THE 1982 SOVIET GRAIN CROP: REASSESSING THE CIA ESTIMATE

• Document Type: CREST
• Collection: General CIA Records
• Document Number (FOIA) /ESDN (CREST): CIA-RDP83B00851R000400070004-2
• Release Decision: RIPPUB
• Original Classification: S
• Document Page Count: 13
• Document Creation Date: December 19, 2016
• Document Release Date: February 20, 2007
• Sequence Number: 4
• Publication Date: December 1, 1982
• Content Type: REPORT

Approved For Release 2007/02/20: CIA-RDP83B00851 R000400070004
flirPptnrito of ecre
'" Intelligence
The 1982 Soviet Grain Crop:
Reassessing the CIA Estimate
Secret
GI 82-10269
December 1982
Copy 4 8 3
Approved For Release 2007/02/20: CIA-RDP83B00851 R000400070004-2
 Approved For Release 2007/02/20: CIA-RDP83B00851 R000400070004-2
Approved For Release 2007/02/20: CIA-RDP83B00851 R000400070004-2
 Approved For Release 2007/02/20: CIA-RDP83B00851 R000400070004-2
Directorate of
Intelligence
5X1
The 1982 Soviet Grain Crop:
Reassessing the CIA Estimate
This report was prepared by the Agricultural
Assessment Branch, Office of Global Issues.
Comments and queries are welcome and may be
directed to the Chief, Agricultural Assessments
Branch, Strategic Resources Division, OGI, on
I I
Secret
GI 82-10269
December 1982
Approved For Release 2007/02/20: CIA-RDP83B00851 R000400070004-2
 D
Approved For Release 2007/02/20: CIA-RDP83B00851 R000400070004-2
Soviet Union: Major Grain Growing Regions
Baltic Sea
Moldova" $SS
Romania
MOSCOW* V
e#orussi Central
ter, and Spring Grains
Trans-
Cyprus
Mediterranean Sea `~ $ tr-a
Lebanon
-.~-, Israel'
Boundary representation is
not necessarily authoritative
( Lake
Balkhash
China
Indian Claim
 Approved For Release 2007/02/20: CIA-RDP83B00851 R000400070004-2
Secret
Overview
Information available
as of 10 November 1982
was used in this report.
The 1982 Soviet Grain Crop:
Reassessing the CIA EstimatiI
Several months ago, we estimated that the 1982 Soviet grain crop would
approximate 165 million tons, somewhat above last year's crop but well
below what Moscow was expecting. Although the USSR has not an-
nounced the size of this year's grain harvest, remarks by Soviet officials
can be interpreted as implying an output as low as 140 million tons or as
high as 190 million tons. In view of this wide range and the slow pace of
Soviet grain purchases, we reexamined our grain crop estimate-takin
another look at the evidence 25X1
The review helped us 2C50t
our assessment, but we found no reason for altering our view on the size of
this year's crop. Because of the many variables involved, the 165-million-
ton figure should be considered our best estimate, but one that is subject to
error. The maximum range of error in our grain crop estimate over the past
four years has been ? 8 percent.
Secret
GI 82-10269
December 1982
Approved For Release 2007/02/20: CIA-RDP83B00851 R000400070004-2
 Approved For Release 2007/02/20: CIA-RDP83B00851 R000400070004-2
Approved For Release 2007/02/20: CIA-RDP83B00851 R000400070004-2
 Approved For Release 2007/02/20: CIA-RDP83B00851 R000400070004-2
Secret
25X1
0
The 1982 Soviet Grain Crop:
Reassessing the CIA Estimate
0
we have reviewed our analysis o this year's grain
crop. The review included an evaluation of what we
know about other Soviet grain crop estimates and the
assessment of Postharvest indicators
1 To
cross-check our estimate, we also used an alternative
methodology that looks at grain production in six key
areas.
A Survey of 1982 Soviet Grain Crop Estimates
Recent estimates of 1982 Soviet grain production
range from about 140 tons to about 190 million tons
(table 1). Estimates attributable to Soviet officials
privy to internal information about the size of this
year's crop cover the entire range of forecasts, thus
providing little conclusive insight regarding the likely
outcome of the harvest.
Approved For Release 2007/02/20: CIA-RDP83B00851 R000400070004-2
 Approved For Release 2007/02/20: CIA-RDP83B00851 R000400070004-2
Approved For Release 2007/02/20: CIA-RDP83B00851 R000400070004-2
 Approved For Release 2007/02/20: CIA-RDP83B00851 R000400070004-2
Secret
Table 3
USSR: Preliminary Procurement Data, 1982
Economic Estimated Percent of Oblasts Date of Amount of
Region Percent of Total Reporting Procurement Procurements a
Grain Production Reporting a
(1971-80 Average)
a Compared with the 1976-80 period.
b Kustanay is the only major grain-producing oblast in Kazakhstan
that has reported as of 2 November. The relationship between yield
and procurements is in agreement when comparing 1981 with 1982.
Our 1981 estimate for yield in Kustanay was 11 c/ha while
procurements were about 3 million tons. A recent Soviet press report
placed this year's yield at 13 c/ha and procurements at 4 million
tons. The record amount of grain sold to the state from Kustanay is
4.7 million tons.
Insufficient data available to make a valid determination.
Grain Procurements
We have completed an analysis of all grain procure-
ments reported in the Soviet press through 2 Novem-
ber. Although the level of state procurements does not
necessarily relate directly to total production, it can
indicate whether a region has experienced a compara-
tively good, average, or bad year. The procurement
data analyzed reflects reporting primarily at the
oblast level. We have aggregated it by region/republic
in table 3. We believe the procurement data, or in
some cases the lack of it, tend to support our current
production estimate:
southern Volga valley are sparse; we believe that
this indicates lower-than-average sales to the state
In years when production was better than
average in these major winter grain-producing
areas, we had seen more procurement reporting by
now. The lack of reporting from the Ukraine is of
particular interest in that this area typically pro-
duces about 40 percent of the total winter grain
crop.
? Conversely, press reports indicate that the level of
procurements in the northern regions of European
Russia (Baltic Republics, Central, Central Black
Earth, and Volga-Vyatka) is much closer to average,
and in some areas even above average. Our analysis
indicates that grain production in these areas will be
near average or slightly above average.
Approved For Release 2007/02/20: CIA-RDP83B00851 R000400070004-2
 Approved For Release 2007/02/20: CIA-RDP83B00851 R000400070004-2
Table 4
USSR: Computations for Six Key Oblasts/Republic Methodology
Average yield
(centners per hectare)
11.5
10.8
27.1
21.4
Implied production d 184.5 182.3
(million metric tons)
Actual production 186.8 181.2
Percent of error -1.2 0.6
8.6
12.7
12.7
2.5
15.7
9.0
16.3
13.5 11.0 b
7.0
5.0
4.5
17.1
16.0
3.9
19.2
12.0 b
19.7
9.5 b 15.0 b
7.0
9.0
15.3
31.7
30.7
18.1
NA
33.0
33.0
18.1 b 27.8
23.0
23.9
17.3
21.9
26.6
19.7
26.8
23.2
25.5
15.1 16.0
20.3
21.4 c
149.6
225.8
205.1
130.3
230.2
216.3
252.0
175.2
195.8
152.7
151.9
168.2
222.5
195.7
140.1
223.7
195.7
237.4
179.2
189.1
158.0
-11.1
1.5
4.8
-7.0
2.9
10.5
6.2
-2.2
3.5
-3.4
a The average yield for all grains grown in these areas correlates
closely with the average all-grain yield for the USSR.
b Estimate.
c In late October the Soviet press reported a yield of 19.8 c/ha for
1982.
d Implied production = average yield (oblasts/republic) X harvest-
ed area (USSR) X 0.943 correction factor.
? Procurement information for the regions east of the
Volga River-normally published by late October-
is still largely unavailable
Oblasts/Republic Methodology
As a check on our current estimating methodology,
we used an alternate method to obtain a total grain
production figure for the USSR. The "Six Key Ob-
lasts/Republic Methodology" suggests a total crop of
about 152 million tons for this year (table 4). The five
oblasts (Altay Kray, Severo-Kazakhstan, Orenburg,
Volgograd, and Dnepropetrovsk) and one republic
(Belorussia) used in this methodology were selected
because they generally experience the types of weath-
er problems and growing conditions that exist in the
principal winter and spring grain regions. The total
average yield for all grains grown in the oblasts/ 25X1
republic correlates closely each year with the national
average yield. 25X1
Approved For Release 2007/02/20: CIA-RDP83B00851 R000400070004-2
 Approved For Release 2007/02/20: CIA-RDP83B00851 R000400070004-2
Secret
Table 5
USSR: Growth Environment in Six Key Oblasts/Republic
Oblast/Republic Key Precipi- ETP b Soil Moisture c
Month tation
Winter and spring
grains
July
July
Belorussia June 75 121 36 35
July 62 146 35 30
Estimated Precipi- ETP Soil Moisture Estimated
Yield tation Yield
(c/ha) Beginning End of (c/ha)
of Month Month
20.3 57 113 53 27 21.4
62 111 27 7
a Precipitation (measured in millimeters of rainfall). Rainfall during c Soil moisture (expressed as a percent of the soil's water-retention
the critical months of June and July are vital to plant growth and de- capacity). A calculated value that incorporates precipitation and
velopment. At this time of year low precipitation usually results in ETP. The lower the soil moisture, the less water available to the
reduced yields. This year, precipitation was low in all spring grain plant's root system. Severe stress usually occurs when soil moisture is
areas. less than 25 percent of total available soil moisture.
b ETP (Evaporation Transpiration Potential). A measurement of
demand placed upon the total plant system. Included in this
parameter are temperature and wind movement. The higher the
ETP values, the more stress the plant endures.
We also have conducted a detailed 1981-82 compari-
son of the growing environment in the key oblasts/
republic during June and July, the critical months
during which winter and spring grains flower and
yield potential is determined (table 5). In general, the
comparison shows that overall growing conditions, as
indicated by the combination of soil moisture, evapo-
transpiration
25X1
Probability of Error Analysis
Based on our track record over the past eight years,
probability analysis shows that CIA estimates are
within 9 percent of actual Soviet grain production two
out of three times (see graph). We are within 12
percent of the reported figure 90 percent,of the time.
However, the number of data points represented by an
eight-year period are statistically too few to generate
a true error probability curve. If applied to this year's
estimate, probability analysis suggests that there is
one chance out of three that we could be off by 15
and precipitation, were only slightly million tons, and one chance out of 10 that we could
improved this year. In the areas where growing be off by 20 million tons. Th:s is overly nessimistic in
conditions were better this year, our published est'- that errors in the early years
mate of yields shows commensurate improvement
25X1
25XI
Approved For Release 2007/02/20: CIA-RDP83B00851 R000400070004-2
 Approved For Release 2007/02/20: CIA-RDP83B00851 R000400070004-2
USSR Grain Production:
Probabilities of Error in CIA Forecasts
were considerabl hi her than those in recent
years. Improvements
Table 6
USSR: Grain Production, Area, and Yield Analysis
1977
1978
1979
1980
1976-80 (average)
1982 Scenarios
Highly unlikely
Production
(million
tons)
Area
(million
hectares)
Yield
(centners/
hectare)
195.7
130.3
15.0
237.4
128.5
18.5
179.2
126.4
14.2
189.1
126.6
14.9
205.0
127.9
16.0
190
122
15.6
180
122
14.8,
170
122
13.9
165
122
13.5
a Production is the unofficially reported Soviet figure for the 1981
crop.
million, or 190 million tons-indicates the following:
? A crop of 190 million tons harvested from an area of
I only 122 million hectares would give a nationwide
have enabled us, on the average, to yield of 15.6 centners per hectare (c/ha)-the
come within 4 percent of actual Soviet grain produc- fourth-highest yield ever. This case is highly
tion over the past four years (assuming that the 1981 unlikely.
crop was 158 million tons). Errors during this period ? A crop of 180 million tons would require a yield of
ranged from -5.6 percent to about 8 percent. A 14.8 c/ha-possible but also unlikely in our
similar range of error suggests that 1982 Soviet grain judgment.
production could fall between 156 and 178 million ? At 1.70 million tons, the yield becomes 13.9 c/ha-a
tons .1 possible achievement.
Historical Yield and Acreage Data
A strong case can be made that this year's grain crop
is not likely to be much larger than 165 million tons, a
figure reached by analyzing historical harvested acre-
age data and average national yields (table 6). This
year's harvested area-estimated to be about 122
million hectares-is the smallest in a decade, and that
constraint alone places upper limits on the amount of
grain that could possibly be produced, given this
year's poor weather. Analysis of three possible
cases-that is, if the crop comes in at 170 million, 180
Assuming that last year's crop was only 158 million
tons, a yield of 12.6 c/ha was obtained from a total
harvested area of 125.5 million hectares. Our pub-
lished estimate of 165 million tons implies a 1982
yield of 13.5 c/ha, which is consistent with the
opinion of most Western experts on Soviet agriculture
that this year's crop is somewhat better than last
year's. The average all rain nationwide yield during
1979-81 was 13.9 c/ha.
25X1
25X1
25X1
25X1
25X1
Approved For Release 2007/02/20: CIA-RDP83B00851 R000400070004-2
 Secret
Approved For Release 2007/02/20: CIA-RDP83B00851 R000400070004-2
Secret
Approved For Release 2007/02/20: CIA-RDP83B00851 R000400070004-2