MODELING SOVIET AGRICULTURE: ISOLATING THE EFFECTS OF WEATHER

• Document Type: CREST
• Collection: General CIA Records
• Document Number (FOIA) /ESDN (CREST): CIA-RDP89T01451R000500520004-0
• Release Decision: RIPPUB
• Original Classification: K
• Document Page Count: 45
• Document Creation Date: December 27, 2016
• Document Release Date: December 3, 2012
• Sequence Number: 4
• Publication Date: August 1, 1988
• Content Type: REPORT

Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89T01451 R000500520004-0
To recipients of DI Paper SOV 88-10054, August 1988, Modeling Soviet
Agriculture: Isolating the Effects of Weather: This unclassified report
documents the development and use of a model to examine past trends in
agricultural productivity, measures the relative contribution of labor and
capital to farm output, and assesses Soviet prospects for meeting the goals
The agriculture model is currently used in CIA's macroeco-
nomic model of the Soviet Union,
25X1
25X1
25X1
25X1
25X1
Comments and queries are welcome and may be addressed to the Chief,
Economic Performance Division, Office of Soviet Analysis, 25X1
25X1
Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89T01451 R000500520004-0
 Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
Directorate of
Intelligence
Modeling Soviet Agriculture:
Isolating the Effects of Weather
RECORD COPY
,Return to SOMA Registry
STAT
SOV 88-10054
August 1988
Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
 Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
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 2012/12/03 : CIA-RDP89TO1451 R000500520004-0
 Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
Directorate of
Intelligence
Modeling Soviet Agriculture:
Isolating the Effects of Weather
STAT
SOV 88-10054
August 1988
Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
 Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89T01451 R000500520004-0
Figure 1
Observed Farm Output and the Model's Prediction, 1965-87a
I I I I I I I I I I I I I I I I I I I I I I
a Net of feed, seed, and waste.
b Preliminary
Model F
predictions
STAT
Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89T01451 R000500520004-0
 Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
Modeling Soviet Agriculture:
Isolating the Effects of Weather
Summary General Secretary Gorbachev needs to improve food supplies dramatically
Information available to bolster popular support for the economic restructuring program. Mos-
as of 1 June 1988 "intensify"
was used in this report. cow's campaign to agriculture, particularly grain production,
has resulted in recent gains. But agriculture still faces serious problems,
and, unless strong measures are taken to stimulate productivity on the
farm, Moscow will become increasingly unable to meet the demand for
more and better food supplies without resorting to substantial hard
currency imports.
Gorbachev has been seeking ways to overcome the gross inefficiencies of
the agroindustrial sector. Agricultural reforms since he came to power
include the creation of the superministry Gosagroprom, endorsement of
collective contracts for farmworkers, enforcement of stable procurement
plans, and promotion of the right of farms to directly market a portion of
planned fruit and vegetable procurement. Gorbachev's call in 1987 for a
special Central Committee plenum to tackle comprehensive agricultural
reform suggests that more policy initiatives in agriculture are on the way.
To evaluate the effects of such initiatives, it is first necessary to isolate the
effects of weather, which often mask the influences of other variables on
agricultural performance.
Isolating Weather Factors
A mathematical model was developed to separate the effects of weather
from the effects of other factors. In developing the model, it became clear
that weather factors alone were not sufficient to explain agriculture's
dismal showing during the 1979-82 period. When capital, labor, and
productivity changes were included in the model, the results tracked closely
actual fluctuations in output (see figure 1).
The rate at which weather-adjusted output is increasing has important
implications for Gorbachev's agriculture policy. Until 1979 weather-
adjusted output increased steadily, reflecting relatively stable growth of
inputs, steady but slow technological progress, and the absence of sharp
swings in government policy (see figure 2). Weather-adjusted output
dropped precipitously in 1979 and continued to decline in 1980 and 1981.
During this time, growth of deliveries to agriculture slowed as overall
SOV 88-10054
August 1988
Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
 Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
Figure 2
Farm Output After Adjusting for Weather, 1968-878
I I I I I I I I I I I I I I I I I I
a Net of feed, seed, and waste.
b Preliminary
Intensive
technology
campaign
Brezhnev
Food Program
industrial growth slowed, and transportation organizations were increas-
ingly unable to keep pace with the growing requirements. In addition,
government policies specific to agriculture were flawed:
? Investment resources going to agriculture were wastefully allocated and
inefficiently utilized. Soviet authors have complained about losses of
agricultural products because construction of storage facilities and rural
roads was neglected.
? Agricultural machinery downtime increased, efficiency in the use of
inputs-especially machinery, equipment, and fertilizers-declined, and
growth in livestock herds outstripped growth in feed availability.
Industrial growth I
slowdown period
STAT
Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
 Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
In 1983, however, there was a remarkable recovery, reflecting improve-
ments in efficiency stemming from the Brezhnev Food Program imple-
mented the previous year. Since Brezhnev's death in late 1982, Gorbachev
has used his influence in the leadership to reshape the program to reflect
more closely his own views and priorities. In addition to measures targeted
at increasing worker productivity, Gorbachev has given the "intensive
technology" program a high priority. Intensive technology, as defined by
the USSR, includes many practices routinely performed in the West-use
of high-yield varieties, planting after fallow where possible, implementing
efficient field operation schedules, and extensive use of agrochemicals. By
1984 and 1985 weather-adjusted agricultural output had nearly returned
to the pre-1979 trend, and performance was clearly back on trend in 1986
and 1987.
Returns to Capital and Labor
The model results also show that the return to capital is lower in
agriculture than in any other producing sector of the economy except fuels,
and thus underscores the burden imposed on the rest of the economy by
agriculture's large share of investment resources. The capital elasticity was
estimated to be 0.17, indicating that a 1-percent increase in the capital
stock results in only a 0.17-percent increase in output. The return to labor
in agriculture, on the other hand, is estimated by the model to be over four
times higher than the return to capital.
These results demonstrate why the Soviets are concerned about productivi-
ty in agriculture. The structure of the model implies that Moscow has three
potential policy options for increasing farm production: increase the capital
stock by accelerating growth in capital investment; increase the number of
workers and/or hours worked per worker, including increases in the
number of part-time workers; and increase productivity. The low return to
capital relative to alternative investments in other sectors of the economy
suggests that increasing capital investment in agriculture is not in the best
interest of the overall economy. Increasing the labor input is not feasible
because the size of the labor force in agriculture is declining as a result of
natural demographic trends, which Moscow cannot change, and the
leadership is opposed to increasing part-time employment in agriculture at
the expense of production in other sectors of the economy. The only
remaining policy option is to increase the productivity of the labor and
capital inputs.
Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
 Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89T01451 R000500520004-0
This can be accomplished if Moscow continues to push for programs and
policies designed to increase worker efficiency. Before significant progress
is possible, longstanding impediments to productivity growth must be
overcome, including:
? A weak link between the size, quality, and costs of harvests and the
financial rewards for farm workers and managers.
? The low quality and inappropriate assortment of farm machinery.
? Rural living conditions that are still too stark to encourage younger,
skilled workers to stay on the farm.
? A rural education system that is inadequate for teaching modern
agricultural practices.
Outlook
The model was used to evaluate prospects for meeting the 1986-90 Five-
Year Plan goal for agricultural output. Farm output for 1988, 1989, and
1990 was projected after making assumptions about capital and labor
growth and simulating alternative outcomes for weather and government
policy. Model simulations indicate that the Soviets would be able to meet
their plan only if the following three conditions prevail:
? At least "average" weather for 1988-90.
? Continued growth of inputs from other sectors at a rate equal to that of
recent years (4 percent in 1986), together with timely deliveries.
? Productivity gains equivalent at least to a 1-percentage-point increase
above that required to offset employment losses.
If any of these conditions are not met, the goal will be out of reach. Even
with good weather, substantial gains in productivity are required to meet
the five-year plan. Regardless of how successful ongoing and new agricul-
tural policies are, however, bad weather-especially if it occurs in two
consecutive years-could spawn an agricultural failure severe enough to
exacerbate current consumer dissatisfaction with food supplies and threat-
en the success of Gorbachev's reform effort. Although the probability that
bad weather will occur in two consecutive years is low, the impact on Soviet
domestic policy-and foreign trade-would be high.
Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89T01451 R000500520004-0
 Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
Capital
16
Material Inputs and Technology
17
A Model of Soviet Agriculture
1
Factors Influencing Performance
1
Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
 Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
Figure 3
Soviet Dependency on Other Countries for Farm Products
Composition of Hard
Agricultural Imports Currency Imports in US $
Meat and dairy products,
animal fats, eggs
Ruble Value of Agricultural
Percent Imports (All Countries)
12
Agricultural
19.4
a The category "Other" includes vegetables, fruits, sugars,
natural fibers, animal byproducts, tobacco, spices, coffee,
tea, wine, and fruit beverages.
b Preliminary.
(average) (average)
STAT
Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
 Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
Modeling Soviet Agriculture:
Isolating the Effects of Weather
Gorbachev Needs a Success in Agriculture
Agriculture will play an important role in determining
how successful General Secretary Gorbachev will be
in revitalizing the Soviet economy. The next few years
will be difficult ones for the economy as a whole as
Soviet managers and workers attempt to cope with the
numerous and wide-ranging elements of the reform
program. Gorbachev has already encountered serious
opposition to the pace of reform, and additional
resistance is expected as implementation spreads. The
General Secretary, who built his career in part as an
agricultural expert, needs a success in agriculture;
failure to improve the food supply will not only be
damaging to him politically, but could also undermine
popular support for the economic restructuring
program.
Increasing productivity in agriculture-increasing
output per unit of inputs-is as important as increas-
ing the food supply because of the high resource cost
of farm production in the Soviet Union. The food
production sector-agroindustrial complex in Soviet
parlance-is immense, claiming roughly one-third of
total annual investment (including related housing
and services) and employing nearly 30 percent of the
labor force.' Direct farm production activity alone
claims about 20 percent of annual investment and 20
percent of the labor force compared with less than 5
percent each in the United States. Despite the huge
investment in agriculture, however, the Soviet Union
must still import large quantities of agricultural prod-
ucts, particularly grain (see figure 3). Productivity
increases in agriculture would enable Gorbachev to
divert resources (labor and capital investment) from
agriculture to the industrial modernization drive and
to reduce outlays of scarce hard currency for farm
products.
' The food production sector includes not only farms but also
several branches of industry supplying farms with materials, such
as tractors and other farm machinery, repair services, and agro-
chemicals, and branches of industry that process food products.
A Model of Soviet Agriculture 2
The impact of government policies to raise agricultur-
al productivity is often hard to detect because weather
effects are so overwhelming that they obfuscate the
influences of policy changes and changes in quality
and quantity of inputs. To properly evaluate any new
program that Gorbachev may implement, it is first
necessary to isolate the effects of each of the main
factors influencing farm production.
Factors Influencing Performance
Any macroeconomic model of the agricultural sector
must account for six broad categories of factors that
influence production: capital stock, labor, material
inputs (such as manufactured fertilizers), weather,
technology, and government policy. In the Soviet case,
some of these factors are completely controlled by
Moscow, whereas others are partially controlled or
completely outside the government's influence. For
example, Moscow controls the flow of capital invest-
ment and material inputs into agriculture through the
planning process. The supply of labor, on the other
hand, is partly determined by demographic trends,
over which Moscow has no direct control. Moscow
can, however, influence the supply and "quality" of
the agricultural work force to some extent through
government policies such as those directed at relocat-
ing labor and at providing incentives to attract skilled
workers to agriculture. Weather, of course, is com-
pletely outside Moscow's control.
Capital Stock and Investment. Since 1970 the stock
of agricultural machinery, equipment, and nonresi-
dential structures has more than tripled. Fixed pro-
ductive capital in agriculture at the beginning of 1987
' The model deals strictly with agricultural output and does not
address other important components of the agroindustrial complex,
such as the food processing industry and the supply of industrial
products to farms.
Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
 Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
Figure 4
Growth of Capital Stock in Soviet
Agriculture, 1965-86
-0- Capital stock
(exclusive of livestock)
S Stock of tractors
used in agriculture
Growth rate (percent per year)
I I I I I I I I I I I I I I I I
0 1965 70 75 80 85
totaled 330 billion rubles, of which 61 percent repre-
sents nonresidential buildings and installations, 17
percent represents agricultural machinery and equip-
ment, 3.8 percent represents transportation equip-
ment, 0.5 percent represents draft animals, 9.5 per-
cent represents productive livestock, and 4.6 percent
represents perennial plantings.' But, while the overall
size of the capital stock has been growing, the rate at
which it is growing has been slowing since the mid-
1970s (see figure 4). Growth of the stock of tractors in
agriculture, for example, has fallen from about 3
percent per year in the mid-1970s to nearly zero in
1986.
Figure 5
Growth of Capital Investment and Employment
in Soviet Agriculture, 1965-86
-S Employment in hours worked
(excluding the private sector)a
I I I I I I I I I I I I I I I I I
70 75 80 85
a The USSR does not report statistics on hours worked in the
private sector, but Western estimates have remained relatively
stable during this time period.
Because technological advances in design and engi-
neering are embodied in new capital, capital invest-
ment is the carrier of much of the new technology
going into agriculture.4 Growth of investment in agri-
culture fell from a high of 15 percent in 1971 to less
than zero in 1984 (see figure 5). In 1986, however,
investment growth rebounded to a rate approximately
equal to that of the mid-1970s (6 percent).
Capital investment in agriculture includes new machinery and
equipment, new construction and installation of new farm buildings
(including new livestock rearing facilities, irrigation and drainage
systems, and agricultural research institutions), net additions to
livestock, and capital repair.
STAT
STAT
STAT
Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
 Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
Labor. The size of the agricultural work force in the
Soviet Union is shrinking slowly, as is its share of total
employment in the economy as a whole. Over 35
million people are presently employed in agriculture,
and many more engage in part-time farm work and
gardening for personal consumption. During the
1970s there was little change in the size of the labor
supply in terms of hours worked. Since 1984, however,
agricultural employment has been decreasing at about
1 to 2 percent per year (see figure 5). Unless the
Soviets do something to spur labor productivity, labor
requirements in the future will exceed the supply and
possibly result in a serious labor shortage in agricul-
ture. Moscow has issued numerous decrees to improve
the productivity of the farm labor force, but the
decrees have not yet had a widespread positive effect.'
Material Inputs. Material inputs are produced by
nonagricultural sectors of the economy for use in the
agricultural sector, exclusive of capital investment
goods. They include chemicals, fuels, electric power,
animal feed supplements (including byproducts from
food processing), and machinery spare parts.
Among the most important are manufactured fertiliz-
ers and agrochemicals. Aided by large imports of
Western equipment and technology during the 1970s,
the Soviet Union is presently the world's leading
producer of manufactured fertilizers (nitrogen, phos-
phate, and potassium). Increases in crop yields since
1960 are directly attributable to the rapid growth in
fertilizer deliveries. After 1975, however, growth of
deliveries to agriculture slowed (see figure 6) because
of lags in expanding production capacities and under-
utilization of existing capacities, which were caused
by shortages of skilled labor, equipment failures, and
transportation problems. Since 1979, growth of fertil-
izer deliveries has fluctuated at about half the rate of
growth of the early 1970s.
Chemical control of insect pests, plant diseases, and
weeds has also been an important factor in increased
'See Ann Goodman, Margaret Hughes, and Gertrude Schroeder,
"Raising the Efficiency of Soviet Farm Labor: Problems and
Prospects," in Gorbachev's Economic Plans, Volume 2 , U.S.
Congress, Joint Economic Committee, Washington, DC: U.S.
Government Printing Office, November 1987, pp. 100-124.
Figure 6
Growth of Fertilizer Deliveries
to Agriculture, 1965-86
-10 1965 70 75 80 85
yields, particularly for grain. Since 1984 the Soviets
have made special efforts to increase purchases of
sophisticated forms of Western herbicides, insecti-
cides, and fungicides. In contrast to fertilizers, more
than half of the pesticides used in the USSR are
imported from the West and from Eastern Europe.
Although the use of chemical pesticides has increased
in the Soviet Union, the average application rate is
still far below that of Western countries.
Technology. Technology in agriculture encompasses
both enhancements to resources, such as new seed
varieties and livestock breeds, and innovations in the
way resources are used, such as crop rotation schemes
and management of livestock facilities. The USSR
pursues research and development in many areas of
Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
 Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89T01451 R000500520004-0
As Moscow has learned, simply increasing supplies of
physical inputs has not been sufficient to meet the
growing demand for agricultural products. Increases
in productivity are also required. Moscow has no
direct control over productivity growth, and must
depend in part on the diffusion of successful techno-
logical innovations. The USSR pursues research and
development in many areas of farm production and is
also incorporating modern aspects of Western agro-
technology in an attempt to improve productivity:
? Plant breeding. Work on wheat breeding alone is
carried out at nearly 50 institutions. The Soviet
wheat breeding program maintains a germ plasm
collection that contains roughly 40,000 wheat speci-
mens, probably the largest collection in the world.
? Agrochemicals. Advanced chemical fertilizers,
growth stimulants, and pesticides specific to soil
and climate conditions in the USSR are being
developed. Facilities for producing modern agro-
chemicals are also being imported from the West.
farm production, including plant breeding, develop-
ment of chemical fertilizers and pesticides, the design
of agricultural machinery, livestock breeding, and
genetic engineering (see inset).
According to Western scientists, agricultural research
facilities in the USSR range from antiquated to state
of the art. The Soviet Union is at least 10 to 15 years
behind the West in developing and applying agrotech-
nologies. As in the rest of the economy, Soviet
agriculture suffers from a serious lag between devel-
opment of technology and its application. This condi-
tion is exacerbated in agriculture because of the lack
of interdisciplinary teamwork. For example, Soviet
plant breeders do not work closely with plant patholo-
gists and entomologists. As a result, real technological
progress is slow.
? New designs for agricultural machinery. Soviet
engineers are developing agricultural equipment
suitable for tillage techniques needed to conserve
moisture and prevent soil erosion, grain combines
and other harvesting equipment to reduce losses
during harvest, more energy efficient drying equip-
ment, and controlled atmosphere storage.
? Livestock research. Soviet efforts in livestock breed-
ing have focused on developing breeds of cattle and
hogs that will be more efcient-more meat or
milk per animal-and have higher reproduction
rates. Research is also conducted on better methods
for rearing livestock, such as ways to increase
production, harvesting, storage, and utilization of
livestock feed, improved animal shelters, and pro-
phylactic care of animals.
? Genetic engineering. Soviet scientists are placing
considerable attention on agricultural application
of genetic engineering. Progress is occurring in
development of hormones, protein supplements,
antibiotics, and improved vaccines.
Weather. Since a large part of Soviet farm production
occurs in risk-prone areas, year-to-year fluctuations in
weather conditions dramatically affect the volume of
farm output. Most of the agricultural area has a
generally harsh and variable climate. Only about 27
percent of the total land area of the USSR is suitable
for farming. Of this, slightly more than one-third is
arable; the remainder is in meadow, pasture, orchard,
vineyard, or is idle.' More than half of this arable land
lacks adequate and reliable moisture. In general,
areas warm enough to foster plant growth tend to
suffer from lack of moisture, and areas with sufficient
moisture are predominantly located in the cold, north-
ern latitudes where the growing season is short.
Livestock production is less influenced by weather
than crop production, but temperature extremes can
STAT
Declassified in Part - Sanitized Copy Approved for Release 2012/12/03 : CIA-RDP89T01451 R000500520004-0
 Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
have an adverse effect on animal health and produc-
tivity, and weather indirectly affects livestock produc-
tion through its effect on feed availability.
Government Policy. Since the mid-1970s, government
programs have emphasized productivity growth as a
means to increase farm output and-at the same
time-conserve on resources going to agriculture.
Moscow has issued numerous decrees over the last
decade that were intended to improve productivity
and to reduce cost, waste, and the need for agricultur-
al imports. Efforts have focused on labor incentives,
planning and organization, changes in the manage-
ment structure, and the restructuring of investment
allocations within the agroindustrial complex. Since
the initiation of the Brezhnev Food Program in mid-
1982 and the recent campaign to "intensify" agricul-
ture, the flow of fertilizers, pesticides, and other
industrial goods to agriculture has accelerated, and
more care has been taken to apply them where and
when they would do the most good.
The Model
A mathematical model was developed to separate the
effects of weather from the effects of other factors. Of
the six broad categories listed above, capital, labor,
weather, and productivity changes resulting from
government policies are accounted for in the model
explicitly. The capital stock variable serves as a proxy
for the two remaining factors-material inputs and
technological progress. The model predicts the value
of net agricultural output, defined as the sum of the
value of total crop production (less seed and waste)
and the net value of livestock production (including
inventory, excluding feed) measured in constant 1982
prices (see appendix B for a more complete defini-
tion).' The model is used to generate an historical
output series that is adjusted for weather; to estimate
economic gains and losses attributable to weather; to
estimate the trend in agricultural growth owing to
' Previous models have been developed to evaluate prospects for
grain production only. See Russell A. Ambroziak and David W.
Carey, "Climate and Grain Production in the Soviet Union," in
Soviet Economy in the 1980s: Problems and Prospects, Part 2,
U.S. Congress, Joint Economic Committee, Washington, DC: U.S.
Government Printing Office, December 1982, pp. 10-12,
nonweather factors alone; and to evaluate prospects
for meeting Soviet plan targets.
The model was developed as an aggregate production
function for agriculture! As in any aggregate produc-
tion function, the factors of production are themselves
gross aggregates. Capital is the value of the capital
stock used in agriculture, excluding livestock. This
includes the undepreciated value of all machinery and
equipment, tools, vehicles, and value of buildings and
structures, measured as a single input denominated in
comparable rubles. Labor is total employment in
agriculture-socialized and private-measured in
man-hours with no regard to skill level or other
aspects of labor quality. Similarly, the weather vari-
ables are also gross aggregates. Two weather variables
are used in the model: the average winter temperature
and the ratio of temperature to precipitation for late
spring and early summer.
Q = a, a2(W) a3(P) KK Lt-P c,
where Q is output; K and L are capital and labor
inputs, respectively; 0 is the capital elasticity parame-
ter; a, is a scale adjustment that reconciles the units
of measure used for Q, K, and L; a2(W) is the weather
function; a,(P) is a function that reflects potential
productivity changes linked to changes in government
policy; and c is a stochastic error term? With this
model specification, the capital-labor ratio establishes
the trend of agricultural output over time, while
fluctuations about the trend caused by weather and
changes in government policy are modeled by upward
and downward shifts controlled by the functions
a2(W) and a,(P). Appendix A includes a detailed
discussion of the model development, and data used to
fit the model are presented in appendix B.
'The model is currently used in CIA's macroeconomic model of the
Soviet Union. See Robert L. Kellogg "Modeling Soviet Moderniza-
tion: An Economy in Transition," Soviet Economy, 4,1: 36-56,
1988.
' Capital elasticity is the percentage change in output that results
when capital is increased 1 percent, holding all other factors
constant.
Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
 Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
Modeling Policy and Productivity Changes
The function a,(P) was created to reflect relative
changes in productivity due to government policy
actions." In a centrally planned economy like the
Soviet Union, productivity changes arise either direct-
ly or indirectly as a result of government policy
actions. However, modeling the impact of government
activity is difficult because-unlike weather, capital,
and labor-policy variables cannot be measured. Nev-
ertheless, a subjective estimate can be made of the
relative changes in productivity expected from gov-
ernment policies.
The function a,(P) was developed in this way to reflect
the likely impact on agriculture of government poli-
cies and programs for the economy as a whole as well
as for specific programs in agriculture. The 1968-78
period was selected as the base period, and productivi-
ty changes for 1979-87 were modeled relative to this
base. It was thus assumed that productivity growth
arising from changes in government policy during
1968-78 was fairly steady year to year. Most of this
period was free of sharp policy changes in
agriculture."
Beginning in 1976, however, Moscow attempted to
shift from an extensive growth pattern to an intensive
growth strategy for the economy as a whole. In doing
so, it precipitated the 1976-82 industrial growth slow-
down." The problems in industry-including those
sectors supporting agriculture-were most severe dur-
ing 1979-82 (see figure 7). In addition, transportation
organizations were increasingly unable to keep pace
with the growing requirements for timely deliveries
of industrial goods to farms and for shipping farm
products to processors." As a result, growth of
10 The general concept of productivity-increased output with no
change in the quantity of inputs used-is appealed to in this
context. The productivity measure to which this concept best
corresponds is total factor productivity (see subsection, "Total
Factor Productivity Adjusted for Weather").
" See David M. Schoonover, "Agriculture and the Grain Trade-
Overview," in Soviet Economy in the 1980s: Problems and Pros-
pects, Part 2, Joint Economic Committee, Congress of the United
States, December 1982, pp. 1-6.
'Z See Gertrude E. Schroeder, "The Slowdown in Soviet Industry,
1976-82," Soviet Economy 1,1:42-74, January-March 1985.
"See Judith Flynn and Barbara Severin, "Soviet Agricultural
Transport: Bottlenecks To Continue," in Gorbachev's Economic
Plans, Volume 2, U.S. Congress, Joint Economic Committee,
Washington, DC: U.S. Government Printing Office, November
1987, pp. 62-78.
Figure 7
Industrial Growth by Sector, 1971-868
Q Industry
Machinery
Basic materials
Chemical
Percent
a Based on estimates of value added
at 1982 factor cost.
b Average annual growth rates.
deliveries of goods and services to agriculture lagged
(see table 1). A statistical test determined that factors
other than capital, labor, and weather were responsi-
ble for a growth slowdown in agriculture during 1979-
82, similar to that observed for industry, suggesting
that the problems in industry extended to agriculture
as well (see inset).
But the slowdown in growth of deliveries from indus-
try was not the only policy-related factor affecting
agriculture during this period. It was clear that
government policies specific to agriculture were
flawed:
STAT
Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
 Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
Table 1
Average Annual Growth Rates of
Selected Inputs to Agriculture
Tractor deliveries to
agriculture
Grain combine deliveries
to agriculture
Fertilizer deliveries to
agriculture b
10-Year
Growth
Recovery
Period Before
Slowdown
and Post-
Growth
Period
Slowdown
Slowdown
(1979-82)
Period
(1969-78)
(1983-86)
2.3
-1.4
3.1
2.0
0.0
0.2
10.1
2.6
7.4
Current purchases include chemical fertilizers, electric power, fuel
and lubricants, machinery repair, and animal feed supplements.
Capital investment goods are not included in current purchases.
Included are a small amount of nutrients used in feed additives.
Phosphate fertilizers include ground phosphate rock.
? Investment resources going to agriculture were
wastefully allocated and inefficiently utilized. The
construction of livestock facilities had been overem-
phasized, for instance, while the share of investment
allocated to rural housing was cut. Soviet authors
have complained, moreover, about losses of agricul-
tural products (20 to 25 percent) because construc-
tion of storage facilities and rural roads was
neglected.
? Agricultural machinery downtime increased, effi-
ciency in the use of inputs-especially machinery,
equipment, and fertilizers-declined, and growth in
livestock herds outstripped growth in feed
availability."
" See Barbara Severin, "Solving the Soviet Livestock Feed Dilem-
ma: Key to Meeting Food Program Targets," in Gorbachev's
Economic Plans, Volume 2, U.S. Congress, Joint Economic Com-
mittee, Washington, DC: U.S. Government Printing Office, No-
vember 1987, pp. 45-61.
Testing for the Effects of the Industrial Growth
Slowdown on Soviet Agricultural Performance
In developing the model, it became clear that weather
factors alone were not sufficient to explain agricul-
ture's dismal showing during the 1979-82 period. A
statistical test was devised to determine ifthe indus-
trial growth slowdown had a depressing effect on
Soviet agriculture independent of capital and labor
inputs and weather factors. The test was conducted
by replacing the function a3(P) by a dummy variable,
which consisted of I's for the years 1979-82 and 0's
for all other years, and reestimating the model. The
results revealed that the coefficient for the dummy
variable was highly significant statistically and had a
negative sign, suggesting that the slowdown in agri-
culture during this period was associated with the
industrial growth slowdown and may have been
caused by it at least in part.
? Producing and marketing farm products was be-
coming increasingly more difficult to synchronize as
the size and interdependence of the economy
increased.
As the difficulties in agriculture intensified, Moscow
promulgated new policies in attempts to reverse the
decline in productivity. The Brezhnev Food Program
of May 1982 was the most comprehensive of these
measures (see inset). Although the Food Program
resulted in some improvements in productivity, it fell
short of the desired results." Since Brezhnev's death
in late 1982, Gorbachev has used his influence in the
leadership to reshape the program to reflect more
closely his own views and priorities. His most recent
strategy to motivate the individual farmworker has
15 See Penelope Doolittle and Margaret Hughes, "Gorbachev's
Agricultural Policy: Building on the Brezhnev Food Program," in
Gorbachev's Economic Plans, Volume 2, U.S. Congress, Joint
Economic Committee, Washington, DC: U.S. Government Printing
Office, November 1987, pp. 26-44.
Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
 Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
The Brezhnev Food Program was unveiled in May
1982. Key features of the program included:
? "Unified management" of food production, which
ultimately resulted in the establishment of the
State Agroindustrial Committee, Gosagroprom, in
November 1985.
? Reallocation of investment resources within the
agroindustrial complex to upgrade the system for
handling, storing, and processing food and to im-
prove housing and living conditions in the
countryside.
? An increase in .financial as well as nonmonetary
incentives intended to attract skilled workers to
agriculture and encourage workers from southern,
labor-surplus regions to resettle in northern areas,
where labor is insufficient to meet demand.
been to expand the use of the collective contract,
which organizes workers into teams operating under
contract to the farm and pays them on the basis of
what they actually produce. A deadline of December
1988 was set for transferring all farm labor to the
collective contract system.
In addition to measures targeted at increasing worker
productivity, Gorbachev has given the "intensive tech-
nology" program a high priority. Intensive technol-
ogy, as defined by the USSR, includes many practices
routinely performed in the West-use of high-yield
varieties, planting after fallow where possible, imple-
menting efficient field operation schedules, and exten-
sive use of agrochemicals. The program commenced
in 1984 on selected test sites scattered over the Soviet
Union. Intensive technology practices were increased
to include almost 17 million hectares in 1985, and
expanded again in 1986 to about 30 million hectares.
In 1987 the intensive technology area included 35
million hectares, and plans call for the program to
encompass 50 million hectares by 1990.
Table 2
Comparison of Actual Data
to Model Predictions
Year Farm Output Annual Growth Rates
(t) (billion rubles) (percent)
Actual Predicted Actual Predicted
(Q1) (QJ (Q/Q,-,) (QQ/Q-J
1968
105.061
106.393
6.1
7.4
1969
100.303
100.732
-4.5
-4.1
1973
121.807
119.841
16.4
14.5
1974
119.629
121.073
-1.8
-0.6
1975
109.410
109.094
-8.5
-8.8
1976
118.060
114.802
7.9
4.9
1984
128.046
129.277
-0.5
0.5
1985
125.992
127.435
-1.6
-0.5
1986
136.287
134.448
8.2
6.7
1987
132.032-
131.575 b
-3.1
-3.5
a Preliminary.
b The predicted value for 1987 was obtained by assuming that the
trend in employment growth during 1984-86 continues
through 1987.
These policy changes were captured in the function
a3(P) by a variable named PRODCHNG (see appen-
dix A for the complete functional form of the model).
PRODCHNG was defined subjectively so as to reflect
the relative impact that changes in government poli-
cies since 1978 might have had on productivity
growth in agriculture. The variable PRODCHNG
was assigned a value of zero for the 1968-78 base
period. For 1979, the variable was assigned a value of
- 1 to simulate a decrease in productivity growth
relative to the base period as the industrial growth
slowdown and flawed agricultural policies began to
affect production. The variable was assigned the
values -2 in 1980 and -3 in 1981 and 1982 to
Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
 Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
Table 3
Decomposition of the Model Into Functional Components
and Calculation of Weather-Adjusted Output
1969
1.6607
0.9468
1.0000
2.0404
31.3971
0.99574
100.732
100.303
105.939
1970
1.6607
1.0348
1.0000
2.0646
31.6964
1.00064
112.464
112.535
108.749
1971
1.6607
1.0155
1.0000
2.1006
31.2490
1.00615
110.707
111.388
109.685
1983
1.6607
1.0169
0.9523
2.5206
32.2409
0.98418
130.706
128.638
126.500
1984
1.6607
0.9571
1.0000
2.5479
31.9192
0.99048
129.277
128.046
133.780
1985
1.6607
0.9506
1.0000
2.5707
31.3996
0.98867
127.435
125.992
132.537
Note: Q represents the model predictions for farm output,
and is equal to a1a2(W)a3(P)KKL1-0. Q is actual farm
output, also equal to Or- Farm output after adjusting for
weather is Q*, equal to Q/a2(W).
simulate a worsening situation. Under the assumption
that the Brezhnev Food Program and subsequent
programs helped to reverse the decline in productivity
growth, PRODCHNG was assigned the values -1 in
1983 and zero again in 1984 and 1985. To simulate
gains from the intensive technology campaign in 1986
and 1987, PRODCHNG was given the value + 1 for
these two years.16
ed using historical data for 1968 through 1986." The
model fits the historical data quite well (see figure 1),
and even predicts historical growth rates closely (see
table 2). All variables were statistically significant at
the 0.0001 level (that is, the probability of falsely
rejecting the null hypothesis that a parameter is zero
is less than 1 in 10,000). In addition to statistical
significance, the signs of the parameters all matched
a priori expectations. Time series of the functional
components of the model are presented in table 3, and
Weather-Adjusted Output
After incorporating the function a3(P) as derived
above into the model, model parameters were estimat-
1' There is potential for multicolinearity between the functions
a,(W) and a,(P). If this were the case, it would not be possible to
distinguish the effects of weather from policy-related declines in
productivity during the 1979-83 period. Analysis included in appen-
dix A, however, demonstrates that there is no empirical evidence
that multicolinearity is a problem in this case.
" Data for 1987 were not used to estimate parameters because
reliable estimates of employment were not available. Weather data
for 1987 and capital available at the beginning of the year were
available and were used in conjunction with the model to calculate
a model prediction for 1987, which was very close to the prelimi-
nary estimate of farm output for 1987 (see table 2).
Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
 Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89T01451 R000500520004-0
An Economic Measure of the Effects of Weather on
Soviet Agricultural Performance
One way to measure the effects of weather on agricul-
tural production is to estimate output for each year
using "average" weather data and contrast it to
model predictions made using actual weather data.
Output corresponding to "average" weather was de-
termined by solving the model using the mean values
of the two weather variables. These mean values were
based on weather data for the past 20 years.
Comparison of the "average-weather" predictions to
`actual weather"predictions reveals how much loss
or gain may have occurred each year as a result of
weather e) ects alone (see figure 8). Overall, losses
exceeded gains by 41.3 billion rubles over the 20 year
period. Weather-related losses in excess of 2 billion
rubles occurred in eight of the 20 years, whereas
weather-related gains of more than 2 billion rubles
occurred in only three years. These results suggest
that weather-related losses can be expected to occur
more frequently than weather-related gains.
Significant weather-related losses were estimated for
each of the last four years (1984-87). Two of the
years-1985 and 1987-were among the three cold-
est winters in the last 20 years, and the two remaining
years (1984 and 1986) were among the five years with
the hottest and driest conditions during spring and
early summer (April-July) (see appendix B).
statistical properties of the .model parameters are
presented in appendix A.
The model can be used to isolate the effects of
weather on agricultural production, and thus reveal
the relationship between farm output and nonweather
factors. One approach is to solve the model using
"average" weather and compare the results to actual
performance (see inset). The approach taken here was
to adjust the output series for weather, thereby
creating a "weather-adjusted" measure of farm out-
put. This weather-adjusted series, Q*, was derived by
dividing actual output by the model's prediction of the
Figure 8
Estimated Agricultural Losses and Gains Due
to Weather, 1968-87
? Gains
_ Losses
year-to-year fluctuations that are due to weather, as
follows (also see table 3):'a
Q*= Q
a2(W)
1' Actual output (Q) is represented algebraically by the model as
follows:
Q = a, a2(V) a3(P) KB L'-e e,
oz (W), we have
Q
a2(W)
a,
a2(W) a3(P) KB Lt-B e
a2(W)
at a3(P) KB Li-a e = Q*
Q
a2(W)
STAT
Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89T01451 R000500520004-0
 Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89T01451 R000500520004-0
Figure 9
Weather-Adjusted Farm Output, 1965-87-
-0-- Weather-adjusted data
S Actual data
T1111I1111I1111I1111I11
90 1965 70 75 80 85 87b
a Net of feed, seed, and waste.
b Preliminary
This adjustment for weather is similar in concept to
the seasonal adjustment applied to many Western
economic aggregates, except that the "season" ex-
tends through 20 years. Because the adjustment uses
only weather variables, the resulting series retains
year-to-year changes stemming from the growth of
inputs-capital, labor, and material-as well as pro-
ductivity growth, including technological progress and
"human factor" effects. The weather-adjusted series
is contrasted with actual farm output in figure 9.
The pattern of year-to-year changes in weather-
adjusted output corresponds to changes in government
policy (see figure 10). The 1968-78 period is marked
by a steady-although very gradual-increase in
output, reflecting relatively stable growth of inputs,
Figure 10
Long-Run Trend in Farm Output
After Adjusting for Weather, 1968-87
I I I I I I I I I I I I I I I I I I 1-i
100 1968 70 75 80 85 878
Note: Closed circles denote the time series
used to calculate the trend.
steady but slow technological progress, and the ab-
sence of sharp swings in government policy. A depar-
ture from this pattern became apparent in 1979, when
weather-adjusted output dropped precipitously.
Weather-adjusted output continued to fall through
1981 and showed only slight improvement in 1982.
This slump in agriculture corresponds to the worst of
the industrial growth slowdown period, discussed pre-
viously. In 1983, however, there was a remarkable
recovery, possibly reflecting improvements in efficien-
cy stemming from enactment of the Brezhnev Food
Program the previous year. By 1984, performance had
nearly returned to the pre-1979 trend, and perfor-
mance was clearly back on trend again in 1986 and
1987.19
318100 8-88 STAT
"This relationship between weather-adjusted farm output and
nonweather factors is not dependent on a,(P). Nearly identical
results were obtained when model parameters were reestimated
after dropping a,(P) from the model and excluding the 1979-82
period from the dataset (see appendix A for more details).
STAT
Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89T01451 R000500520004-0
 Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
The rate at which weather-adjusted output is increas-
ing has important ramifications for Gorbachev's agri-
culture policy. The long-run trend in weather-adjust-
ed output was measured by regressing weather-
adjusted output against time for the 1968-78 period.
The regression equation is (standard errors in
parentheses):
weather-adjusted output = 102.61668
+ 1.94925 X t
(.817536)
(.12054)
where t is time in years (t = 1 for 1968). This
equation estimates that weather-adjusted farm output
has been increasing only 1.9 billion rubles per year.
Although 1984-87 were not included in the regression,
weather-adjusted output in these years conforms
closely to the pre-1979 trend (see figure 10). This gain
in output is nearly offset by increased costs of inputs.
For example, assuming no changes in the growth of
labor and capital or changes in productivity growth,
the Soviets will have to spend about 1.3 billion rubles
in current purchases alone to obtain the additional 1.9
billion rubles of weather-adjusted output.20
Returns to Capital and Labor
The model estimates the return to capital in agricul-
ture by the parameter a-the capital elasticity pa-
rameter. The capital elasticity was estimated to be
0.17 percent, indicating that a 1-percent increase in
capital produces only a 0.17-percent increase in farm
output, holding all other factors constant.21 This mea-
sure of the capital elasticity represents the average
2p This analysis is based on the judgment that, in order to sustain
weather-adjusted output growth at the 1968-78 trend, growth of all
inputs-including current purchases-must also be sustained. The
pre-1979 trend for the value of current purchases increased about
1.3 billion rubles per year. The time trend equation is
where t is time in years (t = I for 1969).
2' The parameter (3 can also be interpreted as the relative share of
the total output contributed by capital. According to this estimate
of (3, capital accounts for 17 percent of the value of farm output.
Using a different estimation method, Diamond and Krueger ("Re-
cent Developments in Output and Productivity in Soviet Agricul-
ture," in Soviet Economic Prospectsfor the Seventies, US Con-
gress, Joint Economic Committee, Washington, D.C.: US
Government Printing Office, June 1973, p. 329) estimated the
relative share of capital in total output to be 15 percent.
Table 4
Capital Elasticity Estimates for Agriculture
and Other Producing Sectors a
.455
.523
.728
.423
.039
.892
a The capital elasticity parameters were estimated by fitting a
modified Cobb-Douglas production function with data on capital,
labor, and output for 1969-85 (see Kellogg, op. cit.). The general
form of the production function was:
Q = output measured in 1982 rubles at factor cost
K = average capital stock in 1973 rubles
L = employment in man-hours
a(t) = scale adjustment and adjustment for 1976-82 industrial
growth slowdown period
(3 = capital elasticity
b output for all groups except agriculture is measured in value-
added units. Output for agriculture is not value added, since it
includes the value of purchases from other sectors (such as fuels and
agrochemicals). Thus, the agricultural capital elasticity is not
completely comparable to the others. Since the value of purchases
from other sectors has been growing faster than the value of farm
output, the capital elasticity in value-added terms would be smaller
than reported here.
Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
 Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
The Don Combine An Attempt To Increase Farm
Productivity Through Capital Investment
Even though the return to additional capital has been
low in agriculture, it may still be rational to attempt
to boost farm output by increasing capital investment.
If old machinery and equipment were replaced by the
right kinds of modern, efficient farm machinery, it
would theoretically be possible to increase the return
to the new capital substantially above returns regis-
tered in the past. However, the Soviet system seems
incapable of making such gains very quickly or
easily. Consider, for example, the case of the Don
grain combine.
Soviet planners in the late 1970s assigned top priority
to modernizing the fleet of grain harvesting combines.
Their intent was to replace their obsolete fleet of
combines-which prolonged the harvest period and
lost substantial quantities of grain during harvest-
ing-with new, modern combines. The new Don 1500
combine, which was to be 50 to 70 percent more
productive than existing models, was designed for use
not only for harvesting grain, but also for harvesting
seed grasses, soybeans, sunflower seeds, and corn. In
his report to the 27th Party Congress, Gorbachev
claimed that the use of this machine in the 12th FYP
period would reduce grain losses by millions of
metric tons and eliminate the need for 400,000
machine operators, equal to nearly 15 percent of the
present force.
Problems in manufacture and delivery have been
extensive, however, and the Don thus far has had
little positive impact on grain harvesting:
? Design flaws made initial models too heavy to
operate in any but the most ideal ground condi-
tions. In subsequent models, engine horsepower was
increased and the weight reduced from 18 tons to
13 tons.
? Parts for the Don were supplied by 500 separate
industrial enterprises, and many deliveries were
late. Moreover, the quality of component parts was
low; tests in 1986 showed that 80 percent of break-
downs were due to flaws in parts and accessories.
? The first large shipment to consumers in June
1987-3,000 combines-consisted largely of ma-
chines that were missing accessories and parts. At
least half had no headers for cutting crops and were
therefore useless.
Nor have Soviet farmers been favorably impressed
with the Don. A July 1987 Pravda article stated that
users were finding the Don too heavy, too costly, and
too complicated to operate and repair. One collective
farm official complained that, of the 18 Dons pur-
chased by his farm, only seven were operating-the
rest had been cannibalized for parts.
Under development since the late 1970s, the Don was
put into series production in September 1986.
return to additional capital over the past 20 years. By
this measure, the return to capital in agriculture is
lower than in any other productive sector of the Soviet
economy except the fuels branch of industry (see table
4). Estimates of capital elasticities in industry (exclud-
ing the fuels sector) are roughly three to five times as
great as in agriculture. Some of the reasons for the
low return to capital are revealed in the difficulties
the Soviets have had introducing a new, modern fleet
of grain combines (see inset).
The return to labor in agriculture, on the other hand,
is over four times higher than the return to capital.
The labor elasticity is estimated to be 0.83 (one minus
the capital elasticity). Unlike capital, however, the
labor input is gradually declining. Thus, the high
return to labor works to the Soviets' disadvantage.
That is, a 1-percent decline in agricultural employ-
ment (holding other inputs constant) produces a 0.83-
percent decline in farm output, which represents a
substantial marginal loss.
Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
 Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
Table 5
Derivation of Total Factor Productivity Index
Value-Added
Capital
Labor
Combined Inputs
Factor Productivity
Weather-Adjusted
Output Index
Index
(K1)
Index
(LI)
Index
Annual
Growth Rate
(percent)
Index
Annual
Growth Rate
(percent)
1968
1.000
1.000
1.000
1.000
1.000
1969
0.993
1.095
0.973
0.993
-0.64
1.000
0.03
1970
1.018
1.174
0.985
1.015
2.16
1.003
0.32
1971
1.018
1.301
0.968
1.018
0.33
1.000
-0.33
1972
1.015
1.444
0.970
1.038
1.94
0.978
-2.16
1973
1.055
1.603
0.978
1.064
2.51
0.992
1.38
1974
1.033
1.793
0.984
1.089
2.42
0.948
-4.44
1975
1.050
2.015
0.973
1.101
1.11
0.953
0.55
1976
1.104
2.238
0.962
1.110
0.82
0.994
4.31
1977
1.059
2.428
0.968
1.132
1.96
0.935
-5.92
1978
1.097
2.650
0.975
1.156
2.08
0.949
1.48
1979
1.005
2.873
0.969
1.166
0.87
0.861
-9.22
1980
0.972
3.095
0.971
1.183
1.46
0.821
-4.65
1981
0.932
3.333
0.974
1.200
1.48
0.776
-5.49
1982
0.949
3.571
0.993
1.234
2.82
0.769
-0.96
1983
1.059
3.841
1.005
1.262
2.27
0.838
9.05
1984
1.128
4.095
0.993
1.263
0.09
0.893
6.49
1985
1.096
4.317
0.974
1.254
-0.74
0.873
-2.18
1986
1.186
4.539
0.959
1.249
-0.39
0.949
8.71
1987
1.149
4.761
0.941
1.239
-0.79
0.926
-2.40
Sources: The value-added, weather-adjusted output index is from
appendix B, table B-5. The capital index was obtained by dividing
beginning-of-year capital by the value for 1968 (see table B-1 for
original capital series). The labor index was obtained by dividing
average annual agricultural employment by the value for 1968 (see
table B-3 for original employment series). The combined inputs
index was calculated as Kl?"Ll-8J. The total factor productivity
index was calculated by dividing the output index by the combined
inputs index.
Note: The combined inputs index included only capital and labor
because the model provided estimates of the factor shares-17
percent for capital and 83 percent for labor. Current purchases-
representing material inputs such as fuels and agrochemicals-
were subtracted from gross output prior to the calculation. Land
was excluded from the calculation entirely; however, much of the
increase in the services from land in the last 20 years is included in
capital because of the huge capital investment expenditures allocat-
ed to land reclamation.
These results demonstrate why the Soviets are con-
cerned about productivity in agriculture. The struc-
ture of the model implies that Moscow has three
potential policy options for increasing farm produc-
tion: increase the capital stock by accelerating growth
in capital investment; increase the number of workers
and/or hours worked per worker, including increases
in the number of part-time workers; and increase
productivity. The low return to capital relative to
alternative investments in other sectors of the econo-
my suggests that increasing capital investment in
agriculture is not in the best interest of the overall
economy. Increasing the labor input is not feasible,
because the size of the labor force in agriculture is
declining as a result of natural demographic trends,
which Moscow is powerless to change, and the leader-
ship is opposed to increasing part-time employment in
agriculture at the expense of production in other
Declassified in Part - Sanitized Copy Approved for Release 2012/12/03 : CIA-RDP89TO1451 R000500520004-0
 Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
sectors of the economy. The only remaining policy
option is to increase the productivity of the inputs, in
particular, the productivity of farmworkers.
Total Factor Productivity Adjusted for Weather
Productivity can be measured in several ways. The
measure used extensively by the Soviets is labor
productivity, which is estimated by dividing total
output by the labor used to produce it. This approach
can be misleading, however, because it fails to ac-
count for the capital cost. Another measure is capital
productivity, determined as the ratio of output to the
value of the capital used to produce it. Similarly,
capital productivity ignores labor as a source of
productivity.
Total factor productivity is a measure that accounts
for both capital and labor growth. It is calculated by
dividing total output by a measure of combined
inputs, as follows:
KK Lt P
where Q, K, and L are indexes (with the same base
year) for value-added output, capital, and labor, re-
spectively, and R is capital's share of total output. By
definition, then, total factor productivity growth in-
cludes all sources of output growth other than in-
creases in labor and capital, including: technological
progress, human factor effects, labor quality changes,
capital quality changes not reflected in the measure-
ment of capital, and even gains and losses attributable
to the weather.
For agriculture, it is useful to refine the calculation
further by adjusting for weather so that productivity
from remaining sources can be examined. This was
accomplished by substituting the weather-adjusted
output series-Q*-for Q in the above equation (see
table 5). Adjustment of Q* to a value-added measure
was made according to the method presented in
appendix B. The model's estimate of R-17 percent-
was used as the relative share of capital.
This weather-adjusted measure of total factor produc-
tivity reveals that the Soviets have made respectable
gains in agricultural productivity in recent years-the
annual growth rate for total factor productivity for
1984-87 averaged 2.7 percent (see figure 11). Not only
Figure 11
Total Factor Productivity in
Soviet Agriculture, 1968-87
Combined inputs
Value-added, weather-adjusted
farm output
S
-- Total factor productivity
Index: 1968=1
I I I I I I I I I I I I I I I I I I I I
0.7 1968 70 75 80 85 87a
Slump
period
318101 8-88 STAT
has weather-adjusted output been increasing since
1982, but the growth of combined inputs (capital and
labor) leveled off in 1983 and 1984 and has since been
gradually declining. Since input growth is likely to
continue to slow, further gains in productivity will be
required to maintain or increase output growth.
Prospects for the Future:
Can the 1986-90 Plan Still Be Met?
The goal for Soviet agriculture as stated in the 12th
Five-Year Plan (FYP) is "that the average annual
volume of agricultural output in 1986-90 should be
Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
 Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
increased by 14.4 percent over the previous five-year
period." 22 In terms of the output measure used in this
study, a 14.4-percent increase represents an average
output of 140.9 billion rubles per year, which substan-
tially exceeds expectations based on past perfor-
mance.23 Average production in 1986 and 1987-
134.2 billion rubles-was well below the goal (see
inset). If the Soviets are to meet their FYP goal,
output during 1988-90 must average 145.5 billion
rubles.
The model was used to evaluate prospects for meeting
the plan goal. Farm output for 1988, 1989, and 1990
was projected after making assumptions about capital
and labor growth and simulating alternative outcomes
for weather and government policy.
Assumptions
Capital. The growth of capital for 1988-90 is assumed
to be 5 percent per year, equal to the average for 1985
and 1986, the most recent years for which data are
available. This level of growth corresponds to what
would be expected if investment in agriculture contin-
ued at about the same level as in 1986, 33.5 billion
rubles per year, and there was no reduction in the
retirement rate.24
Labor. Projections of employment in agriculture are
more uncertain. A decline in the work force is expect-
ed, but how fast it will decline is hard to predict:
? Overall population growth has slowed to less than I
percent per year.
? The working-age population of the European repub-
lics of the USSR is actually declining and will
continue to do so through 1995.
22 "Supreme Soviet Decree on Economic Development," published
in Izvestiya, morning edition, 20 June 1986, p. 1.
23 The average farm output for 1981-85 was 123.19 billion rubles. If
average annual output is to increase by 14.4 percent in 1986-90,
output would have to average 140.90 billion rubles per year.
24 Because of the emphasis Moscow is placing on other components
of the agroindustrial complex, and the slow but downward trend in
capital growth in recent years, holding capital growth steady at 5
percent per year may be optimistic. However, reasonable assump-
tions about slower rates of capital growth had negligible effect on
the projection because of the low return to capital.
Soviet Farm Production in 1986 and 1987:
Not Enough Progress To Meet the Growing Demand
Agricultural performance during the first two years
of the 12th FYP showed considerable improvement
over previous years. Average farm output during
1986 and 1987 was about 5 percent above the average
for 1983-85. The biggest gains were obtained in the
production of grain, sunflower seeds, and livestock
products:
? Grain output for 1986-87 was nearly 14 percent
higher than during 1983-85, exceeding 210 million
metric tons each year.
? Production of sunflower seeds-the USSR's main
source of vegetable oil-was 15 percent higher than
during 1983-85.
? Meat output was 9 percent higher, and milk and egg
production were each 5 percent higher than in 1983-
85.
Production of other major crops, however, was disap-
pointing: production of potatoes and sugar beets
increased only slightly, and output of cotton, vegeta-
bles, and fruit actually declined.
Nonetheless, the improvement in performance was
not sufficient to satisfy consumers. The excess de-
mandforfood was fueled by government policies that
steadily increased disposable income but maintained
stable, relatively low, retail prices for food. Per
capita disposable income grew by about 6 percent
during 1986-87 compared with 1983-85, while overall
per capita availability of farm products increased
only slightly. By 1987, complaints of shortages in
state retail food stores were common; reports of
rationing of meat and butter had increased; and, in
Moscow, collective farm market prices-which are
relatively free to respond to supply and demand-had
risen to record levels.
Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
 Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
? Migration of agricultural workers-especially
skilled labor and the young-to industry is continu-
ing largely as a result of better living standards in
urban areas.
? The share of elderly people in the rural populations
of the European republics and the Russian Republic
(RSFSR) is increasing.
Employment in agriculture declined 2 percent in 1985
and 1.5 percent in 1986. On the basis of 12th FYP
goals for output and labor productivity for socialized
farming, a "planned" rate of decline for labor in
socialized agriculture of about 1.5 percent per year
can be inferred.25 For making projections, this
"planned" rate of decline was applied to total agricul-
tural employment.
Weather. The uncertainties of weather were formally
incorporated into the analysis with stochastic simula-
tion (also called Monte Carlo analysis). Information
about the frequency with which past weather events
occurred was used to generate frequency distributions
for the weather variables specified in the model.
Using these probabilities, agricultural output was
predicted for each year by randomly choosing values
for the weather variables according to a normal
distribution with the appropriate mean and variance."
The model was solved repeatedly (5,000 times), draw-
ing different values for the weather variables each
time, producing a probability distribution of the out-
put. A "most likely" range estimate was then derived
from the probability distribution of the estimated
output, reflecting the likelihood of all possible weather
outcomes. For this study, the most likely range is
defined such that there is a 10-percent chance growth
could be below the lower limit of the range and a 10-
percent chance it could exceed the upper limit.
21 The I2th FYP called for growth of labor productivity in social-
ized agriculture to be 21.4 percent higher in 1986-90 than in 1981-
85. Attainment of both the labor productivity and output growth
goals given the results for 1986 implies that employment must
average 59.9 billion man-hours per year during 1987-90. Assuming
an exponential rate of decline, this is equivalent to an average
annual growth rate of about -1.5 percent for 1987-90.
" Since the two weather variables have been correlated historically
(when one is high the other tends to be high as well), a similar
degree of correlation was incorporated into the simulations.
Material Inputs and Technology. Although purchases
of material inputs are not explicitly accounted for in
the model, there is an implicit assumption that growth
of these inputs be maintained at about the same rate
as in recent years, which was 4 percent in 1986.
Failure to provide sufficient quantities of these inputs
each year will prevent output from increasing unless
substantial efficiency gains in their use occur.
Technological progress is also not explicitly accounted
for in the model, but capital is assumed to capture a
portion of the technological progress while the policy-
related variable PRODCHNG is assumed to capture
remaining sources.
Government Policy. The most uncertain aspect of the
projection is predicting productivity growth stemming
from government policy initiatives. Soviet leaders are,
of course, hoping for a dramatic upsurge in farm
productivity coming from the intensive technology
campaign and recent reform measures. However,
boosting farm productivity will not be easy. Long-
standing impediments to productivity growth must be
overcome before significant progress is possible, in-
cluding: weak links between the size and quality of
harvest and financial rewards for farmworkers; few
incentives for managers to reduce production costs;
low quality and inappropriate assortment of farm
machinery; rural living conditions that are still too
stark to encourage younger, skilled workers to stay on
the farm; and a rural education system that is inade-
quate for teaching modern agricultural practices. It is
not clear that the programs now in place or planned
for agriculture are adequate to the task of substantial-
ly raising productivity in the near term.27 Consequent-
ly, three scenarios were constructed by assuming
alternative degrees of success for these programs.
Scenarios
The first scenario assumes no change in agricultural
policy (that is, PRODCHNG is set equal to 1 for
1988-90, the same value assigned to PRODCHNG
for 1986 and 1987). Total factor productivity growth
Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
 Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
Table 6
Prospects for Meeting the 12th FYP
Goal in Soviet Agriculture a
Productivity Assumptions b
Five-Year
Odds of Meeting
Growth Rate of
Plan c
Output (percent)
Scenario 1:
10.7
Less than I in
No additional productivity
(8.8-12.6)
in 100
growth
Scenario 2:
12.4
1 in 10
Productivity growth sufficient
(10.4-14.3)
to offset expected losses in
employment
Scenario 3: 14.4 1 in 2
Productivity growth sufficient (12.3-16.4)
to sustain the 1968-78
trend (equal to the
average rate for 1985-87)
a Growth rates were calculated by dividing the average 1986-90
output by the average 1981-85 output (123.19 billion rubles), using
actual data for 1986 and 1987. The point estimate (in boldface)
assumes average weather, defined here to be the set of weather
events associated with the 50th percentile (median) level of output.
An 80-percent range estimate, given in parentheses, was derived by
incorporating the uncertainties of weather into the analysis. The
80-percent range means there is a 10-percent chance growth could
be below the lower limit of the range and a 10-percent chance it
could exceed the upper limit of the range. Other assumptions
include 5-percent capital growth and -1.5-percent employment
growth.
b These productivity assumptions were incorporated into the model
by adjusting PRODCHNG as follows:
1988
1.00
1.25
1.70
1989
1.00
1.50
2.10
1990
1.00
1.75
2.50
than the 14.4-percent goal, and would clearly repre-
sent a failure for Moscow. Even extremely favorable
weather would not allow the five-year goal to be met;
taking into account the uncertainties of weather, the
chances of meeting plan are less than 1 in 100.
The second scenario assumes Moscow can stimulate
productivity enough to offset expected losses in em-
ployment (equivalent to a 1.5-percentage-point in-
crease in the growth of total factor productivity). If
this can be done, the five-year increase would be 12.4
percent, assuming "average" weather. While an im-
provement, it still falls short of the FYP goal, and the
odds that weather will be favorable enough to meet
the FYP goal under these conditions are still only 1 in
10 (see figure 12).
In the third scenario, factor productivity growth was
maintained at about 1.4 percent per year, which
produces even odds of meeting the plan. At this rate
of productivity growth, weather-adjusted output
would continue along the 1968-78 trend shown in
figure 10. But even if this rate of productivity growth
is attained-which may be possible if Gorbachev
introduces new programs and policies designed to
increase worker efficiency-there is a 50-percent
chance that unfavorable weather would erode the
positive effect of the productivity gains.
The rate of productivity growth required to ensure
that the plan be met for all but the most severe
weather outcomes was calculated to be nearly 5
percent per year.28 Under these conditions, the most
likely five-year increase would be 17.7 percent, and
odds of falling short of the 14.4-percent goal would be
The goal for Soviet agriculture as stated in the 12th FYP is less than 1 in 30. The only historical precedent for
14.4 percent over the previous five-year period. d
sustalne productlvlty growth of this magnitude oc
actually declines in this case because the projected
decline in employment is not offset by productivity
gains and leads to an even greater decline in output
growth. Under these conditions, farm output for
1986-90 would increase by only 10.7 percent over the
previous five-year period (see table 6) assuming "aver-
age" weather conditions. This is considerably less
-
curred between 1982 and 1984 as agriculture recov-
ered from the preceding slump period. It is highly
unlikely that such productivity gains can be repeated.
" For this calculation, the variable PRODCHNG was set equal to
2.0 for 1988, 3.0 for 1989, and 4.0 for 1990.
Declassified in Part - Sanitized Copy Approved for Release 2012/12/03 : CIA-RDP89TO1451 R000500520004-0
 Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
Figure 12
Projection of Five-Year Growth Rate of Soviet Farm Output
Taking Into Account Weather Uncertainties
Most likely
growth rate
is 12.4%
10% of total area
(10% chance growth
will equal or exceed
the 14.4% planned
growth rate)
8 10 12 14 16 18 20
Percentage growth of 1986-90 average over 1981-85 average
Note: The projection incorporates
actual output results for 1986 and
1987. Assumptions include: (a) 5%
growth of capital stock, (b) -1.5%
employment growth, and (c) total
factor productivity growth
sufficient to offset expected losses
in employment.
318105 6-99 STAT
These model simulations suggest that the Soviets will
be able to attain their 1986-90 goal for agricultural
output only if the following three conditions prevail:
? At least "average" weather for 1988-90.
? Continued growth of inputs from other sectors at a
rate equal to that of recent years, which was 4
percent in 1986, together with timely deliveries.
? Productivity gains not only sufficient to offset losses
in the agricultural labor force, but also equivalent to
an additional one-percentage-point increase in
growth of total factor productivity.
If any of these conditions are not met, the goal will be
out of reach. Bad weather could be potentially devas-
tating to output growth, but good weather is equally
probable. Even with good weather, however, signifi-
cant gains in productivity growth will still be needed
to meet the FYP.
- Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
 Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
Appendix A
Development of the Model
The value of agricultural output was modeled as a
function of labor, capital, and weather. The starting
point for model development is the Cobb-Douglas
production function, denoted as:
Figure 13
Model Predictions Using Only Capital
and Labor, 1968-86
Q = a KK L'-P,
where Q is output, K and L are capital and labor
inputs, respectively, and a and 0 are parameters. This
basic functional form was modified by converting a
into the product of three functions, as follows:
a = a, a2(W) a,(P)
The function a, is a scale adjustment that reconciles
the units of measure used for Q, K, and L. The
function a2(W) contains the weather variables, and
thus measures the effects of weather on agricultural
output. The function a3(P) is an adjustment for rela-
tive changes in productivity originating directly or
indirectly from government programs and policies.
The first step in developing the model was to examine
the relationship between capital, labor, and output
without accounting for any effects of weather or
relative productivity changes. This was done by fitting
the intensive form of the function with a = a,:
Log(Q/L) = Log(a,) + 0 Log(K/L) 29
Results indicated that the model was statistically
significant (see table A-1). The capital elasticity, R,
was estimated to be 0.13. As shown in figure 13,
however, substantial variation still remained
unexplained.
29 The intensive form of the Cobb-Douglas production function is
derived by dividing both sides of the equation by L, logarithmically
transforming both sides, and simplifying.
S Actual data
-?- Model predictions
0.4, -0.2 0 0.2 0.4 0.6 0.8 1.0 . 1.2 1.4 1.6
Log (K/L)
Q - Farm output in billion 1982 rubles.
L - Labor in billion man-hours.
K -Capital in billion 1973 rubles.
318102 8-88 STAT
The second step was to expand the model to include
the effects of weather. Preliminary work indicated
that this effort would be successful only if measures
were taken to isolate the impact of the industrial
growth slowdown on agriculture. The simplest ap-
proach was to exclude the years 1979-82-the worst
of the industrial growth slowdown period-from the
model while searching for the relevant weather
measures.
Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
 Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
Table A-1
Fitting the Model With Capital and Labor Only a
Log (Q/L) = A. + (3Xlog(K/L)
Parameter Parameter Standard Error
Estimate
t for Ho: Probability > [t]
Parameter = 0
Degrees of Sum of Mean F value Probability of R 2 Durbin-
Freedom Squares Square a Greater F Watson D
Model 1 0.07224344 0.07224344 26.791 0.0001 0.5890 1.517
Error 17 0.04584165 0.00269657
Corrected total 18 0.11808509
The objective in selecting weather variables was to were tested in addition to temperature and precipita-
choose a few key variables that reflect overall agricul- tion alone."
tural production, rather than very specific measures
that correspond closely to critical growth stages of J0 The four combinations were:
some particular product, for example, grain. Previous Temperature X precipitation = a measure of hot and wet
research had shown that gross weather aggregates conditions
(weighted according to grain area) such as winter Temperature/precipitation = a measure of hot and dry
conditions
temperature averaged over the six-month period from I /(Temperature X precipitation) = a measure of cold and dry
October to March and spring temperature and precip- conditions
itation averaged over the four-month period from Precipitation/ temperature = a measure of cold and wet
April to July explained a significant portion of the conditions
variation in Soviet grain yields. In the present study, where "X" denotes multiplication and "/" denotes division.
four combinations of temperature and precipitation
I Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
 Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
Table A-2
A Preliminary Model Including Weather,
Estimated Without 1979-82
Log(Q/L) = Ao + (3Xlog(K/L) + A, + A2XHOTNDRY + A,X(1/HOTNDRY) + A4XWINTEMP
Variable
Parameter
Parameter
Estimate
Standard Error
t for Ho:
Parameter = 0
Probability > [t]
HOTNDRY
A,
-14.945366
4.7595395
-3.140
0.0105
1/HOTNDRY
A,
-0.043465
0.0177473
-2.449
0.0343
WINTEMP
A,
0.028784
0.0080357
3.582
0.0050
Degrees of Sum of Mean
Freedom Squares Square
F value Probability of a R z Durbin-
Greater F Watson D
Note: a, = eAo
a2(W) = eA~+A2HOTNDRY+A, (I /HOTNDRY) + A WINTEMP
a The model was initially estimated with only six parameters,
including a parameter for the sum of A. and A,. The parameter for
the sum of A, and A, was determined to be 2.15101 with a standard
error of 0.58376. A. was estimated to be 0.507255 (standard error
= 0.02076) by fitting the following model (excluding the years
1979-82):
Log(Q/L) = A. + Plog(K/L).
After some experimentation, two weather measures
emerged as key variables. The most important is the
ratio of average temperature to cumulative precipita-
tion for the April-July period, named HOTNDRY.
Parameters for both HOTNDRY and its reciprocal
(1/HOTNDRY) had negative signs, indicating that
too much HOTNDRY hurts agriculture and too little
HOTNDRY also hurts agriculture. The second
weather measure was average winter temperature for
the October-March period, named WINTEMP. The
parameter for WINTEMP had a positive sign, as
expected. Fluctuations in these two variables ex-
plained a substantial amount of the year-to-year
variation in agricultural output (see table A-2).
Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
 Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
Table A-3
The Final Model
Log(Q/L) = A0 + (3X1og(K/L) + A, + A2XHOTNDRY + A,X(1/HOTNDRY) + A4XWINTEMP +
A5 X PRODCHNG
Parameter
Parameter
Estimate
Standard
Error
t for Ho:
Parameter = 0
Probability > [t]
HOTNDRY
A,
-15.692849
2.0572757
-7.628
0.0001
1/HOTNDRY
A,
-0.045175
0.0076127
-5.934
0.0001
WINTEMP
A,
0.035548
0.0037783
9.408
0.0001
Degrees of Sum of Mean F value Probability of R2 Durbin-
Freedom Squares Square a Greater F Watson D
Note: a, = eA0
a5(W) = e A,+A2HOTNDRY+A3 (I/HOTNDRY)+AaWINTEMP
a,(P) = eA,PRODCHNG
a The parameter for the sum of A. and A, was estimated to be
2.22947 with a standard error of 0.25126. A, was estimated using
the value for Aa derived in table A-2.
Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
 Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
Total hectarage sown to crops was also considered.
However, statistical tests indicated that this variable
did almost nothing to reduce the remaining unex-
plained variation. The year-to-year variation in sown
area was perhaps too small to measure the effects of
the variable on output in the presence of much
stronger influences like weather.
Figure 14
Model Predictions Using Full Model, 1968-86
S Model predictions
- Actual data
The final step in development of the model was to
account for changes in productivity that occurred
during the 1979-82 period and during 1986-87. For
0
7
.
this purpose, the function a3(P) was created to reflect
11973 Y r' 11985
our subjective estimate of relative changes in produc-
tivity owing to government policy actions. The deriva-
tion of this function is explained in the main body of a-6
this paper. By adding a3(P) to the model, it was
possible to include the years 1979-82 when estimating
parameters. The results are shown in table A-3. All
parameters were highly signincant statistically, and 1968
the R 2 (adjusted for degrees of freedom) was 0.970.
The capital elasticity was 0.17, which is slightly I I i I i I I I i
higher than the estimate made using only information 0.4 -0.2 0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6
on capital and labor. Figure 14 illustrates how closely Log (K/L)
the model predictions correspond to the historical Q -Farm output in billion 1982 rubles.
record. L - Labor in billion man-hours. -
K -Capital in billion 1973 rubles.
318103 8-88STAT
Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
 Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
Q=at a2(W)a3(P)J(@L1 c,
= capital elasticity parameter,
= value of agricultural output, excluding
farm output used within agriculture
(such as feed for livestock and grain
for seed), in billions of1982 rubles,
K = annual capital stock in agriculture at
the beginning of the year, excluding
livestock, in billions of 1973 rubles,
L = total work-hour employment in agri-
culture, in billions of hours,
HOTNDRY = ratio of average temperature (degrees
centigrade) to cumulative precipitation
(millimeters) for April through July,
weighted by total sown area,
WINTEMP = average winter temperature (degrees
centigrade) for October through
March, weighted by area sown to win-
ter wheat,
PRODCHNG = productivity change variable,
A, ... A5 = statistical parameters, and
c = stochastic error term.
There are two factors that can influence agricultural
output that are not explicitly included in the model-
technological progress and material inputs (such as
agrochemicals). Technological advances such as high-
er yielding strains of grain or higher livestock growth
rates resulting from genetic improvements would con-
tribute to higher growth. Similarly, increases in grain
yields can be attributed in part to increased use of
fertilizers and pesticides. Efforts to estimate the mod-
el with an additional time-trend variable representing
technological progress and variables representing de-
liveries of agrochemicals to farms were unsuccessful.
However, the capital input embodies technological
progress to the extent that the value of new machinery
and equipment reflects increased efficiency over the
old machinery and equipment. "Disembodied" tech-
nological progress could also occur as a result of more
efficient management and adoption of new farming
technologies. To the extent that this disembodied
technological progress is an increasing function of
time, the capital input-which is also an increasing
function of time-acts as a surrogate, or proxy, for it.
For the same reason, capital also serves as a proxy for
material inputs.
Testing for the Effects of the Industrial Growth
Slowdown
The model was used to conduct a statistical test to
determine if the industrial growth slowdown during
1979-82 had a detrimental effect on agricultural
performance. The final model presented in table A-3
was reestimated after replacing the function a3(P) by a
dummy variable (DUM) consisting of l's for the years
1979-82 and 0's for all other years.31 A parameter
value for DUM that is not significantly greater than
zero would suggest that nonweather factors other than
capital and labor had little to do with the poor
agricultural performance during this time. As shown
in table A-4, the parameter for DUM was highly
significant statistically (that is, the probability of a
greater t-value was less than 0.0001 under the null
hypothesis that the parameter's true value is zero),
indicating that nonweather factors other than capital
and labor were indeed responsible for the associated
growth slowdown in agriculture during 1979-82.
" A dummy variable is a time-series sequence of l's and 0's. Use of
the dummy variable in hypothesis testing is equivalent to perform-
ing an analysis of variance and testing for significant group
effects-where the two time periods represent two groups-while
simultaneously accounting for variation between the two groups
that is due to differences in capital and labor inputs and weather.
Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89T01451 R000500520004-0
 Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
Table A-4
Testing for Effects of the Industrial Growth Slowdown
Log(Q/L) = A0 + BXlog(K/L) + A, + A2XHOTNDRY + A,X(1/HOTNDRY) + A4XWINTEMP +
A,XDUM
Variable Parameter Parameter
Estimate
1/HOTNDRY A, -0.042382
WINTEMP A, 0.026284
Dummy variable for 1979-82 A, -0.104958
Standard Error t for Ho: Probability > [t]
Parameter = 0
0.01635564 -2.591 0.0224
0.00769052 3.418 0.0046
0.01952918 -5.374 0.0001
Model 5 0.10619527 0.02123905 23.222 0.0001 0.8606 1.552
Error 13 0.01188982 0.00091460
Corrected total 18 0.11808509
a The parameter for the sum of A, and A, was estimated to be
2.11503 with a standard error of 0.53949. A, was estimated using
the value for A. derived in table A-2. .
Comparison of Preliminary and Final Models
To perform simulations with the model, it is impor-
tant that the parameters be measured without signifi-
cant bias. One source of bias common to econometric
models is "multicolinearity." Multicolinearity is a
sample problem for which the sample does not provide
"rich" enough information on the explanatory vari-
ables (such as HOTNDRY, WINTEMP, K, and L) to
prevent one variable from inordinately influencing the
parameter estimate of another variable. In other
words, multicolinearity is a problem when the explan-
atory variables are not sufficiently independent to
meet the requirements of the model.
Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
 Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
Table A-5
Comparison of Preliminary and Final Models
Weather-Adjusted Farm Output
Final
Preliminary
Difference
Final
Model
Preliminary
Model
Model
Model
1968
105.061
106.393
106.242
105.757
105.901
-0.144
1969
100.303
100.732
101.856
105.939
104.788
1.151
1970
112.535
112.464
111.797
108.749
109.428
-0.679
1971
111.388
110.707
110.350
109.685
110.094
-0.409
1972
104.660
104.986
105.826
110.766
109.962
0.804
1973
121.807
119.841
118.830
115.747
116.833
-1.086
1974
119.629
121.073
120.394
115.221
115.994
-0.773
1975
109.410
109.094
108.309
118.221
119.233
-1.012
1976
118.060
114.802
115.811
122.192
121.312
0.880
1977
122.829
123.288
123.185
120.685
120.986
-0.301
1978
126.605
125.758
125.201
124.472
125.252
-0.780
1979
118.927
120.991
126.886
116.730
117.106
-0.376
1980
113.740
113.732
125.816
114.749
114.608
0.141
1981
112.500
111.332
127.930
112.040
113.135
-1.095
1982
120.788
120.174
138.134
114.578
115.672
-1.094
1983
128.638
130.706
136.083
126.500
127.892
-1.392
1984
128.046
129.277
129.442
133.780
133.955
-0.175
1985
125.992
127.435
129.114
132.537
131.169
1.368
1986
136.287
134.448
128.699
142.117
141.792
0.325
1987
132.032
131.575
127.060
139.566
138.048
1.518
In the final model presented above, there is potential
for multicolinearity between the functions a2(W) and
a,(P). One way to determine if multicolinearity is a
problem is to compare parameter estimates of the full
model with parameter estimates for a restricted mod-
el. Such a comparison can be made here by contrast-
ing the preliminary model in table A-2, which ex-
cludes a,(P), with the final model in table A-3, which
includes a,(P). Parameter estimates for the two mod-
els differ very little. Furthermore, there is little
difference in model predictions for years other than
1979-82 and 1986-87, as shown in table A-5. In the
final model, the function a,(P) adjusts for the addi-
tional nonweather factors influencing farm output
during 1979-82 and 1986-87 and thus produces better
predictions for those years. Most important, trends in
the weather-adjusted farm output series created using
the two models are almost identical (see table A-5),
even for 1979-82 and 1986-87. These results indicate
strongly that, if multicolinearity between weather and
the productivity change variable exists, it is not
biasing parameter estimates for the weather variables
to any significant extent.
Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
 Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
Appendix B
Parameters of the model were estimated using data
from 1968 through 1986. The period was not extend-
ed to 1987, because reliable employment data for
1987 were not available at the time of the study, and
only preliminary data on farm output were available.
Table B-1
Productive Fixed Capital Stock and
The Soviet definition of fixed capital includes the
undepreciated value of buildings, structures, convey-
ing equipment, machinery and equipment (including
measurement and control instruments, laboratory
equipment, and computer hardware), vehicles, tools,
and productive and draft livestock of basic herds (but
excluding young livestock, livestock allocated for fat-
tening, and some minor categories such as poultry,
rabbits, and fur-bearing animals). Fixed capital is
broken down into productive and nonproductive capi-
tal. Productive capital is that used directly in the
production process. Nonproductive capital includes
capital in the housing and municipal services sector
and in organizations and institutions of public health,
education, science, culture, art, credit institutions,
and administrative organs.
In fitting the model, nonproductive fixed capital was
excluded, as was productive livestock. The data used
are shown in table B-1.
Agricultural workers fall into four basic categories:
workers and employees on state farms; collective
farmers; persons engaged in private farming; and
temporary workers recruited from nonfarm industries,
the military, and schools to help during peak agricul-
tural periods, primarily the harvest season. The Sovi-
ets report average annual employment statistics for
state and collective farms as well as the number of
Capital Investment in Soviet Agriculture
Beginning-of-Year Capital Stock Capital Invest-
(in comparable 1973 prices) ment (in compa-
Including Excluding rable 1984
Livestock a Livestock b prices) c
1965
72
49
10.600
1966
77
54
11.308
1967
82
58
12.069
1968
87
63
13.466
1969
93
69
14.029
1970
98
74
16.000
1975
154
127
26.100
1976
167
141
27.190
1977
180
153
27.910
1978
194
167
28.895
1979
209
181
29.519
1980
223
195
29.800
1981
238
210
30.500
1982
254
225
30.925
1983
272
242
31.978
1984
288
258
31.000
1985
303
272
31.500
1986
316
286
33.500
1987
330
300
a Narodnoye khozyaystvo SSSR za 70 let., Central Statistical
Administration, Moscow, 1987, p. 100, and other years.
b Based on indexes published in Narodnoye khozyaystvo SSSR za
70 let., Central Statistical Administration, Moscow, 1987, p. 101,
and other years.
c Narodnoye khozyaystvo SSSR za 70 let., Central Statistical
Administration, Moscow, 1987, p. 276, and other years.
Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
 Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
Table B-2
Employment in Socialized Agriculture
Employment (million workers)
Man-Days
per Month
Hours Worked per
Year per Worker
Total Hours Worked (millions)
State
Farms
Collective Recruits
Farms
Total
State Collective
Farms Farms
State
Farms
Collective
Farms
State
Farms
Collective Recruits
Farms
Total
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
(10)
(11)
(12)
1968
8.548
15.782
0.5
1,907
1,580
16,299
24,934
790
42,023
1969
8.725
15.010
0.6
1,890
1,585
16,490
23,798
951
41,240
1970
8.833
14.667
0.6
24.1
22.8
19.2
1,915
1,613
16,917
23,655
968
41,540
1971
9.122
13.478
0.7
23.3
1,924
1,630
17,547
21,973
1,141
40,661
1972
9.244
13.456
0.8
23.5
1,924
1,651
17,782
22,210
1,320
41,313
1973
9.462
13.238
0.9
23.6
1,932
1,669
18,281
22,092
1,502
41,874
1974
9.656
13.044
0.9
23.6
1,932
1,694
18,655
22,097
1,525
42,277
1975
9.787
12.713
1.0
23.5
23.1
20.3
1,940
1,705
18,991
21,678
1,705
42,374
1976
9.970
12.430
1.1
23.5
23.2
20.6
1,949
1,730
19,430
21,509
1,903
42,842
1977
10.180
12.020
1.1
23.3
23.1
20.7
1,940
1,739
19,753
20,900
1,913
42,566
1978
10.387
11.613
1.3
23.3
23.1
21.0
1,940
1,764
20,155
20,485
2,293
42,933
1979
10.481
11.319
1.3
23.1
23.0
21.2
1,932
1,781
20,249
20,157
2,315
42,721
1980
10.693
10.907
1.3
22.9
23.1
21.4
1,940
1,798
20,749
19,606
2,337
42,692
1981
10.817
10.483
1.4
22.7
23.2
21.6
1,949
1,814
21,080
19,020
2,540
42,641
1982
10.978
10.522
1.4
22.9
23.2
21.8
1,949
1,831
21,394
19,268
2,564
43,225
1983
11.098
10.402
1.5
23.0
23.2
22.2
1,949
1,865
21,628
19,398
2,797
43,823
1984
11.102
10.198
1.5
22.8
23.2
22.3
1,949
1,873
21,636
19,103
2,810
43,548
1985
11.095
9.905
1.4
22.4
23.1
22.4
1,940
1,882
21,529
18,637
2,634
42,800
1986
10.968
9.632
1.4
22.0
23.1
22.4
1,940
1,882
21,282
18,124
2,634
42,040
Sources:
Column (1): Narodnoye khozyaystvo SSSR za 70 let., Central
Statistical Administration, Moscow, 1987, p. 86, and other years.
Column (2): column (4) minus column (1) minus column (3). Values
for 1968-69 were taken from Stephen Rapawy, Civilian Employ-
ment in the USSR 1950 to 1983, CIR Staff Paper No. 10, US
Department of Commerce, Bureau of the Census, August 1985, p.
31.
Columns (3) and (4): Narodnoye khozyaystvo SSSR za 70 let.,
Central Statistical Administration, Moscow, 1987, p. 300, and
other years.
Column (5): Narodnoye khozyaystvo SSSR za 70 let., Central
Statistical Administration, Moscow, 1987, p. 292, and other years.
Column (6): Narodnoye khozyaystvo SSSR za 70 let., Central
Statistical Administration, Moscow, 1987, p. 288, and other years.
Column (7): column (5) multiplied by 12 months per year and seven
hours per day. Values for 1968-69 and 1971-74 were derived from
data reported by Stephen Rapawy, Civilian Employment in the
USSR 1950 to 1983, CIR Staff Paper No. 10, US Department of
Commerce, Bureau of the Census, August 1985, p. 29.
Column (8): column (6) multiplied by 12 months per year and seven
hours per day. Values for 1968-69 and 1971-74 were derived from
data reported by Stephen Rapawy, Civilian Employment in the
USSR 1950 to 1983, CIR Staff Paper No. 10, US Department of
Commerce, Bureau of the Census, August 1985, p. 31.
Column (9): column (1) multiplied by column (7).
Column (10): column (2) multiplied by column (8).
Column (11): column (3) multiplied by column (8).
Column (12): column (9) plus column (10) plus column (11).
Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
 Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
Table B-3
Employment in Private Agriculture and Total Employment
Total Hours Worked
in Agriculture
Number of Productive Livestock
(end oj'year, millions)
Sown Area
(million)
hectares)
Total Hours
Worked
(millions)
(millions)
Cattle
Swine
Sheep and Goats
(1)
(2)
(3)
(4)
(5)
(6)
1968
27.3
12.8
34.4
6.77
22,771
64,794
1969
25.0
13.8
31.7
6.78
21,868
63,108
1970
25.0
16.6
33.2
6.73
22,292
63,832
1980
23.0
14.0
30.2
6.16
20,280
62,972
1981
23.4
14.2
30.7
6.15
20,488
63,129
1982
24.2
15.8
31.9
6.16
21,139
64,365
1983
24.6
15.6
33.2
6.16
21,331
65,153
1986
23.7
13.6
33.4
Sources:
Columns (1), (2), and (3): Narodnoye khozyaystvo SSSR za 70 let.,
Central Statistical Administration, Moscow, 1987, p. 253, and
other years.
Column (4): Narodnoye khozyaystvo SSSR za 70 let., Central
Statistical Administration, Moscow, 1987, p. 225, and other years.
Column (5): derived from columns (1), (2), and (3); see text.
Column (6): column (5) plus column (12) from table B-2.
workers involved in temporary seasonal activity. From
this information, an estimate of total work hours in
socialized agriculture can be made (see table B-2).
Although the Soviets report statistics on the number
of workers in private agriculture, the model requires
data on employment in hours worked, which they do
not report. Using a method developed by the US
Department of Commerce, Center for International
Research, an estimate of private employment in hours
worked can be derived from data on the number of
livestock on private farms and the area allocated for
private plots. (see table B-3).32 This is done using labor
32 See Stephen Rapawy, Estimates and Projections of the Labor
Force and Civilian Employment in the USSR 1950 to 1990,
Foreign Economic Report No. 10, US Department of Commerce,
Bureau of Economic Analysis, September 1976, p. 43.
- Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
 Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
Figure 15
Weather Data Used in Model, 1968-87
Ratio of Temperature to Precipitation
April-July (HOTNDRY)
Ratio
11111111111111111111
0-04 1968 70 75 80 85
Average Winter Temperature
October-March (WINTEMP)
Degrees centigrade
I l i 1 i 1 l l i i i l 1 1 1 1 l 1 1
-3 1968 70 75 80 85
coefficients obtained from the Soviet literature, as
follows:
Activity Input Required
(man-days/per unit)
Tending one head of cattle 54.2
Tending one pig 20.6
Tending one sheep or goat 5.6
The total man-days for animal husbandry are in-
creased by 10 percent to allow for labor involved in
tending poultry, horses, and rabbits, which otherwise
would not be included. Man-days are converted to
total hours by multiplying by seven hours per day, the
same daily work rate assigned to state and collective
farms.
Detailed meteorological data from the USSR are
available through the World Meteorological Organi-
zation. As a member, the USSR shares such informa-
tion with foreign countries. These data are part of a
worldwide standardized system that attempts to en-
sure consistent measures of weather parameters from
year to year. Precipitation and temperature data are
available for approximately 1,000 stations located
throughout the grain-growing portion of the USSR."
The data is processed and corrective measured applied
" Summaries of the data for 27 crop regions are reported in
Climate Impact Assessment, Foreign Countries, published by the
National Oceanic and Atmospheric Administration (NOAA). For
the present study, eight additional crop regions were created,
predominantly in Siberia and Kazakhstan.
STAT
Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
 Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
to overcome reporting errors and omissions. Although
the original data set extends to the mid-1940s, the
"corrected" data set begins in 1969. It was possible to
use the "uncorrected" weather data for 1968 and thus
extend the data set an additional year, but attempts to
include years before 1968 in the model were
unsuccessful."
These data were used to calculate monthly precipita-
tion and average monthly temperature for the agricul-
tural area of the USSR. Two weighting schemes were
used to aggregate the data. Precipitation and tem-
perature for the variable HOTNDRY were weighted
according to the area sown to all crops, whereas
temperature data for WINTEMP were weighted ac-
cording to area sown to winter wheat.35 HOTNDRY
is the ratio of average temperature to cumulative
precipitation for April through July. WINTEMP is
the average temperature for October through March.
The data and summary statistics for HOTNDRY and
WINTEMP are shown in table B-4 (also see figure
15). The mean and standard deviation were used to
generate a probability distribution for each variable in
order to conduct the stochastic simulation exercise.
HOTNDRY and WINTEMP are positively correlat-
ed; the Pearson correlation coefficient measured 0.595
(with a standard error of 0.139). That is, when
WINTEMP is high, HOTNDRY is often-but not
always-high as well. Consequently, simulated values
34 The two weather data sets also had different area definitions, and
so it was necessary to link the two series. This was done for 1968
data as follows:
"corrected"
value for 1969
Value for 1968= X "uncorrected" value for 1968
"uncorrected"
value for 1969
" The calculation was made as follows:
35
I[Share of total] X[ Weather data 1= Weighted weather data
;=1 area in area i J IL for area i JJ
Table B-4
Weather Data
1968
0.065680
-0.03
1969
0.052343
-2.60
1970
0.059621
0.15
1971
0.062416
-0.10
1972
0.066987
-1.50
1973
0.059346
0.60
1974
0.052102
0.00
1975
0.085941
2.10
1976
0.058007
-1.90
1977
0.056889
-0.50
1978
0.046413
-0.10
1979
0.059482
-0.30
1980
0.049892
-1.20
1981
0.073477
1.40
1982
0.058703
0.60
1983
0.072209
1.50
1984
0.074187
0.20
1985
0.053265
-2.50
1986
0.069834
-0.60
1987
0.059304
-2.40
Percentiles
99%
0.085942
2.1
90%
0.074187
1.5
75%
0.069834
0.6
50% (median)
0.059483
-0.1
25%
0.053266
-1.5
10%
0.049892
-2.5
1%
0.046414
-2.6
Mean
0.061613
-0.376
Standard deviation
0.0099342
1.36426
Note: Neither of these distributions were significantly different
from the normal distribution. Data for 1968 were excluded from
calculations of summary statistics.
a A percentile represents the probability that a value equal to or less
than the tabled value would be expected to occur, based on the 19
observations in the original frequency distribution. For example, a
value of HOTNDRY equal to or less than 0.053266 (the value for
the 25th percentile) would be expected to occur about once every
four years, on average.
Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
 Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
Table B-5
Derivation of Value-Added, Weather-Adjusted Farm Output
Weather-Adjusted
Farm Output
Gross Weather-
Adjusted Output
Current
Purchases
Value-Added Weather-
Adjusted Output
Actual
(billion rubles)
Index
Including Current
Purchases
(billion rubles)
(billion rubles)
Actual
(billion rubles)
Index
(1)
(2)
(3)
(4)
(5)
(6)
1968
105.757
0.923
120.361
20.1494
100.211
1.000
1969
105.939
0.924
120.568
20.9693
99.599
0.993
1970
108.749
0.949
123.766
21.6942
102.072
1.018
1971
109.685
0.957
124.831
22.7634
102.068
1.018
1972
110.766
0.966
126.062
24.2611
101.801
1.015
1973
115.747
1.010
131.730
25.9391
105.791
1.055
1974
115.221
1.005
131.132
27.5947
103.537
1.033
1975
118.221
1.031
134.546
29.2831
105.262
1.050
1976
122.192
1.066
139.066
28.3749
110.691
1.104
1977
120.685
1.053
137.350
31.1685
106.182
1.059
1978
124.472
1.086
141.661
31.6678
109.993
1.097
1979
116.730
1.018
132.850
32.1303
100.719
1.005
1980
114.749
1.001
130.595
33.1553
97.440
0.972
1981
112.040
0.977
127.511
34.0618
93.450
0.932
1982
114.578
1.000
130.400
35.2400
95.160
0.949
1983
126.500
1.104
143.968
37.8364
106.132
1.059
1984
133.780
1.167
152.254
39.1360
113.118
1.128
1985
132.537
1.156
150.839
41.0018
109.838
1.096
1986
142.117
1.240
161.742
42.7994
118.943
1.186
1987
139.566
1.218
158.839
43.6574
115.182
1.149
Sources: Column (4): current purchases.
Column (1): weather-adjusted output series from table 3. Column (5): column (3) minus column (4).
Column (2): column (1) divided by 114.578, the value of weather- Column (6): column (5) divided by 100.211, the value of value-
adjusted output for 1982. added weather-adjusted output for 1968.
Column (3): column (2) multiplied by 130.4 billion rubles, which is
the 1982 gross value of farm output estimated by extending the
1972 input-output table forward to 1982. It represents complete
coverage of gross output minus interfarm use, as opposed to the net
farm output measure used in this study, which is based on a sample.
Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
 Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0
for HOTNDRY and WINTEMP were created such
that this correlation was preserved; the Pearson corre-
lation coefficient of simulated values was 0.585.
The Soviet measure of gross agricultural output is
inadequate for modeling purposes because no adjust-
ment is made for intra-agricultural use of farm
products (such as seed and animal feed) and because
Soviet gross output statistics include a large element
of waste. The measure of farm output used in this
study-net farm output-is the sum of livestock
production and crop production, minus seed, feed and
waste, valued in average 1982 realized prices. Deriva-
tion of the series has previously been described in
detail." Net farm output is based on a sample of 28
36 See Barbara Severin and Margaret Hughes, "Part III. An Index
of Agricultural Production in the USSR," in USSR: Measures of
Economic Growth and Development, 1950-80, Joint Economic
Committee, Congress of the United States, December 1982, pp.
245-316.
individual crops, 10 livestock products, and four items
of livestock inventory change. These 42 products
account for nearly 95 percent of total farm output net
of intrafarm use of crops.
Value-Added Farm Output
Total factor productivity was calculated using value-
added farm output. Value-added farm output ex-
cludes not only production for intrafarm use, but also
the value of materials and services purchased by
agriculture on current account from nonagricultural
sectors (current purchases). The time series for cur-
rent purchases is based on 10 indexes of material
inputs." Weather-adjusted farm output is converted
to a value-added measure according to the method
presented in table B-5.
" See John Pitzer, "Part I. Gross National Product of the USSR,
1950-80," in USSR: Measures of Economic Growth and Develop-
ment, 1950-80, Joint Economic Committee, Congress of the United
States, December 1982, pp. 88-91.
Declassified in Part - Sanitized Copy Approved for Release 2012/12/03: CIA-RDP89TO1451 R000500520004-0