchange-Rate Adjustment, Trade Policy and Economic
Growth: Experiments with a Dynamic General Equilibrium Model of
Turkey," Discussion Paper No. 76, RPDS, Princeton University.
Dervis, K., 1977b., Planlama Modellerinde Dis Ticaret, Docentlik Tezi,
Istanbul.
Dervis K., and S. Robinson, 1977, "The Sources and Structure of Inequality
in Turkey, 1950-73," July, mimeo.
Ebiri, K., K. Bozkurt and A. Culfaz, 1977, Turkiye Imalat Sanayiinde Sermaye
ve Isgucu, T.C. Basbakanlik Devlet Planlama Teskilati, Yayin
No. DPT: 1624-303.
Ecevit, L. and E. Ozotun, 1975, The Changing Structural'Distribution of
Income and Employment in Turkey and Kuznets' Hypothesis, Ankara,
State Institute of Statistics.
Findlay, R., 1973, International Trade and Development Theory, Columbia
University Press.
Harris, J. and M. Todaro, 1970, "Migration, Unemplyment and Development:
A Two-Sector Analysis," American Economic Review 60.
Herschlag, Z., 1968, Turkey: The Challenge of Economic Growth, E.J. Brill
(Leiden).
International Monetary Fund, International Financial Statistics.
Issawi, C., forthcoming, An Economic History of Turkey, Chicago University
Press.
Karaosmanoglu, A. and M. Durdag, 1977, "Basic Needs in Turkey," World Bank,
mimeo.
Kuran, T., 1978, "Internal Migration: The Unorganized Urban Sector and
Income Distribution in Turkey, 1963-1973," Discussion Paper
No. 82, R.P.D.S., Princeton University.
Krueger, A.0., 1974, Foreign Trade Regimes and Economic Development: Turkey,
New York, Columbia University Press.



- 159 -
Krueger, A.O., 1978, Foreign Trade Regimes and Economic Development:
Liberalization Attempts and Consequences, Ballinger
Publishing Co., New York.
Lancaster, K., 1966, "A New Approach to Consumer Theory," Journal of Political
Econoinm 74, 132-157.
Land, J., 1970, "The Role of Government in the Economic Development of
Turkey, 1923 to 1963," Rice University Program of Development
Studies, Paper No. 8, Fall.
Leamer, E. and R. Stern, 1970, Quantitative International Economics, Chicago,
Aldine Co.
Ministry of Finance, 1977, Budget Revenues Year Book, 1976, Ankara.
Mundlak, Y., 1976, Migration out of Agriculture- Empirical Analysis Based
on Country Data, International Food Policy Research Institute,
Washington D.C.
OECD, 1976, "SUPEMI Continuous reporting system on migration, 1976 Report."
Petri, P., 1976, "A Multilateral Model of Japanese-American Trade," in
J. Skolka, K.R. Polenske, eds., Advances in Input-Output Analysis,
Ballinger Publishing Co., Cambridge.
Pyatt, G., 1978, "Income Distribution and Endogenous Prices in Development
Planning Models," mimeographed, World Bank.
Pyatt, G., and E. Thorbecke, forthcoming, Planning for Growth, Redistribution
and Employment.
Robinson, S. an,d J. de Melo, 1976, "A P:Lanning Model Featuring Trade, Intra-
Sector Product Differentiation and Income Distribution," Research
Program in Development Studies, Discussion Paper No. 69, November.
State Institute of Statistics, 1976, Census of Industry and Business Establish-
ments: Manufacturing Industry, Ankara.
State Planning Qrganization, Annual Programs, Ankara.
State Planning Organization, 1976, Dorduncu Bes Yillik Kalkinma Plani,
1978-1982: Buyume Alternatifleri, T.C. Basbakanlik Devlet
Planlama Teskilati, Yayin No. DPT: 1941-IPD: 395.
State Planning Organization, 1977, Dorduncu Bes Yillik Plan Modeli ve Yan
Calisinalari, T.C. Basbakanlik Devlet Planlama Teskilati, Yayin
No. D:PT: 1602-IPD:401.
Stone, R., 1970, Mathematical Models of the Economy and Other Essays, London,
Chapman and Hall.



- 160 -
Syrquin, M., 1976, "Sources of Industrial Growth and Change: An Alternative
Measure," World Bank, Economics of Industry Division, mimeo.
Taylor, L., ed.,, 1978a. Models of Growth and Income Distribution in Brazil,
forthcoming.
Taylor, L., 1978b, Macromodels for Developing Economies, forthcoming.
T.C. Resmi Gazete, 1978, 1978 Yili Programi, Ankara.
United Nations Department of Economics and Public Affairs, 1975, Towards
a System of Social and Demographic Statistics, New York, United
Nations, ST/ESA/STAT/SER.F/18.
World Bank, 1978, "Historical Rates of Change in U.S.$ GDP Deflators:
1961-1975,", EPD internal memorandum, February.



The Foreign Exchange Gap, Growth
and Industrial Strategy in Turkey: 1973-1983
Kemal Dervis
Sherman Robinson
Appendices A and B
The Equations of the TGT Model
The Data
July 1978



Appendix A
The Equations of the TGT Model
Kemal Derviq
Sherman Robinson
July 1978
A.1 Static Model Equations
A.l.l Prices
A.1.2 Factor Markets and Domestic Supply
A.1.3 Foreign Trade
A.1.4 Income and Nominal Flow-of-Funds
A.1.5 Product Demand and Supply
A.2 Dynamic Model Equations
A.2.1 Investment Allocation
A.2.2 Rural-Urban Migration



- Al -
A.1 Static Model Equations
In this section, we present in summary form the complete set of
equations for the within-period static model. The dynamic, intertemporal
linkage model is discussed in the next section. The endogenous variables
are listed in Table 3. A few notation conventions are followed throughout:
- Endogenous variables are all denoted by capital letters.
Lower case letters, Greek letters, and letters with ^ are all
parameters.
- Any letters with a bar (-) on them are exogenous variables in
the within-period model but are updated as part of the dynamic
model.
- Letters with an * are policy parameters assumed to be set exog-
enously by the government.
- The subscripts i and j are used for sectors. They always range
from 1 to n.
- The subscript k refers to labor categories and ranges from 1 to m.
- The superscript d refers to domestic goods.
- The superscript m refers to imports.
The discussion of the variables and corresponding model equations
will follow the order of the presentation in Table A.l.
A.l.1    Prices
d
The model will solve for the prices of domestic goods, P., en-
dogenously to clear the markets for domestic goods. The market clearing
conditions are given by the product market excess demand equations discussed



- A2 -
in the last part of this section. For the discussion of variables and
equations in the first four parts, it is convenient to assume that domestic
d
prices, PV, are given -- perhaps by some initial guess.
Equation (1) defines the domestic price of imports:
(1)       PMi =    (l + tmi)ER                               n
i~~~~~~
where     is the world price in dollars,
tmi is the tariff rate ad valorem, and
*
ER is the exchange rate.
The price of the composite good, which is a C.E.S. aggregation
of imports and domestically produced goods by the same sector, is given by:
(2)       Pi    [PDi + PM. * M /Di]/f [Mi/Di, 1]              n
where M. is imports,
1
Ddi is domestic demand for domestic production,
fi(-) is the C.E.S. aggregation function.
m d
Note that the C.E.S. aggregation function is evaluated at the point (MiDi, 1).
The composite good price is a function only of the import ratio, not of
the levels of import and domestic demands.
The world price in dollars of domestically produced goods which
are exported   is given by:
(3)       PWE, = PD /[ER* (1 + tei)1                      n
i  i



- A3 -
Table A.l.Engodenous Variables
Equation
Name                             Variable               Reference        Number
1. Prices
Domestic prices                    PD.                     (37)            n
Import prices                      PM.                     (1)             n
Composite good prices              P                       (2)             n
Export prices                      PWEi                    (3)             n
Net prices (value added)           PVi                     (4)             n
Capital stock prices               PKi                     (5)             n
Price level equation                                       (6)
6 n
2. Factor Markets and Domestic Supply
Production functions               Xd                      (7)             n
xi
Capacity utilization               ui                      (8)             n
Labor aggregation function         L1                      (9)             n
Labor demand equations              ki                     (10)            n-m
Aggregate labor demands            LD                      (11)            m
k                      (1
Aggregate labor supplies           Lk                      (12)            m
k
Average wages                      Wk                      (13)            m
3-n + 3-m + n-m
3. Foreign Trade
Export demand                      Ed                      (14)            n
i
Desired imports                    MDi                     (15)            n
i
Domestic supply                    sd                      (16)            n
i
Total imports                      TIM                     (17)            1
Import rationing ratio             RM                      (18)            1
Import demand                      Mm                      (19)            n
4n + 2



Table Al1, continued                - A4 -
Equation
Name                            Variable                Reference      Number
4. Income and Nominal Flow-of Funds
Net labor income                   YL                     (20)             1
Net non-labor factor income        YK                      (21)            1
Government non-factor income       YG                     (22)             1
Total resources                    RT                      (23)            1
Labor resources                    RL                      (24)            1
Capital resources                  RK                      (25)            1
Government resources               RG                      (26)            1
Total investment fund              TINV                    (27)            1
Total private investment           PINV                    (28)            1
Government investment              GINV                    (29)            1
Total investment in stocks         SINV                    (30)            1
Private consumption                PCON                    (31)            1
12
5. Consumption Demand and Supply
Consumption demand                 C                       (32)            n
Stock accumulation by sector
of origin   Si                      (33)            n
Fixed investment by sector of
destination         Y                       (34)            n
Fixed investment by sector of
origin       i                      (35)            n
i
Intermediate demand                Vi                      (36)            n
Domestic demand for domestic       D                       (37)            n
goods                                              6n
Total number of variables: 19n + 3-m + n-m + 14
Total number of equations: 19-n + 3m + nm + 15
of which 19 n + 3-m + n-m + 14 are independent



- A5 -
*
where te. is the export subsidy rate ad valorem. Note that as discussed
in Section 2.6 above, the small country assumption is dropped and the
world prices of l'urkish exports are set by the domestic price. Explicit
export demand functions are introduced below (in Section A.1.3).
Value aLdded per unit of output in each sector, or net price, is given by:
(4)   PV    PE>I   ZP\.aji - ts*,PD,                 n
where aji are fixed input-output coefficients and ts    are indirect tax races.
The price of a unit of aggregate capital by sector is given by:
(5)      PK. = E' b                                     n
where b.. are the fixed shares of capital goods by sector of origin required to
J1
make up one unit of aggregate composite capital by sector of destination. Note that
Eb j=1 for all i.
ji
The overall absolute price level and hence the inflation rate
are set exogenously in the model. The price level equation is given by:
(6)       Ei2 iP  PL                             one
ii
where Qi are the weights for the price index (EMi=l) and
PL is the overall price level.
X.1.2    Factor Markets and Domestic Supply
The production functions are given by:
(7)     Xi   UjAi *    1 (Ki' Li)                  n



- A6 -
where U. is the utilization rate,
A. is the productivity parameter,
- is aggregate sectoral capital stocks,
K.
1
Li is aggregate labor and
gi(-) is the C.E.S. function.
Note that the production function is a two-level function. Capital
is a fixed-coefficients aggregation of capital goods and is assumed to be
immobile across sectors within a period. Intermediate goods are assumed to
be required according to fixed input-output coefficients.
The utilization parameter (Ui) in the production function was
discussed in Section   2.5.    In one variant of the model,it is always assumed
to equal one. In another variant, it is a function of the overall degree of
import rationing in the economy and of the degree of dependence of an in-
dividual sector on imported intermediate goods. The function is given by:
(8)       Ui = (RM)ami                                  n
where RM is the import rationing factor defined in equation (18) below,
a is a parameter, and
mi is the ratio of the value of imported to total intermediate goods
requirements in sector i.
Labor is a C.E.S. aggregation of labor of different skills categories,
with the function given by:
(9)       Li = xi(L i   1, *  *   Lmi)                 n



- A7 -
where Lki is labor of skill category k in sector i and
Ai(-) is the C.E.S. function.
There are three types of labor: agricultural workers and two
types of urban Labor, "organized" and "unorganized." Urban labor does not
work in the agricultural sector and vice versa.
The demand for labor is given by the first order conditions
for profit maximization, that the wage equals the value of the marginal
product.   For organized urban labor (category 2), the wage is assumed to be
fixed. For agricultural labor and unorganized urban labor, the wage is
assumed to adjust to clear the labor market. The equations are:
ax.
cLOa)    PV   axi                             k# 2
laki
mn
ax
(lOb)    PV   Mi      k                       k   2
ki.
where W   is the wage, with W   fixed exogenously.
k                     2
Aggregate demand for labor of different categories is
obtained by solving equations (10) and summing over sectors to get:
D
(11)      Lk = ELki                                          m
The supply of agricultural labor is fixed exogenously:
(12a)     L1 = LA                                            one
As discussed in Section 2.5 , the model has a rather special
treatment of the urban labor market.    The total urban labor



- A8 -
supply is fixed and hence any unemployed organized labor will find
jobs in the unorganized urban labor market. Thus,* the supply of unorganized
urban labor is given by:
(12b)     L3   L  - L2                             one
The employment of organized labor is simply given by-demand:
(12c)     L2 = L2                                  one
Given that the urban organized wage is fixed, there are only two
labor markets for which wages adjust endogenously. Thus, labor market
equilibrium conditions are given by:
D    S
(13a)     L    L  = 0                              k = 1, 3
k    k
(13b)     Wk =W k                                  k = 2
A.1.3    Foreign Trade
The world demand for exports by sector are a function of the
supply price PWE   and the average world price for that sector; Eie   The
i 
equation is:
(14)      Ed = EB   pt   Yi                        n
i~~~~
where EBi is the "normal" demand for exports when    i =PWEi and ni is the
elasticity of demand.



-A8a-
For three sectors (agriculture, mining and textiles) we retain the
small-country assumption and specify that the world price of exports is
not affected by Turkish exports. For these sectors, we instead assume
that the share of domestic production which      exported is a function
of the ratio of the domestic price to the export price. Thus:
Se = EdI/Xd= f(PD./PE )
1    I  i       1 i
where PE, = iT   (1+ te ) ER.  Exports are now given by:
E e  Xd
~  ii
The function f(4) is a symmetric logistic function with lower asymptote
of zero and tipper asymptote of twice the base-year export ratio for each
sector. The equation for value added per unit of output (4) must also
be adjusted to reflect the fact that exports and domestic sales are at
different prdices.
The demand for imports depends on combining imports with domestic
goods in desiLred proportions which are a function of the relative prices
of imports and domestic goods.



- A9 -
Desired sectoral imports are given by:
(15)       MDi =     (    a6  ) i (PD  )ai Sd                  n
MD1     1-6         PM.     ii
i          31
where ai is the elasticity of substitution and
6i is the import share parameter in the C.E.S. function defining the
composite good.
The variable Sd in equation (15) is the supply of the domestically produced
good to the domestic market and is equal to total domestic production
minus exports:
d    d     d
(16)       S  = X  - E                                         n
i    1     i
As discussed in Section 2.3 , we assume that imports are rationed
(in proportion to desired imports) so that total imports do not exceed
available foreign exchange. Total imports in dollars are given by:
-~ed    -      -      -*
(17)       TIM    i Ei  + FK1 + FK2 +K3 -AR                    one
where FK1 is workers' remittances from abroad,
FK2 is long-term foreign capital inflow,
FK is short-term foreign capital inflow, and
*
QR is the target change in foreign reserves.



- A10 -
Imports are determined by multiplying the level of desired
imports by a rationing factor, RM:
(18)       RM = TIM/El    * MD'                               one
ii       i
(19)       Mi = RM * MD'                                       n
i        ~~~i
t.1.4     Income arLd Nominal Flow-of-Funds
Net labor income is given by:
(20)       YL = EIWk * Lki(l - t*)                            one
*
{here tk is the direct tax rate.
Net non-.labor factor income is given by:
(21)     YK =(Pv xd _ WL W       ) * (1 - tk )                one
i    i i    kki ii
*
{here tki is the dlirect tax rate on sectoral non-labor income.
Government non-factor income is given by:
(22)       YG = EEt W L     + Etk   *(PVx    - _EW L     ) + Etm ERM
ik'k kki       i      i i     k kki         i i 
*    *~d      *      d
- I'te iPWEiER Ei + Ets iPDiXi               one
1.           i~~~~
Total resources are defined as the total amount of funds available
:o be spent on gocids, either domestic or imported:
'23)     RT =PV Xd + Itm lmiER Mi m             ER Ed  + Its    D   +(             + 1)ER
ii   i      1      2      3
one



- All -
Labor resources are the total funds available to labor. They
are given by:
(24)      RL = YL + GLTS  - ILAH + FK1ER      PLAR ER              one
where TS  is total transfers
eL  is labor's  share of total transfers,
*
AH  is new money creation (seignorage), and
PL is labor's share of forced holding of new money.
Capital resources are defined as total funds available to recipients
of capital income and are given by:-/
(25)      RK =    - (p K - Ps)AH + OKTS   + FK3ER  - pKAR ER       one
where 1K is capital's share of forced holding of new money,
PS is capital's share of receipts from seignorage, and
OK is capital's share of transfers.
Government resources are given by:
(26)      RG = YG + (1   ps)AH     TS  + FK2ER                     one
As discussed in Section   2.7, the treatment of new money creation,
or seignorage, is designed to permit the model to be extended to include
monetary and inflationary submodels. It can be seen
as a special type of transfer mechanism. The seignorage account takes
*
money, AH , in the form of forced savings from labor and capital income in
the proportions p L and vK (pL+pK=l) and gives it to the government and to
l/ Note that in terms of the Social Accounting Matrix discussed in Section 2.7
this treatment combines the "enterprise" and "capitalist household" accounts.



- A12 -
capitalists in the proportions (l-1is) and pS    Note that changes in foreign
*
reserve (AR ) are assumed to be held by the private sector and are divided
in the same proportions as holdings of new money. Thus, in the model,
they are treat:ed in a manner analogous to that of new money creation.
The transfer account, TS , simply takes money from government
resources and gives it to labor and capital in proportions 0K and 0L
(0 +0 =1). The account is included to capture transfer mechanisms that
K L
are both impl-icit and explicit in the Turkish accounts and to provide an
interesting policy instrument for experiments.
Total nominal investment is given by:
(27)      TINV= SLRL + SKRK + GINV                 one
~~ K
where SL is the average savings rate from labor resources,
SK is the average "savings" rate from capital resources, and
GINV is government investment (defined below).
Total nominal investment outside the central budget is given by:
(28)      PINV = TINV - GINV                        one
Nom:Lnal government investment is either determined residually or
set exogenous:Ly. Thus,
*
(29)      GINV? = RG - GCON                         one
or GIN'J = GINV in which case AH becomes endogenous.



- A13 -
GCON is nominal government consumption which is set exogenously. Al-
ternatively, it can be set endogenously instead of AH when GINV is fixed.
What is required is that some component of the nominal flow-of-funds must
be allowed to adjust endogenously to "close" the accounts. Different closure
rules are discussed in Section 2.7 above.
Total nominal investment in stocks or inventories is given by:
(30)      SINV = EPiSi                             one
ii
where S. is real investment in stocks (determined below). Total nominal
fixed investment equals TINV - SINV.
Total nominal consumption expenditure equals:
(31)      PCON = (1 - SL)RL + (1 - S K)RK          one
We do not distinguish consumption patterns by different categories of con-
sumers and hence treat all consumers as having the same tastes.
A.1.5    Product Demand and Supply
In this section, we present the sectoral demands for products by
consumers, investors and government and derive the excess demand equations
for the product markets. Note that, except for exports, the demands are
for composite goods -- the superscript d is used to denote the domestically
produced good.
Consumption demand is given by:
(32)      Ci = (cpiPCON + cgiGCON)/Pi              n



- A14 -
where cpi are private consumption shares and
cgi are government consumption shares.
Ecp, = Zcg1 = 1.
i      i
These expenditure equations are extremely simple, a linear ex-
penditure system without subsistence m:[nima.
Real investment in stocks, or inventory accumulation, is given by:
(33)      Si = Yi(Xi(t+l) -Xi)
where Yi are inventory coefficients and
-d
Xi(t+l) is projected output next year.
Fixed investment by sector of destination is given by:
(34)      Yi = YSi(PINV + GINV - SINV)/PK1                    n
where YS. are ithe sectoral shares of private investment by sector of
_
destination (EYSi=1).
ii
The shares YSi are determined as part of the investment model and
are a function of sectoral shares in total capital stock and of differential
sectoral profit rates. They can be determined endogenously in the within-
period model and so be a function of current profit rates or can be set in
the between-period dynamic model and so be a function of past profit rates.
Their determination is discussed below in Section A2 as part of the dynamic
model.



- A15 -
Fixed investment by sector of origin is determined by applying
the fixed capital coefficients (bij) to investment by sector of destination:
(35)      Z  = Eb ijYj                                          n
The demand for intermediate goods is given by multiplying the
fixed input-output coefficients by domestic production:
(36)      V. = Eai                                              n
Effective domestic demand for domestic goods is determined by
subtracting sectoral import expenditures from total expenditure on the
composite good by sector.
(37)      Dd = {P.(C. + Zi + V + S)      PM M./PD.              n
i     1  1       i    i         i i    1
Equilibrium in the product markets requires that the excess demand
for domestic goods be zero in every sector (including export demand):
d _d
(38)      D. - S.   0                                           n
The essence of the within-period solution problem is simultaneously
to find a set of prices such that product excess demands are zero, a set 6f
wages such that labor market excess demands - equation (13) - are zero,
and an average import rationing rate (RM) such that the balance of payments
is consistent with desired reserve accumulation. Such a solution
represents a general equilibrium in the markets for products, labor and
foreign exchange that is consistent with the behavioral constraints em-
bodied in the model. Given the specification of the factor markets and of
import rationing, the solution does not represent a neoclassical free-



- A16 -
market equilibrium. It is instead a solution constrained by behavioral
and institutional specifications believed to      represent a picture of
the Turkish economy that is empirically reasonable for a model designed to
explore questLons of trade and industrialization strategies.
A.2 Dynamic ?4odel Equations
Table A.2 presents a list of the exogenous and policy variables
which are updated in the intertemporal model. Of the list, those relating
to investment allocation (YSi) and rural-urban labor supplies (L and L )
are behaviorally the most important. The rest are projected using simple
time trends, growth rates, exogenous projections, or accounting. We will
not consider these variables further, but will instead discuss the models
of investment allocation and of rural-urban migration which determine the
basic structure of the supply of factors in the next period.



- A17 -
Table A.2
Dynamic Model: Variables To Be Updated
Non-policy Variables
world price of imports
e: average world price of sectoral exports
i
PL        :;overall price level
A.           sectoral total factor productivity
EB         : export demand parameter
FK1       : workers' remittances from abroad
FK         : long-term foreign capital inflow
2
FK3        : short-term foreign capital inflow
cp.       : private consumption shares
cgi          government consumption shares
-d
Xi(t+l)      projected sectoral outputs (for inventory investment equation)
Ki         : sectoral capital stocks
YSi        : sectoral investment shares
W2           urban skilled wage
-u
L          : urban labor supply
-A
L          : agricultural labor supply
Policy Variables
*
ER           exchange rate
*
tmin         tariff rates
*
tei       : export subsidy rates
tsi          indirect tax rates
tk           direct tax rates on labor income



- A17 -
Table A.2
Dynamic Model: Variables To Be Updated
Non-policy Variables
world price of imports
e: average world price of sectoral exports
i
PL        :;overall price level
A.           sectoral total factor productivity
EB         : export demand parameter
FK1       : workers' remittances from abroad
FK         : long-term foreign capital inflow
2
FK3        : short-term foreign capital inflow
cp.       : private consumption shares
cgi          government consumption shares
-d
Xi(t+l)      projected sectoral outputs (for inventory investment equation)
Ki         : sectoral capital stocks
YSi        : sectoral investment shares
W2           urban skilled wage
-u
L          : urban labor supply
-A
L          : agricultural labor supply
Policy Variables
*
ER           exchange rate
*
tmin         tariff rates
*
tei       : export subsidy rates
tsi          indirect tax rates
tk           direct tax rates on labor income



- A19 -
The sectoral profit rates are defined as returns to capital when
the entire capital stock is valued in current prices and also includes
capital gains. The equation is:
Pv   xd     EW L
Pit i   -t _ k k ki,t  PK.it - (1 - di)PKi,t-
(40)      Ri,t     PK      K           +         PKt i,t-l
where       i,t-l  is the capital stock at the end of the last period
(and which is used in production in this period),
di      is fixed sectoral depreciation rates, and
PKi     is the capital goods price in equation (5).
When the investment mobility parameter p is zero in equation (39),
investment shares equal last period's capital stock shares. When p is
positive, the sectoral allocation of investment will respond to profit
rate differentials and high profit sectors will attract funds from low
profit rate sectors. Thus, p measures the intersectoral mobility of
investment funds. It is not, however, an index of the degree of perfection
of capital markets. Even if p is zero, the system may move towards equal-
izing profit rates over time, and, if p is too large, it is easy to make
sectoral profit rates oscillate. The parameter p is rather an indicator
of the responsiveness of capital markets to market signals, namely,
differential profit rates among sectors.
A.2.2    Rural-Urban Migration
The rural-urban migration model treats migration as being a function



- A20 -
of the differential between the rural and urban wages. Migrants are
assumed to be attracted by the average urban wage compared to the rural
wage (W1). Total rural and urban labor are also assumed to have exogenously
specified natural rates of growth. The labor supply equations are given
below:
A      A      A
(41)      Lt+1 = Lt (1 + G ) + MIGt
(42)      Lt1 = Lt(l +G)u+ MIGt
(43)      MIGt- =  e      ltLA
(44 )     w e = E(W2L2. + W3L3 )/Lu
As discussed in Section  2.8 above, this approach is similar,
but not identical, to the Harris-Todaro formulation.
It should be clear from the discussion of the dynamic model that
there is considerable scope for flexibility in the choice of which variables
are to be determined in the static and intertemporal stages of the over-
all model. One could, of course, take a lot more "out" of the static model
and put it "into" the intertemporal model. Indeed, price formation could be
modelled dynamically as a disequilibrium adjustment process in the between-
period model. This would dramatically reduce the degree of simultaneity in
the static CGE model! However, such an approach would focus the model
almost exclusively on short-run forecasting issues and would destroy much
of its usefulness for examining issues of medium-term development strategy.






Appendix B
The Data
Kemal Dervis
Sherman Robinson
July 1978






-B1-
Appendix B: The Data
This appendix presents a summary of the data and parameters used
to compute the 1973 base year solution of the TGT model. In common with
other developmetnt planning models, the parameters of the TGT model are based
on a single-year "calibration" of the model supplemented by adjustment in
the dynamic trends of some parameters to fit the historical 1973-1976
path.l/
The base year solution is bui:Lt around the 1973 State Institute
of Statistics (S.I.S.) 63 sector input-output table which also forms the
core data base to Turkey's offical Fourth Five Year Plan (FFYP). While
the data underlying the FFYP are suffic:Lent for a dynamic input-output
modelling exercise, they are not sufficient for the TGT model. We have had
to assemble addlitional data - notably on capital stocks, wages, employment
and government revenues and expenditures - from various other sources. The
most important among these are the following:
- The! S.I.S. Census of Manufacturing Industries.
- The! S.I.S. Annual Surveys of the Manufacturing Industries.
- The S.I.S. 1970 and 1975 (preliminary results) Population
Census.
- The! State Planning Organization's (S.P.O.) Annual Programs,
particularly the 1975 and 1978 Programs.
2/
- The! S.P.O.'s FFYP model-/
1/ Efforts to extend the estimation procedure to formal time-series and
cross-section analysis sill be undertaken under a separate research
project to start in the fall of 1978. (RPO 671-79)
2/ See S.P.O. (1976) and S.P.O. (1977).



-B2-
- A special S.P.O. study on capital and labor in.the manu-
facturing industries.-/
- The Budget Revenues Yearbooks published by the Ministry of
Finance.
2/
We have also drawn on data assembled for the research project:-
"A Comparative Study of The Sources of Industrial Growth and Structural
Change", notably by Merih Celasun of the Middle East Technical University.
In addition, we have used wage-share data assembled for the Bogazici-
Princeton income distribution project.
Table B-1 gives the sectoral aggregation scheme used for the
TGT model. Note that the sectors outside manufacturing are highly aggre-
gated and that the focus is clearly on the manufacturing industries. The
petroleum sector is vertically integrated and includes crude petroleum,
natural gas, refining and other petroleum and coal products. This verti-
cal aggregation was chosen because of the special nature of price formation
and indirect taxes in the sector.
Lightly processed agricultural products such as tobacco products
and ginned cotton are included in the food and textiles sectors respec-
tively. Thus much of cotton and tobacco exports do not appear as "primary"
exports, an important point to remember when interpreting the trade statis-
tics generated by the model.
Infrastructure includes electricity, gas and water as well as
ownership of dwellings, although it is dominated by the transport and
communications sectors. Services includes public services which, in the
1973 S.I.S. input-output table are treated as direct government purchases
of value added.
1/ See Ebiri et al.(1977)
2/ RPO 671-32.



-B3-
Tables B-2 and B-3 give the input-output coefficients and the
capital shares matrices used in the TGT model. The input-output matrix
is derived by straightforward aggregation from the 1973 S.I.S. matrix.
The capital shares matrix was constructed using the composition of invest-
ment data in the S.I.S. Annual Surveys and the capital stock estimates
described in Table B-4. The investment shares data were available only for
manufacturing. Outside the manufacturing sectors, the share coefficients
represent our best guesses, designed to add up to the overall economy-wide
composition of investment as given in the investment column of the 1973
Input-Output Table.
Tables B-4 and B-5 give the data and parameters for the sectoral
production functions4. We have assumed that the elasticity of substitution
between "organized" and "unorganized" labor is unity in all sectors which
employ both categories of labor. The elasticity of substitution between
aggregate labor and capital varies across sectors. For a number of sectors,
it is assumed to be one (i.e., a Cobb-Douglas specification). A CES
specification is used for a number of important manufacturing sectors and
mining. The elasticities of substitution have not been estimated econo-
metrically but are based on informal evidence in Turkey and on estimates from
other countries.
Total domestic output is derived from the 1973 input-output table,
with some adjustment necessary in the treatment of indirect taxes and
subsidies. The export column in the 1973 table does not reflect the fact
that unit export prices are affected by export subsidies. Given a positive
subsidy, the saime physical commodity wi:Ll sell at a lower price in the
export market than in the domestic market. Thus the export "units" are



-B4-
adjusted upward in accordance with the subsidy rates to achieve strict
comparability of units.
To determine employment levels by sector, we used the 1970
Census estimates and the S.I.S. data on the Manufacturing Industries.
First, we determined total sectoral employment levels for 1970 using both
the 1970 population densus figures and the S.I.S. manufacturing census
figures. When there was a conflict, we used the population densus figures
which tended to be somewhat larger. This procedure gives an estimate of
total employment in both the modern, large-scale organized sector and in
the more traditional, small-scale enterprises. From the 1970 manufacturing
census, we also have the total number of workers in large-scale enterprises
alone, where large-scale refers to enterprises with more than ten workers.
It is therefore straightforward to obtain a sectoral ratio of small-scale
to large-scale enterprise employment. We do not have figures on either
total employment or small-scale industry employment for 1973. But the
1973 annual survey of manufacturing does provide the number of large-
scale enterprise workers. Assuming that the small-scale to large-scale
ratio has not changed in the three years from 1970 to 1973, we can there-
fore obtain an estimate of small-scale enterprise employment for 1973.
(given in Table B-4).
For the sectors outside manufacturing, we simply use estimates
based on interpolation between the 1970 and 1975 Census estimates. We
somewhat arbitrarily classifed workers in construction and infrastructure
into labor category 1 ("organized" sector) because they are largely
covered by social security legislation and divided up employment in services
according to the share of workers covered by social security and the share



-B5-
of government employees.l/
To estimate wages and wage payments, we used the same method that
was used for employment, relying on the S.I.S. annual surveys and a parti-
tion between large-scale and small-scale enterprises. The major problem is,
of course, that a large proportion of the employed labor force even in
industry still consists of unpaid family workers and self-employed.-/
The assumption used is that all workers in small-scale enterprises, employees
as well as self-employed and family workers, received the average wage of
small-scale enterprise employees, computed from the 1970 census of small-
scale manufacturing enterprises.3/   No doubt for many family workers this
assumption overstates the return to their labor. On the other hand, many
self-employed may earn considerably more than the average wage of small-
scale enterprise labor. So on average the assumption made is probably not
too far from reality.4/
As is usual when building planning models, the data on capital
stock present the greatest problems. There is no series of investment
by sector of destination in Turkey that is reliable and has sufficient
economy-wide coverage. We have had to rely on imperfect and incomplete
sources to construct our 1973 capital stock estimates. The first source
1/ See S.S.K. Yearbook (1973).
2/ See Ecevit and Ozutun (1975) for a careful breakdown of the labor
force by employment status.
3/ See S.I.S. (1976).
4/ See Kuran (1978) for a more detailed discussion of the urban informal
sector in Turkey.



-B6-
used was the FFYP model document itself.-/ On pages 108-110 it contains
detailed sectoral estimates of marginal capital-output ratios for the FFYP
period. Assuming equivalence between average and marginal capital-output
ratios, once can get a preliminary estimate of 1973 capital stock. While
it is useful to go through this exercise, the marginal ratios may differ
from the average ratios, and it is the average ratios that are needed to
estimate sectoral capital stocks.
More recently, three researchers in the Social Planning Division
of S.P.O. have attempted to provide a consistent series of capital stock
and investment in large-scale manufacturing enterprises based on the S.I.S.
survey data and on the unpublished results of a more restricted S.P.O.
survey of private manufacturing enterprises.-     This study represents
a careful effort to reconcile the available data but it is only addressed
to large-scale manufacturing enterprises. Nevertheless we have taken it
as our starting point for capital stock estimates.
The average capital-output ratios implicit in this study are
significantly lower than those obtained by using the incremental FFYP
estimates and assuming them to be equal to the average ratios.
The capital stock figures provided in Table B-4 are close to
what is implied by the Ebiri et al. S.P.O. study for the manufacturing
sectors. For agriculture, mining, construction, infrastructure, and
services they are weighted averages derived from the FFYP estimates of
1/ See S.P.O. (1977).
2/ See Ebiri et al. (1977).



-B7-
incremental coefficients and assuming that the average coefficient is equal
to the marginal coefficient. The sectoral profit rates reported in Table
B-5 provide a useful check on the "reasonableness" of the estimated
sectoral capital stocks. The magnitude and variation in sectoral profit
rates do, in fact, seem reasonable and support our use of the estimated
average capital-output ratios.
Table B-6 gives the presentation of the base year trade data.
The sectoral breakdown is based on the 1973 S.I.S. input-output table.
Note that trade and transportation margins appear in infrastructure and
services and are not distributed to the individual sectors. The Armington
substitution elasticities and the export demand elasticities constitute
the "best guesses" and represent a compromise between the extreme elasticity
pessimism common in Turkey and the much more optimistic views of many
outside researchers.  They also represent an intermediate position between
the "perfect complementarity" assumptions made by fixed coefficients models
and the "perfeci: substitutability" assumptions of pure trade theory.
Table B-7 gives a number of miscellaneous sectoral parameters.
Perhaps the most: important are the productivity growth or technological
shift parameters.  iney are essentially based on running the model for the
1973-1977 periodl and varying the technological shift parameters until
growth in the major sectors projected by the model equalled growth that
actually took place during that period. We did not have data on sectoral
output growth rates at the two digit level and in the absence of such
knowledge we did not attempt to specify differential rates of technical
progress within manufacturing. While it is quite clear that technical



-B8-
progress has been much slower in the service and infrastructure sectors
than it has been in the manufacturing sectors, we lack knowledge on
relative performance within manufacturing. If it is true that technical
progress is going to be more rapid in the basic intermediate and capital
goods industries than in the "light" consumer goods sectors, this would
have a significant impact on the design of optimal long run industrializa-
tion policies.-/ But the little evidence there is on this issue is con-
flicting and it was preferable to assume essentially uniform technical
progress rates rather than introduce a variation that would add to the
complexity of the results without being based on conclusive evidence.
The world-trade growth rates are somewhat higher than has
materialized in the 1970's, reflecting a more optimistic view than is
currently fashionable. However, one must take into account Turkey's
geographical location near the booming Middle Eastern market where demand
growth has been, and will continue to be, much more rapid than world
averages.
The sectoral consumption shares, indirect tax rates and tariff
rates are based on the 1973 S.I.S. input output table and the S.P.O.'s
FFYP model document.-/ We have also checked these estimates against
tariff revenue data from the Finance Ministry's Budget Revenues Yearbook
and the TGT model's projections are very close to the revenues actually
collected during the 1973-1976 period. The same is true for direct and
1/ See De Melo and Dervis (1977) for an exploration of the impact of
differential technical progress on optimal trade policy.
2/ SPO (1974).



-B9-
indirect tax rates that have been estimated by fitting a linear trend to
the proportional rates so as to reproduce tax collections for the 1973-
1976 period.
Finally the growth rate of world prices is projected in terms of
nominal dollars and includes an assumption of a very small but continuous
downward driEt of the dollar against European currencies, yielding a 9.0
percent dollar denominated world inflation rate. Given that many of our
porjections are reported in nominal dollar terms, this 9.0 percent world
inflation assumption is important to keep in mind, although the real
variables of the model are not affected by the world inflation rate as
such but only by the difference between domestic Turkish inflation and
world inflat:Lon, given the specified time path of the exchange rate.



Table H-1
Definition rdl Composi tioii of the 19 Sectors Used in the TCT Model.
91 OCCLUL                                           ¼ ..4._ .,a-zr
1.  AgricuLltutre                    01.  Agriculture  02.  Animal Husbandry
03. Forestry 04. Fishing
2.  Mining                           05.  Coal Mining
07.  Iron Ore Mining   08.  Non-ferrous,
Ore Mining. 09. Noni-Metallic Minerals
10. Stone Quarrying
3.  Food                             11.  Slaughtering, Preparing, Preserving Meat
12. Cannitng & Preserving Fruits & Vegetables
13.  Mfg. of Vegetables and Animal Oils
& Fats
14. Grain Mill Products
15. Sugar
16. Manufacture of Other Food Products
17. Alcoholic Beverages
18. Soft Drinks
19. Tobacco Mlanufactures
4.  Textiles                         20.  Cotton Ginning
21. Manufacture of Textiles (excluding ginning)
5.  Clothing                         22.  Mfg. of Wearing Apparel
23. Leather & Fur Products
24. Manufacture of Footwear
6.  Wood & Wood Products             25.  Mfg. of Wood & Wood Products
26. M4fg. of Wood Furniture & Fixtures
7.  Paper & Printing                 27.  Mfg. of Paper & Paper Products
28. Printing & Publishing
8.  Chemicals                        29.  Manufacture of Fertilizers
30. Manufacture of Drugs & Medicines
31.  Manufacture of other Chemical Products
9.  Rubber-& Plastics                34.  Mfg. of Rubber Products
35.  Mfg. of Plastic Products
49. Other Manufacturing Industries
10.  Petroleum & Pet. Prod.           06.  Crude Petroleum and Natural Gas
32. Petroleum Refineries
33.  Mfg. of Petroleum & Coal Products
11.  Non-Metallic Mtineral Prod.      36.  Mfg. of Glass & Glass Products
37.  Mfg. of Cement
38.  Other Non-Metallic Mineral Products
12.  Basic Metals                     39.  Iron and Steel
40. Non-ferrous Metal Industries
(Copper, etc.
13.  Metal Products                   41.  14fg. of Fabricated Metal Prodtucts
(Struct. & Heating Equlipment, Tube &
Sheet Iron, Kitchen Utensils,
Products, etc....)
14.  Non-Electrical Machinery         42. M4fg. of Machincry except Electrical
43. Agricultural Maclhinery
(Note: includes refrigerators,
washing machines)
15.  Electrical Machinery             44.  Mfg. of Electrical Machinery
16.  Transport Equipment              45.  Shipbuilding & Repairing
46. Mfg. of Railroad Equipment
47.  Mfg. of Motor Vehicles
48.  Mfg. of Other Transp. Equipment
17.  Construction                     52.  Building Construction
53. Other Construction
18.  Infrastrutcture                  50.  Electricity
51. Gas & Water
56. Other land Transport
58. Water Transport
59. Air Transport
60. Cosmsunication
64.  Ownership of Dwellings
19.  Service3                         54,  Wholesale & Retall Trade
55. Restaurants &. Hotels
61.  Financial Institurions
62. Pers. & Prof. Services
63. 7lu1lic Services



boillo    Liisla   E26010    S906.1 0  6LL s  0  Ooos:O   blDg,o    IOL910   ocqslo           hrS  0 2
Lssolo    L72solo  929clo    296010    Y2&O'G   ?690,0    egLO-10   SOL010   coto,o      S3JIACjS  61
2610'0    v2i0IO   Dsiolo,   6920'0    s2io.0   J.920"O   q2t)OIO   eLLO-10  ?,Gsllc     tlbvbjsI  Si
000010    uooolo   000010    oooolo    0000,0   oooolo    0000"O    DOW D    Qocollo     Incl
tcO  it
iocolo    ISPOIO   iDijolo   f?O92'O   sloo'O   SIqu'-O   Oooo-10   Oluo"O   Occo"O               91
Eooolo    zicolo   Lczolo    600010    ztoilo   ELOO10    Llu"110   iloo"O   2POO'C
000010    LOOO*O   921010    000000    9IOQ u   Si6l"O    Louo,.u   000010   6000"C      V6
313,111-
eloolo    socooo    lotolo   0600'0    6feo o   o2I0'O    96iolo    22UO'O   9000,10     ctclvl-;
  El
0000 10   0 0 0 0  0  ii I? II 0  jr,2 I  o  9 L s 2 1 0  oij I 1
 o  i Oa 0 " 0  I i 1 t7  o  9 0 1 C.- 0  1 v I 3
s v s  2 1
uo o o I o  $ 2 0 0  0  92 u I  0  q 2 0 0  0  jQ00'. o  s 2 o O', 0  4200, u  6 I 0 o  o  I ns 0 1 0  C LleIN i;
CN  I I
6soolo    iis I  0  5 0 1 0  0  ef7oo 10  sioo '. o  z, 9 o 0 " o  sso %) I 0  Le, 10:0  i 6 t o'. C  L:6c , 
fl 26  o I
S20010    yi r? 2 0  u  I 0 0 0  0  1 6 u 0 1 0  l6uO
 0  fqOO,O  S600'.u  0 0 0 0   0  f I o o I 0  Vld
L2,`  6
s 2 o 0 1 o  a I 0 0  0  2 0 I 0  0  i 0 1 0  0  bsaolo  a 6 o o '. o  9biolo  i j u 0 I 0  s6avc  9
9soolo    snooso   0 0 0 0  0  1 0 0 0  0  2EOU"O  DI oo" C  0 n o o  0  aouo',o  i i 0  0        L
0 1 o o I o  9 a 0 0 0 0  svgolo  6 9 0 0  0  L90010  t) q r, o I o  LL 0 0  0  u I oo" o  2 0 0  0  Q CC 0  (i
000060    200060   '0000,0   100010    000010   0000"O    ioio'-o   Qouo',O  0000"o     -IctiH10114  s
920010    200010   000010    200000    2000'0   0000"O    gloo"u    jOuO',O  scoo"o      Sillix . 31  0
992010    aloolo   000010    oooolo    oooolo   gooolo    uooollu   oooolo   Ocuo"O          QCCJ  f
900010    ttiolo   i6OOIO    600040    9OU010   aouolo    6100.,u   LqLo"o   c2LO"o               2
oboo-o    iq(jolo  itoolo    S000.0    2000"O   9100,10   looo'O    iouolo   lnoole
pua
3r,61SNOO  dCJSNViii 'HOVH
313  Vi,0313tN  Qbe1Vj3i%. 1VOKSve  CtidNIWLIN  Nwflico
6171;!?'O  7sbg,o  iissoo    2 Z 0 9 I 0  SS69*0  SGSIO   1.  L I 0  S62,'O   q -4 T I 0  2 o 6 2  0     6  02
Lifolo    029110   W2W  O    r;04110   vn,
010  996010    996('O    W64olu   iqil-o    Etiolo      j3DJAb
S  61
102010    I6*0'O   I2vo,o    9hsolo    2090.0   $610"O    igio'la  iopo'lu   siso"O    ovio"o              el
000060    OOoOIo   000010    000010    U00010   0000 10   oooo'lo   0000"O   000010    0000"O      OrtIVNL
  It
200010    ttoolo   000010    oooolo    0OU010   0000,0    00001,0   200olu   ('910"o   i2OO'O      cr.;"(.NVbl  91
iooolo    000010   uooolo    2OUDIO    000010   000010    1000"O    iouo',u  iVuo"o    000010     04Z'vf
;313  St
600010    ltoolo   SOOOID    200010    2DO010   oooolo    t'loo"o   
000,u   2?7io"o   Looo"o      Vv;;3-13LN  f7l
stoolo    2610*0   621010    9 0 0 0 40  ssoolo  Of?OOIO  0 o o o', o  gnoo"o  C06U0.0  6 2 0 0 1 0  Chelvi3w  Si
efoolo    oqiolo   600010    6200'0    socolo   O
oc I10  oloolo    2soo"o   1610"O    000010      IvI3
SvE  21
200010    efoolo-  29(010    1700040   2OU0,0   0000.0    iooollo   117(io"u  020t)"O  000010     Ctc-NikLh  11
9 2 21) I 0  9 I 0 o I o  0 L I 0 1 u  zillolo  ?S0010  V20O'O  1?1.01)"o  17 q 0 0IU  02t?o 10  I I I u I C  Cec,013d  ot
tooolo    29ROOO   1.7qjolo  220010    Ytoolo   9LOO -'O  L200 0   aloo'lo   LGQO'O    sloolo     vld
E6ritf 6
egoolo    f7t7o?IO,  199210  jqS010    6L2080   9610"O    6LLG.'O   jjoO O   tito'-o   lojo'-o     lv;lk3ho  9
oooolo    911000   scioso    LSEVO     2t?oooo  C-20040   92000     onlo'-o  souo',o   oooclo         dJdvd  L
LOOOIO    iioooo   iloolo    f?ooo,o   2SQ110   iooo"o    flooo,o   oloolo   uooo"o    oloolo          UOLM  9
o o o o I o  o 0 o o I o  0 0 0 0 1 0  0 0 0 0 a 0  2 0 0 0 I 0  V96110  I o o O'. o  oooo'-u  Snuo -10  o o 0 O'. o  !)NIHIC13  S
000010    IOLOIO   900000    12000     iloolo   iLolo     SqU210    ssao"u   s2Qo"O    iotiolo    S31lixii  0
o o o o I o  2 0 o o 0 o  L 6 1 0 I 0  e0oolo  0 0 0 0 0  oisolo  I o o o', o  0  I-10  I 0 0 0 I 0  6 2 1 0 1 0  fjoci  f
;00010    zooo,o   2;iO!O    
0000     iooo"o   w2oo.-o   400c"U    Divu.-V  OSOO"O    oooo'c-      
.-N!Njk  2
100010    M OD     100060    issolo    121i'so  21"0'0    t."L2,-o  SIZOIC,  ZS20'O    CiLl"O     11rolbov 'I
C b'g     V-14+-QiAb 'IV'.lK3H3  6204    OOOM   OhIH4012 6311IX31     coci    
NJNJW   ilrloldov       Nwrilco
vy  lxjblvK   indinC   jriNj
Z-Ei zlizvi



OOOO't      OOOO't     0000 M      0000'1     oooo't      0000-1     0000,1      0000'1     oooo-t              hC6 02
OSWOO       OSWOO      Os fo Io    ost°o I    OOs fO Io   o b0 -     O5bO'0      Osbo Io    O Sa-         6:1c       et ?JA3 
0520'0      05C'SOa0   OS2010      OSZO'O     oseoo eI5zOS21,o       us a o I  0  52Q'0 Io  o5e0|0        t lSVoj,I B I
0999'0     0021'0      OiSO'O     0192'0      09?2.0      OLWO;      0X912"O     OhrOnO     oOst"O        V;bl"!c: Li
OOwO*O      OOsto1     c s iO Io  of? O Io    oqio'O      0090>      ObeO;       ossoBy o  a et'  ocio"o  ct;35b.Vh  9I
0260'0     O 9 I0 O'  o7'Q     oa O'       0060eo      O0$6;O     °ee°;U      ov:O 0     oteO;O        OV h3t3 5 i
OSSO'O 1    0410'0        CIZ' I I  04  O     Ob795'0     ooas 0     Q29S5X°        09 o° 0'  05r,9;0     vWtC et
020010      OIOO       ovoo Io     O8Co' o    0  oXO      Osoo'0     USCoWo      0010'0o    o to o o       bli        1
OOU'        Q°O' u0000' 010  LOI0  0000o o    0 0 0 0;,      0 o00,00; 1)0 QnOO'  O OQu' cO  c 0000       l v 35 9    l
oOoo'Ol     uonoo      coooOFo     oooobo.    oooolQ      oOoo o0    uooollo     u000;o     oeoo'c       tCcc"i,C-, tl
OuouO'      UGuO'O     vuuulo     OOUO'Ol     0000'l u     00 .O 0   0000 °      0000;0     OOoO;C        Cbc tljd Cl
0200'0     of'oolo     oooobo     O0oloO      0200'U.     otoo'o     V200'O      0100"o     oeOQ'O        vICbE6rt e
a0000'0 o   0000'0      0 000'0    0 0 00;.0  0 0 00'0    0000. 0    U000 °       ou' o     ovoO, OOC        a lVIh3!: e 
QOO'Ol      0000'0     OOoO'o      Q000'°0    0000'0      000O%0     OooCO 0     odOvO,"    OOQC'O            b35dVd L
o O oI     0 OoOO1     o o u  o   Ooo I'O    o o o "     oo o   ou  0000'°      oOQO'O     C:CC': I           'O-Ct 9
ucooo       o0o0t0     toQooo'    OOoOwO.     oooQ'o      oooo o     GouO"o      OCOO'O     OCOC'O        ShlIhIo1: 5
oOoO'O      voIolo     0ooooo      oooolO     OOoO'0      Gooooso    OooO-o      oOooo1     OCOC'G        63'1ix-'   t
oooOlO     vooooo      o00OO1     o000O       00o0.O     Ooooo;o     0000;0      oooooo     OQOO'O             coci E
oooo o     0000 O       ooOlOO    oOoQO      o 001        OOo0O o    uouoo °      001 o°°Ol                          2 OCO5hswC
00o010      0oooo       ooo oooooo            ooooo0 OOO'O 00O'      uooollo     0oooXo     oooolo        lltr31etv I
I                         ~~~~~~~~~~~~~~~~S33iAt3S  tlSVbdAh I  ',rt6NGO  dC3SNVtl  vn3  vw;313LN LCt l.v 13WI 1t3etBv 0CbaN 'CIN          Iw:C 
OSaO'"O    OsaO'O      oSaoO01    oslO Io     osanWo1     ost?o ,    OSaO'O      °sDo-o     05osowe      oSoO' ,C     S3;iA8bS elI
0520'0     0520'0      052010      0520'O     OSZO'O      0520,0     0s2'0       °s20'°     0520'10     0520'0        b1svdiNI el
C20t'O      O g II o        olo    ossl 09siO oq6 2'0     09 1;0     o 6 a2'      $   0 o90S,O u92'0o   OC.S-c .O     ;f("c IS,. 3 't 
O09uO'00    O O I O O    eo'o0     0 02 0'a0  0 q990'10   O 2 tO 0   Oid- O'O    o9go0;0   0o60 I       OOeO'O        dc. :S-Nv c  9 1
o0s'o O    O O)0 0    oseolo      or I I     02at,110    O;O I      0060o;0     06Wo°;     ooso" 0e      v oxo'      eV-03-3 SI
0 ot e 'o     2 O  S2g I0    0        .7q 0   U1 I oaa0 9'09 unbZ    09s@   OSs% 0  06050   0OS         00ee' cb~ O   v20~ 131sc', f" 
o£~oQO      Ooeoo'     oqo 090'       Q2 U  0 O oO O Q    02 05 0    G200'0      °tQO IUS( O goo        Oco s  oO     C t:d Iv I  1 i
L)0o ooo'o   oOo' O     so'        oooO'o 10  00  oooO'O  ooo'       ooo O 0 O  o  0O O     00   . O     c o  Ooci  ocoC  iIV1HSve 2 I
oOoO'O      Cooo'Q     ooo 0000    01oooo     oOOO'O      OOoo'      oOoo  O     Q000;0     0000;-0     uCCO;O        Ltc;litkL  It
0(oooo'Q    oooolo     oOoolO      oeDoco     uooOo       ooooO       o QOOO    0O"O-     O(UUOo      oOOO'O        Ccc-tlad 01
0200'0      0900'0     090G'0      Ot.QLO'O   02Q0'O      090;0       Suo  0900"O o   0ol   020010      oloO'o        vic4iEfra e
UOOO'O      COOO'O     0000'O     0000'0      0000'0     0000'0      0000'0      °000;0.    0O0.u0      OCOCO"       .lVZl^3KO 9
O03uO 0      oO'Ol    o0oo'O      0000'0     0000'O      0000'O     0000" C      001       000 ° ooolo  cOoOo'c         b3evd L
OO00'0     O 0OUO'     0°0 0'0,0  0 u0 00'0   0000'0       oolo      c 000'  COOC 0c o O     0G'  OQO'O1  Q C O            C.0  9
ooovo O      o OO O    uD(ooo       O 01O u        ooo0 Fo  Onf OOuu  UOoO'o        OooO u'  oo u ;o O  OCO'          St. 1011 S
0oe0 '0     oooo'o     0000oo      oouolo     U0OO'O      0000'0     OoO8O       oo ooou    OQo;uo     oCoO'G        S3111-xai a
0000' 0     0000       0°U00 0     OOQO O     0000 °      oOooO u    oOo0Oo      QouOO °    JQI 0000l 0O- OOUOEc 
oooOlo     oooo'O      OOOo*O      0000O0     u000'O     ooOo'O      oooo'o      Ooo0Xo     oouooo      ooeooo          -ssb2
00OVoo'      O OO  OOGO'O      oOOlO'    COOOO 0     on(Io o'O   00 0OOoO'O       0 0000'0I  OOUO' ,o;            ;'r3iSv t
1Vn9  tsAs3ss;S3r- nxrlI¢v-~~~~C t
C-9 3a9YIa



-B13-
Table B-4
The Arguments and Parameters of the Production Function, 1973
Produc-    Capital    Labor   Labor    Labor      Subst.a
tion        Stock       1       2        3       Elast.-
1.  Agriculture               103225     146300     9600      --       --       1.00
2.  Mining                      4049       8383      --      106       --       0.50
3.  Food                       45813      16034      --      145       53       1.00
4.  Textiles                   24999      14999      --      142       47       1.00
5.  Clothing                    8904       4274      --       41      288       1.00
6.  Wood & Wood Prod.           6592       4614               14      123       1.00
7.  Paper & Printing            5311       6373      --       25        9       1.00
8.  Chemicals                  10803       8858      --       35        2       0.75
9.  Rubber & Plastics           5870       3404      --       27       17       0.75
10.  Petroleum & Pet. Prod.     16094      28165      --       11       --       0.25
11.  Non-Met. Min. Prod.         6646      10035      --       44       11       0.75
12.  Basic Metals               15200      25840      --       47       --       0.75
13.  Metal Products              6894       3447      --       34       89       1.00
14.  Non-Elec. Machinery         9953       4777      --       38       17       1.00
15.  Elec. Machinery             4774       2387      --       20       15       1.00
16.  Transp. Equipment          12047       5421      --       49       50       1.00
17.  Construction               28570      11428      --      448       --       1.00
18.  Infrastructure             59477     309280      --      439       --       0.75
19.  Services                  107778     105000      --     1500     1580       1.00
Total                482999     719019    9600     3166     2301        --
- Elasticity of substitution between aggregate labor and capital.
Units:
Labor: lOOO's workers
Production and capital stock: millions of 1973 TL.



-B14-
Table B-5
Sectoral Wages and Profits
Wages:
Labor       Labor    Labor     Average    Profit
1           2        3       Wage        Rate
1.  Agriculture                4.57         --        --       4.57       23.96%
2.  Mining                      --        19.50       --      19.50        6.28
3.  Food                        --        19.05    12.17      17.20       42.33
4.  Textiles                     --       20.17     10.79     17.86       12.76
5.  Clothing                    --        18.27      6.30      7.78       39.99
6.  Wood & Wood Prod.            --       16.23      8.36      9.17       17.67
7.  Paper & Printing             --       29.83     11.40     24.76       20.19
8.  Chemicals                    --       29.45     11.25     28.39       47.73
9.  Rubber & Plastics            --       21.74      9.21     16.95       32.71
10.  Petroleum & Pet. Prod.       --       38.00       --      38.00       25.71
11.  Non-Met. Min. Prod.          --       23.51      9.71     20.75       15.36
12.  Basic Metals                 --       29.82       --      29.82       14.06
13.  Metal Prod.                  --       22.48      8.23     12.14       34.55
14.  Non-Elec. Machinery          --       26.30      8.68     20.74       89.81
15.  Elec. Machinery              --       26.44      9.86     19.46       63.26
16.  Transp. Equipment            --       31.03      8.78     19.85       65.34
17.  Construction                 --       18.50       --      18.50       66.23
18.  Infrastructure               --       20.00       --      20.00       12.28
19.  Services                     --       24.00      9.00     16.30       45.50
Average                4.57       22.36     8.72       8.94      23.57
Units:
Wages: 1000 TL per worker
Profit rate: percent



-B15-
Table B-6
The Foreign Trade Parameters
Arming-
Domestic Import    ton   a!          Export     Export
Imports   Demand  Ratio     elast.-  Exports  Ratio      Elasticit3
Agriculture              1213    101093    1.2%     2.00     2132     2.1%     F.W.P.b/
Mining                 -  284)     3703    7.7      0.50      346     8.5       F.W.P.
Food                      420     39276   1.1       0.65     6537    13.7       2.00
Textiles                  373     19843   1.9       0.65     5156    20.5       F.W.P.
Clothing                  188      8126   2.3       0.65      778     8.3      2.00
Wood & Wood Prod.,         27      6551   0.4       0.65       41     0.6       2.00
Paper & Printing          491      5282    9.4      0.65       29     0.5       2.00
Chemicals                7364     10598  69.8       0.33      205     1.8       2.00
Rubber & Plastics        1038      5787  18.0       0.33       83     1.3       2.00
Petroleum & Pet. Prod.   4679     15258   31.2      1.50      836     5.0       2.00
Non-Met. Min. Prod.       395      6291    6.3      0.65      355     5.1       2.00
Basic Metals             4512     14740  30.8       0.50      460     2.9       2.00
Metal Products           1155      6695   17.4      0.50      199      2.8      2.00
Non-Elec. Machinery      9465      9841   96.5      0.33      112     1.1       2.00
Elec. Machinery          2652      4755   56.0      0.33       19     0.4       2.00
Transp. Equipment:       3918     12031   32.8      0.75       16     0.1       2.00
Construction              --      28570    0.0      0.20      --      0.0       2.00
Infrastructure            993     55069    1.8      0.20     4408     7.2       1.25
Services                 1569    102761    1.5      0.20     5017     4.5       1.25
Total
a/Elasticity of substitution in use between imports and domestic goods.
b/Fixed world price.
Jnits:
Imports, domestic demand, exports: millions of 1973 TL.



Table B-7
Miscellaneous Sectoral Data
1977
Technical       World          1977         Indirect          1977        1977
Progress         Trade      Consumption        Tax           Tariff      Export
Sector                     Rate          Growth        Shares          Rate            Rate       Subsidy
1.  Agriculture                     4.5%           4.0%         22.8%           0.7%           19.9%        1.2%
2.  Mining                          5.0           10.0           0.3            1.1             5.8        20.0
3.  Food                            4.0            7.0          14.9            8.0            20.7        11.0
4.  Textiles                        4.0            5.0           5.3            9.2            56.7         6.0
5.  Clothing                        4.0            7.0           2.9            0.1             6.2        20.9
6.  Wood & Wood Prod.               4.0            8.0           1.1            0.8            12.9        21.7
7.  Paper & Printing                4.0            8.0           0.8            3.6            40.6        26.5
8.  Chemicals                       4.0            9.0           3.0            3.0            35.6        20.1
9.  Rubber & Plastics               4.0            8.0           1.2            2.7            38.4        16.5
10.  Petroleum & Pet. Prod.          4.0            8.0           1.0            0.0            29.7        11.0
11.  Non-Met. Min. Prod.             4.0           10.0           1.2            7.5            40.6        17.2
12.  Basic Metals                    4.0            8.0           0.0            3.5            12.3        25.0
13.  Metal Products                  4.0            9.0           0.2            0.3            18.5        29.1
14.  Non-Elec. Machinery             4.0            9.0           1.7            0.1            30.3        27.2
15.  Elec. Machinery                 4.0            9.0           1.4            3.0            17.4        19.4
16.  Transp. Equipment               4.0            9.0           1.3            5.6            31.0        32.2
17.  Construction                    4.0            7.0           0.0            0.5
18.  Infrastructure                  3.0            7.0          17.1            3.7
19.  Services                        2.0            7.0          22.0            5.7
Note:  Indirect tax, tariff and export subsidy rates are actual collection rates, not legal
ad valorem rates. World Trade should be interpreted as a weighted concept where geographical
areas closer to Turkey (such as the Middle East) have larger weights.



RECENT PAPERS IN THIS SERIES
No.                      TITLE OF PAPER                            AUTHOR
.279      India - Occasional Papers                          M. Ahluwalia, J. Wall
S. Reutlinger, M. Wolf
R. Cassen (consultant)
280      Educational Effects of Class Size                  W.D. Haddad
281      Relieving Traffic Congestion:  The Singapore       P.L. Watson
Area License Scheme                                 E.P. Holland
282      World Trade and the International Economy:         B. Balassa
Trends, Prospects and Policies
283      Urban Land Policy Issues and Opportunities         H. Dunkerley, A. Walters,
J. Courtney.
W. Doebele, D. Shoup,
M.D. Rivkin (consultants)
284      Pakistan:  Forestry Sector Survey                   S.A. Draper, A.J. Ewing.
J. Burley, G. Grayum
(consultants)
285      The Leisure Cost of Electric Power Failures        M. Munasinghe
286      Shadow Pricing and Power Tariff Policy             M. Munasinghe, J. Warford
287      Wages Capital Rental Values and Relative            S. Guisinger (consultant)
Factor Prices in Pakistan
288      Educational Reform in the Soviet Union:             I. Blumenthal,
Implications for Developing Countries              C. Benson (consultants)
289      Petroleum and Gas in Non-OPEC Developing           R. Vedavalli
Countries: 1976-1985
290      Major Reforms of the Swedish Education System      A. Heidenheimer
(consultant)
291      Industrialization, Technology and Employment -      T.G. Rawski (consultant)
China
292      Developmient and Income Distribution - Zambia      C. Blitzer
293      World Potash Survey                                W. Sheldrick, H. Stier
294      The Economic Dimensions of Malnutrition in         M. Selowsky
Young Children
295      The Technology of Rural Development                J.P. McInerney
(consultant)



No.                      TITLE OF PAPER                             AUTHOR
296      The Financial Cost of Agricultural Credit:          C.D. Datey (consultant)
A Case Study of Indian Experience
297      Agricultural Sector Planning Models:                A.C. Egbert
A Selected Summary and Critique
298      Textbooks and Achievement:  What We Know            S.P. Heyneman.
J.P. Farrell,
A. Sepulveda-Stuardo
(consultants)
299      An Economic and Social Analysis of the Chao Phya    C. Bruce
Chao Phya Irrigation Improvement Project II         Y. Kimaro
300      Two Studies of Development in Sub-Saharan           S. Acharya
Africa                                              B. Johnston (consultant)
301      The Intermediate Sector, Unemployment, and          W.F. Steel
The Employment-Output Conflict:  A Multi-Sector     Y. Takagi (consultant)
Model
302      The Economic Theory of the Household and Impact     D. Chernichovsky
Measurement of Nutrition and Related Health
Programs
303      Trade Restrictior.s and International Price         M. Bale
Instability                                         E. Lutz
304      Intergovernmental Fiscal Relations in               R. Bird (consultant)
Developing Countries
305      A Programming Approach to Fertilizer Sector         A. Choksi
Planning                                            A. Meeraus



PUB HG3881.5 .W57 W67 no.306
Dervis, Kemal.
The foreign exchange gap,
growth and industrial
strategy in l'urkey,
PUB HG3881.5 .W57 W67 no.306
Dervis, Kemal.
The foreign exchange gap,
growth and industrial
strategy in Turkey,