REVISITING SOVIET ECONOMIC PERFORMANCE UNDER GLASNOST: IMPLICATIONS FOR CIA ESTIMATES

• Document Type: CREST
• Collection: General CIA Records
• Document Number (FOIA) /ESDN (CREST): CIA-RDP89T01451R000600690001-4
• Release Decision: RIPPUB
• Original Classification: K
• Document Page Count: 32
• Document Creation Date: December 27, 2016
• Document Release Date: January 10, 2013
• Sequence Number: 1
• Publication Date: September 1, 1988
• Content Type: REPORT

Declassified in Part- Sanitized Copy Approved for Release 2013/01/10: CIA-RDP89T01451R000600690001-4
Declassified in Part- Sanitized Copy Approved for Release 2013/01/10: CIA-RDP89T01451R000600690001-4
STAT
Declassified in Part -Sanitized Copy Approved for Release 2013/01/10: CIA-RDP89T01451R000600690001-4
Intelligence
Revisiting Soviet Economic
Performance Under Glasnost:
Implications for CIA Estimates
0079996
S0V*10068*88
SOVA*16574188
1512-1512
FILE COPY/SOURCED COPY (A-2/1-1:0SWR1
CPAS/MC/CONTROL BRANCH
ROOM 7G07
HQS (NO ADHOC REQUIREMENTS)
SOV 88-10068
September 1988
Declassified in Part- Sanitized Copy Approved for Release 2013/01/10: CIA-RDP89T01451R000600690001-4
Declassified in Part- Sanitized Copy Approved for Release 2013/01/10: CIA-RDP89T01451R000600690001-4
This publication is prepared for the use of US Government
officials, and the format, coverage, and content are designed to
meet their specific requirements. US Government officials may
obtain additional copies of this document directly or through
liaison channels from the Central Intelligence Agency.
Requesters outside the US Government may obtain subscriptions to
CIA publications similar to this one by addressing inquiries to:
Document Expediting (DOCEX) Project
Exchange and Gift Division
Library of Congress
Washington, D.C. 20540
or: National Technical Information Service
5285 Port Royal Road
Springfield, VA 22161
Requesters outside the US Government not interested in subscription
service may purchase specific publications either in paper copy or
microform from:
Photoduplication Service
Library of Congress
Washington, D.C. 20540
or: National Technical Information Service
5285 Port Royal Road
Springfield, VA 22161
(To expedite service call the
NTIS Order Desk [phone])
Comments and queries on this paper may be directed to the DOCEX
Project at the above address or by phone [phone]), or the
NTIS Office of Customer Services at the above address or by phone
(703-487-4660). Publications are not available to the public from the
Central Intelligence Agency.
Declassified in Part- Sanitized Copy Approved for Release 2013/01/10: CIA-RDP89T01451R000600690001-4
Declassified in Part- Sanitized Copy Approved for Release 2013/01/10: CIA-RDP89T01451R000600690001-4
Intelligence
Reverse Blank
Revisiting Soviet Economic
Performance Under Glasnost:
Implications for CIA Estimates
SOV 88-10068
September 1988
Declassified in Part- Sanitized Copy Approved for Release 2013/01/10: CIA-RDP89T01451R000600690001-4
Declassified in Part- Sanitized Copy Approved for Release 2013/01/10: CIA-RDP89T01451R000600690001-4
Scope Note
Reverse Blank
Revisiting Soviet Economic
Performance Under Glasnost:
Implications for CIA Estimates
This paper assesses the implications of recent glasnost-inspired critiques of
the USSR's official economic statistics for CIA's estimates of Soviet
economic performance. The CIA estimates, although predicated on the
belief that Moscow's macroeconomic measures are unreliable, are based on
a variety of official Soviet data. In particular, the paper focuses on what
the recent criticisms have to tell us about the accuracy of CIA's estimates
of the growth and structure of Soviet gross national product (GNP).
Although determining the size of the USSR's GNP relative to those of the
United States and other countries is also important to an assessment of the
USSR's economic performance, recent Soviet critics of the official statis-
tics have had little to say on this matter. As a result, this subject is not dis-
cussed in this paper but is deferred to a subsequent study.
This paper draws heavily on the results of an unclassified conference on the
impact of Gorbachev's policies on Soviet economic statistics, sponsored by
the Office of Soviet Analysis in December 1987. STAT
STAT
111
SOV 88-10068
September 1988
Declassified in Part- Sanitized Copy Approved for Release 2013/01/10: CIA-RDP89T01451R000600690001-4
Declassified in Part- Sanitized Copy Approved for Release 2013/01/10: CIA-RDP89T01451R000600690001-4
Summary
Revisiting Soviet Economic
Performance Under Glasnost:
Implications for CIA Estimates
Mikhail Gorbachev's glasnost policy has encouraged Soviet critics of
official economic statistics to embark on what amounts to a broad-based
depreciation of the USSR's past economic performance. Their writings
have generally confirmed what CIA and most Western Sovietologists have
long believed about Moscow's macroeconomic data?that they are ex-
pressed in prices that fail to reflect the relative costs of producing goods
and services and that they greatly inflate real growth. The alternative
estimates of overall economic growth that these critics present seem
unrealistically low, in our view, but some of the particulars of their
criticisms highlight the need for improvements in important parts of CIA's
alternative measures of Soviet economic performance.
Soviet Unofficial and CIA Views
CIA and the Soviet critics of the USSR's published statistics agree in the
sense that both reject Moscow's claims of economic growth as inflated and
its data on the allocation of resources as misleading and flawed. CIA's
estimates have consistently suggested that the structure of Soviet GNP has
been substantially different from that shown by data in official prices and
that the growth of GNP has been substantially lower than official data
imply (see figure 1). CIA's estimates are much closer to those of the critics
than to those of the state statistical authorities. Nonetheless, some of these
critics have advanced alternative estimates that suggest Soviet growth has
at times been appreciably lower than estimated by CIA.
Assessment of the Charges
Soviet critics of the official statistics have identified two sources of upward
bias in the disaggregated Soviet data on which CIA's indexes of GNP
growth are based: overreporting of the physical quantities of goods
produced and hidden inflation in the statistics reported in allegedly
constant ruble terms. We rely overwhelmingly on statistics in physical
units because we agree that Soviet value statistics reported in supposedly
constant prices reflect a considerable degree of inflation. We believe the
physical production statistics are reliable. Detailed Western studies of the
quality of data on physical production suggest that the degree of overre-
porting is small and has not increased over time. Also, critics of the official
statistics, while complaining of such overreporting, have themselves used
physical measures in developing their alternative estimates.
Declassified in Part- Sanitized Copy Approved for Release 2013/01/10: CIA-RDP89T01451R000600690001-4
Declassified in Part- Sanitized Copy Approved for Release 2013/01/10: CIA-RDP89T01451R000600690001-4
Figure 1
A Comparison of Offical Soviet Statistics on
and CIA Estimates of the USSR's Economic Growth
(1950=100)
Total Output
1,200 I
1
1
1.
Soviet
( National
ipnrcoodmuceed)
CIA
(GNP)
1,000
?,
ll
800
? ?
?,
IN
600
.--
r400
1.01111111111
200
!
0 1950 55 60
Industrial Production
65
70
75
80
85
86
87
1,600
i 1
1
Soviet
1,400
1----0
,
1,200
1
Eli
1,000
800CIA
r?I
n ?1F--'
?I
400
1-7 ?1
1
200
I
!
1950
55
60
65
70
75
80
8
86
87
Machine-Building Output
6,000
il
I
1
_
Soviet
CIA
5,000
?
1
?
ii
4,000
I
3,000
2,000
1,000
T ir---'
7
i
il
0 1950
55
60
65
70
75
80
85
Official Soviet figures are from Narodnoye khozyaystvo SSSR, various years.
CIA estimates are in 1982 factor-cost prices.
86
87
318818 9-88
vi
Declassified in Part- Sanitized Copy Approved for Release 2013/01/10: CIA-RDP89T01451R000600690001-4
Declassified in Part- Sanitized Copy Approved for Release 2013/01/10: CIA-RDP89T01451R000600690001-4
Although we have more reason to be concerned about official Soviet value
statistics, even in this case the biased data affect only portions of our
estimates:
? In the estimates of growth on the sector-of-origin side, the share of ruble
value data in our sample is fairly small (11 percent). In our judgment, any
overestimation resulting from their use is likely to be offset by underesti-
mation resulting from our greater reliance on physical measures that fail
to reflect improvements in product quality.
? On the end-use side, because our indexes of the growth of investment and
consumption rely more heavily on official ruble value series and current-
price statistics deflated with official price indexes, the inflationary biases
in the official data may have a larger impact.
In some rough sensitivity testing, when allowance is made for likely
downward as well as upward biases in our indexes, our estimates of the av-
erage annual growth of GNP during the years 1951-87 remain virtually
unchanged. Yet, because the growth of some key end uses could have been
different from what we estimate, the same sensitivity testing suggests that
by 1987 the share of investment in GNP could have been about one-sixth
less than we estimate and the share of consumer goods in GNP could have
been smaller by about one-tenth. Because defense expenditures are esti-
mated independently of other end uses by a method requiring little reliance
on officially published data, our estimates of their level and their share of
GNP remain the same.
On balance, the glasnost-inspired criticisms of official Soviet statistics,
while reassuring with regard to CIA's rejection of the official estimates,
add weight to earlier indications that the CIA estimates of some compo-
nents of GNP may require revision and refinement. We are at present
focusing on the machine-building and metalworking sector of origin and
the new fixed investment end-use category. Still, additional data, not yet
available in official reporting despite Gorbachev's professions of openness,
are likely to be required before the necessary changes can be made with
confidence. The estimates provided to date by the critics are of little help in
this regard. Some of the critics are establishment figures?for example,
Gorbachev's adviser Abel Aganbegyan?who might have access to the
great volume of data required to produce reliable alternatives to the official
measures of economic performance. Yet all the critics have displayed the
vii
Declassified in Part- Sanitized Copy Approved for Release 2013/01/10: CIA-RDP89T01451R000600690001-4
Declassified in Part- Sanitized Copy Approved for Release 2013/01/10: CIA-RDP89T01451R000600690001-4
same inattention to documentation and vagueness on methodology as the
statistical agencies they fault. The published criticisms, moreover, have a
pronounced political coloration. They serve to blacken the record of
Gorbachev's predecessors as well as to set that record straight.
Continuing Need for Alternative Estimates
The need for independent estimates of Soviet economic performance like
those of the CIA is likely to increase. While Soviet critics are debunking
official claims of past growth, the state statistical agency is not backing
away from these claims and is further distorting the recent record. In the
absence of estimates derived without reliance on Soviet macroeconomic
data, US policymakers would have to accept the flawed official mea-
sures?at the risk of seriously overstating the USSR's ability to raise living
standards and provide for future growth and understating its commitment
of resources to defense.
viii
Declassified in Part - Sanitized Copy Approved for Release 2013/01/10: CIA-RDP89T01451R000600690001-4
! Declassified in Part- Sanitized Copy Approved for Release 2013/01/10: CIA-RDP89T01451R000600690001-4
Contents
Page
Scope Note
iii
Summary
V
Recent Soviet Critiques of Official Statistics
1
Antecedents and Common Themes
1
CIA Estimates?An Alternative to Soviet Measures
4
Estimating GNP in 1982, the Base Year
4
Estimating the Growth of GNP
5
CIA Estimates Versus Official Measures
7
CIA Estimates and the "New" Statistics
10
Unclear Basis of Alternative Estimates
10
Inflation and CIA Estimates of Growth
11
Use of Quantity Data
12
Estimating the Allocation of GNP
14
Some Sensitivity Testing
16
Growth of Soviet GNP
16
Distribution of GNP by End Use
17
Conclusions and Prospects
18
Reverse Blank
ix
Declassified in Part- Sanitized Copy Approved for Release 2013/01/10: CIA-RDP89T01451R000600690001-4
! Declassified in Part- Sanitized Copy Approved for Release 2013/01/10: CIA-RDP89T01451R000600690001-4
Revisiting Soviet Economic
Performance Under Glasnost:
Implications for CIA Estimates
Recent Soviet Critiques of Official Statistics
The Gorbachev years have been a time of criticism
and reassessment of the USSR's economic statistics
and much of its "official" historical record. The
General Secretary's glasnost has led to the airing of
often sensational charges of inaccuracy and falsifica-
tion on the part of the state statistical agencies and of
sharp complaints about the inadequacy of official
measures of economic performance. Gorbachev has
encouraged such criticism by his frequent references
to the late Brezhnev years as a "period of stagnation"
while official statistics depict the period as a time of
respectable growth.
Other Soviet officials and economists have made more
explicit charges. For example:
? Gorbachev's economic adviser Abel Aganbegyan
asserted in a recent book that Soviet economic
growth since 1965 has been substantially lower than
officially claimed and that, during the period 1981-
85, economic growth was close to zero (see figure 2).
? In an article in the literary journal Novyy mir,
economist Grigoriy Khanin and journalist Vasiliy
Selyunin charged that Soviet national income did
not increase 90-fold during 1928-85, as official
statistics show, but only sixfold or sevenfold.'
? Many economists and journalists have denounced
the pricing system for failing to reflect the relative
resource costs of producing goods and services and
Soviet indexes of retail prices for understating the
actual increase in the-cost of living.
? Dramatic accounts have been published in the press
about the falsification of cotton production figures
in Soviet Central Asia during the Brezhnev years,
and scores of officials have been prosecuted for
involvement in this and similar frauds.
' This and other notes are appended.
1
For many Soviet readers these revelations have doubt-
less amounted to a hard-to-swallow rewriting of the
USSR's economic history. For Western analysts, who
long have discounted the reliability and utility of
official Soviet claims of economic growth and have
constructed independent estimates, the glasnost-
inspired charges have been less surprising. Nonethe-
less, they do raise questions about the adequacy of the
corrections for such inaccuracies and distortions in
Western estimates. This paper assesses the implica-
tions of recent Soviet criticisms of Moscow's official
macroeconomic measures for the most widely used
Western substitute for them: CIA's independent esti-
mates of the growth and structure of Soviet GNP.
Antecedents and Common Themes
CIA's decision to compile independent estimates of
Soviet economic performance reflects the fact that
Western observers have for decades rejected the
statistics on economic growth reported by the USSR's
statistical authorities.' Western experts concluded
many years ago that there was an inherent bias in the
way Soviet statistics were collected and reported and
that there were serious gaps in the published statistics.
Published Soviet complaints about official statistics
came later. Present-day denunciations of the official
statistics, with high-level support, were preceded by
less dramatic charges. In the 1960s, for example,
critiques of the officially reported growth rates for
Soviet national income by Al'bert Vaynshteyn were
published in the USSR and attracted considerable
attention in the West.' Similarly, the validity of the
published price indexes for the machine-building sec-
tor of industry was questioned by Soviet scholars in
the 1960s and 1970s.4 In some instances the economist
critics offered alternative price indexes that showed
rates of inflation substantially greater than officially
acknowledged. Soviet statistics on the growth of
machine-building output subsequently were criticized
as inflated in articles by V. K. Fal'tsman, A. Kornev,
K. K. Val'tukh, and B. L. Lavrovskiy.5
Declassified in Part- Sanitized Copy Approved for Release 2013/01/10: CIA-RDP89T01451R000600690001-4
Declassified in Part- Sanitized Copy Approved for Release 2013/01/10: CIA-RDP89T01451R000600690001-4
Figure 2
Growth of Soviet National Income Utilized,
by Five-Year Periods, 1966-2000
Percent
50
41

u
0
0
0
1
1
Projection
Official Soviet
40
,
?
21
0
0
0
_g.
30 ?
2
0
0
0
22
: data
n Aganbegyan's
?.--
2-0
10:5
0
0
_a
calculations
10
0
0
0
0
1966-70
1971 75
1976 80
Source: Abel Aganbegyan, The Economic Challenge of Perestroika
(Bloomington: Indiana University Press, 1988)
1981-85
1986 90
1991-2000
Average annual, in percent
8
i
Published
rates
?
Aganbegyan
1
rates a
4
_
-
2
1966-70
1971-75
a Adjusted, according to Aganbegyan, using a "more realistic"
index of price change.
1976-80
1981 85
318819 9-88
2
's
Declassified in Part- Sanitized Copy Approved for Release 2013/01/10: CIA-RDP89T01451R000600690001-4
! Declassified in Part- Sanitized Copy Approved for Release 2013/01/10: CIA-RDP89T01451R000600690001-4
These and other articles from Soviet scholarly jour-
nals and monographs, when combined with the recent
discussions of official statistics in the popular press,
represent a sizable body of criticism. Although the
particulars of the criticism vary greatly, certain com-
plaints have occurred frequently. For example, there
has been widespread agreement that state-controlled
official prices (which we call established prices) distort
the true structure of production and present a serious-
ly flawed picture of how resources are allocated to end
uses. On the sector-of-origin side, the most common
charge is that the prices of agricultural goods and
consumer services are set too low, thus understating
their share of annual production and causing the
manufacturing and processing industries to appear
larger than deserved. More generally, the real re-
source contributions of various sectors are misstated
because, under Soviet pricing, profits do not ade-
quately reflect capital charges and economic rent. On
the end-use side, the discounted, often below-cost
prices at which some goods are sold and the inclusion
of large turnover (sales) taxes in the prices of others
prevent a true picture of resource flows. In neither
case is falsification of the statistics at fault; rather,
the distortions are systemic.
With regard to the growth of Soviet output, both
Soviet and Western critics of the official statistics
have charged that the supposedly constant or, in
Soviet parlance, "comparable" prices in which pro-
duction is reported actually incorporate substantial
"disguised" or "hidden" inflation and that, as a
result, real growth is much less than claimed. Most
often, this inflation is said to take the form of the
overpricing of new products that differ little if at all
from their predecessors but are assigned substantially
higher prices. The abuse of new product pricing,
which helps to meet plan targets, is said to be
especially common in industries such as machine
building, where claims of product innovation can be
made much more easily than in agriculture or the
extractive industries. Directly related to this charge is
the complaint that the official indexes of prices fail to
reflect the inflation that actually occurs.
The official index of retail prices has come in for
particular criticism under glasnost. According to the
index, consumer prices rose at an average rate of only
3
0.3 percent a year during 1961-87. Soviet economist
Oleg Bogomolov, however, has claimed that over the
same general period the cost of the shopping basket of
goods bought by the average city dweller has in-
creased at an average annual rate of 1.5 to 2.5
percent.' Other Soviet authors have cited even higher
estimates of the growth in retail prices.
Inflation in the official investment statistics has not
been discussed as much in the popular press as
inflation in the consumer sector, but some Soviet
economists say this is a serious problem. In a recent
survey, Vladimir Kontorovich and Boris Rumer sum-
marized the views of those scholars and journalists
who allege this to be the case as suggesting "a
consensus range of 2 to 3 percent inflation in invest-
ment in 1971-75 and 2 to 4 percent in 1976-80. The
inflation rate increased from the 1960s to the 1970s?
but this trend of inflation apparently was not contin-
ued in the early 1980s." '
In contrast to the widespread complaints about the
reliability of measures of Soviet output expressed in
comparable ruble terms, official indicators expressed
in terms of physical quantities?such as tons and
kilowatt-hours?have come in for much less criticism.
Nonetheless, the reported falsification of statistics on
cotton production in Central Asia and other published
accounts of overreporting have led some critics to
question the reliability of physical indicators of
growth as well as of value measures. This subject is
discussed in the section "Use of Quantity Data."
Another charge made by critics of the official statis-
tics is that, whether expressed in terms of physical
quantities or ruble values, statistics on the production
of goods and services are inadequate indicators of the
satisfaction of demand, because the quality of the
goods produced is so poor and the assortment so ill
suited to consumer needs. It has, for example, become
fashionable for both Soviet leaders and economic
specialists to point out that, while the Soviet Union is
among the world's leading producers of shoes, con-
sumers are loath to buy domestically manufactured
Declassified in Part- Sanitized Copy Approved for Release 2013/01/10: CIA-RDP89T01451R000600690001-4
Declassified in Part- Sanitized Copy Approved for Release 2013/01/10: CIA-RDP89T01451R000600690001-4
Soviet Versus Western Measures of Economic Growth
The Soviets usually measure their economic perfor-
mance using a Marxian concept that we call net
material product (NMP). This measure differs from
Western GNP in that it excludes depreciation and
services that do not contribute directly to material
output (for example, passenger transportation, educa-
tion, health, and government administration). Pub-
lished NMP growth rates?reported by the Soviets as
both national income produced and national income
used?are further biased because inadequate adjust-
ment is made for inflation. The NMP and GNP
measures also differ considerably in some years
because the weight used for agriculture in Soviet
accounting differs from that used in Western esti-
mates. For years such as 1986, when agriculture's
performance was unusually good, the two measures
are closer together. Both GNP and NMP measures,
however, show that economic growth has slowed in
the USSR since the mid-1970s.
footwear, and Soviet shoes have no export market.'
Similar charges have been leveled about data on the
production of investment goods, and the general con-
clusion is that statistical indicators of growth often
reflect, in practice, a growing waste of resources. This
topic is discussed in the section "CIA Estimates and
the 'New' Statistics."
CIA Estimates?An Alternative to Soviet Measures
In an effort to correct the distortions in official Soviet
macroeconomic data and to measure Soviet economic
growth in terms familiar to a Western audience, CIA
has constructed independent estimates of the USSR's
GNP since the 1950s (see inset).9 Specifically, the
CIA attempts to replicate as closely as possible the
concepts used by the United States and other Western
countries.rn Estimates of the size and growth of Soviet
GNP are based on a wide array of data and informa-
tion culled from official Soviet statistical handbooks,
the press, and specialized economic publications. The
CIA procedure is to develop a detailed set of GNP
accounts for a base year (currently 1982) and then to
move the components over time with indexes that
reflect changes in real output.
Estimating GNP in 1982, the Base Year
To estimate GNP in 1982 in established prices, we
first construct four basic national income accounts?
two showing the incomes of the public and household
sectors and two showing the expenditures of those
sectors on final goods and services in that year (see
figure 3). GNP equals the sum of total incomes (or
expenditures) of the public and household sectors.
Total GNP is then disaggregated to obtain value
added in 30 sectors of origin (such as branches of
industry, agriculture, and transportation) and values
for 26 end uses (such as various categories of con-
sumption, investment, and government services).
The values obtained at this point in the estimating
process, if correctly calculated, reflect the actual
prices received by producers and paid by consumers in
1982. Because of the peculiarities of Soviet price
setting, however, these values do not measure the
relative real resource costs of producing the various
goods and services. Serious distortions result from
heavy turnover taxes on many consumer goods and
large price subsidies for other goods and some ser-
vices; other major distortions arise because the profits
reflected in current product prices stem from profit
margins that are arbitrarily set and do not measure
actual returns on capital. To correct for these major
distortions, we modify the sectoral values estimated in
actual prices by subtracting turnover taxes, adding
subsidies, and replacing profits with values that re-
flect a uniform rate of return on capital (currently
taken to be 12 percent). These adjusted ("factor-cost")
valuations of outputs of producing sectors and of end
uses of GNP provide the base-year weights that are
used in estimating changes in GNP and its major
components over time."
4
Declassified in Part- Sanitized Copy Approved for Release 2013/01/10: CIA-RDP89T01451R000600690001-4
! Declassified in Part- Sanitized Copy Approved for Release 2013/01/10: CIA-RDP89T01451R000600690001-4
Figure 3
The CIA's Estimation of Soviet GNP
Base-Year Estimates Growth of Components
Household
income
Household
outlays
Public sector
incomes
Public sector
outlays
GNP in estab ished prices
Sector of End use
origin
PIPASS,
GNP at factor cost
Sector of
origin
End use
Sample of products
Base-year x Given-year
price quantity
Equals
Constant-price value
Growth of components
44?Wftii..:;z;
Growth of GNP
(including subtotals)
Growth of GNP
(factor cost)
FTZMPflAt*-14W1)-:
Meanwhile, the role of the "second economy" contin-
ues to bedevil both Western and Soviet measures of
national output (see inset, page 6). CIA includes
second-economy activity in the base-1982 value of
GNP but cannot track the changes in second-economy
production over time.
Estimating the Growth of GNP
To measure the growth of GNP, we move the factor-
cost values for each producing sector over time with
quantity indexes designed to reflect real changes in
5
3188209-88
output. The growth of GNP as a whole is obtained
from the sums of the values for all sectors in each
year.
Sector-of-Origin Indexes. The indexes of growth in
the individual producing sectors necessarily are based
for the most part on data for samples of products or
activities. For some sectors, such as agriculture and
transportation, the sample coverage is complete or
nearly so. For other sectors, such as the chemical and
Declassified in Part - Sanitized Copy Approved for Release 2013/01/10: CIA-RDP89T01451R000600690001-4
Declassified in Part- Sanitized Copy Approved for Release 2013/01/10: CIA-RDP89T01451R000600690001-4
The Second Economy and GNP
According to a mass of evidence, a thriving "second
economy" exists in the Soviet Union alongside the
formal economy. Gregory Grossman defines the sec-
ond economy as comprising "all production and
exchange activity that fulfills at least one of the two
following tests: (a) being directly for private gain or
(b) being in some significant respect in knowing
contravention of existing law." a There are fairly
reliable data for the former, but not for the latter. In
estimating the size and growth of Soviet GNP, the
CIA attempts to include the full range of economic
activities that are measured in GNP statistics in the
West. This standard calls for the inclusion of all
legal private production and also of some activities
that are legal in the West but illegal or semilegal in
the USSR.
Our estimates of Soviet GNP in 1982 include second-
economy activities that account for 11 percent of
consumption and 6 percent of GNP. We count all of
the legal private production that official Soviet data
permit us to identify?production on private farming
plots and private investment. We also include esti-
mates of a variety of privately produced services,
without regard to their legality. These estimates,
which probably are too low, are based on fragmentary
and often conflicting information drawn from the
Soviet press and from Western surveys of the family
budgets of Soviet emigrants. Lack of reliable data
precludes estimation of illegal production of consum-
er goods. For the most part, we have assumed that
the second-economy activities counted in GNP grow
in real terms as do the sectors in which they are
included. There are no credible data with which to
assess changes in the volume of illegal activities.
"The 'Second Economy' in the USSR," Problems of Communism
(September-October 1977): pp. 25-40.
nonferrous-metals branches of industry, the coverage
is much smaller. Although we believe our samples to
be reasonably representative overall, we take every
opportunity to improve them when new data are made
available.
In most instances, data for the products in the sample
measure output in physical units (such as tons of
wheat, liters of vodka, or ton-kilometers of freight
carried), as shown:
Physical Units
Percentage Share of
Sample Data
Value Terms
GNP by Sector
of Origin
89
11
Of which:
Industry
82
18
Agriculture
100
0
Services
89
11
Whenever possible, we adjust these physical quantity
series to account for changes in product mix and for
quality. In the case of automobiles, for example, our
data capture changes in the numbers of different
models produced; and, in the case of coal, they reflect
changes in the quantity of the various grades extract-
ed. Growth in many of the services (such as education
or health) is measured by work-hour employment?a
practice that does not capture growth based on pro-
ductivity gains but one that is commonly employed in
Western national income accounting because of the
difficulty of measuring productivity trends in these
sectors.
But CIA has not been able to dispense completely
with ruble measures of output. Published Soviet mea-
sures in comparable prices are used to estimate
growth in output for products that make up about 20
percent of the value of the sample of industrial
products. In the machinery branch, some 40 percent
of the sample consists of such data series. Although
there are reasons to doubt that these official measures
have been properly deflated, we employ them to
obtain more complete coverage and to include new
and complex products for which physical quantities
either are not available or would fail to capture
changes in quality and mix. The likely net effects of
this reliance on official numbers are assessed in the
section "Some Sensitivity Testing."
6
Declassified in Part- Sanitized Copy Approved for Release 2013/01/10: CIA-RDP89T01451R000600690001-4
! Declassified in Part - Sanitized Copy Approved for Release 2013/01/10: CIA-RDP89T01451R000600690001-4
End-Use Indexes. Indexes of growth are also calculat-
ed for each of the 26 major end uses of GNP. For
consumption?the largest component of GNP?our
sample of products and services covers more than
nine-tenths of estimated total consumption, including
state expenditures for education and health. The
indexes used for investment in principle cover the
entire sector, as do nearly all of the indexes that
measure government administrative services.
The quantity indexes used to measure change over
time in the 26 end-use sectors are based mainly on
data in physical units, as shown:
Physical Units
Percentage Share of
Sample Data
Value Terms
Consumption
78 22
Investment
58 42
The indexes of consumption of food and soft goods
rely almost entirely on published Soviet data on per
capita consumption of major goods and on production
data adjusted whenever possible to take account of net
imports, inventory change, and changes in mix and
quality. Because of a lack of reliable data for alterna-
tive approaches, our indexes for consumer durables
and personal services are based on sales data deflated
by official price indexes, which probably understate
real price changes. The indexes for all other services
rely mainly on work-hour employment or other physi-
cal indicators. Although we rely on physical quantity
data to measure the growth of the construction com-
ponent of investment, we employ official Soviet value
statistics?said to be in constant ("comparable")
prices?to assess the growth of the machinery and
equipment component of new fixed investment.
CIA estimates the magnitude and growth of Soviet
military expenditures in rubles independently of
GNP. We employ a direct-costing (building-block)
7
approach that requires the identification and enumer-
ation of the physical elements constituting Moscow's
defense effort over time and the application of cost
factors to them. These ruble estimates?except for
research, development, testing, and evaluation activi-
ties?are aggregated by military service and military
mission. Estimates of total defense spending in rubles
are compared over time with CIA measures of
GNP?in factor-cost prices?to obtain measures of
the burden of defense in the USSR. That burden is
currently estimated to be about 16 percent of GNP.12
Because we consider the data available to estimate
growth of GNP by producing sector to be more
reliable than those used for the end uses, we set the
values of total GNP by end use in each year equal to
those obtained for GNP by producing sector. Sub-
tracting the total value estimated for the 26 end uses
in each year from the total GNP thus obtained leaves
a residual value, which contains a portion of defense
expenditures as well as net exports, inventory change
(including the inventory writeoff due to waste in the
distribution system), and any statistical discrepancy.
Additional defense expenditures are believed to be
scattered among other components, notably in invest-
ment and in research and development expenditures.
Attempts to estimate net exports and inventory
change in constant prices have been blocked by the
lack of suitable price deflators.
CIA Estimates Versus Official Measures
By using these methods to estimate the structure and
growth of Soviet GNP, CIA historically has obtained
results that differ markedly from the USSR's official
macroeconomic measures. As can be seen from a
comparison of the structure of Soviet GNP in estab-
lished prices and in adjusted factor-cost prices, for
example, factor-cost measures also give a quite differ-
ent picture of the contributions of the various produc-
ing sectors to overall economic output and of the
Declassified in Part- Sanitized Copy Approved for Release 2013/01/10: CIA-RDP89T01451R000600690001-4
Declassified in Part- Sanitized Copy Approved for Release 2013/01/10: CIA-RDP89T01451R000600690001-4
Table 1
Distribution of Soviet GNP in 1982,
by Sector of Origin
Percent
Established
Prices
Factor-Cost
Prices
GNP
100.0
100.0
Industry
50.4
33.4
Construction
6.7
8.0
Agriculture
15.4
20.2
Transportation
8.4
9.9
Communications
0.9
0.9
Trade
4.7
6.7
Services
11.3
18.5
Military personnel
1.6
2.0
Other branches
0.7
0.6
Because of rounding, subtotals may not add to 100.
Table 2
Distribution of Soviet GNP
in 1982, by End Use
Percent
Established
Prices
Factor-Cost
Prices
GNP
100.0
100.0
Consumption
52.7
52.7
Consumer goods
41.4
34.7
Food
24.7
24.6
Soft goods
11.1
6.5
Durables
5.6
3.6
Consumer services
11.3
18.0
Investment
27.9
31.2
New fixed investment
21.7
24.3
Machinery and equipment
8.6
9.3
Construction and other
13.0
14.7
Net additions to livestock
0.2
0.3
Capital repair
6.2
6.9
Other government expenditures a
19.4
16.1
a Includes most of defense spending. Some defense outlays are also
reflected in other end uses such as investment.
allocation of resources to major end uses in the base
year than would data in official prices (see tables 1
and 2).
Similarly, CIA's reliance on production indexes based
largely on disaggregated physical data yields esti-
mates of the growth of GNP and its components
substantially lower than those implied by the Soviet
macroeconomic measures that are their closest coun-
terparts (see table 3). For the last two five-year plans,
in fact, CIA estimates show a rate of growth roughly
half that of the official series on national income
produced. For industry and machine building as well
as for national income, the differences between the
estimated and reported rates of growth have been
substantial (see table 4). Nonetheless, both the esti-
mated and reported series are alike in indicating that
Soviet economic growth has been slowing gradually
but persistently over the past two decades.
The differences between CIA estimates and official
claims of growth on the end-use side are especially
striking with regard to consumption and defense. In
the case of consumption, CIA estimates show marked-
ly smaller gains-an average annual growth of 2.9
percent per capita during 1951-87-than the corre-
sponding Soviet measure "real per capita income of
the population," which grew at an average of 4.5
percent a year during the same period. In the case of
defense, Soviet officials have acknowledged in recent
years that the only data officially released-a single-
line entry in the annual state budget-include only a
portion of actual defense expenditures. To judge from
our estimates, the portion of actual defense expendi-
tures included in the announced defense budget is
small. Furthermore, although the announced budget
has been roughly constant since 1965, even when
8
Declassified in Part- Sanitized Copy Approved for Release 2013/01/10: CIA-RDP89T01451R000600690001-4
Declassified in Part- Sanitized Copy Approved for Release 2013/01/10: CIA-RDP89T01451R000600690001-4
Table 3
Comparison of CIA Estimates of and Official Statistics
on Soviet Economic Growth
Average annual growth in percent
CIA Estimates
Official Statistics b
GNP
GMP
National
Income Produced
National
Income Used
1951-60
5.1
6.7
10.3
NA
1961-65
4.8
4.9
6.5
5.7
1966-70
5.0
5.3
7.8
7.1
1971-75
3.1
3.3
5.7
5.2
1976-80
2.2
2.2
4.3
3.9
1981-85
1.8
1.7
3.6
3.1
1986-87
2.2
1.8
3.2
2.4
1951-87
3.8
4.2
6.7
NA
1966-87
2.9
3.0
5.2
4.6
a CIA estimates of GNP have been adjusted to make them roughly
comparable in coverage to the official estimates by excluding
services that do not contribute directly to material output. The
resulting estimates of gross material product (GMP), however, still
include depreciation, which is excluded from Soviet national in-
come measures.
b National income produced is the sum of value added (except
depreciation) in all branches of material production; national
income used is smaller than national income produced by losses
(from natural disasters and the like) and the value of net exports of
material goods.
Table 4
Comparison of CIA Estimates of and Official Statistics
on Soviet Economic Performance
Average annual growth in percent
Industry
Machine Building and Metalworking
Official
Statistics a
CIA
Estimates b
Difference
Official
Statistics c
CIA
Estimates b
Difference
1951-60
11.7
8.7
3.0
15.4
7.5
7.9
1961-65
8.6
6.5
2.1
12.4
7.0
5.4
1966-70
8.5
6.3
2.2
11.8
7.1
4.7
1971-75
7.4
5.4
2.0
11.6
6.6
5.0
1976-80
4.4
2.6
1.8
8.2
3.7
4.5
1981-85
3.7
1.8
1.9
6.2
1.3
4.9
1986-87
3.2
2.1
1.1
5.9
1.4
4.5
a Industrial output in comparable prices. From Narodnoye kho-
zyaystvo SSSR, various issues.
b In 1982 factor-cost prices.
c Machine-building and metalworking output in comparable prices.
From Narodnoye khozyaystvo SSSR, various issues.
9
Declassified in Part- Sanitized Copy Approved for Release 2013/01/10: CIA-RDP89T01451R000600690001-4
Declassified in Part- Sanitized Copy Approved for Release 2013/01/10: CIA-RDP89T01451R000600690001-4
Figure 4
Soviet Defense Expenditures: Published Budget
Versus CIA Estimate
Billion rubles
140
CIA estimate
(Constant 1982
rubles)
Published budget
(Current rubles)
120
....
L,
100 '
-----
woon......-- I
_
N
60
40

20
. .._--,
j ............----'
0 1965
70
75
80
85
87
expressed in constant-price terms, our estimates indi-
cate that actual defense expenditures have consistent-
ly increased (see figure 4).
CIA Estimates and the "New" Statistics
CIA and recent critics of the USSR's official econom-
ic data have similar views of Soviet economic statistics
in the sense that both reject Moscow's claims of
economic growth as inflated and its data on the
allocation of resources as misleading and flawed.
CIA's estimates are much closer to those of the critics
than to those of the state statistical authorities.
Nonetheless, some of these critics?in particular,
318821 9-88
Selyunin and Khanin?have advanced alternative es-
tifnates that suggest that Soviet growth has at times
been lower than CIA maintains (see table 5).
Unclear Basis of Alternative Estimates
As participants in a recent CIA-sponsored conference
on Soviet statistids noted, however, Soviet critics of
the state statistical authorities have generally dis-
played the same lack of methodological rigor and
inattention to documentation as the regime agencies
they fault.? As a result, it is often difficult to
determine exactly how their estimates were construct-
ed. Even the Selyunin-Khanin estimates have serious
10
Declassified in Part- Sanitized Copy Approved for Release 2013/01/10: CIA-RDP89T01451R000600690001-4
Declassified in Part- Sanitized Copy Approved for Release 2013/01/10: CIA-RDP89T01451R000600690001-4
Table 5
Alternative Measures
of Soviet Economic Growth
Average annual growth
in percent
Official
Soviet
Statistics a
Selyunin-
Khanin
Estimates b
CIA
Estimates
1951-60
10.3
7.2
5.1
1961-65
6.5
4.4
4.8
1966-70
7.8
4.1
5.0
1971-75
5.7
3.2
3.1
1976-80
4.3 .
1.0
2.2
1981-85
3.6
0.6
1.8
a National income produced in comparable prices. From Narodnoye
khozyaystvo SSSR v 1985, pp. 34, 38.
b National income produced in comparable prices. From Richard E.
Ericson, "The Soviet Statistical Debate: Khanin vs. TsSU," Harri-
man Institute for Advanced Study of the Soviet Union Occasional
Paper No. 1, unpublished (May 1988).
c GNP in 1982 factor-cost prices.
problems of this sort, and they are perhaps the best
documented of the recent glasnost-inspired alterna-
tive estimates of Soviet economic growth by virtue of
Khanin's publication of methodological articles in
Soviet academic journals:4 Professor Fyodor Kush-
nirsky, formerly of the Ukrainian Gosplan and now at
Temple University, reports, for example, that when he
followed the methods Khanin outlined and used the
types of statistics to which Khanin referred, he ob-
tained results that were strikingly different from those
that Khanin described:5 Kushnirsky's conclusion was
not that Khanin was wrong, but that his estimates fail
to meet the test of being "reproducible."
In the absence of clear evidence on the methods and
sources underlying the alternative Soviet estimates of
the USSR's economic growth, we can only speculate
about the reasons for the differences between them
and the CIA estimates. On the sector-of-origin side,
two shortcomings of the disaggregated Soviet data on
which CIA relies to construct indexes of growth could
be responsible for the differences: hidden inflation in
the series expressed in ruble prices and overstatement
of the data on the physical quantities produced. As we
have already indicated, however, the share of ruble-
value data in our overall sample is small, and we
11
believe that any overestimation resulting from their
use is likely to be at least partially offset by the
underestimation resulting from our reliance on physi-
cal indexes that fail to reflect improvements in prod-
uct quality and increased labor productivity (see
section "Some Sensitivity Testing"). In the case of
data on physical quantities, we believe the evidence
that overreporting of these data has increased over
time is inconclusive and note that even the critics of
official Soviet statistics make heavy use of such series
in constructing their alternative estimates.
Inflation and CIA Estimates of Growth
Inflation of Soviet ruble-value series has the most
potential for affecting CIA measures for the industri-
al sector, particularly machinery production. As men-
tioned above, about 40 percent of the value of CIA's
sample of machinery products consists of official
measures in comparable prices. Although most West-
ern and many Soviet experts have long maintained
that official data on machinery's growth are inflated,
the recent critics of the official statistics generally say
that the rate of inflation has been greater and the real
growth of machinery less than estimated by CIA (see
table 6).
It should be borne in mind, however, that, although
they account for 40 percent of our machinery sample,
potentially inflated machinery production data ac-
count for only 13 percent of the sample from which
the growth of industry is estimated and less than 5
percent of the sample used in estimating the growth of
GNP. Any upward bias in our estimates due to hidden
inflation, moreover, is offset to some degree by the use
of quantity data in the remaining 60 percent (by
value) of the machinery sample. These data often do
not reflect improvements in the average quality of the
product. In some machinery sectors, such as those for
automobiles and weapons, the indexes probably con-
tain little bias since they are based on highly disaggre-
gated and nearly complete samples in which changes
in assortment are reflected. In other machinery sec-
tors, however, our indexes probably understate growth
because for them we use more aggregated physical
series that do not account for changes in product mix
and quality improvements over time.'6 To the extent
Declassified in Part- Sanitized Copy Approved for Release 2013/01/10: CIA-RDP89T01451R000600690001-4
Declassified in Part- Sanitized Copy Approved for Release 2013/01/10: CIA-RDP89T01451R000600690001-4
Table 6
Alternative Measures of Growth of Soviet Machinery
Output, 1961-85
Average annual growth in percent
CIA a
Treml b
Selyunin and
Khanin
Val'tukh and
Lavrovskiy d
Official Soviet e
1961-65
7.0
NA
8.7
10.7
12.4
1966-70
7.1
6.3
5.5
5.9
11.8
1971-75
6.6
4.5
5.1-6.2
4.7
11.6
1976-80
3.7
2.3
2.0-3.1
1.2
8.2
1981-85
1.3
NA
0.2-1.2
NA
6.2
1961-85
5.1
NA
4.3-4.9
NA
9.4
1966-80
5.8
4.4
4.2-4.9
3.9
10.5
a Growth of total machinery, calculated from indexes of gross
output by subbranch, weighted by value added in 1982 prices.
b Growth of civilian machinery, calculated from indexes of gross
output by subbranch, weighted by gross output in 1982 prices. See
Vladimir G. Treml, "Weighted Aggregate Price Index for Soviet
Machinery (1965-82)," manuscript (1986), p. 12.
e Growth calculated by dividing official index of total machinery
output by Selyunin and Khanin's index of disguised inflation,
presumably for total machinery. See Vasiliy Selyunin and Grigoriy
Khanin, "Lukavaya tsifra," Novyy mir (February 1987): pp. 187,
194.
this is true, any upward bias due to hidden inflation
caused by using value series is at least partially offset.
As for the rest of industry, there probably is little or
no bias in the production data for the basic-materials
branches and fuel branches-except possibly coal and
gas where caloric content per unit has declined-
because product characteristics tend to be stable. (For
coal and gas we use reported production in standard
fuel units rather than in tons.) Biases in either direc-
tion could be present in the remaining industrial-
branch indexes, but they probably are small.
Downward bias in our indexes of the growth of Soviet
services might also offset at least some of the upward
bias resulting from the use of inflated ruble data in
other indexes. In a recent Ph.D. dissertation, Mark
Prell maintains that our reliance on physical quantity
data to estimate the growth of housing construction
and on labor inputs to estimate the growth of several
other services results in underestimation because
these data fail to reflect improvements in quality or
labor productivity.'' Similarly, Michael Boretsky, af-
ter applying our estimating procedures to US and
a presumably, growth of civilian machinery. See K. K. Val'tukh
and B. L. Lavrovskiy, "Proizvodstvennyy apparat strany: ispol'zo-
vaniye i rekonstruktsiya," Ekonomika i organizatsiya promyshlen-
nogo proizvodstva (February 1986): p. 29.
e Calculated from indexes given in various issues of Narodnoye
khozyaystvo SSSR.
West German data and obtaining results that showed
lower growth than officially reported for several ser-
vices, maintained that we underestimate the growth of
Soviet services as wel1.18 We have acknowledged that
our estimates of the growth of some Soviet services
are understated 19 but do not believe the degree of bias
is as great as Prell and Boretsky claim.
Use of Quantity Data
Another possible source of error in CIA estimates is
that Soviet industrial and agricultural production
data in physical units may be bogus. This issue has
been raised recently by Prof. Richard Ericson of
Columbia University, who-after reviewing the argu-
ments made by critics of official Soviet statistics-
asserts that physical data "have become increasingly
subject to fabrication and falsification." According to
an unidentified source cited by Ericson, Soviet au-
thorities estimate that 1.5 to 3 percent of reported
physical output is fictitious.2?
12
Declassified in Part- Sanitized Copy Approved for Release 2013/01/10: CIA-RDP89T01451R000600690001-4
!' Declassified in Part- Sanitized Copy Approved for Release 2013/01/10: CIA-RDP89T01451R000600690001-4
,
,
Quantity Data Probably Reliable. We first consider
the general reliability of quantity data. A consider-
able body of research conducted over a number of
years by Western and Soviet experts suggests that
disaggregated Soviet data in physical units provide a
generally reliable measure of reality:
? A comprehensive study of Soviet physical output
statistics in the industrial sector in the 1950s by
Gregory Grossman found that such data "generally
meet certain rough tests of internal and external
consistency wherever such tests are possible and
have been tried." According to Grossman, distortion
in the data is kept well in check by the interests of
customers, the difficulties of concealing large inven-
tory shortages, controls over the distribution of
products and the allocation of inputs, the attention
of authorities to this key segment of the economy,
and possibly severe penalties!'
? A study of Soviet statistics by Stephen Shenfield of
the Center for Russian and East European Studies,
University of Birmingham, United Kingdom, based
on interviews with former Soviet statistical person-
nel, concluded, "The scope for statistical overreport-
ing (pripiski) varies appreciably from field to
field... but appears usually to be relatively
modest." 22
? Using information from the Soviet Interview Pro-
ject, Prof. Susan J. Linz of Michigan State Univer-
sity recently analyzed report falsification by manag-
ers in the USSR. In the study, managers describe
such practices as relatively inconsequential: "Fac-
tories always keep a 1-percent margin" because
"nobody complains about small errors." But, these
managers said, "Falsifications on a grand scale are
dangerous." Specifically, according to the manag-
ers, "The director would be fired," "expelled from
the Party," or "taken to prison as well." 23
Changes in the Extent of Cheating Are Key. It is
necessary to bear in mind that, even if production
levels are exaggerated (or underreported), estimates of
growth based on these levels will remain accurate if
the extent and direction of misreporting do not change
over time." The key question, therefore, is whether
the degree of cheating has changed, particularly in
13
recent years. Anders Aslund argues in a recent paper
that such fraud "is likely to increase from year to
year" and to be more frequent "when a national
campaign is waged, and personnel shaken out on an
almost national scale, as is the present case." 25 A
more convincing argument can be made, however,
that fraud and cheating have not increased in recent
years. The increased emphasis given to the discipline
and anticorruption campaign by Gorbachev, for in-
stance, together with the creation of a new quality
control system in civilian industry and the call for
better and more rigorous statistical procedures, could
be discouraging the kind of cheating that makes its
way into Soviet economic reporting. And, while coun-
terarguments may be made with regard to the Brezh-
nev years, when corruption was reportedly rampant,
even such critics of official statistics as Selyunin and
Khanin have relied on physical production data in
constructing their estimates of growth for that period.
Indeed, Selyunin and Khanin, while arguing that
overreporting has affected some physical measures
published by the Soviet authorities, explicitly note
that useful estimates can be constructed on the basis
of such data.26
The Problem of Waste and Losses. Even if official
data on the physical quantities of goods produced are
not falsified, however, our indexes of production can
overstate the GNP available for final end use (other
than inventory change) if substantial losses occur
between production of the goods and their delivery to
consumers. In agriculture, for example, we adjust the
official Soviet data on production of grain and sun-
flower seeds to account for the waste that occurs at
harvest time. But our sector-of-origin indexes do not
further discount the official data to account for the
reportedly large losses that occur in transporting and
storing the harvest. Rather, these indexes are de-
signed to measure the growth in production, not final
disposition of what is produced.
From an accounting standpoint, such losses should be
reflected on the end-use side in the GNP residual?
the difference between total GNP and those portions
allocated to investment, consumption, defense, and
other end uses that we explicitly estimate. Indeed, the
Declassified in Part - Sanitized Copy Approved for Release 2013/01/10: CIA-RDP89T01451R000600690001-4
Declassified in Part- Sanitized Copy Approved for Release 2013/01/10: CIA-RDP89T01451R000600690001-4
Inflation in the Soviet Index of Investment
residual should also account, in theory, for the build-
up of inventories of goods unwanted by consumers?
another problem about which critics of official Soviet
statistics have had a great deal to say. In practice,
however, the degree to which losses and inventory
change are captured in the residual is limited by the
accuracy of the estimates of growth in major end uses,
which is discussed in the following section.
Estimating the Allocation of GNP
In contrast to its estimates of the growth of the major
producing sectors, CIA's estimates of the growth of
investment and consumption rely more heavily on
officially published growth measures in ruble terms
and official indexes of prices. As a result, the infla-
tionary biases of the Soviet data may affect these
estimates to a greater extent. If so, the estimated
growth of the end-use residual?which includes net
exports, inventory change (including postproduction
waste), and a portion of defense expenditures?is
understated. The estimated growth of total defense
outlays, however, would be unaffected by a change in
the residual's growth because it is calculated separate-
ly. Moreover, as with the estimates of growth on the
sector-of-origin side, there is reason to believe that at
least some of the inflationary bias in the estimates of
the growth of some end uses is offset by likely
underestimation elsewhere.
Possible Overstatement of Investment Growth. CIA's
estimates of the growth in new fixed investment are
the sum of separately calculated indexes for construc-
tion and for machinery and equipment. The index of
the construction component is based on physical
inputs of materials. The index of investment in ma-
chinery and equipment, however, is based on officially
published data that are said to be in constant prices
but are widely believed to have been improperly
deflated (see inset).
Because the components of the estimated index of
growth in new fixed investment are based on different
types of data, their likely biases differ. Our measure
of the construction component of investment (which is
about 60 percent of total new investment in 1982 and
is based on material inputs measured in physical
units) would understate growth if there were improve-
ments in the quality of those materials or a decline in
Long before the advent of glasnost, Soviet economists
alleged, directly or by implication, that officially
published investment data were biased upward.a A
sizable body of recent Western literature also ad-
dresses this issue, focusing mainly on investment in
machinery and equipment.b Western scholars general-
ly agree that some inflation is present in the officially
published investment data, but estimates of its size
differ greatly. Abram Bergson's analysis puts the
likely upward bias (if any) in investment data at
perhaps "a fraction of a percentage point" annually
during 1976-80. He concludes that the construction
component of investment probably contains little or
no inflation and states, "It would. . . be surprising if
the resultant overstatement overall in annual produ-
cers' durables investment growth was more than 1
percentage point for 1971-75 and 3 percentage points
for 1976-80."C
But other Western scholars estimate that the rates of
inflation are significantly higher, and the issue re-
mains unresolved. Peter Wiles, for example, esti-
mates that during 1966-76 "Inputs into investment
grew about 3 percent per annum less rapidly than
whatever rate is indicated by the current money
values. . . as officially deflated. "d
a See the references provided by Vladimir Kontorovich and Boris
Rumer, Inflation in the Soviet Investment Complex (Princeton
Junction, Ni: Command Economies Research, Inc., May 1988).
b The Western articles include: Abram Bergson, "On Soviet Real
Investment Growth," Soviet Studies (July 1987): pp. 406-424;
Stanley H. Cohn, "Response to Alec Nave," Soviet Studies (April
1981): pp. 296-299; Philip Hanson, "The CIA, the TsSU and the
Real Growth of Soviet Investment," Soviet Studies (October 1984):
pp. 571-581; Hanson, "Soviet Real Investment Growth: A Reply to
Bergson," Soviet Studies (July 1987): pp. 425-430; Alec Nove, "A
Note on Growth, Investment and Price Indices," Soviet Studies
(January 1981): pp. 142-145; Nove, "Reply to Stanley H. Cohn,"
Soviet Studies (April 1981): pp. 296-299; Nove, "Has Soviet
Growth Ceased?" Manchester Statistical Society Bulletin (15
November 1983); Nove, "Soviet Real Investment Growth: Are
Investment Volumes Overstated? A Reply to Bergson,"Soviet
Studies (July 1987): pp. 431-433; and Peter Wiles, "Soviet Con-
sumption and Investment Prices and the Meaningfulness of Real
Investment," Soviet Studies (April 1982): pp. 289-295.
c Bergson, "On Soviet Real Investment Growth," p. 420.
Wiles, "Soviet Consumption and Investment Prices and the
Meaningfulness of Real Investment," p. 292.
14
Declassified in Part- Sanitized Copy Approved for Release 2013/01/10: CIA-RDP89T01451R000600690001-4
rDeclassified in Part- Sanitized Copy Approved for Release 2013/01/10: CIA-RDP89T01451R000600690001-4
the material-intensiveness of construction. Our index
of construction increases somewhat more slowly than
the comparable Soviet measure-5.6 percent per year
versus 6.9 percent during 1951-87. If we are substan-
tially overstating the growth of that component, the
growth in the physical production of the relevant
materials is equally overstated. But, as the works of
Grossman, Shenfield, and Linz suggest, the trends in
Soviet data expressed in physical units do not seem to
be overstated.
CIA measures of the machinery and equipment com-
ponent?nearly 40 percent of total new investment?
would overstate real growth to the extent that the
official values that underlie our estimates are not fully
corrected for price changes. Inflation could come
either from failure to fully deflate the values of
domestic production of investment goods or from
failure to allow for price increases of imported equip-
ment, or both. Until now, the CIA has considered the
official data on the machinery component of invest-
ment acceptable because their growth, excluding im-
ports, matches fairly closely the growth of the index of
production of producer durables (excluding exports),
which is calculated independently by CIA.27 The
respective average annual growth rates are 7.7 per-
cent and 7.9 percent for the period 1951-87. Thus, if
there is substantial inflation in domestically produced
machinery, both our producer durables production
index and our investment index are also overstated. If
this is so, however, we believe they probably are not
overstated by much.
Inflation could also be present in the official machin-
ery investment data if the Soviet state statistical
authorities record imported machinery in current
prices. Soviet imports of machinery have increased
rapidly and have become a large share of total
machinery investment?probably more than one-
fourth in 1986 if imports are recorded at their ruble
cost. During 1971-80, for example, the volume of
machinery imports rose by 128 percent and their
prices by 77 percent, according to calculations from
data published by a Soviet economist." We do not
know what methodology the statistical authorities use.
If, however, they did not correct for increases in prices
of imported machinery during the 1970s, average
15
annual growth of investment in machinery could be
overstated by 1 percentage point or more on this
account alone.
On balance, then, we judge that our estimates of
investment are probably biased upward by some
amount. Inflation in imported machinery prices is
probably a partial reason, and some overstatement
may be caused by the use of data on the annual
growth of the machinery and equipment portion of
new fixed investment reported by Soviet statistical
authorities.
Possible Overstatement of Consumption Growth.
CIA's index of the growth of consumption, as noted
above, consistently shows markedly smaller real gains
than does the corresponding Soviet measure." For
1951-87, the average annual rates of growth given by
the two measures (per capita) are 2.9 percent and 4.5
percent, respectively. Nonetheless, as was the case
with investment, our estimates of consumption growth
contain some error because of inflation that cannot be
accounted for and a failure to allow for improvements
in the quality of consumer services.
Although CIA's measure of consumption growth is
not obtained by deflating current-price values by an
estimated consumer price index, we calculate a so-
called alternative retail price index. It is based on a
comparison of the growth of the published current
values of retail purchases with their growth as mea-
sured in constant 1982 prices by our index of con-
sumption of goods." This index shows an average
annual rate of price inflation of 1.5 percent during
1961-87 (a total increase of 51 percent); the official
Soviet retail price index shows 0.3 percent annual
inflation (a total of 8 percent).
CIA's alternative consumer price index, however, does
not indicate that the rate of inflation has been as high
as some recent Soviet studies suggest. A study by the
USSR Price Committee and Ministry of Trade, for
example, found that the average level of retail prices
rose by an average of 2.8 percent a year during 1971-
83.3' Moreover, Prof. Richard Ericson has maintained
that "it seems generally believed (in the USSR) that
Declassified in Part- Sanitized Copy Approved for Release 2013/01/10: CIA-RDP89T01451R000600690001-4
Declassified in Part- Sanitized Copy Approved for Release 2013/01/10: CIA-RDP89T01451R000600690001-4
the price of a typical bundle of consumer goods has
been rising 3 to 5 percent a year, at least over the past
decade." 32
Regardless of whether these claims are accurate, the
CIA alternative index undoubtedly records lower
rates of inflation than would be revealed by a West-
ern-style consumer price index. The index cannot
capture the price creep that results from the "washing
out" of cheaper grades of goods, leaving the consumer
with only expensive grades to buy. Moreover, the
accuracy of the "alternative" index depends on the
accuracy with which we have measured changes in
real consumption of goods (food, soft goods, and
durables)."
Our measures of food and beverage consumption,
which are based almost entirely on quantity data
adjusted for quality and mix when possible, are
probably the most reliable. An alternative price index
for food alone shows a greater price change than does
the overall alternative index and far more than the
official retail price index for food and beverages. It
rises by 63 percent during 1961-87, while the official
index increases by 22 percent.
Our measure for soft goods may overstate growth,
however, because it relies on Soviet value series for
production of clothing and a few retail sales indexes,
both of which may be inflated. On the other hand, the
measure also relies heavily on physical series, which
may not properly reflect changes in quality and mix,
and thus may understate growth.
For consumer durables, we rely on a measure of
purchases in current prices deflated by an estimated
official Soviet price index for similar goods. We do
this because the quantity data needed for another
approach, like that used for food and soft goods, are
not available. The estimated official price index un-
doubtedly understates price change; there is as yet no
good way to determine how much.
Our measures of the growth of services are subject to
both overstatement and understatement. Growth of
personal services probably is overstated because the
index relies on Soviet-published value-of-sales data in
constant prices; there are good reasons to be skeptical
of the reliability of those data. On the other hand, our
measures for housing, health, and education probably
understate growth because of insufficient allowance
for improvement in quality.' The indexes for educa-
tion and health as end uses, however, do reflect some
improvements in quality by incorporating a measure
of purchases of materials by these institutions.
Some Sensitivity Testing
In an effort to assess the likely impact of these
different biases in the data used in our indexes of
growth on both the sector-of-origin and end-use sides,
we did some sensitivity testing.
Growth of Soviet GNP
To measure the likely understatement or overstate-
ment of CIA's measure of growth of Soviet GNP by
sector of origin, the following assumptions were made:
? In the industrial sector, the ferrous metals index was
assumed to understate growth by as much as 1
percentage point a year because of failure to capture
changes in the assortment and quality of steel
products such as cold-rolled sheet and seamless
pipes and tubes.
? For the machinery branch, we constructed an index
that splits the difference between our current index
and an index of civilian machinery production com-
piled by Prof. Vladimir Treml of Duke University."
(Treml's index is based entirely on samples reported
in physical units for various machinery sectors, and
his sample is highly aggregated. As a result, it
probably provides a "rockbottom" estimate of the
actual growth of machinery output.) The net effect
of this change was a reduction in overall machinery
growth by almost 1 percentage point a year.
? For the chemicals branch, a national average under-
statement of 1 percentage point per year for the
aniline dye, rubber products, and three synthetics
sectors was assumed; a 1-percentage-point over-
statement was assumed for mineral chemicals, the
single value series.
16
Declassified in Part- Sanitized Copy Approved for Release 2013/01/10: CIA-RDP89T01451R000600690001-4
Declassified in Part- Sanitized Copy Approved for Release 2013/01/10: CIA-RDP89T01451R000600690001-4
Table 7 Average annual growth
Alternative Estimates of Growth in percent
of Soviet GNP, by Sector of Origin,
195187a
Current CIA
Estimate
Adjusted
Estimate
GNP
3.8
3.9
Industry
5.5
5.3
Agriculture
1.6
1.6
Services
3.2
4.1
Other b
5.2
5.2
a In 1982 factor-cost prices.
b Construction, transportation, communications, trade, military
personnel, and other branches.
? In the services sector, in line with Mark Prell's
thesis that our estimates of growth are understated
because we do not take quality gains into account,
several adjustments were made.36 CIA's and Prell's
growth rates were averaged to obtain a substitute
index for housing. For the rest of the services, Prell's
estimates of growth of services were used instead of
the CIA estimates.
The net effect of all the adjustments was a negligible
increase in the growth of GNP (see table 7). The
growth rates of industry and services-the sectors
affected by the adjustments-changed somewhat
more."
Distribution of GNP by End Use
As with the sector-of-origin measures, a number of
assumptions were made to test the sensitivity of CIA's
estimates of GNP by end use to changes in the end-
use indexes:
? The consumer durables index was adjusted to incor-
porate a rate of inflation of 1 percent per year, a
rate equal to the difference between the growth of
our index, which is based on deflated retail sales,
and the growth of Treml's sample of consumer
durables, which is based on physical production.
? For the machinery and equipment component of
investment, we adopted the same discount as that
used to adjust the index of machinery's growth on
the sector-of-origin side.
17
Table 8
Alternative Estimates of Growth
of Soviet GNP, by End Use,
1951-87 a
Average annual growth
in percent
Current CIA
Estimate
Adjusted
Estimate
GNP
3.8
3.9
Consumption
3.8
4.3
Consumer goods
3.7
3.6
Consumer services
4.0
5.0
Investment
7.0
6.8
New fixed investment
6.5
6.3
Capital repair
10.6
9.5
Defense b
2.7
2.7
Other government
expenditures c
1.0
2.6
Residual a
...e
In 1982 factor-cost prices.
b The CIA estimate of defense expenditures does not rely on Soviet
statistics and was assumed not to be biased.
c Administration, other services, and civilian research and
development.
d Includes net exports, inventory change (including waste), losses,
and statistical discrepancy.
c Declines from 9.2 billion rubles in 1950 to -17.3 billion rubles in
1987.
I Declines from 9.2 billion rubles in 1950 to -13.3 billion rubles in
1987.
? The discount for capital repair was carried over
from the capital repair sector of machinery output.
? Adjusted indexes for end-use services were obtained
by applying the factors used for adjusting the
sector-of-origin services indexes.
The results of this exercise are presented in table 8.
Rates of growth for GNP, of course, are the same as
those in table 7 because the sector-of-origin estimates
provide the control totals for GNP. The main differ-
ences between CIA's current estimate and the adjust-
ed estimate are a somewhat slower growth for con-
sumer goods in the adjusted estimate because of
slower growth of consumer durables; substantially
faster growth of consumer and government services;
and slower growth of new fixed investment and
capital repair.
, Declassified in Part- Sanitized Copy Approved for Release 2013/01/10: CIA-RDP89T01451R000600690001-4
Declassified in Part- Sanitized Copy Approved for Release 2013/01/10: CIA-RDP89T01451R000600690001-4
Table 9
Distribution of Soviet GNP by End Use
in 1987 Based on Alternative Estimates
of Hidden Inflation a
Percentage New Official Data on GNP
shares
Current CIA
Estimate
Adjusted
Estimate
GNP
100.0
100.0
Consumption
53.7
56.8
Consumer goods
35.0
32.2
Consumer services
18.6
24.6
Investment
29.6
24.4
New fixed investment
23.8
20.6
Capital repair
5.8
3.9
Defense
16.0
15.2
Other government
expenditures b
3.2
5.3
Residual C
?2.4
?1.8
a In 1982 factor-cost prices.
b Administration, other services, and civilian research and
development.
c Includes net exports, inventory change, and statistical
discrepancy.
The effect of these adjustments is, however, far from
negligible. Indeed, if these differential rates of infla-
tion had prevailed over the period 1951-87, the distri-
bution of GNP in 1987 would have been appreciably
different from that in our current estimate. The
composition of consumption and the investment share
of GNP are markedly different, for instance (see
table 9).
Conclusions and Prospects
On balance, the glasnost-inspired criticisms of official
Soviet statistics are generally reassuring with regard
to the accuracy of CIA's estimates of overall Soviet
economic growth. However, they also lend support to
earlier indications that the estimates of growth in new
fixed investment and the output of producer and
consumer durables may require adjustment. Efforts to
accomplish this as well as to improve the estimates of
growth in civilian and military machinery and ser-
vices, the treatment of foreign trade, and the coverage
of activities from the "second economy" are currently
in progress.
In mid-1988, the Soviet Government published for
the first time a measure of the size and growth of the
economy using the Western concept of GNP. (The
narrower concept of national income produced ex-
cludes depreciation and most services.) The new data
include a figure for GNP in 1987 (825 billion rubles)
and indexes of its real growth during 1981-87.a The
rates of growth reported for Soviet GNP show consid-
erably better performance than does the narrower
measure of national income traditionally used?an
average annual growth of 3.9 percent, as compared
with 3.5 percent during 1981-87. Like the traditional
measures, the new official GNP growth rates are
markedly higher than those estimated by the CIA.
We estimate that average annual growth of GNP
during 1981-87 was 1.9 percent.
The new Soviet measures suffer from the same faults
as the traditional ones used to assess economic
performance. Evidently, the state statistical authori-
ties have used the official indexes of net material
product to measure growth of that component of GNP
(roughly 75 percent of the total). The rest of GNP?
mostly services and depreciation?increases at a rate
of about 5 percent annually during 1981-87; a rough-
ly comparable CIA measure of that component in-
creases at less than half that rate.
a SSSR v tsifrakh v 1987 godu (USSR in Figures in 1987), p. 14.
In many instances, however, additional data from
either official sources or the writings of Soviet critics
will be required before such improvements can be
made with confidence. Glasnost may result in the
publication of such data, although to date the
glasnost-inspired critics have, unfortunately, been
more intent on exposing the unreliability of official
18
Declassified in Part- Sanitized Copy Approved for Release 2013/01/10: CIA-RDP89T01451R000600690001-4
Declassified in Part- Sanitized Copy Approved for Release 2013/01/10: CIA-RDP89T01451R000600690001-4
data than on providing well-documented and authori-
tative alternatives. Still, now that it has begun, the
campaign of criticism will be difficult to halt because
it clearly serves the current leadership's purpose of
blackening the record of its predecessors. Therefore,
the unofficial reassessment of Soviet economic history
is likely to continue.
The need for alternative estimates of Soviet economic
performance like those produced by CIA will continue
as well, because, while critics debunk past official
claims of growth, the Soviet Government continues to
use them and has even further distorted the recent
record. Its recent publication for the first time of
Western-style estimates of GNP to supplement the
traditional Marxian measure net material product is a
case in point (see inset). The new official measures,
like the old, imply growth rates substantially higher
than those indicated by the CIA estimates, and the
cause of the differences is likely to lie, in our view, in
the deficiencies of the Soviet value data.
Reverse Blank
19
In the absence of reliable alternatives to Moscow's
macroeconomic measures, US policymakers would
have to accept the inconsistent and largely undocu-
mented Soviet figures or refrain completely from
assessing actual Soviet performance. Because assess-
ing that performance is crucial to assessing the
USSR's ability to raise living standards, provide for
future growth, and maintain its heavy commitment of
resources to defense, either course would be unwise.
Declassified in Part- Sanitized Copy Approved for Release 2013/01/10: CIA-RDP89T01451R000600690001-4
rDeclassified in Part- Sanitized Copy Approved for Release 2013/01/10: CIA-RDP89T01451R000600690001-4
Notes
1. Vasiliy Selyunin and Grigoriy Khanin, "Lukavaya
tsifra" ("Cunning Figures"), Novyy mir (February
1987): pp. 181-201.
2. Prof. Abram Bergson was a pioneer in the construc-
tion of GNP accounts for the Soviet Union. His work
was further developed by scholars at the Rand Corpo-
ration. The result of their efforts was a set of accounts
for 1928, 1937, 1940, 1944, and 1948-66 in current
and constant rubles, published in the 1950s and
1960s. A number of other scholars have also pub-
lished estimates of Soviet GNP. For a survey of this
voluminous literature, see Joint Economic Committee
of Congress, USSR: Measures of Economic Growth
and Development, 1950-80 (Washington, DC: US
Government Printing Office, 1982), pp. 11-12.
3. Al'bert L. Vaynshteyn, Narodnyy dokhod Rossii i
SSSR (The National Income of Russia and the
USSR) (Moscow: Nauka, 1969). See also Frederick G.
Denton, "A Recent Soviet Study of Economic
Growth, 1951-63," Soviet Studies (April 1968): pp.
501-509.
4. Ya. Kvasha and V. P. Krasovskiy, "Kapital'noye
stroitel'stvo i problema vozmeshcheniya" ("Capital
Construction and the Problem of Replacement"),
Voprosy ekonomiki, No. 11(1964): p. 12; Kvasha and
Krasovskiy, "Kapital'noye stroitel'stvo i nakopleniye"
("Capital Construction and Accumulation"), Voprosy
ekonomiki, No. 7 (1965): pp. 4-5; Kvasha and Kra-
sovskiy, "Perspektivnoye planirovaniye i khozyayst-
vennye izmereniya" ("Perspective Planning and Eco-
nomic Measurements"), Voprosy ekonomiki, No. 4
(1968): P. 71; Krasovskiy, Planirovaniye i analiz
narodnokhozyaystvennoy struktury kapital'nykh
vlozheniy (Planning and Analysis of the Economic
Structure of Capital Investments) (Moscow: Ekono-
mika, 1970), p. 242; and N. M. Mitrofanova, "Ten-
dentsii dvizheniya kontraktnykh tsen v torgovle stran
SEV" ("Trends in Contract Prices in CEMA Trade"),
Voprosy ekonomiki (August 1978): pp. 101-106 (her
calculations pertain to prices of exported machinery).
21
5. K. K. Val'tukh, "Investitsionnyy kompleks i inten-
sifikatsiya proizvodstva" ("The Investment Complex
and Intensification of Production"), Ekonomika i
organizatsiya promyshlennogo proizvodstva, No. 3
(1982): pp. 4-31; V. K. Fal'tsman, "Narodnokho-
zyaystvennyy zakaz na novuyu tekhniku," ("The
Economy's Order for New Equipment"), Ekonomika
I organizatsiya promyshlennogo proizvodstva, No. 7
(1983): pp. 3-19; Fal'tsman and A. Kornev, "Reservy
snizheniya kapitalo-yemkosti moshchnostey promysh-
lennosti" ("Potential for Reducing the Capital Inten-
siveness of Industrial Capacity"), Voprosy ekonomiki,
No. 6 (1984): pp. 36-45; Val'tukh and B. L. Lavrov-
skiy, "Proizvodstvennyy apparat strany: ispol'zovan-
iye i rekonstruktsiya" ("Production Facilities of the
Country: Their Use and Reconstruction"), Ekono-
mika i organizatsiya promyshlennogo proizvodstva
(February 1986): pp. 17-32.
6. Literaturnaya gazeta (16 September 1987): p. 12.
7. Vladimir Kontorovich and Boris Rumer, Inflation
in the Soviet Investment Complex (Princeton Junc-
tion, NJ: Command Economies Research, Inc., May
1988), p. i.
8. Economist Nikolay Shmelev has been especially
vocal on this subject. For example, see his interview
with a Hungarian correspondent as translated in
Foreign Broadcast Information Service, Daily Report
(SOV 88-112, 10 June 1988), p. 89; or his subsequent
interview with a Bulgarian correspondent as translat-
ed in Daily Report (SOV 88-120, 22 June 1988), pp.
67-68.
9. CIA's methodology is explained and documented in
Joint Economic Committee of Congress, USSR: Mea-
sures of Economic Growth and Development,
1950-80.
Declassified in Part - Sanitized Copy Approved for Release 2013/01/10: CIA-RDP89T01451R000600690001-4
,
Declassified in Part- Sanitized Copy Approved for Release 2013/01/10: CIA-RDP89T01451R000600690001-4
10. Western specialists have written many books and
journal articles on the selection and refinement of
estimating methods. An annotated bibliography of
much of the literature is provided in the notes at the
end of each chapter in Dan Usher, The Measurement
of Economic Growth (New York: Columbia Universi-
ty Press, 1980).
11. In making this adjustment, we rely on the pioneer-
ing work done by Abram Bergson. See, for example,
Bergson, The Real National Income of Soviet Russia
Since 1928 (Cambridge, Massachusetts: Harvard
University Press, 1961).
12. For a discussion of CIA's methodology, see Joint
Economic Committee of Congress, Allocation of Re-
sources in the Soviet Union and China-1983 (Wash-
ington, DC: US Government Printing Office, 1983),
pp. 347-351.
13. Central Intelligence Agency, The Impact of Gor-
bachev's Policies on Soviet Economic Statistics (Di-
rectorate of Intelligence Conference Report SOV 88-
10049, July 1988), pp. 2-3.
14. In addition to "Lukavaya tsifra" (see note 1),
Khanin previously published several articles critical of
official Soviet data. See, for example, "Alternativnyye
otsenki rezul'tatov khozyaystvennoy deyatel'nosti
proizvodstvennykh yacheyek promyshlennosti" ("Al-
ternative Estimates of the Results of the Economic
Activity of Industrial Production Units"), Izvestiya
akademiy nauk, seriya ekonomicheskaya (June
1981): pp. 62-71; and "Puti sovershenstvovaniya infor-
matsionnogo obespecheniya svodnykh planovykh na-
rodnokhozyaystvennykh raschetakh" ("Ways of Im-
proving the Information Supply of Consolidated
National Economic Plan Accounts"), Izvestiya aka-
demiy nauk, seriya ekonomicheskaya (March 1984):
pp. 58-67.
15. Fyodor Kushnirsky, "New Challenges to Soviet
Official Statistics: A Methodological Survey," in
Central Intelligence Agency, The Impact of Gorba-
chev's Policies on Soviet Economic Statistics, pp. 12-
13, 15-16.
16. Construction and road machinery is a case in
point. A study undertaken by the CIA using the so-
called hedonic method analyzed the behavior of prices
in this sector during 1960-73 to determine the extent
to which increases were justified by improvements in
product quality. Using this technique, the relative
influence of the different physical characteristics in
the determination of machinery prices can be judged
by their respective regression coefficients. The results
indicated that, although the increment in prices of
new products often exceeded changes in quality, in
most cases the technical characteristics proved to be
statistically significant. In other words, for all of the
types of construction machinery tested?scrapers,
bulldozers, rollers, graders, and excavators?the re-
search suggests that average quality improved over
the period analyzed. More recently, Kushnirsky's
research on the Soviet electrotechnical and automo-
bile industries also concluded that the quality of
products in these branches of machine building trend-
ed upward. Central Intelligence Agency, An Analysis
of the Behavior of Soviet Machinery Prices, 1960-73
(ER79-1063b, December 1979), and Kushnirsky,
"New Challenges to Soviet Official Statistics," pp.
19-20.
17. Mark A. Prell, The Role of the Service Sector in
Soviet GNP and Productivity Estimates, Ph.D. dis-
sertation, Massachusetts Institute of Technology
(1987). Prell's estimates of average annual growth are
higher than ours for 1951-84 (by the percentage
points shown) for housing (2.8), education (1.1), health
(1.5), government (0.7), and municipal services (2.0).
His estimate for the growth of science over the same
period is 0.7 percentage point less than ours.
18. Michael Boretsky, "The Tenability of CIA Esti-
mates of Soviet Economic Growth," Journal of Com-
parative Economics (December 1987): pp. 526-27,
532.
19. Joint Economic Committee of Congress, USSR:
Measures of Economic Growth and Development,
1950-80, pp. 112-113, 344-345, 349.
20. Richard E. Ericson, "The Soviet Statistical De-
bate: Khanin vs. TsSU," paper presented at the
Hoover Institution?Rand Corporation Conference on
22
Declassified in Part- Sanitized Copy Approved for Release 2013/01/10: CIA-RDP89T01451R000600690001-4
Declassified in Part - Sanitized Copy Approved for Release 2013/01/10 : CIA-RDP89T01451R000600690001-4
the Defense Sector and the Soviet Economy, Stanford
University (23-24 March 1988, revised May 1988), p.
10. In another paper presented at the conference-
"How Small Is the Soviet National Income?" (p.
12)-Anders Aslund writes that "according to some
calculations of economists, fraud accounts for 5 to 25
percent of the production a year" in the "raw material
producing sectors." He cites as the source Aleksey
Sergeyev, "Prestige of the Honest Ruble," Sovetska-
ya Rossiya (18 March 1987): p. 1. But the instances of
figure padding Sergeyev discusses do not include the
overreporting of physical production. Instead, he dis-
cusses stealing of state resources, failure to report
hours worked accurately, and-in general-"remu-
neration for work that has not been performed" and
materials that have not been used in production.
21. Gregory Grossman, Soviet Statistics on Physical
Output of Industrial Commodities (Princeton: Prince-
ton University Press, 1960), pp. 123-134.
22. Stephen Shenfield, "State Statistical Work in the
USSR: Findings from Interviews with Former Soviet
Statistical Personnel," (Washington, DC: The Nation-
al Council for Soviet and East European Research,
April 1986), pp. 107-122.
23. Susan J. Linz, "Managerial Autonomy in Soviet
Firms," Soviet Studies (April 1988): p. 187.
24. Alec Nove discusses the "law of equal cheating"
in "The Pace of Soviet Economic Development,"
Lloyd's Bank Review (April 1956): p. 3.
25. Aslund, "How Small Is the Soviet National
Income?" pp. 12-13.
26. Selyunin and Khanin, "Lukavaya tsifra," p. 200.
27. Joint Economic Committee of Congress, USSR:
Measures of Economic Growth and Development,
1950-80, pp. 203-207.
28. V. Seltsovskiy, "Some Economic Statistical Meth-
ods of the USSR Foreign Trade Development Analy-
sis in the Ninth and Tenth Five-Year Plan Periods,"
Foreign Trade, No. 5 (1982): pp. 37-42.
Reverse Blank
23
29. When assessing changes in living standards, CIA
uses its indexes of components of consumption calcu-
lated in established (official) prices because they
reflect the prices actually paid by consumers and
institutions.
30. The derivation and an assessment of this index are
given in Gertrude E. Schroeder and Barbara S.
Severin, "Soviet Consumption and Income Policies in
Perspective," in Joint Economic Committee of Con-
gress, Soviet Economy in a New Perspective (Wash-
ington, DC: US Government Printing Office, 1976),
p. 631.
31. V. Rutgayzer, Yu. Shevyakhov, and L. Zubova,
"Sovershenstvovaniye sistemi planovikh pokazateleii
dokhodov nasileniya" ("Perfecting the System of
Planned Indicators of Income of the Population"),
Voprosy ekonomiki, No. 1 (1988): p. 33.
32. Ericson, "The Soviet Statistical Debate: Khanin
vs. TsSU," p. 4.
33. For a discussion of CIA's index, see Gertrude E.
Schroeder and M. Elizabeth Denton, "An Index of
Consumption in the USSR," in Joint Economic Com-
mittee of Congress, USSR: Measures of Economic
Growth and Development, 1950-80, pp. 317-401.
34. Prell, as noted above, made a case for a substan-
tial understatement of our indexes of these services.
Even in the much maligned health sector, the educa-
tional qualifications of doctors and other employees
have increased, and the quantity and quality of
equipment available to each employee have risen
substantially.
35. Vladimir G. Treml, "Weighted Aggregate Price
Index for Soviet Machinery," unpublished paper (Jan-
uary 1986).
36. Prell, The Role of the Service Sector in Soviet
GNP and Productivity Estimates.
37. For a full discussion of this exercise see James
Noren, "The New Look at Soviet Statistics: Implica-
tions for CIA Measures of the USSR's Economic
Growth," in Central Intelligence Agency, The Impact
of Gorbachev's Policies on Soviet Economic Statis-
tics, pp. 73-79.
Declassified in Part- Sanitized Copy Approved for Release 2013/01/10: CIA-RDP89T01451R000600690001-4