PROSPECTS FOR COMPUTERS IN THE SOVIET ECONOMY
Document Type:
Collection:
Document Number (FOIA) /ESDN (CREST):
CIA-RDP79-00927A005600030002-8
Release Decision:
RIFPUB
Original Classification:
S
Document Page Count:
11
Document Creation Date:
December 19, 2016
Sequence Number:
2
Case Number:
Content Type:
SUMMARY
File:
Attachment | Size |
---|---|
![]() | 810.08 KB |
Body:
Approved For Release 2006/05/01 : CIA-RDP79-00927A0t
DIRECTORATE OF
INTELLIGENCE
Secret
No Foreign Dissem
WEEKLY SUMMARY
Special Report
Prospects for Computers in the Soviet Economy
Secret
6 January 1967
No. 0271/67A
Approved For Release 2006/05/01 : CIA-RDP79-00927AO05600030002-8
Approved For Rq se 2006/05/01 : CIA-RDP79-00927A0005600030002-8
Secret
ci.Il Beport:: are supplements to the Current Intelligence: Week-
is.sr_Iecl [w the Office of Current Intelligence. The Special Re-
is are published separately to ocrmit more comprehensive treat-
>i it ' nrreiect. The are prepared by the Office of Current
-ilrgencc tie' Office of Research and Reports. or the Directorate
.e and Technologv. Special Reports are coordinated as
late artkong the Directorates of CIA but, except for the
ial suOstaliuve exchange with other agencies at the working
have not been coordinated outside CIA unless specifically in-
RNING
s document contains information affecting the national defense
Unit eci states, within the ineirnin4y of Title 18, sections 793
its contents to or receiilt by an unauthorized person is
crcuiucut MUST NOT BE RELEASED TO FOREIGN GOV-
vlh',NTS. If marked with Specific dissemination controls in
dance whir the provisions of DUD 1/7, the document must be
ci.f xf within the framework of the limitation so imposed.
Secret
Approved For Release 2006/05/01 : CIA-RDP79-00927A005600030002-8
Approved For Release 2006/05/01 : CIA-RDP79-0092705600030002-8
SECRET
PROSPECTS FOR COMPUTERS IN THE SOVIET ECONOMY
The USSR is far behind the West in the produc-
tion and use of computers for the general economy,
primarily because of its concentration on military
and scientific computers. The Soviets only recently
recognized the value of computerizing their anti-
quated and inefficient system of processing economic
data, and must now solve a multitude of technolog-
ical problems before much progress can be made. At
present, Soviet computers are not tailored closely
enough to the needs of individual consumers, and
obsolete equipment is used even after new models are
developed. Installation and service must be improved,
and qualified personnel trained. Moreover, a host
of support equipment, especially important to the
solution of economic problems, must be developed and
produced. The ultimate goal of many Soviet econo-
mists is a completely computerized economy, but
great controversy surrounds this proposal. A less
ambitious project, the development of a computerized
information system, has been planned for the next
decade.
Early Development
Soviet computers used to
solve economic problems are tech-
nologically behind those in the
US by about five years. The USSR
is even further behind in computer
production. The US installed its
first thousand computers about
eight years before the USSR did,
and it now has about ten times as
many in use. The US has automated
much of its economic data system
with the help of more than 15,000
computers, but the Soviets use
fewer than 500.
Research institutes under the
Academy of Sciences began computer
development in the USSR, primarily
to promote scientific research.
Few resources were allocated for
the design or production of com-
puters especially adapted to com-
mercial or economic use, and
little research was directed to-
ward the development of the aux-
iliary equipment necessary for
data processing. As a result,
the USSR now uses less than 15
percent of its computers for eco-
nomic work. By comparison, the
US employs nearly 60 percent of
its total inventory in the econ-
omy.
Computers are generally de-
signed either for scientific or
economic use. Scientific problems
typically involve a large number
of computations and a relatively
small amount of data. Data proc-
essing problems, on the other
hand, involve a small number of
SECRET
Page 1
Approved For Release 2006/05401 : CIA-RDP79-00927AO05600030002-8
Approved For Rese 2006/05/01 : CIA-RDP79-00927A600030002-8
COMPARISON of US and SOVIET PRODUCTION of
COMPUTERS and DATA-PROCESSING EQUIPMENT
2,500
2,000
500
1958 59 60 61 62 63 64 65 Est. 1966
Estimates of production do not include special-purpose military computers in either country.
Rubles have been converted to dollars at the rate of 1 ruble = US $0.15
Approved For Release 2006/05/01 : CIA-RDP79-00927A005600030002-8
Approved For Release 2006/05/01: CIA-RDP79-00927A 105600030002-8
t~ 14W
SECRET
simple calculations on a tremen-
dous volume of data, thus requiring
computers with large storage
capacity but less speed in compu-
tation.
Technological Problems
Typically, the Soviets em-
phasized serious production of a
minimum assortment of models
rather than the production of com-
puters tailored to the requirements
of individual consumers. Moveover,
even when a computer has become
obsolete and a replacement model
more responsive to user needs has
been developed, the obsolete model
is often used for a long time be-
cause plant managers are reluctant
to change models--the changeover
usually leads to reduced bonus
payments for both management and
work force. The delay in the ap-
pearance of improved models is
also partially the result of the
bureaucratic separation of the
users from the designers and pro-
ducers of the equipment.
Quick and responsive service
by American computer manufacturers
has been a major factor in the
widespread adoption of computer
systems in the US, but the Soviet
manufacturer generally neglects
his customers. The customer re-
ceives brief, inadequate instruc-
tions in programming, operating,
and maintaining his computer, and
the manufacturer takes no respon-
sibility for its maintenance.
Spare parts are often unobtain-
able and the user must employ
makeshift expedients to keep his
machine in operation. These may
change the operating character-
istics of the computer enough to
prevent sharing of programs
among users of the same models.
An All-Union Technical Planning
and Installation Association was
set up last March and it may
have alleviated some of these
servicing problems.
The USSR is short of quali-
fied computer specialists. Al-
though technical and economic
institutes have established an
increasing number of computer
courses, the supply of program-
mers, operators, repairmen, and
other specialists is far below
demand and will remain so for
many years.
In general, Soviet computers
are incompatible in that one pro-
gram cannot be used on different
computer models. Thus, programs
must be re-done for different
machines and personnel retrained
when transferred to different
installations or to new machines.
The Soviet "command economy,"
ironically, has been unable to
resolve this problem while US in-
dustry achieves tremendous econ-
omies through standardization.
Support Equipment and
Special Techniques
A dearth of support equipment
also has plagued the Soviet com-
puter industry and is a direct
consequence of the long delay in
recognizing the need for computers
in processing economic data. Sup-
port equipment typically includes
language compilers, program
libraries, and input-output equip-
ment.
SECRET
Page 3 SPECIAL REPORT 6 Jan 67
Approved For Release 2006/05/01 : CIA-RDP79-00927A005600030002-8
Approved For Rel se 2006/05/01 : CIA-RDP79-00927AO05600030002-8
S U ,C, R ET
Special computer languages
eliminate the complicated task of
transcribing instructions into
abstract number ccdes that the
computer can read. The Soviets
have done little work with these
special languages for general and
economic use and have concentrated
their limited effcrts on scientific
languages. They may try to shorten
the path of development followed
in the US by modifying one of
their algebraic languages to meet
the requirements of economic data
processing, but are unlikely to
succeed. To program a large cyber-
netic system directly into machine
language without the use of pro-
gramming languages would require
more trained people than could be
made available.
Packaged programs can be pre-
pared when certain computer opera-
tions are performed repeatedly
and these can then be placed in
a central repository for all to
use, a common practice in the US.
The USSR, however, has made little
effort to create such program
libraries. A central repository
for algorithms and computer pro-
grams was establis.aed in October
1965, but it is intended mainly
for scientific use. Nevertheless,
it should provide valuable experi-
ence eventually for similar ef-
forts in the economic area.
Efficient inp-,at-output equip-
mnent is particularly important to
the solution of economic problems.
Since these problems typically in-
volve a large volume of data, all
of which cannot be stored inside
the computer unit, the controlling
factor in efficiency becomes the
speed with which data can be se?-
lected and transferred from ex-
ternal storage to the computer
unit. The Soviet input-output
capability has greatly improved
in recent years, but it still
lags behind the development of
computers.
Input-output operations
utilize punch-cards, tapes, and
storage units. The USSR often
uses punch-cards as a direct in-
put for economic data. The US
practice is to transcribe the
cards to magnetic tape, a much
faster and more efficient system.
The Soviets use magnetic drums
for external storage; the more
efficient disk files, widely used
in the US, are just beginning to
be developed. Input-output equip-
ment has been improved recently,
but it still hinders the expansion
and effective use of computer
systems.
Accounting and Data Processing
The USSR could reap tremen-
dous economic gains from moderniz-
ing its accounting system. At
present, about 3 million peo-
ple are employed in accounting
work, most of which is done by
abacus and desk calculator with
the aid of some punch-card equip-
ment. Soviet academician Glushkov
dramatized the inefficiency of
this system when he stated that
by 1980, using the present tech-
niques, the volume of paperwork
will have increased by 3,600 per-
cent and the entire population
of the USSR will be needed to
compile and analyze economic data.
The need for fast, accurate proc-
essing of accounting information
SECRET
Approved For Release 2006/05/01 : CIA-RDP79-00927AO05600030002-8
Approved For Release 2006/05/01: CIA-RDP79-009?7A005600030002-8
SECRET
will become even more critical as
the economy completes the trans-
fer to the new rules for economic
reform in industry begun in 1965.
A government decree of Novem-
ber 1964 provided a better atmos-
phere for the introduction of
computer techniques by ordering
the centralization of accounting
functions within each enterprise.
Some experiments were conducted
by the regional economic councils
several years ago in centralizing
the accounting work of a number
of separate firms. The Central
Statistical Administration (CSA)
originally planned to build 630
machine-calculating stations to
handle the centralized accounting
work of groups of enterprises by
the end of 1965, but progress on
these stations has not been re-
ported recently.
Statistical Collection and
Compilation
The Central Statistical Ad-
ministration, consisting of 800
statistical centers and 2,400
smaller bureaus, is the basic sta-
tistical collection and processing
center for the Soviet Government.
The CSA records the fulfillment
of state plans, computes national
income accounts, conducts cen-
suses, reports on economic sub-
jects, and publishes statistical
bulletins and handbooks. In addi-
tion, the CSA is required to sup-
ply statistics to Gosplan (State
Planning Committee) and many other
agencies. However, some other
organizations--the Ministry of
Finance, Gosbank (State Bank), and
the individual production minis-
tries--also collect and compile
statistics, sometimes duplicat-
ing and confusing the CSA's work.
This unwieldy system,
largely hand-operated, possesses
an enormous potential for com-
puterization. Experimental
computer systems have already
been installed in Moscow and
Latvia. Last March, the USSR
announced plans for a nation-
wide computer network to be
started during the current five-
year plan (1966-70). This net-
work, which is to collect and
process economic information
for the whole country, will not
be completed for many years,
well after 1970.
Planning at Enterprise and
Branch Level
During the last few years
the USSR has developed various
methods for applying computers
to the decision-making process
at enterprise and branch level.
Mathematical methods, especially
linear programming,* have been
used to solve a wide range of
problems in agriculture, indus-
try, transportation, and con-
struction.
*A method for calculating the best plan for achieving
stated objectives--such as maximum profit or output--
in cases where resources are limited.
SECRET
Page 5 SPECIAL REPORT 6 Jan 67
Approved For Release 2006/05/01 : CIA-RDP79-00927A005600030002-8
Approved For Rel a 2006/05/01 : CIA-RDP79-00927AO05600030002-8
SL~;CRE"l'
The Soviets have used linear
programming in agriculture to help
determine crop distribution and
rotation, livestock production,
and allocation of labor and mate-
rials. Industrial computer tech-
niques are also being developed,
mainly in the fields of metallurgy,
chemistry, and electric power gen-
eration. However, the returns
from practical applications have
been small since the generally low
level of Soviet plant technology
makes the installation of auto-
matic control equipment unprofit-
able.
Computerized control of trans-
portation in the US3R is still in
the pioneering stage. A nation-
wide electronic computer system
for rail transport Ls planned, and
the first stage in :his system--
a computer center for the Moscow
;Railroad Administra :ion---has been
built. A number of problems re-
main to be solved, however, if
computers are to be used profitably
by the transport industry. For ex-
ample, transmission of data by
telegraph is too slow and experi-
ments with new data transmission
equipment are proceeding unsatis-
factorily. Also, the whole con-
cept of using linear programming
in the transport industry, as op-
posed to nonlinear types of pro-
gramming, has been questioned.
The USSR began using mathe-
matical methods and computers in
the construction industry in 1963.
.1:n one technique, developed in the
US and called the Critical Path
Method (CPM), a pro_ect is broken
down into all the separate tasks
involved and the time required to
complete each is calculated. The
tasks on the "critical path"--
those which would delay the proj-
ect if not completed on time--are
supervised more closely. The suc-
cess of the first 18 Soviet proj-
ects converted to CPM techniques
in 1964 led to an expansion to al-
most 300 projects in 1965. For
1966, 365 new projects, accounting
for. 12 percent of the total vol-
ume of industrial construction
were scheduled. The Soviets are
studying methods to implement CPM
on a larger scale, and plan to
build a computer network for the
construction industry during 1966-
70.
Macro-Planning: Over-all
Economic Control
Economic planning in the USSR
is a highly complex, imprecise,
unwieldy, and time-consuming proc-
ess. Gosplan, the primary agency
charged with the planning respon-
sibility, computes a detailed na-
tional balance sheet for about
1,000 major products and a less
precise plan for about 16,000 ad-
ditional ones. The planning proc-
ess is a year-round operation and
is done almost completely in quan-
titative terms--prices play only
a minimal role. Working from the
statistics collected by the CSA,
Gosplan negotiates "material bal-
ances" with the production minis-
tries culminating in the distri-
bution of output and
to the enterprises.
omy of the USSR has
and complexity, the
ance method has
inefficient way
product mix.
SECRET
supply plans
As the econ-
grown in size
material bal-
become an extremely
to determine the
Approved For Release 2006/05/01 : CIA-RDP79-00927AO05600030002-8
Approved For Release 2006/05/01 : CIA-RDP79-00927405600030002-8
IWAV,
SECRET
It would be difficult to use
computers in the present planning
process. They would have to be
fed exact instructions for plan
formulation and the present system
is too imprecise for this to be
done. Although a study of ways
to computerize the system might
force managers to adopt more pre-
cise procedures, Gosplan has not
yet taken any major steps toward
putting even routine tasks on com-
puters and such work is still only
in the experimental stage. Linear
programming methods and input-out-
put analysis have been used ex-
perimentally in national planning
with some success, but Gosplan of-
ficials have reported that this
type of work is being "poorly in-
troduced into the practice of
planning."
The ultimate goal, or dream,
of many Soviet economists is an
all-inclusive cybernetic system,
consisting of a complicated mathe-
matical model and a large network
of computers to control every im-
portant aspect of the economy.
The system would receive data from
all levels of the economy, synthe-
size it, and devise a number of
alternative plans. The regime
would choose one plan that would
then be disseminated, monitored,
and corrected by the system.
Controversy has long sur-
rounded this idea. Probably the
most fundamental dispute concerns
the amount of central control that
this type of cybernetic system.
would generate. The recent eco-
nomic reforms, although decreas-
ing the amount of control held by
the central government agencies,
have retained the principle of
central planning. Soviet econo-
mists who fear that decentraliza-
tion will creep into the planning
process cling to the proposal of
a large cybernetic system as a
means of effectively retaining
central control. They argue that
economic plans can best be made
from the top, especially if a
computer network rapidly proc-
esses masses of data and reduces
the numerous complexities to a
manageable size.
Opponents of the large cyber-
netic system argue that it is both
impossible and undesirable to di-
rect the entire economy by com-
puters. They favor a small com-
puter system to determine the
general plan and a price and in-
centive system to guide enterprise
managers in making decisions con-
sistent with the planned goals.
In addition, they stress that a
computer-directed economy is a
long-range solution--15 to 20
years away at the earliest--
whereas a smaller system could be
achieved in the near future.
The problems of developing
a comprehensive cybernetic system
are staggering. The construction
of a model capable of controlling
the entire Soviet economy would
require mathematical tools that
have not been developed even in
theory. Although two models are
in the first stages of develop-
ment, it will be many years before
either can be made operational.
Both seek to maximize "satisfac-
tion of social requirements," a
nebulous phrase that includes
personal consumption, defense,
investment, and science. The
model will not utilize a single
SECRET
Approved For Release 2006/05/01 : CIA-RDP79-00927AO05600030002-8
Approved For Rise 2006/05/01 : CIA-RDP79-00927
SECRET
criterion for weighing the end
products in the ma;cimization prob-
:Lem since this selection will be
Left to the "competent organs,"
i.e., the party leaders.
The proposed computer net-
work, to be divided into three
Levels, would requ:.re a tremendous
number of computers, auxiliary
equipment, and tra:.ned personnel.
The lowest level would consist of
many small compute:: centers at
individual firms o:- groups of en-
terprises. The second level would
include a set of 50 to 80 regional
computer centers, each connected
with its own local centers and
with the other req:-onal centers.
The third level would be a huge
center in Moscow. A Soviet com-
puter expert has estimated that
this network would require at
least 10 quadrillion operations
a year. Using this figure and
the average speed of the new se-
ries of Soviet computers, it is
estimated that 14 , 000 computers
would be required---compared with
the 500 now used in economic work.
A greatly strengthened communica-
tions network would also be re-
quired for this proposed computer
system.
Planned Computer Network
The size and complexity of
tshe obstacles to the development
of a comprehensive cybernetic sys-
tem have probably 'orced the So-
viets to postpone serious efforts
toward this end and to concentrate
on a less ambitious goal--the de-
velopment of a computerized infor-
mation system, which is provided
for in a decree of March 1966.
This system, to be started during
the current five-year plan, clearly
is not intended to be a compre-
hensive cybernetic operation to
direct all economic planning and
it will not automate Gosplan's
planning system.
The March decree gave the
authority to operate the proposed
network to the Central Statistical
Administration, and it will be
formed through an expansion and
modernization of the existing CSA
network. In addition, sectoral
and departmental systems of plan-
ning, accounting, control, and
data processing, are to be created
as necessary, and are to be con-
nected with the state network.
Gosplan is charged with the ini-
tial planning of the computer cen-
ters and with projecting the pro-
duction of computer equipment.
The requirements of the pro-
posed network are not nearly as
demanding in quantity and sophisti-
cation as those needed for com-
plete cybernetic control of the
economy. Even so, the Soviets
admit that their computer industry
cannot provide all of the equip-
ment needed for the proposed net-
work in 1966-70. Nevertheless,
they propose to make a good start
during this period. Moreover, if
they eventually decide to proceed
with a complex cybernetic system,
it is likely that most of the sys-
tem's developmental work would
parallel that for the computer
network. In all probability, they
will work on the more localized
and immediately profitable com-
puter applications for the next
decade and then examine further
the case for a cybernetic system.
(CONFIDENTIAL)
(Prepared by the Office of Research and Reports)
SEX,CRET
Approved For Release 2006/05/01 : CIA-RDP79-00927AO05600030002-8
Approved For Release 2006/05/01 : CIA-RDP79-00927AO05600030002-8
Secret
Secret
Approved For Release 2006/05/01 : CIA-RDP79-00927AO05600030002-8