• TABLE OF CONTENTS
HIDE
 Front Cover
 Half Title
 Title Page
 Table of Contents
 List of Tables
 Foreword
 Acknowledgement
 Summary
 Food insurance as a means of alleviating...
 Estimation of the cost and distribution...
 Conclusion
 Appendix 1. Overview of past approaches...
 Appendix 2. Projected demand, trends,...
 Appendix 3. World wheat price...
 Appendix 4. Supplementary...
 Appendix 5. Country list
 References to Appendix 1
 Back Matter
 Back Cover














Title: Food security
CITATION THUMBNAILS PAGE IMAGE ZOOMABLE
Full Citation
STANDARD VIEW MARC VIEW
Permanent Link: http://ufdc.ufl.edu/UF00085370/00001
 Material Information
Title: Food security
Physical Description: Book
Creator: Konandreas, Panos A.,
Publisher: International Food Policy Research Institute,
Copyright Date: 1978
 Record Information
Bibliographic ID: UF00085370
Volume ID: VID00001
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: 04498802 - OCLC

Table of Contents
    Front Cover
        Page 1
        Page 1a
    Half Title
        Page 2
        Page 3
    Title Page
        Page 4
        Page 5
    Table of Contents
        Page 6
    List of Tables
        Page 7
        Page 8
        Page 9
    Foreword
        Page 10
        Page 11
    Acknowledgement
        Page 12
    Summary
        Page 13
        Page 14
        Page 15
        Page 16
    Food insurance as a means of alleviating food insecurity
        Page 17
        Page 18
        Page 19
        Page 20
        Page 21
        Page 22
        Page 23
    Estimation of the cost and distribution of benefits of the proposed food insurance scheme
        Page 24
        Page 25
        Page 26
        Page 27
        Page 28
        Page 29
        Page 30
        Page 31
        Page 32
        Page 33
        Page 34
        Page 35
        Page 36
        Page 37
        Page 38
        Page 39
        Page 40
        Page 41
        Page 42
        Page 43
        Page 44
        Page 45
        Page 46
        Page 47
        Page 48
    Conclusion
        Page 49
        Page 50
        Page 51
        Page 52
        Page 53
        Page 54
        Page 55
        Page 56
        Page 57
        Page 58
        Page 59
    Appendix 1. Overview of past approaches to the instability of grain markets
        Page 60
        Page 61
        Page 62
        Page 63
        Page 64
        Page 65
    Appendix 2. Projected demand, trends, and variability of cereal production
        Page 66
        Page 67
        Page 68
        Page 69
        Page 70
        Page 71
        Page 72
        Page 73
        Page 74
        Page 75
        Page 76
        Page 77
        Page 78
        Page 79
    Appendix 3. World wheat price generation
        Page 80
        Page 81
        Page 82
    Appendix 4. Supplementary tables
        Page 83
        Page 84
        Page 85
        Page 86
        Page 87
        Page 88
        Page 89
        Page 90
        Page 91
    Appendix 5. Country list
        Page 92
        Page 93
    References to Appendix 1
        Page 94
        Page 95
    Back Matter
        Page 96
    Back Cover
        Page 97
        Page 98
Full Text
RESEARCH REPORT





Food Security:
An Insurance Approach
by
Panos Konandreas
Barbara Huddleston
Virabongsa Ramangkura


4




w r Nfl


The International Food Policy Research Institute is an independent, nonprofit organization that conducts
research on policy problems related to the food needs of the developing world. IFPRI's research is directed
toward policy makers at the national and international level and is distributed to those concerned with food
policy issues.


























FOOD SECURITY:

AN INSURANCE APPROACH





































































The research presented in IFPRI's Research Reports is conducted at the institute; however, the interpreta-
tions and views expressed are those of the authors and are not necessarily endorsed by the institute or the
organizations that support its research.


























FOOD SECURITY:


AN INSURANCE APPROACH




Panos Konandreas
Barbara Huddleston
Virabongsa Ramangkura















Research Report 4
International Food Policy Research Institute
September 1978













































Copyright 1978 International Food Policy Research Institute.

All rights reserved. Sections of this report may be reproduced with
acknowledgement to the International Food Policy Research Institute.






Library of Congress Cataloging in Publication Data
Konandreas, Panos, 1946-
Food security.

(Research report-International Food Policy Research Institute ;4)
Includes bibliographical references.
1. Underdeveloped areas-Food supply.
2. Insurance, Food supply. I. Huddleston, Barbara, 1939- joint author.
II. Virabongsa Ramangkura, 1943- joint author. IIl. International Food
Policy Research Institute. IV. Title. V. Series: International Food Policy
Research Institute. Research report-international Food Policy Research
Institute ;no.4. HD9000.6.K63 338.1'9' 1724 78-11309


ISBN 0-89629-005-0












CONTENTS


Summary

Food Insurance as a Means of Alleviating Food
Insecurity


3. Estimation of the Cost and Distribution of
Benefits of the Proposed Food Insurance Scheme 24

4. Conclusions 49

Appendix 1: Overview of Past Approaches to the
Instability of Grain Markets

Appendix 2: Projected Demand, Trends, and
Variability of Cereal Production 66

Appendix 3: World Wheat Price Generation 80

Appendix 4: Supplementary Tables 83

Appendix 5: Country List 92

References to Appendix 1 94












TABLES


1. How a food insurance scheme operates as a compen-
satory financing mechanism for a hypothetical country
during a five-year period 22

2. Aggregate projected demand, trend production, trend
imports and import variability of food deficit, develop-
ing countries, 1978 to 1982 25

3. Probability distribution of the total costs of the scheme
operating as a compensatory financing mechanism for the
five-year period 1978 to 1982 27

4. Probability distribution of current cost of the scheme
operating as a compensatory financing mechanism for
the middle year (1980) of the five-year period 29

5. Estimated percentage share of benefits from the scheme
operating as a compensatory financing mechanism 30

6. Expected present value of the total cost of the scheme
operating with and without a grain reserve for the five-
year period 1978 to 1982 34

7. Probability that varying grain reserve levels of the scheme
will cover production shortfalls during high-price years,
and expected grain required and released, 1978 to 1982 36

8. Estimated percentage share of benefits when the scheme
combines a 20 million metric ton grain reserve and a com-
pensatory financing mechanism 38

9. Expected present value of the cost of the scheme operating
under alternative insurance levels 42

10. Expected present value of the cost of the scheme operat-
ing as a compensatory financing mechanism in conjunc-
tion with a 20 million metric ton grain reserve under
alternative grain release rules 44








11. Probability of grain reserve availability during high-price
years in a scheme operating as a compensatory financ-
ing mechanism in conjunction with a 20 million metric
ton grain reserve under alternative grain release rules 45

12. Expected present value of cost components of a scheme
operating as a compensatory financing mechanism in
conjunction with a 20 million metric ton grain reserve
under alternative discount rates and per unit carrying
costs 47

13. Expected present value of total cost of a scheme operat-
ing as a compensatory financing mechanism in conjunc-
tion with a 20 million metric ton grain reserve under
alternative discount rates and per unit carrying costs 48

14. Funds required to attain a given probability that the
scheme will achieve its objectives under alternative
methods of operation 50

15. Estimated percentage share of benefits and present value
of expected total withdrawals by country, 1978 to 1982 53

16. Parameters used in cereal demand projections 67

17. Estimates of cereal production trends and variability
from trend by country 72

18. Matrix of correlations between percentage deviations
from cereal production trend for countries by region
76

19. World wheat price distribution 82

20. Probability distribution of present value of the total
cost of the scheme operating as a compensatory financ-
ing mechanism for the five-year period 1978 to 1982,
excluding India 83

21. Probability distribution of the current cost of the scheme
operating as a compensatory financing mechanism for
the middle year (1980) of the five-year period, excluding
India 84








22. Probability distribution of the present value of payments
made through the compensatory financing mechanism
under alternative grain reserve levels 85

23. Probability distribution of current payments made for
the middle year (1980) of the five-year period through
the compensatory financing mechanism under alterna-
tive grain reserve levels 87

24. Probability distribution of the present value of payments
made through the compensatory financing mechanism
under alternative insurance levels 88

25. Probability distribution of present value of payments
made through the compensatory financing mechanism
operating in conjunction with a 20 million metric ton
grain reserve, under alternative grain release rules 90












FOREWORD


The International Food Policy Research
Institute's early research efforts have shown
dramatically what an immense food problem
the Third World and the globe generally
face over the next few decades. Solution to
that problem requires a major commitment
of resources and the political will to back
up that commitment on the part of develop-
ing and developed countries alike. But, even
if that commitment is made, it is becoming
increasingly apparent that lack of purchasing
power in the hands of the poor not only
stands in the way of reaching the objective
of adequate diets for all, but may also pre-
vent latent demand from making itself felt
through remunerative prices. Lacking the
incentive of remunerative prices, production
may not grow at the rates required to meet
the true food needs of the world's popula-
tion a decade or two from now, even though
resources and technology are available to
meet this objective.
That situation has prompted the Insti-
tute's research into policies for increasing
incomes and food consumption of low in-
come people through food subsidies, em-
ployment subsidies and market forces. It
is of course the lower income majority who
spend 50 or 60 percent of additions to their
income on grain alone, as compared to the
5 or 10 percent of added income spent on
grain by the upper income deciles. Therefore
they are hurt more when supply variation
forces prices up.
It has become apparent that for govern-
ments to make the political and economic
commitments essential to increase effective
demand for food by the lower income


people, there must be a reasonable degree
of certainty that this increased demand can
be consistently met. Given the inevitable
vagaries of weather and the variability of
world prices, the high cost of national
schemes of food storage and the intense
demand for resources for development it-
self, it is essential that a viable international
system of food security be devised. In re-
sponse to this need, the International Food
Policy Research Institute has carried on a
research program dealing with various
aspects of food security. This Research
Report reflects part of this work. It suggests
an approach based on insurance principles,
by which the international community can
contribute to the food security of food def-
icit developing countries without having to
create large buffer stocks and stabilize world
grain prices. The objective of such a scheme
would be to permit developing countries to
stabilize cereal consumption at a relatively
stable cost.
The reserve proposal which is in the fore-
front of international discussion at the mo-
ment is that being considered by the Inter-
national Wheat Council under the auspices
of the United Nations Conference on Trade
and Development. Whether countries can
agree on a formula for a system of nationally
held, internationally coordinated buffer
stocks which would be operated in response
to price movements remains to be seen. If
no agreement is reached, the importance of
considering a more limited approach to the
food security problem of low income, food
deficit developing countries is obvious. Even
if agreement is reached, however, the food








security needs of many developing countries
will still not be fully met.
Operation of an international reserve
designed to deal with price instability in
world markets does not eliminate the need
for consumption adjustment in participating
countries when prices exceed the release
trigger. Developed countries with substantial
livestock sectors can more easily make this
adjustment than can developing countries in
which a substantial portion of the popula-
tion lives at the subsistence margin. Further,
even if grain is released onto the market in
years of relatively high prices, developing
countries would still have to bid competi-
tively with other importers to obtain needed
supplies. In such years, developing countries
with exceptionally poor harvests, foreign ex-
change constraints, and relatively small
orders would be relatively weak bidders.
Thus, in order to avoid curtailing consump-
tion, they might still have no choice but to


pay excessively high food import bills in
relation to what their economies can afford.
Finally, release of grain in response to price
triggers would not take into account the fact
that countries with less flexibility in making
transport and handling arrangements for
substantially increased imports would be at
a disadvantage in obtaining access to this
grain.
This report is part of an ongoing project
of IFPRI's Trade Program, under the general
direction of Alberto Valdes. Other aspects
of this project include a quantitative assess-
ment of the sources and magnitudes of food
security in developing countries; the comple-
mentarity of grain reserves, food aid and
food insurance; and a series of country
studies assessing in detail the characteristics
and magnitude of food insecurity and possi-
bilities of implementing alternative
approaches in Asia, Africa and Latin
America.


John W. Mellor


Washington, D.C.
September 1978


























ACKNOWLEDGEMENTS

We wish to acknowledge earlier contributions of D. Gale Johnson, Dale Hathaway and
Shlomo Reutlinger, which led to the approach followed in this study. Several individuals have
read an earlier draft of this study. We are grateful to the comments by D. Bigman, R. Fox,
D. G. Johnson, D. Morrow, S. Reutlinger, A. Sarris, R. Simpson, A. Valdes, and J. Zapata.
In addition, thanks are due to Marcia Mason for her programming skills, J. Hayssen for his
assistance in assembling and handling the data, Kate Hathaway and Philippus Willems for their
editorial comments and Lucy McCoy for her careful typing of the manuscript.










1


SUMMARY


This paper suggests an approach by which
the international community can contribute
to the food security of food deficit, develop-
ing countries without having to create large
buffer stocks and stabilize world grain
prices. The international community is
considering an agreement which would es-
tablish minimum and maximum indicator
prices for wheat, and use buffer stocks to
defend them. But important differences
among major participants about commodity
coverage and the use of production adjust-
ment and trade policy measures make agree-
ment unlikely. A more acceptable alternative
may be a scheme based on insurance princi-
ples specifically designed to assist food defi-
cit, developing countries. The objective of
such a scheme would be to permit develop-
ing countries to stabilize cereal consumption
within a range of projected demand at a
relatively stable cost.
Two alternative insurance schemes were
evaluated for sixty-five food deficit, develop-
ing countries for the period 1978 to 1982:
a purely compensatory financing mecha-
nism, and a financing mechanism combined
with a physical wheat reserve. In the latter
case, stocks would be released only during
very high-price years (price above $200 per
metric ton (MT) or $5.45 per bushel of
wheat) and only to countries experiencing
a production shortfall of more than 5 per-
cent during those years. On the basis of pres-
ent market conditions it was assumed that a
physical reserve could be acquired from this
year's crop at the prevailing price. Reserve
levels of 4, 8, 12, 16, and 20 million tons
were considered.


Five years was considered a reasonable
period for making realistic statistical es-
timates. An agreement could be negotiated
for a five-year period, but the anticipated
duration would have to be considerably
longer for the scheme to operate effectively.
Compensation from the scheme would be
permitted whenever a developing country's
cereal import bill exceeded a specified per-
centage of the trend import bill (e.g., 110,
120, or 130 percent of trend). The consump-
tion level defended by the scheme at a given
year depends on a country's cereal produc-
tion during that year. Thus, if its domestic
production shortfall is more than 5 percent
below trend, the country's actual bill would
be calculated for the quantity of imports
that would maintain 95 percent of projected
demand. If the shortfall is between 95 and
100 percent of trend production, consump-
tion would be maintained at the same per-
centage of projected demand. Finally, if
production exceeded trend, consumption
would be maintained at 100 percent of pro-
jected demand. These rules would ensure
that each country could maintain a con-
sumption level between 95 and 100 percent
of projected demand in all years, depending
on the performance of its own cereal pro-
duction. However, consumption is not
restricted to the guaranteed level. A country
can maintain its own supplementary reserves
and/or allocate additional foreign exchange
to food imports if it wishes to support a
higher consumption level.
Because of the sharp fluctuations in do-
mestic production in many developing coun-
tries, there would be at least some countries








eligible for compensation each year. Once a
country's domestic production for a given
crop year was known and enough informa-
tion was available to permit an estimate of
the expected world market price, the prob-
able cost of import requirements could be
estimated. Funds could then be made avail-
able to eligible countries. Some mechanism
would have to be devised to tie these funds
to the purchase of the specified quantity of
grains for food consumption. Since in most
years the world price would not be high
enough to trigger release of grain from the
reserve, the assistance to eligible countries
would be in the form of compensatory fi-
nancing. The wheat reserve would be drawn
upon when world market prices exceeded
the release price, and would be used only
to compensate for production shortfalls of
5 percent or more below trend production
in those years. This released grain would be
valued at the prevailing world price, and the
amount subtracted from the total compensa-
tion for which each country is eligible. The
remainder of the payments due to countries
would be provided through compensatory
financing. In the event that production
shortfalls and/or price increases were so ex-
treme that resources would be exhausted be-
fore meeting the requirements of all eligible
countries, the consumption level defended
by the scheme would be reduced for each
country and the additional consumption ad-
justments necessary would be shared propor-
tionately by all participating countries.
The scheme is least costly with a grain
reserve of about 8 million tons; with a re-
serve of about 16 million tons the cost is
equal to that of a scheme operating solely
as a compensatory financing mechanism.
However, the differences between total
expected costs at various reserve levels are
so slight that they have no practical signifi-
cance as a measure for deciding which alter-
native is preferable. More important is the


effect of the reserve on the distribution of
the cost of the scheme: the larger the grain
reserve held in the system, the lower the
probability of very high cost. Consequently,
for a given level of funding, the larger the
grain reserve, the greater the probability of
achieving the objectives of the scheme. In
addition, a larger grain reserve provides a
higher and more equitable probability over
the years of covering production shortfalls
of 5 percent or more below trend produc-
tion during high-price years. For these
reasons if a grain reserve was established in
conjunction with a compensatory financing
mechanism, 20 million tons of grain would
be the suggested reserve level for developing
countries.
In addition to these technical considera-
tions, a scheme with both funds and stocks
has the advantage of providing a supply
guarantee to back up the financial insurance,
and is likely to be preferred by potential
developed country contributors. Without a
physical reserve, the additional purchasing
power acquired by developing countries
could, in periods of particularly short sup-
ply, pressure developed countries to make
politically unacceptable adjustments in their
own domestic consumption or cause the
scheme to fail because of the imposition of
export controls. The physical reserve could
also provide an outlet for surplus stocks
which tend to accumulate in certain export-
ing countries.
Alternative insurance levels for a scheme
with a 20 million ton wheat reserve were
evaluated and their costs compared to those
for a scheme involving a compensatory fi-
nancing mechanism only. The financial ca-
pacity needed to attain a given probability
that the scheme will achieve its objectives
under various alternatives is also calculated.
Two important findings are worth mention-
ing here. First, a given level of funds pro-
vides a higher probability that the scheme









will achieve its objectives when the scheme
involves grain reserves in addition to com-
pensatory financing. Second, the marginal
funds needed to increase the probability
of the scheme achieving its objectives are
larger the higher the probability level.
It should be noted that for a given level
of funding, higher insurance coverage im-
plies a greater risk of depleting funds during
the last years of the five-year period; there-
fore, benefits are unevenly distributed over
time and among countries, favoring coun-
tries that happen to draw early during the
five-year period. For this reason, for a given
level of funds the insurance coverage should
be set at a level that would imply a high
probability for the scheme achieving its
objectives over the five-year period. Assum-
ing that for most countries concerned a
cereal import bill up to 130 percent of trend
during some years would not be a formid-
able obstacle in meeting their consumption
targets, a fund of $3.7 billion would cover
the expected cost of a food insurance
scheme. Such a level of funding would as-
sure at least a 75 percent probability of the
scheme achieving its objectives. Additional
funding of about $2.4 billion would increase
this probability to about 90 percent.
The share of each country's benefits
from the system has also been estimated
based on the expected total withdrawals for
a given country and the expected total cost
of the scheme for the five-year period. India
receives more than 20 percent of the bene-
fits, followed by Morocco, Mexico, and
Turkey. Six or seven countries account for
almost 50 percent of the cost of the system.
Crucial components in the success of any
food security scheme are the source and


management of funds needed for its opera-
tion. One option is a scheme self-financed by
the member countries. For example, if India
had to pay an equal annual premium for the
next five years based on its expected with-
drawals, the premium would range from
about $206 million to about $245 million,
depending on the level of insurance provided
by the scheme. In practice, however, most
low income countries could probably not
afford to participate unless the scheme were
subsidized by developed countries. De-
veloped countries could make their contri-
butions either bilaterally or through a multi-
lateral mechanism. Ideally, developed coun-
try assistance would be most effective and
desirable from the point of view of recipient
countries if it was a collective contribution
that would be paid through an agreed upon
cost-sharing arrangement. Donor countries
could subsidize the premium payments of
low income countries or make concessional
food aid available to lower the cost of im-
ports for recipient countries and thereby
help them meet their food security pre-
miums. Alternatively, funding countries
could simply agree on the level of financial
commitment they were willing to undertake
collectively, and a scheme could be designed
accordingly.
Whether funded by a schedule of pre-
mium payments or by a schedule of donor
country contributions, the scheme should
be funded to cover at least its expected cost
at a given insurance level. Borrowing might
be necessary if its cost exceeds the amount
of funds available.
Finally the scheme could be funded
through the compensatory financing facility
of the International Monetary Fund (IMF)


This figure also includes the cost of $1.1 billion needed for the operation of the 20 million ton grain
reserve.








by including cereal import expenditures in
the IMF's existing compensatory financing
facility for commodity exports or by creat-
ing a new facility. Funding the scheme
through the IMF may be the most desirable


approach, both because the IMF could
handle the necessary financing arrangements
and because the bureaucracy needed to oper-
ate it already exists.








2


FOOD INSURANCE AS A MEANS
OF ALLEVIATING FOOD INSECURITY


Much of the debate in recent years over
how to attain food security has concentrated
on the desirability of accumulating large
buffer stocks with the objective of stabiliz-
ing the world price of grains.* A number of
developed countries, both importers and ex-
porters, advocate such an approach for
reasons having to do with the operation of
their domestic food and agricultural policies.
For developing countries, price stability re-
sulting from buffer stocks would mean mod-
eration of the often wide fluctuations in
foreign exchange expenditures required to
purchase food imports. But developing
countries would have no special advantage
over other buyers in obtaining grains from
the buffer stock in years of short supply.
Moreover, even with grain price stabilization,
a country will still have a food insecurity
problem to the extent that its domestic
production fluctuates.
Apart from the degree of food security
provided, several problems are inherent in
such an approach. Although discussions are
continuing in the International Wheat
Council, it is not clear that nations could
agree on the size, location, price bands and
administration of the buffer stocks that
would form the backbone of a price stabili-
zation scheme. There are important differ-
ences among major participants about what
production adjustment and trade policy
measures should be used if buffer stocks are
not adequate to keep prices within the speci-


fied range. If a group of countries reached
agreement on buffer stocks without inclusive-
participation, member countries would want
to exclude non-member countries from the
benefits of stable prices. In order to do this,
members in the scheme would have to de-
sign some system of export controls or dis-
criminatory pricing to bar non-members
from reaping benefits from the partial buffer
stocks. While it might be possible to bar
non-members, such action would invite
political or trade retaliation that could ul-
timately break the system.
Even if a group of nations could agree in
principle on a price stabilizing arrangement,
agreement on a cost-sharing formula would
be difficult to work out. Although under
certain assumptions one can infer which
nations stand to gain the most from stable
prices, it is not possible to estimate ac-
curately the actual monetary costs and ben-
ifits. Moreover, a price stabilizing arrange-
ment does not allow individual countries to
select the degree of food security they desire
and does not allow them to reduce their
share in the cost. In some countries food
production is highly variable; in others, it
is relatively stable. In addition, some nations
have enough foreign exchange to absorb vari-
ations in domestic production by buying
grain in the international market. Therefore,
the degree of food security desired will vary.
Finally, pursuit of short-run self-interest
by grain exporters and importers thwarts the


A brief review of the literature on the welfare implications of price stabilization and the experimen-
tal work on grain buffer stocks is presented in Appendix 1.








negotiation and implementation of a price
stabilization arrangement. During years of
good harvests, when low grain prices and
ample supplies encourage the initiation of a
buffer stock system, importing nations see
no urgent need for one, because they realize
that major grain producing nations will
unilaterally store excess supplies for release
in years of short supply anyway. On the
other hand, when poor harvests have forced
grain prices up, exporting countries show
little interest in buffer stocks that would
set an upper limit on the international
grain price.
This obstacle to grain stocks clearly
would disappear if policy makers took
longer-term considerations into account.
However, the usually short planning horizon
of governments and the political pressures of
interest groups in both exporting and im-
porting nations bias decisions toward im-
mediate gains. As a result, proposals for
grain buffer stocks face serious obstacles
caused by the divergence of short-run inter-
ests between importing and exporting coun-
tries.



Description of the Insurance Approach

Guidelines and objectives

By limiting the scope of an agreement
and specifically addressing the needs of food
deficit, developing countries, the interna-
tional community could avoid some of the
problems associated with a full-scale market


price stabilization agreement. One way
would be to design an agreement along in-
surance principles.
Johnson* has proposed a food insurance
scheme which calls for the United States and
other industrial countries to cover produc-
tion shortfalls in developing countries when
cereal output drops below a given percentage
of their production trends. Although specific
to developing countries, this proposal treats
only one of the two aspects of food inse-
curity.
The two sources of food insecurity in de-
veloping, food deficit countries are: (1) a
temporary reduction in domestic food pro-
duction and (2) a temporary increase in
international foodgrain prices. The Johnson
proposal deals only with the first source of
food insecurity and implicitly assumes that
food deficit, developing countries can
overcome the second source by varying their
foreign exchange expenditure on food im-
ports. But even if such variable spending of
scarce foreign exchange were possible, for
many developing countries it could severely
hamper overall economic development.t
Therefore, to protect developing countries
from both sources of food insecurity, a food
security scheme must deal with fluctuations
in their food import expenditures.
Reutlingert has redefined food insecurity
of developing countries by recognizing the
importance of fluctuating foreign exchange
requirements. His study estimates the cost of
financing food imports needed to maintain
consumption in developing countries when
their food import bills exceed normal.


D. Gale Johnson, "Increased Stability of Grain Supplies in Developing Countries: Optimal Carry-
overs and Insurance," World Development 4 (1976):977-987.

t Dale E. Hathaway, "Grain Stocks and Economic Stability: A Policy Perspective," in Analysis of
Grain Reserves, a Proceedings. ERS-634. (Washington, D.C.: U.S. Department of Agriculture and the
National Science Foundation, 1976).

t Shlomo Reutlinger, Food Insecurity: Magnitude and Remedies, World Bank Working Paper No. 267.
(Washington, D.C.: World Bank, July, 1977).








This study follows Reutlinger's approach
in principle, but adds a number of refine-
ments. First, rather than treating develop-
ing countries as a group, this study estimates
the level and variability of grain production
and the import requirements of each coun-
try individually, taking production and con-
sumption growth into account. Second, it
clearly specifies operating rules at the coun-
try level, allowing an objective estimate of
benefits received by each country. Finally,
this study estimates the cost of an insurance
scheme for a period of five consecutive years
rather than for a single independent year.
The objectives of the food security
scheme analyzed here, stated more explic-
itly, are: (1) to permit developing countries
to stabilize food consumption within a range
of projected food demand, and (2) to per-
mit developing countries to stabilize their
food import bill within a range of their
trend import bill while stabilizing food con-
sumption. Adjusted target consumption, the
consumption level defended by the system,


varies between 95 and 100 percent of the
projected cereal demand for each country.
Although a country's actual consumption
levels might be higher or lower than adjusted
target consumption, adjusted target con-
sumption constitutes the basis for determin-
ing year-to-year transactions between a
country and the system. *
In years of very high world prices and/or
very low domestic production, a country
could be eligible to draw from the system's
resources in order to maintain its adjusted
target consumption level. Withdrawals from
the system depend on the variability of each
country's cereal import bill and on the de-
gree of insurance desired. t The system
would make payments only for the amount
by which a country's actual import bill ex-
ceeds a certain percentage of its trend cereal
import bill. Should a country desire a high
level of insurance, it could set this uninsured
import bill level very low, and thus draw
frequently from the system. This discussion
assumes an uninsured level of 110 percent


Adjusted target consumption in the ith country during a particular year (Cit) is determined on the
basis of actual grain production level in the country for that year. It is set lower than projected demand
( it) when a country experiences a production shortfall. Thus, if actual production for a particular
year (Qit) is below 95 percent of trend production (Qit), then

Cit = 0.95 Cit

If Qit is between 95 percent and 100 percent of Qit, then Cit is set at the same level i.e.,



it 0 it
Qit -
Cit = i- cit

and finally, if Qit is above Qt then C = Cit. Appendix 2 describes the analytical expressions used
-in obtaining projected demand and trend production.

t Throughout this paper, the term cereal is used to refer to wheat, rice and feedgrains. The composi-
tion of cereal Imports varies from country to country and even within countries, depending on consump-
tion preferences, and the variability of the composition in its domestic cereal production. For purposes
of this study, wheat is taken as the basic staple commodity, and the value of cereal imports is computed as
if these imports were composed entirely of wheat. The term grain is used in the context of discussions of
physical reserves which could be comprised either of wheat only or of wheat and rice. The grain reserve
discussed in this paper is assumed to be wheat.








for all countries. In other words, if a coun-
try's target food imports in a given year
(computed as adjusted target consumption
minus current production), valued at the
current world price, exceeded 110 percent
of its trend import bill, then the country is
eligible for compensation from the system
for the excess bill over 110 percent of trend.
Reducing the consumption level that the
system defends when domestic production
falls below trend production provides a
country with incentives to maintain produc-
tion at least at its trend level. If the system
provided coverage for 100 percent of each
country's projected demand, then there
would be no means of preventing a country
from drawing excessive short-term benefits
by deliberately under-reporting domestic
production and/or reducing resources allo-
cated to food production. $
Over time, under-reporting would result
in a decrease in the rate of growth of domes-
tic production, and would thus lead to an
increase in trend imports. Because the sys-
tem compensates a country based on the dif-
ference between its actual import bill and a
percentage of its trend import bill, an in-


crease in trend imports caused by under-
reporting would reduce the compensation
for which the country qualifies.
An international food insurance scheme
that would provide member countries the
resources to defend their consumption can
operate in either of two ways: (1) the
scheme can serve as a purely financial mech-
anism that provides member countries with
funds to cover overruns in their cereal im-
port bill; or (2) the insurance scheme can op-
erate a limited size grain reserve in addition
to the financial mechanism; grain would be
released to eligible member countries when
world food supplies are very short. Both of
these alternatives will be examined.


A Food Insurance Scheme as a
Compensatory Financing Mechanism

A food insurance scheme operating as a
purely financial compensatory mechanism
implies that the system provides member
countries with the funds necessary to cover
overruns in their cereal import bill. Recipi-
ent countries use these funds to import food


Two alternative levels of 120 and 130 percent are also evaluated.

t Trend import bill equals the product of trend imports (projected demand minus trend production)
and the average world price.

t There is another reason for not providing 100 percent coverage: the primary source of income in
most of the developing countries is agricultural production. Thus a high correlation exists between rural in-
come levels and grain production. This correlation implies that in years of low grain production, effective
demand for grain Is below projected demand (based on a smooth normal growth in domestic income).
Therefore, aside from providing proper disincentives to countries in cases of below-trend food production,
an additional reason for setting adjusted target consumption below projected demand is due to an actual
reduction in effective demand because of income Icsses in part of the population.
It can be argued that the purpose of a food security system should be to prevent such a reduction in
effective demand, particularly among the rural poor, who are most likely to suffer nutritionally from the
reduction in consumption. However, making available food supplies in excess of those implied by effective
demand could have undesirable consequences unless governments are willing to make appropriate adjust-
ments in their internal food distribution system in favor of the poor majority (including creation of domes-
tic reserves to supplement this scheme). If such measures exist, and the resources of the food security
scheme permit, the minimum target consumption level to be defended by the scheme could be set at higher
than 95 percent of trend consumption. In these ways it should be possible to minimize, if not eliminate,
the hardships caused by domestic production shortfalls.









to meet their consumption targets.
Table 1 presents several possible trans-
actions between a hypothetical country and
the system. In year 1, actual production in
the country in question is 90 percent of
trend. Thus the rules of the system set
adjusted target consumption at 95 percent
of projected demand or 10.45 million metric
tons. The difference between adjusted target
consumption and actual production, multi-
plied by the current world price, yields the
size of the target import bill. Because the
target import bill for year 1 ($137.8 million)
is less than 110 percent of the trend import
bill ($165 million), the country receives no
compensation from the system. In year 2,
actual production is 98.06 percent of trend
and adjusted target consumption is set at
11.18 million metric tons (i.e., 98.06 per-
cent of projected demand). With the target
import bill smaller than 110 percent of
trend, the country again receives no compen-
sation in year 2. In the third year, actual
production exceeds trend, so adjusted target
consumption equals projected demand. The
combination of a good domestic crop and a
low world price during this year results in a
target import bill considerably lower than
the trend import bill. The country is again
ineligible for compensation from the system.
In year 4, low domestic production and high
world price result in a much higher import
bill. The target import bill equals $395.0
million, considerably more than 110 percent
of trend import bill (which is equal to
$214.5 million). Thus, the country is eligible
for compensation in the amount of $180.5
million ($395.0 million minus $214.5 mil-
lion). Finally, in year 5, although the coun-
try has a better than average crop of its own,
a continued high world price results in an
import bill that still exceeds 110 percent of
trend import bill. Thus, the country receives
compensation in the amount of $3 million.


A Food Insurance Scheme as a
Combined Grain Reserve and
Compensatory Financing Mechanism


Thus far the discussion described the op-
eration of a food insurance scheme as a com-
pensatory financing mechanism only. A sys-
tem that used a grain reserve in conjunction
with a compensatory financing mechanism
might provide a more cost-effective way of
achieving the same objectives. Moreover, the
presence of a grain reserve adds an important
supply guarantee to the system. In high price
years commercial stocks are ordinarily very
low or nearly depleted, so that merely pro-
viding financing in these years cannot guar-
antee availability of supplies. In fact, com-
pensatory payments may bring only higher
prices and greater financial strain on food
importing countries as buyers bid for scarce
supplies on the world market. Exporting
countries may also prefer a system with a
grain reserve, since the reserve would protect
them against consumption adjustments or
the need to impose export controls during
high-price years. Appropriately managed, a
relatively small reserve could help to guaran-
tee supplies and conceivably reduce the cost
of the system while minimizing the disrup-
tive effects on the international grain mar-
ket.
A food security system with a limited
reserve would have to specify price levels at
which it would acquire and release grain.
The correct choice of these price levels is
crucial to the system's success. On the one
hand, a low acquisition price and a high re-
lease price mean less interference with the
world grain market. On the other hand,
there should be a reasonable probability
that the world price will trigger acquisition
and release sometime during the period for
which an international insurance scheme is










Table 1- How a food insurance scheme operates as a compensatory financing
mechanism for a hypothetical country during a five-year period


Year 1 Year 2 Year 3 Year 4 Year 5


Current world price of
wheat ($/MT) a/

Cereal production trend
(million MT)

Projected cereal demand
(million MT)

Trend cereal imports
(million MT) b/

Trend cereal import bill
(million $) c/

110 percent of trend import
bill (million $)

Actual cereal production
(million MT)

Actual production as
percentage of trend
production (%) d/

Adjusted target consump-
tion (million MT) e/

Target cereal imports
(million MT) f/


Bill for target imports
(million $) g/

Compensation to
country (million $) h/


95.0 140.0


1.0 1.1


150.0 165.0


165.0


181.5


9.0 10.1


90.0



10.45


1.45


98.0



11.18


1.08


137.8 151.2


0.0 0.0


110.0


10.7


11.9


1.2


180.0


198.0


11.0


102.8



11.90


0.9


250.0


11.2


12.5


195.0


214.5


10.3


92.0



11.88


1.58


395.0


0.0 180.5


a/ For this illustration only, assume that the average world price of wheat equals $150/metric ton.
b/ Computed as the difference between row 3 and row 2.
c/ Computed as the product of row 4 and the assumed average world price of wheat ($150/metric ton).
d/ Computed as the ratio between row 7 and row 2.
e/ Computed by adjusting projected demand (row 3) on the basis of the values of row 8 as explained in the
text.
f/ Computed as the difference between row 9 and row 7.
g/ Computed as the product of row 10 and row 1.
h/ Computed as the difference between row 11 and row 6 when the former is larger than the latter.


180.0


11.7


13.1


210.0


231.0


11.8


100.9



13.1


1.3


234.0








negotiated.
If, for example, a food insurance scheme
has an initial projected life of five years, a
release price reached in only one year out
of thirty would be unrealistic, since there
would be a very low probability that the sys-
tem would use its grain reserve. Conversely,
a release price that is reached in two years
out of four would interfere unnecessarily
with the free market. With these considera-
tions in mind, a release price level which
will be exceeded in one year out of five and
an acquisition price which will be exceeded
in four years out of five have been assumed.
Given the world price distribution for wheat
(see Appendix 3), an acquisition price of
$105 per metric ton (about 80 percent prob-
ability of being exceeded) and a release price
of $200 per metric ton (about 20 percent
probability of being exceeded) are speci-
fied. Using these prices, the system could
expect to acquire grain in one year out of
five and to release it in one year out of five.
Once the world price reaches the release
price, production levels in member countries


will determine which ones are eligible for
compensation in the form of grain. When
high prices trigger the grain release mech-
anism, countries whose production falls
short of trend by more than 5 percent are
eligible to bring their grain supplies up to
95 percent of trend by drawing upon the
system's reserves, to the extent that such
reserves are available. t
Thus, consider -again the hypothetical
country illustrated in Table 1. During year 4,
the high-price year, domestic production in
the country in question is 92 percent of
trend production, so the country is eligible
to draw from the grain reserve facility of the
system. The 95 percent level of trend pro-
duction equals 10.64 million metric tons, so
the country receives 0.34 million metric tons
(10.64 minus 10.30) of grain. t Use of the re-
serves reduces the compensatory financing
received by the country in question for that
year by the value of the grain released to it,
(i.e., by $85 million [0.34 x 250]). The
country receives the remaining $95.5 million
(180.5 minus 85) in compensatory funds.


The effect of an alternative release price of $170/metric ton is also analyzed. The probability that
world price would exceed this level in a given year is about 32 percent.

t An alternative level of this country-specific trigger for grain release is also considered. In this case,
grain Is released to countries that experience a production shortfall greater than 3 percent during high price
years and in amounts (to the extent available in the system) that will bring the physical availability of grain
in the country to 97 percent of trend production.

If the grain available in the grain reserve facility of the system is not enough to cover all requests,
then each country's allotment is reduced in proportion to the available reserve.








3


ESTIMATION OF THE COST AND DISTRIBUTION OF

BENEFITS OF THE PROPOSED FOOD INSURANCE SCHEME


Components of the Simulation Model
Used in the Evaluation of the
Proposed Food Insurance Scheme


To evaluate the proposed scheme a model
was built which simulates projected demand,
trend production and production variability
on a country by country basis, as well as the
variability of the world price of wheat.
A country's projected cereal demand is
obtained on the basis of past per capital
cereal consumption and projections of popu-
lation growth, GNP growth, and income elas-
ticities of demand. Appendix 2 describes the
analytical expression used and the data on
which demand projections are based.
Analysis of each country's past produc-
tion provides cereal production trends and
production variability. Trend production is
based on a logarithmic time trend derived
from data on total cereal production for the
period 1960 to 1975. Past variability of pro-
duction above and below trend is estimated,
and cross-country correlation in production
is also analyzed. Appendix 2 presents the
methodology used and results obtained from
this analysis.
Cereal import trends and the variability
of imports by country are derived by sub-
tracting simulated future production out-
comes from projected demand. Table 2
presents aggregate figures for projected de-
mand, trend production, trend imports and
variability of imports. Aggregate projected
demand of food deficit, developing countries
grows by about 3.3 percent annually, where-


as domestic grain production grows by about
2.9 percent annually. As a result, aggregate
trend import requirements will grow from
about 42.0 million metric tons in 1978 to
about 52.4 in 1982.
Production and consequently, import re-
quirements fluctuate around trends, and due
to the assumed multiplicative nature of pro-
duction variability (see Appendix 2), the ab-
solute size of fluctuations increases over
time. Thus, the standard deviation of pro-
duction increases from 9.6 million metric
tons in 1978 to about 10.6 million metric
tons in 1982. On the basis of these standard
deviations of production, a range in the vari-
ability of import requirements can be com-
puted (see Table 2).
Growth and variability of cereal produc-
tion vary greatly from country to country,
as shown in Table 17, Appendix 2. In addi-
tion, though percentage deviations from
grain production for the countries con-
sidered in aggregate are rather low (standard
deviation of about 3 percent), percentage
deviations on an individual country basis
are much greater. Cereal production variabil-
ity expressed as a percentage of trend is
highest in North Africa and the Near East
and lowest in Southeast Asia. However,
Southeast Asia's production variability in ab-
solute levels is much greater because its pro-
duction level is much higher.
The world wheat price generating func-
tion is presented in detail in Appendix 3.
Analysis of past correlation between total
developing countries' cereal imports and
world wheat price reveals the causal effect
















C,

ar


o














oc
a =
3

L. (t




.0



3



































0
-e





















j,
(0p


o00















n


3.











0.



0O







a.


*o







Q.



0
04
3
M,


w) w N)
0 0 %V








Uw W Ul
4b. w "s
00 %0 00
w -_ 00


b\ oo o s




-J -4 0Ch 0h 0a

ts) K) 0 0 00


n 0
o 0 2D


a


Q. S
aao



- .0
C CD






CCL





aa~
o 3
C C '
P




CD a,
3 3 ^
















CL c
2. S: o
S. Sa










m




V a
o o.
0. 5
















o
CCD
a
0.









a 0
a a
0 I
Q. 0
a 0

a. a





*0 0-

N) )(
N)


z '0 Z 0 ;z 0 '8 0
00 00 00 -J -4
t) -. 0 '0 00























o -4 %0
" 00 00 i0 0











N) 0 0. W M















m w Q 000 h


a.
3o
c













0
30








o










04
0 .

































SU
CL


















o a








0.
a5





























^I%


C,
0







(3

00



00


c?3






















C,
U'.




















o
0.







-4
o

















0.
-.







C,






0








of these imports on world price and the
price generating function incorporates this
effect. The serial correlation observed in
time series price data is also taken into con-
sideration by incorporating a lagged price
into the price generating function. All other
factors that have an impact on world wheat
price formation are treated as a random vari-
able. The probability distribution obtained
for world wheat price is depicted in Table
19, Appendix 3. The resulting expected
price equals $155.8 per metric ton.



Estimated Cost and Distribution of
Benefits of the Scheme Operating as a
Compensatory Financing Mechanism


A food insurance scheme operating as a
compensatory financing mechanism would
provide member countries the foreign ex-
change needed to defend their adjusted
target consumption when their import bills
for that level of consumption are excessive.
The simulation model generates 300 five-
year (1978 to 1982) cereal production se-
quences for each member country, and 300
corresponding five-year world wheat price
sequences. As explained earlier, the cereal
import requirement for a given year is com-
puted as the difference between adjusted
target consumption and the production level
randomly generated by the model. The
cereal import requirement, multiplied by the
corresponding world price for that year, pro-
vides the target cereal import bill for the
country in question. This import bill is then
compared with the country's trend cereal


import bill.* When its target cereal import
bill is above the uninsured level (e.g., 110
percent of trend import bill), the country re-
ceives compensation from the system for the
amount of the excess.
The summation of all payments to mem-
ber countries for a given year constitutes the
cost of the scheme for that year. The cost of
the scheme for a five-year sequence is ex-
pressed in present value by discounting fu-
ture expenditures The outcomes of all 300
sequences are summarized by probability
distributions.
Table 3 presents the probabilities of dif-
ferent costs of operating the scheme. For ex-
ample, the probability that the total cost of
the scheme for the five-year period will be
$3 billion or less is 53.8 percent. Similarly,
there is a probability of about 95 percent
that the total cost will be less than $16 bil-
lion. The expected cost of the scheme repre-
sents the mean of the present values of the
costs generated for the 300 five-year se-
quences. This expected cost for the five-year
period equals $5.1 billion. Because the cost
distribution is rather skewed, (median cost
less than $3 billion) the probability that the
cost of the scheme will exceed the expected
cost of $5.1 billion is about 30 percent. In
other words, if the scheme were to be set
up with a financial capacity equal to its ex-
pected cost of $5.1 billion, this amount
would provide about a 70 percent guarantee
of meeting all needs for the entire five-year
period. Higher degrees of assurance would
require substantially more funds. For ex-
ample, a 95 percent guarantee would require
about $16 billion.
One important qualification should be


The trend cereal import bill is computed as the difference between projected demand and trend pro-
duction for a given year, multiplied by the expected world wheat price ($155.8/metric ton).
A discount rate of 8 percent per annum is assumed here. Alternative discount rates are also analyzed
in a later section of this paper.









Table 3-Probability distribution of the total cost of the scheme operating as a
compensatory financing mechanism for the five-year period, 1978 to
1982

Present Value of Relative Cumulative
Total Cost a/ Frequency b/ Frequency c/
($ billion) (%) (%)


0-1 20.0 20.0

1-2 21.8 41.8

2-3 12.0 53.8

3-4 7.7 61.5

4-5 8.0 69.5

5-6 5.0 74.5

6-7 4.8 79.3

7-8 2.2 81.5

8-9 2.1 83.6

9-10 3.3 86.9

10-11 1.5 88.4

11-12 1.4 89.8

12-13 1.5 91.3

13-14 1.4 92.7

14-15 0.4 93.1

15-16 1.1 94.2

16-17 1.1 95.3

17-18 1.1 96.4

greater than 18 3.6 ...


a/ The expected present value of the cost equals $5.1 billion. A discount rate of 8 percent has been as-
sumed here. The system is assumed to cover the excess over 110 percent of trend in each country's
current cereal import bill.
b/ The relative frequency of a cost range is the probability that the scheme's total cost will fall in that
range.
c/ The cumulative frequency of a cost range is the probability that the scheme's total cost will fall in or
below that range.







made here. The cost figures presented in
Table 3 represent the total cost of the
scheme for the five-year period. Yearly costs
are much lower. For example, the distribu-
tion of the cost of the scheme for the middle
year (1980) is presented in Table 4. The fig-
ures in this table indicate that there is ap-
proximately a 95 percent probability that
the cost of the scheme in 1980 will be $5
billion or less. The expected cost for that
year is $1.2 billion.
An alternative way of expressing the
yearly cost of the scheme is in the form of
an average annual cost. The average annual
cost represents an equal annual expenditure
in each of the scheme's five years that would
yield a present value equal to the expected
cost. An average annual cost of $1.183 bil-
lion yields a present value of $5.1 billion as
the scheme's expected cost.
Another important dimension of the an-
nual cost of the scheme is how this cost
grows over time. As shown in Table 2, the
standard deviation of cereal import require-
ments increases in absolute terms with time,
because a given percentage variability ap-
plied to increased total production yields
greater absolute variations in output and in
import requirement. In addition, the price
distribution also becomes more disperse with
time, reaching its equilibrium distribution by
1981, as the impact of the initial price on
the world price generating function is mini-
mized (see Appendix 3). Although both of
these factors are responsible for an increase
in the current cost of the scheme from $912
million in 1978 to $1,445 million in 1982,
only about $332 million of the increase
(about $83 million a year) is attributed to
the increasing variability of cereal import
requirements.
Obviously, the benefits from the opera-
tion of the scheme are not distributed


equally among member countries. In general,
the larger and more variable the domestic
production of a country, the more it bene-
fits from the scheme. Each country's share
of total benefit has been computed (Table 5)
by dividing the expected present value of
payments received by the country by the
expected present value of the total cost of
the scheme. The figures of this table are
rather interesting. India alone receives more
than one-fifth of the total payments made
by the system. Only seven countries-India,
Morocco, Mexico, Turkey, Republic of
Korea, Arab Republic of Egypt, and
Nigeria-account for more than 50 percent
of the total payments made by the system,
and 20 countries account for about 80 per-
cent.
India's very large share of expected bene-
fits suggests that India might also be largely
responsible for the skewed distribution of
the total cost of the scheme (Tables 3 and
4). It is therefore of interest to investigate
the cost of the scheme when India does not
participate. This analysis is presented in
Table 20, Appendix 4. As anticipated, the
expected cost of the scheme is reduced from
the $5.1 billion of Table 3 about 20 per-
cent, a reduction equal to India's share.
More important, the cost distribution has
become considerably less skewed. For ex-
ample, the probability that the present value
of the total cost of the scheme will exceed
$12 billion is only 6.5 percent, compared
to about 10 percent when India is included.
Table 21, Appendix 4 presents the distribu-
tion of current cost of the scheme for 1980,
the middle year of the five-year period, with-
out India. Again, one can observe an ex-
pected cost about 20 percent lower than
that expressed in Table 4 and a much less
skewed distribution of payments.








Table 4-Probability distribution of current cost of the scheme operating as a
compensatory financing mechanism for the middle year (1980) of the
five-year period



Cost a/ Relative Cumulative
($ billion) Frequency Frequency
(%) (%)


8-9


9-10


10-11


greater than 11


a/ The expected current cost equals $1.2 billion. The excess over 110 percent of trend in each country's
current cereal import bill is covered by the scheme.












i- 0 r- I- o N o %D w I r;- 0 0 0 $
40 V v f c Cl cI Ci In Cn .o CC I C.

(3.



.U









0



C
0.







E
0
r.










4-
46
o








C-,

0





4)
E
4u

n
E


4-












0.
(U
E
I




45




U


*o





I-c


?1






iCO
o





























0

r-























c
0
Ul








































0-

CO





QB


m~ Cn k C) 00 It 0 0 0 N- %O It
N- m 0 00 00 R 00 00R N 4 4
Ci O i n ^ CA oO r O O O 1- \0 t





i-~ 0-. r' ~ >o ^ o; 0o Co oqCA 0






~ 0 0
c m M E



ECr .
m 0 C C 6 o .C 0
S. N IP I I- I C ( I D

rN oo 0 0 '- (N CC) l v4 io r% oo C; o ri
- - (N N ( CN (Ne ( C (N (N N (N m Mn en







o 40 (N m oo C r- C) r- (- N 0< *
0oq N 0 ( (N o 0 Cq C OI 0I E 00tq
o r-N Cn e C' CC)n cn en N (N 0 -
(N







4-

0
d. ..


0 0
SCO IC1
o 4 .-4 0
0 0 "c a c N
0 v b 0 0 O

c4 4 0 d 0 6 a; 6 c 4 %. 60 0
>


Qi
0

C m
a, c g 2 S
00 .pCO>
IV E
0 = 0 E E o
0 00 m m a
wmNU -
4 C; %6 : 00 0; 6
en CS en M en C en IC


d.
0




0 to
E >= o *


C C4 W % o- 0
r t I* 't v t v



















-.



0-





OQ








c






o
0


0.













r*
ao











o
a

































CD
0
a













o

0

















































C
a































a
r*
c



0p
a


FU
a



a


U,




a







a


oU
C


lU LA LL A LA Vi ^
4' LJ 0 -. 0 LO

z z n c M >























Sa U L











00 0 -


5 .E I I


























0% 0t 0a 00 0%


n. U, a a i-
U,










g M 0 0 0


n
0
:

















SU
3















rto


41a
7aa













0
In
o S















ro
























0
co
r








U,
aa
00









o
i-
















on

P
a,






a0











ea








Estimated Cost and Distribution of
Benefits of the Scheme Operating as a
Combination of Grain Reserves and
Compensatory Financing


The objective of the food security scheme
can be achieved, alternatively, by a system
using a grain reserve in conjunction with a
compensatory financing mechanism. As dis-
cussed earlier, the system would accumulate
grain when the world price fell below $105
per metric ton. When the world price rose
above $200 per metric ton, the system
would release grain to countries that exper-
ienced a shortfall of more than 5 percent be-
low trend production for that year.
To realistically analyze the system with
grain reserves, the present situation in the
world grain market should be taken into
consideration. Current wheat prices are be-
low the system's acquisition price level and
supplies are ample. Thus, immediately after
being set up the system could acquire the
desired grain reserve. With an initial ex-
pected life of five years for the scheme it
would then be unnecessary to make another
acquisition of grain when part of the initially
acquired grain is drawn or the stock is de-
pleted. If additional grain were acquired, it
is highly unlikely that it would be drawn
upon in the remainder of the five-year pe-
riod. Given the world price distribution, the
probability of price fluctuating from low to
high twice within the five-year period is
almost zero.
Therefore, the cost evaluation of this
scheme assumes a one-time grain acquisition
at the beginning of the five-year period. This


grain is acquired at the current world wheat
price, stored on behalf of the scheme at cer-
tain agreed upop localities at a certain an-
nual storage rent, and released for the use of
member countries in years when price ex-
ceeds $200 per metric ton. *
The cost of this scheme is again estimated
by simulation analysis. The scheme operates
exactly as described in the previous section
when world wheat prices are below $200 per
metric ton. When the world wheat price is
above $200 per metric ton, the system
makes part of the compensation to eligible
countries by releasing grain (to the extent
available) in order to bring the physical avail-
ability of grain in eligible countries up to
95 percent of trend production. The released
grain is valued at the prevailing world wheat
price of that year, and the value of the grain
is subtracted from the monetary amount
for which each country is eligible. After
grain transactions have taken place, compen-
satory financing is made available for the
residual. The total compensation paid to
countries eligible for grain during a high-
price year is monetarily equivalent to that of
the scheme that provides compensatory fi-
nancing only.
In addition to the cost of the compensa-
tory financing mechanism, there are three
other cost components which have to be
taken into account in computing the total
cost of this system. First, the acquisition
cost at the beginning of the period; it is
assumed that grain is bought at $90 per
metric ton, which was approximately the
world wheat price in October 1977. Second,
the carrying charge for the grain stored on
behalf of the scheme is assumed to equal


If the stock is depleted during this five-year period, and if the prospects of renegotiating a new agree-
ment at the end of this period are good, then the authorities of this scheme could replace the depleted grain
during the next low-price year to provide continuity for the new agreement. This complication was thought
unnecessary for the purpose of this analysis.







$10 per metric ton annually.* Finally, the
salvage value of the grain remaining in the
system at the end of its projected five-year
life has to be subtracted from the system's
cost. The salvage value of this residual grain
is assumed to equal the residual quantity
multiplied by the market price prevailing
in the final year.
Again, the present value calculated by
discounting future expenditures and rev-
enues expresses the cost of the scheme for
the five-year period of its operation. Five
grain reserve levels have been analyzed and
associated costs are presented in Table 6.
The existence of a grain reserve in the sys-
tem has some effect on the total cost of the
scheme. For example, compared with the
scheme that provides only financing, a
grain reserve of 8 million metric tons reduces
the expected present value of the cost of the
system by about $59 million, while a grain
reserve of 20 million metric tons increases
the cost of the system by $47 million.
Figure 1 depicts the effect on costs of
various levels of grain held by the system.
One can observe that there are decreasing
marginal savings to the compensatory fi-
nancing mechanism as the size of the grain
reserve increases. The first 4 million metric
tons of grain result in a reduction in the cost
of compensatory financing of $319 million;
the second 4 million metric tons have a
smaller effect, $280 million. The last incre-


ment (from 16 to 20 million metric tons)
decreases the compensatory financing cost
of the system by only $118 million. Mar-
ginal savings exceed marginal costs up to a
grain reserve level of about 8 million metric
tons. Thus, a level of about 8 million metric
tons of grain is the most cost effective. Also,
from Figure 1 it can be observed that a
scheme with a grain reserve of about 16.5
million metric tons is equal in cost to a
scheme operating as a purely compensatory
financing mechanism.
The striking finding from this analysis is
that the total costs of operating the system
at various grain reserve levels are not signifi-
cantly different.t Thus, criteria other than
expected cost should determine the choice
of the appropriate size of the grain reserve,
if any. One such criterion is how well the
presence of grain in the system raises the
physical availability of grain in every mem-
ber country to 95 percent of trend produc-
tion during very high-price years. Table 7
shows how different reserve levels fulfill the
objective of attaining supplies equal to 95
percent of trend during high-price years. For
example, a reserve level of 8 million metric
tons implies that there is a 77.5 percent
probability that every member country
attains the objective during the first year of
operation. By 1982, this probability falls
to about 19.5 percent. A reserve level of 20
million metric tons provides much greater


This figure is based on what the United States Department of Agriculture (USDA) pays United
States grain producers to cover their storage cost for farmer-held stocks. Currently the USDA is required to
pay producers twenty cents/bushel/year which will be raised to twenty-five cents/bushel/year in order to
assure full implementation of the grain reserve program. Twenty-five cents/bushel/year translates into about
$9.25/metric ton/year. The actual storage cost of the grain reserve to be held by the food security system
proposed here will be a function of the location of these stocks, the existence of underutilized capacity,
etc., factors which cannot be assessed at this stage. For this reason a sensitivity analysis is performed in a
later section on the cost of this scheme under alternative storage costs.
t The cost reduction of a system operating with 8 million metric tons of grain versus a system without
a grain reserve expressed as a percentage of the total cost is only 1.16 percent, very insignificant. Similarly,
the cost increase of a system operating with 20 million metric tons of grain versus a system without a grain
reserve expressed as a percentage of total cost is only 0.9 percent, again very insignificant.



















a = a li
CS o V
a O
0. =
CD a -


I '




=03


S0
a o4





0 3e

o o
a





















q- 0
=0-4





, a ..











Sn.
3- a


















os '
o ,,
aq
Q:




















o C



ra
Co
T11




o-?


N> PI 4,





4 -J 0









jJ -1 .P, o
-4 UL .4, 00
",3 "j 0


LA 0 LI
uif w 00 0


-~ 0.-~m 0
.0 4) O
'.0 u' 0


4 s
3

-*

a


O
0
a


0


-I







I







0-
O*
00






(DC




o.
CTD


























ro
(D

-h.
-4.


0





3-

(D








-5














-I
r1

(D






-I


rC
as
P,



S{








Figure 1-Expected present value of the cost of a scheme, operating as a
combination of grain reserves and a compensatory financing mechanism



-.........--------------- V Total cost of the operation
-- of the system
5.0-









U Z
I







STotal cost of
I compensatory
I financing



Sf 4.0 -


Grain reserve
(million metric tons)









U)

I-

I-


0.
















0

I-





E
0
I




E














o6


I.
I-









Is-




W BO
m5



.$*


w


0..


X
W
B3




a
u.1
54)


440
4)i


It 00 0 c U)








%0 %0 ko %0 %0
Ni N C CM
60 60 %0 .0 %0










o on r-









SoaMrM


C4 m 0 D









4i' w U
040% 0 N C C



e % Go o 0 m








) %q 0en 00
m 4o U 0N







- eq)

CC
0 0







*C
o





Sa"







-c -
o S



sa
0'5
0.






o. C



*.- 4- -
o1
0)












0 .
















03 M
ESC



5
0 E
-00


c=E
4- 4M
1E





o S =

a -
0.C


.oC *0







4) ~ .
4- =
sO1W

4- 4- 4







guarantees, satisfying this objective with a
probability of 98.5 percent in 1978 and 62.5
percent in 1982.
This analysis suggests that a reserve level
of 20 million metric tons provides high prob-
abilities of supply guarantees for food def-
icit, developing countries during high-price
years. Should greater guarantees be desired
or should supply guarantees be set at dif-
ferent levels for different countries, the
model could provide the grain reserve level
required to meet these alternative objectives.
As expected, the larger the grain reserve
held, the lower the cost of the compensa-
tory financing component of the scheme.
Thus, a system with 20 million metric tons
of grain results in an expected present value
of $4.0 billion for the compensatory financ-
ing component, $1.1 billion less than the
present value of the scheme operating solely
as a compensatory financing mechanism.
In addition, the distribution of this cost
becomes less skewed, meaning that ex-
tremely high compensatory financing pay-
ments are less likely (see Table 22 in Appen-
dix 4). Thus there is a 95 percent probability
that the total compensatory financing cost
of a scheme with 20 million metric tons of
grain will be less than $12 billion. Adding
the $1.1 billion cost for the grain reserve
still yields a figure well below the compa-
rable cost of about $16 billion for a scheme
operating solely as a compensatory financing
mechanism. Similarly, Table 23, Appendix
4, compares the distributions of the com-
pensatory financing cost of the system for
the middle year of its operation (1980)
under alternative grain reserve levels. The ex-
pected payment during 1980 is $916 million
when 20 million metric tons are held in the
system compared to $1,229 million when


the scheme operates as a compensatory fi-
nancing mechanism only. Furthermore, with
the reserve, availability of less than $3.5 bil-
lion will be adequate to cover all require-
ments 95 percent of the time. Such coverage
would require about $5 billion if the system
holds no grain.
The share of benefits to member coun-
tries has been estimated and presented in
Table 8. Since there are two facilities in this
food security scheme, there are two calcula-
tions of benefit shares: one for the compen-
satory financing facility and the other
for the grain reserve facility of the scheme. *
India, for example, accounts for about 17
percent of all compensatory financing paid
by the system, and receives, on the average,
about one-third of the grain reserves released
during high-price years. Morocco ranks
second after India and receives about 11.55
percent of the grain released. Benefits from
the grain reserve facility of the system are
clearly more concentrated than benefits
from the compensatory financing facility.
Only three countries-India, Morocco and
Mexico-receive more than 50 percent of
the grain, and 12 countries receive more
than 80 percent. Under compensatory fi-
nancing, seven countries receive 50 percent
and 20 countries receive 80 percent of the
benefits.
Table 8 provides also an indication of
the severity of the food problem in different
countries. Recall that the system's grain
reserve facility releases grain during high-
price years. Should the need for grain in
those years be the same for all countries,
then each country's share of benefits from
the grain reserve facility would be the same
as its share of benefits from the compensa-
tory financing facility. This is hardly the


Total benefits received by a member country under this scheme expressed in monetary terms are
exactly the same as benefits received under the scheme operating as a compensatory financing mechanism
only. The composition of benefits is different between the two schemes.











'^ w l' -. 0 '0 00 .1 0% U) .A iA) is3 ...





i0 e %I
< J O 4 U

an w mo u -I *
N a o 0 a< -



















?~~~ . ' S r 3
> 0



































:0 0% to b :o il) cr o )j bo bo
0* w- -j w
o o. 0% b% L-1













o w -D





N NO~ a W a C








ON)




0t





a% 00 w) 0% %D 00 Ch 0 %D hi 4m'0 '0 0












00 w) U) -4 vi 00 '0 %0 LA L-1 0h hi U)


0


























5. ;?

r3.
09


a

















<-





5 S

C aOQ
o


5. 3 "




-a~


a l-
0

R'l

(A9


o


S -4
a.










3=








c
0









M












@"
a-
(U







0





CD















OI
.9
3-

























9u



M,











-N P.- 11 LI) W W W o W Uw W W LI Il


CD 0 n .
> l 0 m


W W
c 3. uS S. g Bl oa a 3 B o
3 0 fflC 1 Z
X c












W 6 W i 0 o -o u -.
4 A V, o LA N. o m U3














NJ C %D 4. bj -J 0. m pk m0














W (
0 < o 0. 0.. 0












%0 LA 00 LA m^ -4 I) w. N3 k 'o 00 00 iw













-. O p. -P, %D 0 -t 0 0
0 t.b.b ~.'0 '0 LJ 0


0
C















10



rt
-a









Ca






lu
X Q0















co
0 3-


































o-i











.- 8
--- 3



-c o
w p A )


a. a 0.
r 3-






a a
o0 0










O g
08
0 0






a


t h
0
3 E
Q c








0 I

3 S
a











CQ CD


-. 0
I



i n
ca c

o a
0. p


I. |
3. S

K a

3 r*

3 a
I 0






3, S
a a

-b
%. i
a


- 0 w -

c F ? ?








a -
C a















g' .LA
ko t9,


0, 0) D 0


am a
0























b0b b *
h i-h.) W l


0 a













C,
ao










on
5'


"3














a





-a





0*







case, however. Some countries are in greater
need of grain during high-price years either
because their own production shortfalls
push world prices up or because the high
variability of their production increases the
probability of a shortfall during high-price
years. India's share of benefits from the
grain reserve is much greater than its share
of benefits from the compensatory financing
facility because a major shortfall in India
will very likely result in high world prices,
thus making India eligible to draw from the
grain facility of the system. On the other
hand, an equal percentage shortfall for a
small country does not cause higher world
prices, so the probability of concurrently
high prices is much smaller.


Cost of the Scheme Under Alternative
Rules of Operation


Alternative insurance levels


The criterion for a country's eligibility
for compensation from the scheme in a
particular year is the comparison of its cur-
rent cereal import bill with its trend import
bill. The preceding analysis assumed an
uninsured import bill of 110 percent for all
countries; that is, when a country's current
import bill exceeds 110 percent of its trend
import bill the country is eligible for the
amount of this excess. To examine the sen-
sitivity of cost to the level of insurance pro-
vided by the scheme, alternative levels of
120 and 130 percent for all countries have
been evaluated. Under these alternatives
each country is eligible for compensation
from the scheme when its cereal import


bill exceeds 120 percent and 130 percent,
respectively, of its trend bill.
Table 9 presents the cost of the scheme
under these two alternative insurance levels.
It also includes, for comparison purposes,
some of the cost figures from Table 6 (110
percent uninsured level). The figures show
that a higher uninsured level reduces the cost
of the scheme substantially. Operating the
scheme as a compensatory financing mech-
anism at a 120 percent uninsured level re-
sults in a cost reduction of $760 million
compared to operating the scheme at a 110
percent level. A 130 percent level further
reduces the cost by $658 million. The cost
reduction from higher uninsured levels for
a scheme operating with a 20 million metric
ton grain reserve in addition to the compen-
satory financing mechanism are practically
the same as those just mentioned. It should
be noted, however, that in this case, almost
all of the reduction in the total cost of the
scheme comes from reductions in its com-
pensatory financing component. In other
words, the grain released from the system
is the same under all three alternatives. This
implies that even in the case of a 110 per-
cent uninsured level, where countries receive
compensation for the excess over 110 per-
cent of trend import bills, countries that
receive grain from the scheme as part of
their compensation are, in fact, facing a
cereal import bill greater than 130 percent
of their trend import bill.
The distribution of compensatory financ-
ing payments is presented in Table 24, Ap-
pendix 4. The higher the uninsured level, the
less skewed this distribution becomes, al-
though the higher uninsured levels do not
significantly reduce the probability of very
large payments. This persistence of large


Although it would be possible under this scheme for countries to select the insurance level desired,
this analysis does not compute the cost of a scheme which assigns different insurance levels to different
countries.











C > 0. 0- -g n -I
- 0 -.- - 0








3 o o Q ;T ;
c a M g

O c .. v












,E < l r*3 -r C S 3 = 3 -r1
. o 0
.a C, CD
S o C3

o -0 0 -

:le a .
| g 5
0
CD n 3 -9




















o L
2 E

EL Tvo
S 0 VO
wCD
MO 0 00 . . 0 CD

00cr oo 5



















~ 0 t -. -5.CO
Co 00 =0
w0 . 0
CO I 4


U, 4, & COs?,, 1 S
- -. eT ^-^* 3^
lu 0. 00

MI I
wCO- 3



B CD0< 0 U0 w
= -;> :.j S : : : -A R B
0L = Lu 00 . . 00 C a





0 5 ~ CO ^ - W O
C 0 su CO0C







I *0*! 3*| :^S 1g
CD 0.









I ~ .. .i *,X-
CD 33 S




1. CI.! 1
-4 o< CD- 3
:z CO CM 0O












COO _B M 0 ^
-k Igl CO


0 0 co 00
^ CD O
co 0 9
-C 00 CD
EO. CiO~ 4
vi to 0.. 0C

*1' 0 ~ -. -.w0 -9 *


CO -D
Al COCO

% 0
0riC -4 W-l)Q
-4 J 0=?0I


CD







payments should be expected since increas-
ing the uninsured level reduces payments
received by each country by fixed amounts.
These reductions are significant at low pay-
ment levels but become less and less signifi-
cant at higher payment levels.


Alternative rules for grain release


In the preceding analysis two factors trig-
ger the release of grain from the grain reserve
facility: (1) the world price of wheat; and
(2) individual country production shortfalls.
The system releases grain only in years when
price rises above $200 per metric ton and
then only to countries that experience a
shortfall of more than 5 percent during
those years. Although these release rules
favor grain release when it is most needed,
they might be considered conservative, since
there is only 20 percent probability that the
world wheat price will rise above $200 per
metric ton. Moreover, the probability is
very low that some countries will concur-
rently experience shortfalls greater than 5
percent and high grain prices.* Thus, it is
interesting to explore alternative release
rules for the system. Two alternative grain
release rules have been analyzed. In the first,
the release price is reduced to $170 per
metric ton. (There is a 32 percent proba-
bility of exceeding this price.) In the second,
the release price is kept at $200 per metric
ton, but grain is released to countries with a


shortfall greater than 3 percent. Such coun-
tries will receive grain, to the extent avail-
able in the system, in amounts that bring the
physical availability of grain to 97 percent
of trend production.
Tables 10 and 11 show the results of this
analysis. Changing the release rules has very
little impact on the expected total cost of
the scheme. There is, however, a significant
reduction in the compensatory financing
component of the cost, particularly with a
reduced grain release price of $170 per met-
ric ton, which reduces this cost by almost
half a billion dollars.t This occurs because
more grain is released at a lower release
price. Even though the net cost of the grain
reserve facility is now higher due to the sub-
stantial reduction in the salvage value of
the left-over grain at the end of the five-
year period, there is still some saving in the
total expected cost of the scheme (about
$150 million). Moreover, changing the re-
lease price trigger from $200 to $170 per
metric ton results in a higher probability
that grain will be available in later years
than changing the production shortfall
trigger from 5 to 3 percent. Thus, if a choice
between these alternative grain release rules
should need to be made, reduction in the
release price would be preferable.
Finally, Table 25, Appendix 4 depicts the
probability distribution of compensatory fi-
nancing payments. As noted above, there are
significant differences in the payments re-
quired under alternative grain release rules.
As the figures of this table show, however,


For example, a country with a 10 percent standard deviation of production variability has a 31 per-
cent probability of experiencing a production shortfall greater than 5 percent. Assuming that this is a small
country and thus, its shortfall does not affect the world price, then the probability that its shortfall will
coincide with high world price is only 6.2 percent (31 percent x 20 percent).

t This does not imply that a further reduction in the release price will further reduce the cost. Given
the probability distributions, there exists an optimum release price (maximizing the revenue of the grain
release). This analysis simply implies that the optimum release price is closer to $170 per metric ton than it
is to $200 per metric ton.






% Table 10-Expected present value of the cost of the scheme operating as a compensatory financing mechanism in conjunc-
tion with a 20 million metric ton grain reserve under alternative grain release rules


Release Rules

Cost Price Greater Than $200/Metric Ton Price Greater Than $170/Metric Ton Price Greater Than $200/Metric Ton
Components a Shortfall Greater Than 5 Percent b/ Shortfall Greater Than 5 Percent Shortfall Greater Than 3 Percent


(million dollars)

Cost of compensatory 4,006 3,548 3,874
financing

Cost of grain reserve
Acquisition cost 1,800 1,800 1,800
Carrying cost 775 713 769
Salvage value 1,426 1,068 1,323

Total cost of grain 1,149 1,445 1,237
reserve


Total cost of the scheme 5,155 4,994 5,111



a/ A discount rate of 8 percent has been assumed. The excess over 110 percent of trend in each country's current cereal import bill is covered.


b/ This column is duplicated here from Table 6 for comparison.







Table 11-Probability of grain reserve availability during high-price years in a scheme operating as a compensatory
financing mechanism in conjunction with a 20 million metric ton grain reserve under alternative grain release
rules a/

Release Rules

Price Greater Than $200/Metric Ton Price Greater Than $170/Metric Ton Price Greater Than $200/Metric Ton
Year Shortfall Greater Than 5 Percent b/ Shortfall Greater Than 5 Percent Shortfall Greater Than 3 Percent


(percent)

1978 98.5 99.1 98.4
1979 95.8 97.0 94.0
1980 93.3 91.5 84.7
1981 70.8 69.5 59.1
1982 62.5 52.2 46.3
Overall probability of
grain reserve availability 82.7 81.9 74.2


Expected Grain Required and Released


(million metric tons)

Total five-year
expected grain requirements 6.26 9.96 7.83
Total five-year
expected grain released 5.18 8.16 5.81


a/ The excess over 110 percent of trend in each country's current cereal import bill is covered.
b/ This column is duplicated from Table 7.








these differences are a result of probability
shifts for low levels of payments. The prob-
abilities of high payments are practically
the same under all alternatives considered
here.


Sensitivity of the Results to
Changes in the Discount Rate and
the Carrying Cost


All the analyses of the preceding sections
have assumed a discount rate of 8 percent,
and a carrying cost for grain of $10 per met-
ric ton per year. These particular values
were to some extent arbitrarily chosen.
Obviously, more precise figures could be
obtained if some important questions of
implementation were resolved. For example,
the discount rate is a function of the inter-
national lending rate, the financial manage-
ment of the scheme, the degree of involve-
ment of international financial institutions
and donor countries, etc. Similarly, carrying
costs depend primarily on the location,
quality, and extent of underutilized capacity
of storage facilities.
It is possible that alternative assumptions
on the values of these parameters could
modify the conclusions reached. For ex-
ample, a lower discount rate reduces the op-
portunity cost of the initial expenditure
for grain acquisition. Thus, the optimum
level of grain reserve might be higher. Sim-
ilarly, a lower carrying cost implies lower
total cost for the system as a whole as well


as an increase in the optimum size of the
grain reserve. The opposite effects should be
expected for higher discount rates or higher
carrying cost. Because, as shown earlier,
the level of the grain reserve has very little
effect on the expected total cost of the sys-
tem, the effect of changing the above two
parameters will be explored only for a sys-
tem operating with 20 million metric tons of
grain in conjunction with a compensatory
financing mechanism.
The cost components and the total cost
of this scheme under alternative discount
rates and carrying costs are presented in
Tables 12 and 13. As expected, the higher
the discount rate and the lower the carrying
cost, the lower the present value of the total
cost of the scheme. It is instructive to
express the effect of these two parameters in
marginal terms. From the figures in Table
13 it follows that, on average, each dollar
increase in carrying cost per metric ton
results in approximately an $82 million
increase in the present value of the total cost
of the system. Similarly, the present value of
the total cost of the scheme is reduced by
approximately $42 million for each percent-
age point increase in the discount rate.
Although altering the values of these
parameters causes some variation in the cost
of the scheme, from a practical point of
view, this refinement of the results is not
crucial at this stage. Other issues, more qual-
itative and political in nature, are far more
important in the implementation of this
scheme.








Table 12-Expected present value of cost components of a scheme operating as a compensatory financing mechanism in
conjunction with a 20 million metric ton grain reserve under alternative discount rates and per unit carrying
costs a/


Cost of the Grain Reserve


Cost of
Compensatory
Financing


Acquisition
Cost


Total Carrying Cost Under Alternative Per Unit
Carrying Costs ($/MT/Year)


Salvage Value


(million dollars)


4,559

4,360

4,177

4,006


1,800

1,800

1,800

1,800


1,074

1,036

1,002


1,792

1,658

1,536

1,426


a/ The excess over 110 percent of trend in each country's current cereal import bill is covered.


Alternative
Discount
Rates


(percent)







oo Table 13-Expected present value of total cost of a scheme operating as a compensatory financing mechanism in con-
junction with a 20 million metric ton grain reserve under alternative discount rates and per unit carrying
costs a/


Alternative Per Unit Carrying Cost ($/MT/Year)
Alternative Discount
Rates 2.5 5.0 7.5 10.0 12.5



(percent) (million dollars)


2 4,782 4,996 5,211 5,426 5,641


4 4,709 4,917 5,123 5,331 5,539


6 4,640 4,841 5,041 5,242 5,442


8 4,574 4,769 4,963 5,155 5,350



a/ See Table 12 for the components of total cost. The excess over 110 percent of trend in each country's current cereal import bill is covered.









4


CONCLUSIONS


Before discussing methods of funding a
food insurance scheme of the type des-
cribed in this paper, the different cost fig-
ures derived in the previous sections must
be put into proper perspective. One impor-
tant distinction needs to be made at the
start. Automobile, property, or health in-
surance plans involve a very large number of
participants, and the risk factor is con-
tinuous over time.* Thus, the probability
that the actual cost will equal expected cost
within a given period is extremely high. An
international food insurance scheme, on the
other hand, involves a limited number of
participants and a non-continuous risk fac-
tor.t As a result, the actual cost of operating
the scheme for a finite period might differ
significantly from its expected cost.
Whatever the form of its financial re-
sources (cash in hand, scheduled premium
payments, borrowing capacity, or standby
pledges), a food insurance scheme of the
type described in this paper will not be
established unless it can function at a cost
that seems reasonable to its funders. If a
high degree of certainty is desired that funds
will be available to meet the objectives of a


food security scheme, the amount of these
funds will have to be very large, even though
there is little likelihood that the scheme will
actually cost that much. If a lower degree
of certainty is acceptable, the potential
cost of the insurance scheme will be con-
siderably lower, and the reduced initial
capital requirements are more likely to seem
reasonable to probable funders of the
scheme. In the event that the scheme's
funds do not fully cover requirements in
some years, target consumption will have
to be adjusted proportionately for all par-
ticipating countries. Since funders will
certainly set an upper limit to the financial
capacity of any food insurance scheme, the
question then becomes, what is the best use
of a given level of funds?
To indicate the range of options which
funders might want to consider, Table 14
and Figure 2 present the probabilities as-
sociated with different cost levels above the
expected cost for six scenarios: three insur-
ance levels without reserves, and three insur-
ance levels with a grain reserve of 20 million
tons. These clearly show that the higher the
probability that the objectives of the scheme


This is so because the probabilities of an automobile accident, a house catching fire, or an individual
having a heart attack are not correlated among themselves and among different insurance policy holders.
The fact that an individual has an expensive medical bill during a given month does not mean that his
next-door neighbor will also have a high bill during the same month. Furthermore, if several individuals
insured in the same insurance company happen to get sick at the same time, this would not affect the price
of medical services, and therefore the per unit cost to the insurance company would not rise.
SCrop production outcomes have a time cycle of one year, and are to some degree, correlated among
neighboring countries. In addition, favorable or unfavorable weather worldwide during a particular year has
an impact on cereal prices, and thus, the per unit cost to a food insurance scheme can vary substantially
from one year to the next.








Table 14-Funds required to attain a given probability that the scheme will achieve its objectives under alternative
methods of operation a/
Probability of the Scheme
Alternative Scenarios Achieving its Objectives
on the Operation of Expected (percent)
the Scheme Cost
70 75 80 85 90 95


(billion dollars)
Scheme operating as a compen-
satory financing mechanism
only, with uninsured import bill
of

I. 110% of trend 5.11 5.1 6.1 7.3 9.4 12.1 16.7
II. 120% of trend 4.35 4.1 4.9 6.0 7.9 10.7 14.5
III. 130% of trend 3.69 3.3 3.9 4.8 6.4 8.7 12.5

Scheme operating as a compen-
satory financing mechanism in
conjunction with 20 MT of
grain reserve, with uninsured im-
port bill of b/

IV. 110% of trend 5.15 5.1 5.6 6.1 7.0 9.5 13.2
V. 120% of trend 4.40 4.1 4.4 4.8 5.7 7.7 11.2
VI. 130% of trend 3.74 3.3 3.6 3.9 4.7 6.1 8.9

a/ The scheme attains its objectives when it has adequate funds to compensate member countries under its rules of operation.
b/ The cost figures under scenarios IV,V and VI include also the net cost of the grain reserve (about $1.1 billion). Grain is released in years when world price
exceeds $200 per metric ton, to countries that experience greater than 5 percent production shortfalls in those years and in amounts that will bring the
physical availability of supplies (to the extent available in the system) to 95 percent of trend production in all countries.







17 Figure 2-Probability of the scheme achieving its objectives

16

15- 1

14-

13 Io

12

11- ... .e

10

1 0 financ*ng ocason
9.9
-e
8-

7-

6-- -* -----

5 110% .........................


........ I Uninsured import bill equals 120% of trend
3 130% ......... I I Uninsured import bill equals 130% of trend
I i I Scheme operating as a compensatory
2 I financing mechanism only
...... Scheme operating as a compensatory
1 -- financing mechanism in conjunction
I with 20 million tons of grain reserve


70 75 80 85 90 95
SProbability (percent)
a/ For definition of alternative scenarios for operating the scheme see Table 14.








will be met, the higher the marginal cost of
food insurance. For example, under sce-
nario II (scheme operating solely as a com-
pensatory financing mechanism with an un-
insured import bill of 120 percent of trend)
an additional financial capacity of $0.8 bil-
lion increases the probability level from 70
to 75 percent, whereas increasing this prob-
ability from 90 to 95 percent requires $3.8
billion.
There is, however, a trade-off between
the level of insurance provided and the prob-
ability that the objectives of the scheme will
be met. For a given financial capacity, a
high degree of insurance implies a high risk
of depleting funds during the last years of
the five-year period, thus favoring countries
that happen to draw from the scheme early
in the period. Therefore, lower insurance
coverage (uninsured import bill of 130 per-
cent of trend) with a high probability (90
to 95 percent) of defending the scheme's
stated objectives should be preferable to
high insurance coverage (uninsured import
bill of 110 percent of trend) with a lower
probability (70 to 75 percent) of achieving
the objectives.
Another consideration is the presence or
absence of a grain reserve in the system. Al-
though a grain reserve has virtually no effect
on the probability that the scheme will meet
its objectives if funding is provided at the
expected cost level, additional funding above
the expected cost makes a scheme with a
grain reserve preferable since it provides
a higher probability of achieving its objec-
tives. This effect is greater the larger the size
of the reserve. In addition, as shown earlier,
a larger reserve provides a higher probability
of maintaining supply availability in eligible
countries at 95 percent of trend production
during high-price years. For these reasons,
if a grain reserve were established in conjunc-
tion with a compensatory financing mecha-
nism, a grain reserve level of 20 million met-


ric tons for developing countries is sug-
gested.
If it is reasonable to assume that most
countries can manage import bill overruns
up to 130 percent of the trend import bill
in some years, then the expected present
value of the cost of the scheme amounts to
about $3.7 billion, of which about $1.1
billion is the cost of a 20 million metric ton
grain reserve. This level of funding implies
a probability of at least 75 percent that the
scheme will successfully defend its objec-
tives. A probability level of 90 percent might
require up to $2.4 billion more during the
next five-year period.
A crucial component in the success of
this scheme is the source of funds needed
for its operation. Three alternative ap-
proaches to financing the scheme are dis-
cussed: (1) self-financing by participating
countries, (2) financing primarily by devel-
oped countries, with some contributions
from, developing countries, either collec-
tively or through a scheme that would in-
corporate existing food aid programs, and
(3) financing through a compensatory fi-
nancing facility of the International Mone-
tary Fund (IMF).
Under the first alternative, member coun-
tries pay annual premiums in proportion to
their expected benefits during the five-year
period. Countries could pay equal premiums
every year or, depending on their year-to-
year overall foreign exchange position, they
could pay variable annual premiums, larger
when foreign exchange position is good,
smaller when it is poor. Based on expected
total withdrawals (Table 15), annual prem-
iums by country can be computed. For ex-
ample, if India had to pay an equal annual
premium for the next five years, that prem-
ium would range from about $206 million to
about $245 million, depending on the level
of insurance provided by the scheme. The
funding provided by this approach would









Table 15-Estimated percentage share of benefits and present value of expected total withdrawals by country, 1978-82 a/


Uninsured Import Bill Uninsured Import Bill

130% of Trend 110% of Trend 130% of Trend 110% of Trend

y Share of Expected Share of Expected Share of Expected Share of Expected
y Benefits Total Benefits Total Country Benefits Total Benefits Total
Withdrawal Withdrawal Withdrawal Withdrawal


(%) ($ millions) (%) ($ millions) (%) ($ millions) (%) ($ millions)


India 24.07 900.2 20.80 1071.2 Venezuela 1.49 55.7 1.99 102.5
Morocco 9.72 363.5 7.76 399.6 Cuba 1.46 54.6 1.94 99.9
Mexico 5.74 214.7 5.02 258.5 Iraq 1.62 60.6 1.84 94.8
Turkey 4.87 182.1 4.24 218.4 Sudan 2.34 87.5 1.79 92.2
Korea, Rep. of 3.27 122.3 4.23 217.8 Peru 1.31 49.0 1.75 90.1

Egypt, Arab Rep. of 3.07 114.8 4.08 210.1 Zambia 2.31 86.4 1.75 90.1
Nigeria 3.23 120.8 4.01 206.5 Tunisia 1.85 69.2 1.71 88.1
Indonesia 3.63 135.8 3.63 186.9 Sri Lanka 1.31 49.0 1.63 83.9
Bangladesh 2.67 99.9 3.17 163.3 Chile 1.11 41.5 1.44 74.3
Syrian Arab Rep. 3.44 128.7 3.13 161.2 Philippines 1.00 37.4 1.19 61.3
Algeria 2.44 91.3 2.92 150.4 Burma .95 35.5 .98 50.5
Iran 2.29 85.6 2.77 142.7 Lebanon .63 23.6 .84 43.3
Brazil 2.56 95.7 2.41 124.1 Tanzania .72 26.9 .81 41.7








Table 15-Continued


Uninsured Import Bill Uninsured Import Bill

130% of Trend 110% of Trend 130% of Trend 110% of Trend

Share of Expected Share of Expected Country Share of Expected Share of Expected
ounry Benefits Total Benefits Total Benefits Total Benefits Total
Withdrawal Withdrawal Withdrawal Withdrawal


(%) ($ millions) (%) ($ millions) (%) ($ millions) (%) ($ millions)


Malaysia .58 21.7 .80 41.2 Dominican Rep. .27 10.1 .36 18.5
Malawi 1.03 38.5 .80 41.2 Cyprus .28 10.5 .34 17.5

Libya .53 19.8 .70 36.0 Yemen Arab Rep. .24 9.0 .31 16.0
Afghanistan .63 23.6 .67 34.5 Mali .26 9.7 .30 15.4
Colombia .54 20.2 .66 34.0 Malagasy .23 8.6 .30 15.4

Upper Volta .72 26.9 .64 33.0 Niger .23 8.6 .28 14.4
Senegal .51 19.1 .61 31.4 Ivory Coast .22 8.2 .27 13.9

Jordan .47 17.6 .55 28.3 Trinidad .19 7.1 .25 12.9
Ecuador .29 10.8 .38 19.6 El Salvador .30 11.2 .24 12.4
Bolivia .28 10.5 .37 19.1 Angola .23 8.6 .23 11.8
Zaire .29 10.8 .37 19.1 Ghana .27 10.1 .23 11.8
Cameroon .35 13.1 .36 18.5 Uganda .28 10.5 .22 11.3
Jamaica .27 10.1 .36 18.5 Costa Rica .15 5.6 .20 10.3









Table 15-Continued


Uninsured Import Bill Uninsured Import Bill

130% of Trend 110% of Trend 130% of Trend 110% of Trend

Share of Expected Share of Expected Share of Expected Share of Expected
Country Benefits Total Benefits Total Country Benefits Total Benefits Total
Withdrawal Withdrawal Withdrawal Withdrawal


(%) ($ millions) (%) ($ millions) (%) ($ millions) (%) ($ millions)


Nicaragua .18 6.7 .17 8.8 Rwanda .09 3.4 .09 4.6

Honduras .13 4.9 .16 8.2 Liberia .07 2.6 .08 4.1

Paraguay .21 7.9 .16 8.2 Sierra Leone .06 2.2 .08 4.1

Guatemala .12 4.5 .16 8.2 Benin .05 1.9 .06 3.1

Panama .11 4.1 .14 7.2 Gambia .02 0.7 .02 1.0

Haiti .09 3.4 .11 5.7 Chad .03 1.1 .02 1.0

Guinea .11 4.1 .11 5.7


Note: Countries are listed in descending order of benefits drawn at the 110 percent uninsured levels.

a/ Expected cost with: 110 percent uninsured import bill $5,150 million
130 percent uninsured import bill $3,740 million







cover only the expected cost of the scheme,
so there would be about 25 percent proba-
bility that the scheme's funds would run
out. If funds ran out, then the scheme would
either have to borrow the cost overruns up
to the limit that had been set, or it would
have to adjust compensation to eligible
countries so that all shared the burden
proportionally.
The longer the life of the scheme, the
closer the cumulative disbursements at any
point of time will be to the cumulative pre-
miums paid in; over a sufficiently long time
the scheme itself would be able to repay past
debts or to meet by itself higher than ex-
pected current costs. However, if the
scheme's life is abruptly terminated after
the initial five-year period, the scheme will
very likely be left with either a surplus or a
deficit both in overall receipts versus dis-
bursements and in the distribution of bene-
fits to individual countries. If the scheme has
a balance other than zero on termination,
then surpluses should be refunded to and
deficits paid by member countries according
to their net contributions to the scheme.
In practice, most low income countries
are not able to participate in a self-financed
scheme without help from developed coun-
tries. A food security scheme designed to
maintain consumption in developing coun-
tries when world prices are high would per-
mit these countries to rely on commercial
imports for an increasing percentage of their
growing food requirements. Therefore it
would be to the advantage of grain exporters
in particular to contribute to such a scheme.
Under the second funding alternative,
developed countries could provide assistance
in a number of ways. If they chose to do so,
they could finance the scheme along with
the richer of the developing countries with-
out reference to the distribution of benefits
among recipient countries. In this case,
donor countries could simply make a collec-


tive commitment to finance the scheme at
an agreed upon level. Contributions could
be made either in the form of money or in
the form of grain to set up a reserve on be-
half of the scheme. The cost of setting up
the grain reserve could perhaps be the re-
sponsibility of the major grain exporters.
Another possibility would be for donor
countries to make commitments to pay a
specific percentage of the premiums required
from low income countries. A collective
effort by developed countries to subsidize
the scheme in this way would be desirable
from the point of view of recipient countries
because it would stipulate a regular flow of
financial assistance on terms known in ad-
vance. But donor countries might not, for
various economic and political reasons want
to commit themselves to this approach.
Many of them might prefer to integrate
existing food aid commitments into the
scheme rather than or in addition to provid-
ing direct financial subsidies to eligible par-
ticipants. At present, food aid flows are not
always responsive to the requirements of
recipient countries for which they are in-
tended. Incorporating variable food aid com-
mitments in a food security scheme could
remedy this situation.
For countries willing to contribute to
the financing of a food insurance scheme
through their food aid programs, there are
several options.Food aid can be used instead
of funds to subsidize the annual premium
payments of low income countries. In cer-
tain years, the donor nation might make a
food aid contribution to a beneficiary coun-
try in lieu of a dollar payment to the food
insurance scheme. Assume, for example,
that a country is obligated to pay an annual
premium of $10 million to the food security
scheme in order to guarantee a certain de-
gree of domestic food security, and that a
donor country has pledged a contribution of
$5 million (half of the annual premium) on








behalf of this country. Suppose that during
a normal year the country needs to import
$20 million worth of food to meet its con-
sumption targets. Thus, during a normal year
this country will have to spend a total of
$25 million for food imports and its par-
ticipation in the food security scheme. If
during a particular year an excessive grain
supply in the donor country makes it more
desirable for the donor to make its pledged
contribution in the form of grain, it can do
so by contributing $5 million worth of grain
directly to the beneficiary country. The
beneficiary country will then have to import
only $15 million, and this $5 million savings
on its normal import bill plus the $5 million
that it has allocated for its participation in
the insurance scheme will cover its required
premium for that year.
Food aid can also contribute directly to
the stabilization of domestic food supply in
the recipient country, so that the country
will seek a lower degree of protection from
the food insurance scheme and thus reduce
its annual premium payments. Because
domestic supply fluctuations in the recip-
ient country will not in general coincide
with the capacity of the donor country to
provide food aid, such aid could be used
when available to build up a reserve in the
recipient country. If stocks created from
food aid contributions provided protection
against a certain level of domestic produc-
tion shortfall, then this country should seek
coverage from the food insurance scheme in
excess of the insurance it provided itself.
The premiums needed in this case could be
considerably lower than those for full
coverage.
Whatever the financing arrangement, not
all of the funds would have to be provided
in the form of paid-up capital. Donor coun-
try commitments could take the form of
securing the scheme's borrowing capacity.
A practical way to initiate negotiations


would be for a group of interested countries
to present a food insurance proposal to the
International Wheat Council and/or to the
Grain Subgroup of the Multilateral Trade
Negotiations, presenting the proposed
scheme as a means of making special pro-
vision for developing countries within the
more general agreement being negotiated.
A third alternative for funding a food
security scheme is through a compensatory
financing facility within the IMF. This could
be done by creating a new facility specifi-
cally to cover overruns in cereal import bills.
Or the cereal import bill could be taken
into account along with fluctuations in
commodity export earnings in determining
whether a country is eligible for compensa-
tory financing from the existing facility.
Funding through the IMF's compensatory
financing facility is a reasonable and just
solution because it takes into consideration
the foreign exchange position of a country
in a given year. More important, the mech-
anisms that could put such a scheme into
operation have already been established and
tested, thus precluding the need to establish
a vast new bureaucracy.
In addition to funding arrangements
already discussed, several other issues will
require attention. The foremost problem to
be resolved is the identification of reliable
data sources to be used in determining
which countries are eligible for compensa-
tion and the exact amounts they are eligible
for each year. For the initial five-year
period, production trends and consumption
projections could be estimated with a high
degree of accuracy. With some refinement
the figures developed in this analysis could
be used. The only remaining problem as-
sociated with data will then be the accuracy
of future annual production figures. An
international authority would have to deter-
mine whether a reported production short-
fall is the result of genuine adverse weather







conditions, or whether it results from a
country cutting back resources in cereal pro-
duction or deliberately under-reporting its
production performance. Rather than creat-
ing a new bureaucracy, an authority already
competent to assess cereal production and
consumption data submitted by participat-
ing nations should take on administrative
responsibility for the scheme. Over time, a
country's consistent falling behind in pro-
duction performance would be reflected in
lower production trend and consequently,
higher import trend. Since each country is
itself responsible for covering a certain
percentage of its trend import bill, it would
be discouraged from under-reporting, know-
ing that under-reporting would effect its
eligibility for compensation in future years.
Although this effect might not be very im-
portant within the time period of one agree-
ment, it would have a bearing on the nego-
tiation of future agreements.
Another issue deals with the administra-
tion of grain, if grain stocks are held in the
system. Because current world prices are
relatively low, it was suggested that the sys-
tem could start with the desired grain re-
serve in stock. A delay in grain acquisition
is almost certain to mean higher acquisition
costs. A $200 per metric ton price has been
assumed to trigger the release of grain, and
only to countries that experience more than
5 percent production shortfalls during high-
price years. An alternative trigger price of
$170 per metric ton and an alternative pro-
duction shortfall level of 3 percent were
also explored. Although these alternatives
did not have a significant effect on the ex-
pected cost of the scheme, they resulted in
a slightly less skewed distribution of com-
pensatory payments.
The location of a reserve is another issue
that usually emerges in discussions of inter-
nationally held grain reserves. Ideally, the
grain reserve should be located as close as


possible to countries that are more likely to
call upon it, assuming that high quality
storage facilities exist. Alternatively, on the
assumption that grain exporting countries
would share in the cost of the scheme either
by collectively providing part of the grain
reserve or by integrating into the scheme
their current food aid programs as discussed
earlier, grain exporters could hold part of
the grain reserve on behalf of the scheme in
existing storage facilities. Major grain ex-
porters might be willing to waive carrying
costs by including them in the cost of
achieving their domestic agricultural policy
objectives. Thus, though the location of
grain reserves is a politically sensitive issue
for reasons of self-reliance, it appears that
from a practical point of view negotiations
should focus on a few storage sites close to
large contributors and to large expected
users of the grain. Countries could minimize
the disadvantages of distant location of the
international reserve by holding small re-
serves of their own to supplement those of
the scheme.
Finally, food deficit, developing countries
will have to develop internal policy instru-
ments to insure that the steady flow of grain
imports obtained by a country's participa-
tion in the food insurance scheme reaches
the poorest part of its population. In most
developing countries agriculture generates a
large portion of the income of the poor,
so years of poor harvests imply lower buying
capacity for a large portion of the popula-
tion. If the reduction in consumption caused
by this decline in effective demand is greater
than that tolerated by the food security
scheme, countries will need to operate in-
ternal food subsidy programs during these
years. In addition, since the added supply of
grain from an external source will keep
prices lower than they would have been
otherwise, farm incomes will be rather low
(a combination of poor harvests and prices







depressed by added supplies). Thus, aside
from providing assistance to consumers,
domestic policy would have to provide in-
come compensation to grain producers to
negate possible adverse effects on their
earnings.
In summary, this paper has presented a
food insurance scheme to deal with the food
insecurity problem of food deficit, develop-
ing countries; a scheme whose cost proves
to be within the reach of the international


community. The world food problem is
no temporary phenomenon, and the earlier
a comprehensive effort, such as the one sug-
gested here, is made, the better. The environ-
ment for initiating this effort may never be
more favorable in terms of currently avail-
able grain supplies. Should this effort prove
fruitful, more countries will undoubtedly
seek to join, and the chances of alleviating
mankind's ultimate problem will be greatly
enhanced.








Appendix 1


OVERVIEW OF PAST APPROACHES TO
THE INSTABILITY OF GRAIN MARKETS


Welfare Effects of Commodity
Price Stabilization


Much of the early work on the desirabil-
ity of commodity price stabilization brought
out the conflict of interest between pro-
ducers and consumers. Welfare analysis was
used to determine benefits and costs from
price stabilization and their distribution be-
tween the two groups within a closed
economy framework. Massel (1969), build-
ing upon the original work by Waugh (1944)
and Oi (1961), assumed demand and supply
linear and subject to additive disturbances,
and showed that the elasticities of demand
and supply and the source of the instability
determine which group gains from stabilized
prices. Despite the divergence of economic
interests between producers and consumers
with regard to stabilized prices, the gainers
can compensate the losers so that both
groups would be better off with than with-
out price stability. The same conclusions
were reached independently by Samuelson
(1972).
Since 1972 there have been a number of
new developments in the economic analysis
of price stabilization. First, less stringent
assumptions have been introduced to make
analysis more realistic for today's world;
and second, the international dimensions of
the problem have been brought into perspec-
tive. In this context, the conflict of eco-
nomic interests from price stabilization is
not only a domestic issue but also an inter-
national one.


Heuth and Schmitz (1972) extended the
Waugh-Oi-Massel closed economy model to
an international trade mcdel. Their conclu-
sions are basically similar to those of Massel.
That is, whether or not an individual coun-
try gains from stabilization basically depends
upon the source of the instability. A country
gains from price stabilization if the source
of instability is domestic (either demand or
supply) but loses if the source of instability
is foreign. Domestic producers gain from
price stabilization if they are the source of
the instability, and they lose if domestic
demand and/or foreign demand are the
sources of instability. On balance, both
countries gain from price stabilization, pro-
vided that international and domestic com-
pensations are made from gainers to losers.
The question of who gains and who loses
from price stabilization can only be an-
swered by empirical investigation.
Konandreas and Schmitz (1978) have em-
pirically tested the Heuth and Schmitz
model for the United States for the period
of 1954 to 1972. On the basis of their econ-
ometric analysis, United States producers
and the U.S. as a whole would gain from
feedgrain price stabilization, and perhaps
also from wheat price stabilization, although
this is less clear. These results are consistent
with U.S. policy before 1972, when the
main source of variability in grain markets
was U.S. grain production, and the U.S.
government held large stocks to maintain a
stable price floor in the producer's interest.
The Konandreas and Schmitz study does
not fully capture the effect of increased







variability in foreign demand since 1972
because such high foreign demand fluctua-
tions were not experienced during the period
of their analysis.
These analyses assume that demand and
supply curves are linear and that random
disturbances are additive. Gray (1974)
noted the significance of the shape of the
demand curve for the welfare implications
of price stabilization when he speculated
that the demand curve for grains may grow
steeper at higher price levels. Just et al.
(1977) have analyzed the non-linear case
with an international trade model. As in the
linear case, exporters and importers taken
together gain from price stabilization.
Broadly speaking, importers tend to gain
while exporters tend to lose, and benefits
seem to shift from producers to consumers
in both countries. The welfare implications
of price stabilization have been further in-
vestigated by incorporating expectations
into the supply function. Turnovsky (1974)
experimented with adaptive and rational
expectation schemes and found a theoret-
ically ambiguous distribution of benefits in
this case.
Reutlinger (1976) employed an aggregate
simulation model to calculate the effective-
ness of an international wheat stock on sta-
bilizing prices. The model is run at dif-
ferent stock levels, using quantitative trig-
gers and estimates the probability of price
stabilization within specified ranges. For
example, a reserve stock target of 30 million
tons, acquired when production exceeds 355
million tons and released when it falls below
335 million tons, stabilizes price within a
range of $115 to $188 with a 78 percent
degree of assurance and a mean price of
$138 per ton. But stock levels that are
optimal in terms of direct costs and welfare
benefits are likely to be too low to afford a
satisfactory degree of price stabilization. The
cost of storage to the economy rises rapidly


above stock levels of 5 to 10 million tons.
Further, a high degree of stabilization seri-
ously imbalances the distribution of gains
and losses in favor of consumers and against
producers. In the above example, consumers
would have to be taxed fifty cents per ton
to cover the net loss of the storage opera-
tions, and producers could expect to lose
thirty-five cents per ton.
Reutlinger notes that both producers
and consumers might be willing to pay these
costs in order to obtain stabilization of their
effective incomes. Also, if more stable prices
led to better resource allocation and reduced
inflation, or if they reduced the cost of
quantitative food aid commitments, the
costs to the economy might be worth bear-
ing. He also cautions that the cost/benefit
results can be strongly affected by the se-
quence of good and bad harvests after the
initial investment, and by the elasticities
assumed for different segments of the
demand curve, neither of which can yet
be specified reliably. Finally, the effective-
ness of buffer stock programs operated with
insurance-oriented storage rules was also
investigated; it was shown that with such a
scheme adequate protection against extreme
world grain production shortfalls could be
achieved.
Using a kinked demand curve, and adjust-
ing for country and world market price dif-
ferentials resulting from trade barriers, Sarris
et al. (1977) also experiment with the price
stabilization effects and costs of different
sizes of wheat and feedgrain stocks. On the
basis of their results, they argue that an
international wheat reserve of up to 15 mil-
lion tons could stabilize prices within a
wide band and imply greater welfare benefits
than associated costs. This reserve level
would reduce price fluctuations by about
one-third, keeping them within a band of
roughly $110 to $180 per ton at current
prices.







For modeling purposes they assume that
an international buffer stock agency would
manage the reserve through buying and sell-
ing operations whose precise rules would not
be made public, because of the need to pre-
vent offsetting market interventions by
other buyers and sellers. Over time, wheat
importers would be slight gainers while
exporters would be slight losers, but the
amounts involved are minimal. A feedgrain
reserve would reverse the country distribu-
tion of net gains and losses, with importers
losing and exporters gaining. Thus, it might
be advantageous to create a joint reserve,
even though feedgrain prices would be much
less responsive to the operation of a buffer
stock than wheat prices.


Simulation Analysis Experiments
on International Grain Reserves

Another body of literature employs simu-
lation analysis with mechanical models that
approach the reserves problem from the
standpoint of supply availability. These
models do not deal with price stabilization
from the point of view of its welfare implica-
tions as such; rather, they treat reserves as
a means of ensuring that certain consumer
groups will not be cut off from supplies
when harvests are low. This was a logical
reaction to the 1972-74 food crisis.
One of the first analytical responses to
that crisis was the study by the Economic
Research Service of the United States
Department of Agriculture, by Bailey et al.
(1974). This study concerns itself with the
level of reserve stocks required to provide
various degrees of supply security, and es-
timates the storage cost of each alternative


considered. It points out that when the
ratios of United States wheat and corn carry-
out to total use fall below 20 percent, there
is a significant increase in prices, but it does
not consider price stabilization as an objec-
tive, nor does it discuss the specific stabiliza-
tion effects of reserve stocks as such.
The study tries a number of different cal-
culations to estimate the effect of a United
States reserve program on supply availability
during the period 1950-69, had such a pro-
gram been in effect. The size of the reserve
is fixed as a percentage of U.S. trend pro-
duction, and rises gradually over time. Cal-
culations are made with stock level targets
ranging from 8 to 12 percent of U.S. trend
production. At 8 percent, for example, the
target size of the reserve would have risen
from 9.9 to 16 million metric tons, whereas
at twelve percent it would have risen from
13.9 to 23 million metric tons. Reserve
stocks would be accumulated each year up
to the established percentage limit, to the
extent permitted by the availability of other
stocks plus the level of current production.
Out of a total world shortfall of 179.4 mil-
lion metric tons for this period, 79 percent
would have been covered by the smaller
reserve (8 percent of trend production) at
a cost of $4.52 per metric ton, and 95 per-
cent by the larger (12 percent) reserve at a
cost of $5.60 per metric ton.
This study also identifies some important
policy issues and raises practical questions
about how to implement a grain reserve pro-
gram. Policy questions that the authors raise
include: what kind of demand should be
supported (domestic, commercial export,
foreign aid)? How much uncertainty and ad-
justment should be allowed? What price level
would be protected?


* World shortfall is computed as the summation of actual shortfalls in major exporting countries
(United States, Canada, Argentina, and Australia) and shortfalls in excess of 5 percent of trend production
in all other countries. For the latter, it is assumed that policies will be adjusted so that production varia-
tions in excess of 5 percent of trend will be absorbed by adjustments in consumption.






A follow-up of this approach is contained
in a later USDA publication, World Food
Situations and Prospects, which presents
data for the period 1960-73 showing that if
wheat, rice, and coarse grains are taken sep-
arately, a combined reserve of 80 million
metric tons would have been needed to
cover 95 percent of world production short-
falls. If all grains are assumed to be perfectly
substitutable for each other, the reserve
requirement would have been 56 million
metric tons. Substantially smaller amounts
would have been needed if the reserve were
used only to cover shortfalls in trend im-
ports, and it is estimated that no more than
10 million metric tons would be needed for
a famine relief reserve. Like the earlier
USDA study, a number of practical ques-
tions are addressed.
This study argues that economically op-
timum stock levels likely to be held by the
private sector will not be large enough to
provide adequate supply security, and that
while several different cost-sharing formulas
might be applied, a reserve stocks system
would operate most efficiently if the stocks
were located in only a few major exporting
and importing countries. The idea that im-
provement in the market information system
might help to reduce speculative activity is
stressed, but the possible stabilizing effects
of freer trade policies on world grain mar-
kets is not mentioned.
Eaton et al. (1976) estimate the stock
levels needed for various degrees of food
security in a world where trade, transport,
or distribution barriers do not exist and
grains are perfectly substitutable. Data for
the period 1960-74 are used. The authors
note that because of these ideal assumptions,
the reserve requirements computed are un-
realistically low. Reserves of 58 million
metric tons would be sufficient to cover 98
percent of production shortfalls in a free
trade world between now and the end of the


century, and only 5 million metric tons
would cover 95 percent of probable future
shortfalls.
Johnson (1976) believes that in an ideal
free trade world there would be little need
for grain reserves in most years. However,
since governments do generally try to shield
their domestic markets from the full price
effect of production fluctuations he feels
that some special arrangements are needed
to help developing countries stabilize grain
consumption within a reasonable range.
His views on the kind of arrangement needed
appear in a collaborative work by Danin
et al. (1975). They propose a reserve scheme
which would provide developing countries
with a specified level of insurance against
their own production shortfalls. Their study
shows that optimal carryovers for individual
developing countries would be greatly re-
duced if an international insurance scheme
were in effect. On the basis of their findings,
they propose that the United States and
other developed countries guarantee cover-
age of production shortfalls in developing
countries greater than a given percentage of
their production trends.
Following Gustafson (1958), they define
optimal carryover as that amount for which
the expected marginal cost of holding an
additional ton of grain equals the expected
marginal gain which would result from pre-
venting a price increase. A linear demand
curve with a price elasticity of -0.1 was as-
sumed for all grains, and actual production
for the years 1948-73 was used to calculate
trend production. Nonvariability of trade
outside the country or region was assumed
in order to show the implications of letting
stocks bear the entire burden of adjustment
of their own production variability. No dis-
tinction is made between foodgrains and
coarse grains, on the assumption that they
are highly substitutable for each other.
The level of country disaggregation af-







fects the results. For example, without an
insurance scheme in effect, the expected
annual amount of carryover in four develop-
ing country regions, treated separately,
would equal 13.7 million metric tons in
1975, with a 95 percent probability that the
maximum would not exceed 34.5 million
metric tons. However, if developing coun-
tries are treated as a single, free-trade region,
the expected yearly carryover is only 5 mil-
lion metric tons, and 95 percent of the time
the maximum does not exceed 15 million.
In collaboration with Cochrane, Danin
(1976) has extended the work he began with
Johnson into a full-scale grain reserve model.
He projects the reserve stock levels and de-
gree of price stabilization for the period
1975-85, estimating that by 1985 the degree
of price variation around a target price of
$100 per ton could be reduced from 27 per-
cent to as low as 10 percent with the reserve
growing gradually to 75 million metric tons
by that time. A reserve level of 39 million
tons in 1985 would reduce price variation to
around 19 percent. A beginning reserve
stock would produce an immediate reduc-
tion in price variability, if desired, but for
the long-run a beginning grain reserve is not
necessary to achieve this objective. Danin
also stresses the importance of the price
elasticity assumed, i.e., the more elastic the
demand, the more a given reserve level re-
duces price variability.
The significance of this work is that it
shows the degree to which pooling reserves
among countries reduces the aggregate op-
timum carryover level. Further, when de-
veloping countries reserve supply guarantees
covering a percentage of their production
shortfalls, these optimum carryovers are sub-
stantially reduced. Finally, this study shows
that free trade among developing countries
would be the most effective policy for re-
ducing maximum carryover levels.
An insurance scheme which would have


covered shortfalls in excess of 6 percent of
trend production in four developing country
regions from 1954 through 1973 would have
made zero payments 65 percent of the time,
and would have paid less than 2.5 million
metric tons 80 percent of the time. The
maximum payment required would have
been 9.2 million tons, an amount that would
either have to be carried in an international
reserve or made available to developing
countries by purchases in commercial mar-
kets if the scheme were to produce the in-
tended results. Had the scheme been in ef-
fect, 95 percent of the time the maximum
optimal carryover for the four regions would
have been less than 23 million tons, and for
developing countries treated as a single free-
trading region less than 14 million tons.
Walker and Sharpies (1976) investigate
the consequences of a reserve system op-
erated unilaterally by the United States.
Two systems are evaluated, one operated
with price triggers and the other operated
with quantity triggers.
For the reserve system based on price
triggers, the price range for wheat is set
at $80 to $140 per metric ton, and for corn
at $70 to $120 per metric ton. For the sim-
ulation period 1976 to 1982, the average
wheat price is the same ($110 per ton)
whether or not the government accumulates
stocks, but the degree of price variability
is reduced by half when government stocks
are held. For feedgrains, the average price
drops from $100 to $95 per ton if buffer
stocks are held, and the degree of variability
is again reduced by almost half. Slightly
more wheat moves into export with stocks
than without, and slightly less feedgrain. The
average size of the combined stock is 41 mil-
lion metric tons (18 million wheat and 23
million feedgrains), although the maximum
stock size could occasionally rise to more
than 80 million metric tons. The net annual
cost for running this system is $389 million,








though in some years it could exceed $2 bil-
lion. A size limit is not imposed on stocks.
Such a limit would reduce costs, but would
increase the chances of extremely high prices
from time to time.
For the reserve system based on quantita-
tive triggers, two rules are evaluated: one in
which stocks are accumulated and released
when world production deviates more than
two percent above or below trend, and the
other when production deviates by more
than five percent. In both cases, net storage
costs are about the same as with price trig-
gers, but less price stabilization is achieved.


Finally, Trezise (1976) proposes a wheat
and coarse grain reserve of 64 million met-
ric tons, rising to about 80 million metric
tons by 1980. This reserve would insure
against all sources of instability: a wheat
shortfall in importing countries, a poor
coarse grain harvest in the United States, or
a serious failure of the world rice crop. He
concludes that to be workable, a reserve
system probably must be multilaterally
financed and managed, and recommends
ways in which a multilateral reserve arrange-
ment might be operated without interfering
with normal market forces.









Appendix 2


PROJECTED DEMAND, TRENDS
AND VARIABILITY IN CEREAL PRODUCTION


Projected cereal demand for each country
is obtained on the basis of population
growth, GNP growth, and income elastic-
ities. The expression used is as follows:
(1)
Cit = Nit [oi 1975 (1 + ri goi) t -1975 +

Cei 1975(1 + ri ge t-1975

where

Cit = projected cereal demand of the ith
country in year t;

Nit = projected population of the ith
country in year t;

Coi 1975 cei 1975 = per capital trend
foodgrain and feedgrain consumption, re-
spectively, of the ith country in 1975.

ri = annual growth rate of real per capital
GNP of the ith country.

goi gei = income elasticities of foodgrains
and feedgrains, respectively, of the ith
country.

Per capital trend consumption for food-
grain and feedgrain were obtained by fitting
logarithmic time trends using actual 1960-75
data. The values of these trends as well as
the values of the other variables of the con-
sumption function and their source are pre-
sented in Table 16.


Production trends for total cereals have
been estimated for each country by fitting
a logarithmic time trend to actual produc-
tion data for the period 1960 to 1975. The
trend equations obtained are in the form of

(2)
- b. (t-1960)
Qit = io e t = 1960, 1961,...
where Qit io are trend production for the
ith country of year t and the base year
1960, respectively; and bi is the annual rate
of growth of cereal production for the ith
country. Annual cereal growth rates and the
1975 trend production levels are presented
in Table 17.
The variability of cereal production above
and below trend has been analyzed for each
country. Actual production levels for each
year from 1960 to 1975 have been expressed
as a percentage deviation from trend, as

(3)

Oit- it 100
q 100
qit it

where Qit is the actual production level in
year t. The standard deviations of this pro-
duction variability measure are also pre-
sented in Table 17.
Production variability of individual coun-
tries was correlated with that of other coun-
tries within the same geographical region.
Countries which showed significant correla-
tion were then grouped together. This re-













-Im I1- I 5-g-m -
"a
-. 0. w- o

-.~ 'A a-n 'A

I~ ?- I 5
00
a~~r


0D
0
s,,
Co







ha k





0a% 00 0 0 WO O Go













00 0 000 00 000
0O 0' 0. a^ aob c o -' n t


vi
.n
* . 0


S0
* . O0


w
o p o A
o oC -


0 0 0 w


- o 00
W 4^ LU U


n
0









c





















rr~
-I



iJ a
a -5
0.n-


3
^ SP



S"
(u


"j
0
8o




'3.














0

06-


0
-c






3 *
0.(







00 Table 16-Continued


Annual Per Capita Income Elasticity b/ Per Capita Trend Population Projections
Country GNP Growth Consumption (1975) c/ (thousands) d/
Country Rate (%) a/ Foodgrain Feedgrain
Foodgrain Feedgrain 1975 1980 1985
(kg/person)


Cyprus

Jordan

Lebanon

Morocco

Syrian Arab Rep.

Tunisia

Afghanistan

Yemen Arab Rep.

Sudan

Nigeria

Angola

Cameroon

Ghana

Ivory Coast


.. 527.5

1.120 133.8

.760 168.9

321.1

1.170 255.4

1.040 210.9

.. 225.3

... 99.4

... 136.5

1.130 127.1

97.9

... 126.6

.. 95.7

.. 160.5


673

2,688

2,869

17,504

7,259

5,747

19,280

6,668

18,268

62,925

6,353

6,398

9,873

4,885


714

3,177

3,360

20,384

8,536

6,562

22,038

7,741

21,420

72,596

7,181

7,088

11,446

5,579


755

3,752

3,956

23,788

10,081

7,537

25,207

9,000

25,147

84,400

8,188

7,987

13,395

6,399






Table 16-Continued


Annual Per Capita Income Elasticity b/ Per Capita Trend Population Projections
GNP Growth Consumption (1975) c/ (thousands) d/
Country Rate (%) a/ Foodgrain Feedgrain
Foodgrain Feedgrain 1975 1980 1985
(kg/person)


Liberia 1.6 .500 ... 87.0 0.0 1,708 1,937 2,199

Senegal 0.5 .244 .840 235.7 1.1 4,418 4,989 5,642

Zambia 1.7 .100 .970 242.2 8.2 5,022 5,874 6,920

Chad 0.5 .500 ... 5.9 0.0 4,023 4,473 4,978

Benin 0.7 .483 ... 78.9 0.0 3,074 3,534 4,070

Gambia 2.8 .500 ... 68.2 0.0 510 563 624

Guinea 0.5 .483 ... 89.5 0.0 4,416 5,014 5,718

Malagasy 0.5 .100 ... 199.7 0.0 8,020 9,329 10,909

Malawi 2.9 .089 ... 253.9 0.0 4,916 5,577 6,369

Mali 0.9 .350 ... 136.2 0.0 5,697 6,470 7,374

Niger 0.5 .200 ... 179.9 0.0 4,592 5,272 6,077

Rwanda 0.5 .500 ... 66.6 0.0 4,200 4,865 5,654

Sierra Leone 1.2 .500 ... 132.6 0.0 2,983 3,392 3,870


Tanzania 2.0


.324 162.5 0.0


15,438 18,052 21,142






S Table 16-Continued

Annual Per Capita Income Elasticity b/ Per Capita Trend Population Projections
GNP Growth Consumption (1975) c/ (thousands) d/
Country Rate (%) a/ Foodgrain Feedgrain
Foodgrain Feedgrain 1975 1980 1985
(kg/person)


Uganda 1.4 .274 ... 119.3 0.0 11,353 13,222 15,423

Upper Volta 0.5 .291 ... 162.4 0.0 6,032 6,774 7,639

Zaire 2.0 .500 ... 38.0 0.0 24,485 27,952 32,139

Brazil 3.0 .100 .480 151.5 76.2 109,730 126,389 145,082

Ecuador 3.0 .420 .880 80.0 9.8 7,090 8,303 9,689

Mexico 2.5 .100 .610 225.1 69.3 59,204 69,965 82,803

Venezuela 3.0 .190 .460 137.1 49.9 12,213 14,134 16,326

Bolivia 1.9 .450 .860 104.0 32.0 5,410 6,162 7,013

Chile 1.3 .000 .690 199.6 46.3 10,253 11,235 12,303

Colombia 2.0 .268 .690 77.7 17.3 25,890 30,215 35,050

Paraguay 1.5 .099 ... 157.8 0.0 2,647 3,062 3,540

Peru 1.5 .370 .770 125.9 41.1 13,526 17,711 20,424

Costa Rica 2.2 .178 .640 109.7 31.6 1,994 2,286 2,611

Cuba 0.5 .218 ... 195.5 0.0 9,481 10,533 11,660







Table 16-Continued


Annual Per Capita Income Elasticity b/ Per Capita Trend Population Projections
GNP Growth Consumption (1975) c/ (thousands) d/
Country Rate (%) a/ Foodgrain Feedgrain
Foodgrain Feedgrain 1975 1980 1985
(kg/person)


Dominican Rep. 2.3 .500 .850 67.0 23.5 5,118 6,052 7,171

El Salvador 1.4 .298 .700 125.6 37.8 4,108 4,813 5,643

Guatemala 2.5 .155 .700 152.5 11. 6,129 7,100 8,210

Haiti 0.5 .500 ... 80.4 0.0 4,552 4,956 5,441

Honduras 1.2 .175 .650 134.5 14.2 3,037 3,595 4,242

Jamaica 2.7 .211 ... 144.4 0.0 2,029 2,172 2,316

Nicaragua 2.2 .250 .610 138.2 33.3 2,318 2,732 3,218

Panama 3.1 .180 .700 120.1 21.2 1,678 1,930 2,217

Trinidad 3.0 .037 ... 212.6 0.0 1,009 1,062 1,116


a/ A low income growth assumption is made here. These growth rates were arrived at from the 1976 World Bank Atlas and other World Bank materials.
b/ Income elasticities were largely derived from FAO (Agricultural Commodity Projections 1970-1980, Vol. 2, 1971), adjusted to accommodate the low income
growth assumption.
c/ Obtained from a logarithmic fit to actual per capital data for the period of 1960 to 1975 provided by the USDA.
d/ United Nations population estimates and medium variant projections are used here.









Table 17-Estimates of cereal production trends and variability from trend by
country


Cereal Production a/

Base Year Annual Variability
Country 1960 Growth from
(thousand metric tons) Rate (%) Trend b/ (%)


Korea, Rep. of 5,046.9 2.083 5.87

Malaysia 615.7 5.433 5.08

Bangladesh 9,587.1 1.496 7.18

India 65,055.8 2.721 6.62

Indonesia 10,238.3 3.863 7.13

Philippines 3,470.1 3.965 6.12

Burma 4,861.0 1.327 7.09

Sri Lanka 623.9 3.327 14.14

Algeria 1,753.2 -0.628 30.32

Iran 4,130.5 2.836 9.87

Iraq 1,745.3 1.370 24.62

Libya 125.9 0.757 29.30

Egypt, Arab Rep. of 5,083.3 2.385 4.58

Turkey 12,055.3 1.979 9.51

Cyprus 127.5 -0.022 41.86

Jordan 158.6 1.053 59.95

Lebanon 85.1 -2.437 25.50

Morocco 2,301.9 4.682 31.14

Syrian Arab Rep. 1,264.2 1.376 42.22

Tunisia 415.8 5.543 33.74

Afghanistan 3,374.1 1.560 7.06

Yemen Arab Rep. 603.1 -0.855 7.59

Sudan 1,345.8 3.873 17.14









Table 17-Continued


Cereal Production a/

Base Year Annual Variability
Country 1960 Growth from
(thousand metric tons) Rate (%) Trend b/ (%)


Nigeria

Angola

Cameroon

Ghana

Ivory Coast

Liberia

Senegal

Zambia

Chad

Benin

Gambia

Guinea

Malagasy

Malawi

Mali

Niger

Rwanda

Sierra Leone

Tanzania

Uganda

Upper Volta

Zaire

Brazil


7,714.1

445.6

661.4

358.6

289.0

91.8

527.8

621.2

18.9

296.4

20.3

230.0

998.2

813.8

1,071.2

841.9

266.2

259.3

1,801.2

868.7

852.3

284.8

13,731.3


-0.023

1.831

0.706

5.148

4.642

1.173

1.953

4.142

1.612

-1.268

0.947

2.706

2.502

2.648

-2.285

-0.406

0.427

1.919

1.669

3.017

0.623

5.171

3.756


10.66

12.53

14.24

9.88

10.79

15.40

18.23

29.34

21.76

7.20

16.06

11.02

8.27

12.84

11.18

9.91

10.91

5.23

10.76

7.54

18.08

7.94

5.38









Table 17-Continued


Cereal Production a/

Base Year Annual Variability
Country 1960 Growth from
(thousand metric tons) Rate (%) Trend b/(%)


Ecuador 381.9 0.964 8.54

Mexico 7,959.5 4.652 9.28

Venezuela 528.6 3.446 13.30

Bolivia 375.0 1.902 5.22

Chile 1,630.9 -0.295 12.44

Colombia 1,310.7 3.049 7.13

Paraguay 147.3 5.935 4.12

Peru 1,003.4 1.888 5.89

Costa Rica 107.7 1.899 8.54

Cuba 208.6 3.197 25.82

Dominican Rep. 117.7 4.399 5.87

El Salvador 275.4 5.200 11.86

Guatemala 589.2 2.858 5.30

Haiti 283.0 0.399 7.66

Honduras 349.3 0.855 9.68

Jamaica 5.6 1.878 34.57

Nicaragua 196.4 3.999 12.61

Panama 149.5 0.816 13.62

Trinidad 8.5 3.354 11.31


a/ The logarithmic trend on total grain production was fitted using data from 1960 to 1975 provided by
the USDA.

b/ Measured as the standarddeviation of percentage variations of actual production (Q it) from trend pro-
duction (Qt), i.e., (Qit-t) Qit for the period 1960-75.
d cin(it








gional classification of countries according
to their degree of correlation in production
variability is given in Table 18. Countries
whose production did not show a significant
positive correlation with any other country
within their geographical region were as-
sumed independent.
For the simulation analysis the generation
of production levels for each country took
into consideration the observed regional
correlation in production by jointly generat-
ing production deviation vectors. Thus, the
percentage deviations of production of the
n countries in the jth region were ob-
tained from a multivariate normal distribu-
tion, i.e.,

(4)

qlt 0
q2t 0

-N



nt ( 00
I


for all t, where j. is the estimated vari-
ance-covariance matrix of observed per-
centage deviations of the n countries in
the jth region. Assume that qi is one
generated value from the above process for
the ith country, corresponding to year t.
Then, the corresponding production level is
obtained by solving (3) for Qit, i.e.,


Q = (1+100 it


where Qit is the trend production as ob-
tained from (2).
Production variability as specified above
is multiplicative in nature which means that
although production deviations from trend
expressed as a percentage do not grow over
time, the absolute levels of production fluc-
tuations increase with time.







a% Table 18-Matrix of correlations between percentage deviations from cereal production trend for countries by region

South Asia

1 2


1. India

2. Sri Lanka


Far East Asia

2


1. Bangladesh

2. Burma

3. Indonesia

4. Philippines


East Africa


1. Malagasy

2. Malawi

3. Tanzania

4. Uganda

5. Zambia


1

0.34


0.54
0.54 1







Table 18-Continued


West Africa

5 6


7 8 9


10 11 12


1. Cameroon 1

2. Chad 0.29

3. Benin 0.15

4. Gambia 0.12

5. Guinea 0.64

6. Ivory Coast 0.62

7. Liberia 0.11

8. Mali 0

9. Niger 0.35

10. Senegal 0.26

11. Sierra Leone 0

12. Upper Volta 0.60


1

0 1


1

0.51 1

0 0

0.24 0.58

0 0.62

0 0.50

0 0

0 0.45


1 2


3 4


1

0 1

0.49 0







0 Table 18-Continued


North Africa and Near East Asia

1 2 3 4 5 6 7


1. Afghanistan 1

2. Iran 0.39 1

3. Iraq 0.46 0.32

4. Jordan 0 0

5. Lebanon 0 0

6. Syrian Arab Rep. 0.12 0

7. Yemen Arab Rep. 0 0

8. Cyprus 0 0

9. Algeria 0.12 0.13

10. Egypt, Arab Rep. of 0 0.22

11. Morocco 0 0.45

12. Libya 0.55 0.36

13. Tunisia 0 0.24


8 9 10 11 12 13


1

0.17 1


0.19 0.32

0.45 0.45

0 0.13

0.19 0.57

0.47 0

0 0

0.37 0

0.47 0.13

0 0


1

0.61 1


0.35 0.21

0.34 0.53

0 0.37

0 0

0 0

0.28 0.45

0 0.34


1

0 1

0 0.24 1

0 0 0.48 1

0 0.19 0.59 0.69 1

0 0.16 0.11 0 0

0.59 0 0.42 0.11 0


0 1







Table 18-Continued

Central America


Costa Rica
Cuba
Guatemala
Jamaica
Mexico
Panama
Nicaragua
Trinidad


South America


Brazil
Chile
Colombia
Ecuador
Paraguay
Peru
Venezuela








Appendix 3


WORLD WHEAT PRICE GENERATION



In order to obtain cereal import expendi-
tures which are the basis for estimating com-
pensations paid to each food deficit develop-
ing country, a way of generating world
wheat prices is needed.* One approach to
this is to analyze past world price distribu-
tion and on the basis of this distribution
devise a random number generating function
which would approximate the observed price
data. This approach would be appropriate
to use if one were concerned with a single
small importing country whose import de-
mand had no practical effect on world
price. However, this analysis, dealing as it
does with countries that account for more
than 40 percent of total world wheat im-
ports, cannot ignore the causal effect of
their demand on world price.
Thus, the first attempt to explore this
causal relationship is to regress world wheat
price on total cereal imports of food deficit
developing countries (FDDC) to see whether
a relationship exists. Using data from 1960
to 1975, this regression yielded the follow-
ing equation:

Pt = -30.1 + 0.003445 Mt R2 = .67

(5.2)

where Pt is the world wheat price, Mt total
cereal imports of all FDDC's. Since the co-


efficient of Mt proves significant, FDDC im-
port demand should enter the price generat-
ing function. Obviously, import demand
from the rest of the world also has a causal
effect on the world price of wheat. This
analysis, however, simulates import demand
of a subgroup of importers only, and treats
the effect on price of the demand by the
rest of the world as the random component
of the price generating function.
Another factor to be incorporated into
the random price generating function is the
serial correlation observed in price time
series. Very high wheat price in year t-1 is
not followed by a very low price in year t,
and vice-versa. This is the effect of carryover
stocks on prices. Thus, if year t-1 is a poor
harvest year worldwide, prices will rise and
accumulated stocks will be depleted. If year
t is also a poor crop year, depleted stocks
Will force current prices even higher. If year
t is a good crop year, price will drop but not
to very low levels, because the additional
demand to replace part of depleted stocks
keeps price from falling. This reasoning sug-
gests that a lagged price should be included
in the price generating function.
Finally, in order to avoid any trend in
real price, both imports and world price are
introduced as normalized variables and re-
gression on 1960-75 actual data provides
the following relationship:


The composition of cereal imports varies from country to country depending on consumption pref-
erences and even for the same country from year to year depending on the variability of the composition
in its domestic cereal production and changes in taste. The composition of cereal imports will not be of
concern in this study. Instead, wheat is taken as the basic staple commodity, and the value of cereal imports
is computed as if these imports were composed of wheat.








P M
loge = -0.89028 + 0.96268 -=-+
P* Mt
(2.0)

0.86181 loge +ut
p*()
(4.6)


SE = 0.17378 R2 = 0.75

where Pt is world wheat price in year t
(in constant 1972 dollars);

P* is the average world wheat price for the
pre-1972 period ($74.1/MT) which is as-
sumed to approximate the cost of produc-
tion in exporting countries;

Mt Mt are total food deficit developing
countries' actual and trend cereal imports,
respectively, in year t ; and

ut is the random component of the price
generating function in year t.


The log-linear specification used above is
meant to reflect the non-linear price be-
havior ascribed to the international wheat
market.
The price generating function used in
the simulation was based on the estimated
relationship. P* was adjusted to reflect cur-
rent production costs in exporting countries.
Thus, P* is set at $85.0 per metric ton
(about $2.31 per bushel) which is approxi-
mately the current level of the United States


government's wheat loan rate and thus re-
flects current production costs. As an initial
value for the lagged price (Pt1), the average
wheat price for the last three years (1975-
77) is used. This price, a simple average of
United States and Canadian wheat prices
expressed in 1977 dollars, equals $137.0 per
metric ton.
A price outcome is computed from this
function after the generation of an import
outcome (the difference between total pro-
jected cereal demand and total randomly
generated cereal production, as described in
Appendix 2) and an outcome of the random
component ut. The distribution obtained
from 300 five-year production sequences is
depicted in Table 19.
The initial value of the lagged price af-
fects somewhat the price distribution. Thus,
the distribution of the first year (1978) is
not as dispersed as it would have been if
the initial price was allowed to take a ran-
domly generated value. This is designated
as "first year distribution" in Table 19. The
impact of the initial assigned value is con-
siderably less by the second and third years
and is practically non-existent thereafter.
In other words, from the third year on, the
price distribution obtained with an assigned
initial value is practically the same as the dis-
tribution obtained with an initial price
randomly generated. This is designated as
"equilibrium distribution" in Table 19. The
expected prices obtained are $146.7 per
metric ton for 1978, $150.7 per metric ton
for 1979, $153.6 per metric ton for 1980,
and $155.8 per metric ton for the years
thereafter.










Table 19-World wheat price distribution a/


First Year Distribution (1978)

Price Relative Cumulative
Interval Frequency Frequency
($/MT) (%) (%)


Less than 70


70-90


90-110


110-130


130-150


150-170


170-190


190-210


210-230


230-250


0.67


5.00


15.00


17.33


17.33


14.33


13.33


7.33


4.33


2.00


250-270


270-290


290-310


310-330


330-350


Greater than 350


0.67


5.67


20.67


38.00


55.33


69.67


83.00


90.33


94.67


96.67


98.67


99.67


100.00


100.00


100.00


Equilibrium Distribution

Relative Cumulative
Frequency Frequency
(%) (%)


1.67 1.67


7.69 9.36


13.05 22.41


16.00 38.41


15.50 53.91


13.93 67.84


9.76 77.60


7.02 84.62


4.74 89.36


92.84


95.44


96.80


97.57


98.42


98.94


a/ Obtained by the procedure described in the text. The expected prices for 1978, 1979, 1980 and
thereafter are $146.7 per metric ton, $150.7 per metric ton, $153.6 per metric ton and $155.8 per
metric ton, respectively.










Appendix 4 SUPPLEMENTARY TABLES

Table 20-Probability distribution of present value of the total cost of the
scheme operating as a compensatory financing mechanism for the
five-year period 1978 to 1982, excluding India a/

Cost b/ Relative Cumulative
($ billion) Frequency Frequency
(%) (%)


0-1 26.9 26.9

1-2 20.7 47.6

2-3 13.9 61.5

3-4 8.0 69.5

4-5 5.4 74.9

5-6 3.3 78.2

6-7 3.6 81.8

7-8 3.3 85.1

8-9 2.9 88.0

9-10 1.5 89.5

10-11 2.5 92.0

11-12 1.5 93.5

12-13 0.3 93.8

13-14 1.1 94.9

14-15 0.7 95.6

15-16 0.0 95.6

16-17 0.8 96.4

17-18 0.3 96.7

18-19 0.4 97.1

Greater than 19 2.9 ...

a/ The excess over 110 percent of trend in each country's cereal import bill is covered by the scheme.

b/ The expected present value of the cost equals $4.096 billion. A discount rate of 8 percent has been
assumed here.









Table 21-Probability distribution of the current cost of the scheme operating as
a compensatory financing mechanism for the middle year (1980) of
the five-year period, excluding India a/


Cost b/ Relative Cumulative
($ billion) Frequency Frequency
(%) (9%)


0-1 74.5 74.5


1-2 10.6 85.1


2-3 5.1 90.2


3-4 4.7 94.9


4-5 1.8 96.7


5-6 1.1 97.8


6-7 1.1 98.9


7-8 0.4 99.3


Greater than 8 0.7



a/ The excess over 110 percent of trend in each country's cereal import bill is covered by the scheme.

b/ The expected current cost equals $0.960 billion.






Table 22-Probability distribution of the present value of payments made through the compensatory financing mechanism
under alternative grain reserve levels a/


Grain Reserve Levels
(million MT)

Payments c/
($ billions)


0-1

1-2

2-3

3-4

4-5

5-6

6-7

7-8

8-9

9-10

10-11

11-12

12-13


0b/


4


Relative Cumulative Rel. Cum. Rel.
Frequency Frequency Freq. Freq. Freq.
(%) (%) (%) (%) (%)


20.0 20.0

41.7 21.5

55.1 16.3

64.8 10.6

72.5 8.0

78.0 4.7

81.3 2.2

83.6 2.2

86.0 2.9

88.4 1.4

90.1 1.5

91.1 1.0

92.4 0.8


8


12 16 20


Cum. Rel. Cum. Rel. Cum. Rel.
Freq. Freq. Freq. Freq. Freq. Freq.
(%) (%) (%) (%) (%) (%)


20.0

41.5

57.8

68.4

76.4

81.1

83.3

85.5

88.4

89.8

81.3

92.3

93.1


20.0 20.0

42.2 22.5

59.3 16.8

69.5 10.5

80.0 10.6

84.7 5.4

87.3 1.5

89.5 1.8

90.9 2.5

92.4 1.1

93.5 0.8

94.2 1.4

95.3 1.1


Cum.
Freq.
(%)







Table 22-Continued


Grain Reserve Levels
(million MT)

Payments c/
($ billions)



13-14

14-15

15-16

16-17

17-18

Greater than 18

Expected Present
Value of Total
Payments
($ billion)


0b/ 4 8 12 16 20

Relative Cumulative Rel. Cum. Rel. Cum. Rel. Cum. Rel. Cum. Rel. Cum.
Frequency Frequency Freq. Freq. Freq. Freq. Freq. Freq. Freq. Freq. Freq. Freq.
(%) (%) (%) (%) (%) (%) (%) (%) (%) (%) (%) (%)


1.4 92.7 1.2 93.6 1.1 94.2 0.7 94.2 1.1 96.4 0.4 96.4

0.4 93.1 0.4 94.0 0.9 94.9 1.4 95.6 0.3 96.7 0.3 96.7

1.1 94.2 1.1 95.1 1.1 96.0 0.4 96.0 0.0 96.7 0.0 96.7

1.1 95.3 0.9 96.0 0.4 96.4 0.4 97.4 0.4 97.1 0.4 97.1

1.1 96.4 0.4 96.4 0.0 96.4 0.7 97.1 0.4 97.5 0.4 97.5

3.6 ... 3.6 ... 3.6 ... 2.9 ... 2.5 ... 2.5 ...


5.108


4.789


4.509


4.290


4.124


4.006


a/ The current cereal import bill in excess of 110 percent of trend is covered for each country.
b/ This corresponds to a system operating as a compensatory financing mechanism only.
c/ In computing present values a discount rate of 8 percent has been assumed.







Table 23-Probability distribution of current payments made for the middle year (1980) of the five-year period under
alternative grain reserve levels a/


Grain Reserve Levels
(million MT)

Payments /
($ billions)



0-1
1-2
2-3
3-4
4-5
5-6
6-7
7-8
8-9
9-10
10-11
11-12
Greater than 12

Expected Payment
($ billion)


0 b/


Relative Cumulative
Frequency Frequency
(%) (%)


4

Rel. Cum.
Freq. Freq.
(%) (%)


69.5 69.5
81.1 12.7
87.3 6.9
91.7 4.4
95.3 1.8
96.4 1.8
97.5 1.1
98.2 0.0
98.2 0.0
98.6 0.4
99.3 0.7
99.3 0.0
0.7


1.229


1.158


8

Rer. Cum.
Freq. Freq.
(%) (%)


69.5 70.9
82.2 13.5
89.1 6.5
93.5 2.6
95.3 2.9
97.1 1.5
98.2 0.4
98.2 0.0
98.2 0.0
98.6 0.4
99.3 0.7
99.3 0.0
0.6


12 16 20

Rel. Cum. Rel. Cum. Rel. Cum.
Freq. Freq. Freq. Freq. Freq. Freq.
(%) (%) (%) (%) (%) (%)


0.9 71.6
4.4 13.1
0.9 6.9
3.5 4.0
6.4 1.5
7.9 0.7
8.3 0.4
8.3 0.0
8.3 0.7
8.7 0.0
9.4 0.4
9.4 0.0
... 0.2


1.084


71.6 72.4
84.7 13.1
91.6 8.4
95.6 2.6
97.1 1.1
97.8 0.4
98.3 0.4
98.3 0.7
99.4 0.4
99.4 0.0
99.8 0.4
99.8 0.0
... 0.2


1.012


0.951


72.4 72.4
85.5 13.1
93.9 8.7
96.5 2.2
97.6 1.5
98.0 0.4
98.4 0.7
99.1 0.4
99.4 0.0
99.4 0.0
99.8 0.4
99.8 0.0
... 0.2


0.916


a/ The excess over 110 percent of trend in each country's current cereal import bill is covered.

b/ This corresponds to a system operating as a compensatory financing mechanism only.
4 c/ In computing present values a discount rate of 8 percent has been assumed.







0 Table 24-Probability distribution of the present value of payments made through the compensatory financing mechanism
under alternative insurance levels


Scheme Operating as a Compensatory Scheme Operating as a Combination of a Compensatory
Financing Mechanism Only Financing Mechanism and 20 MT of Grain Reserve

Uninsured Import Bill Level Uninsured Import Bill Level

110 Percent b/ 120 Percent 130 Percent 110 Percent 120 Percent 130 Percent
Payments a/ Relative Cumulative Rel. Cum. Rel. Cum. Rel. Cum. Rel. Cum. Rel. Cum.
Frequency Frequency Freq. Freq. Freq. Freq. Freq. Freq. Freq. Freq. Freq. Freq.

($ billions) (percent)

0-1 20.0 20.0 25.5 255 31.6 31.6 20.0 20.0 25.5 25.5 36.4 36.4

1-2 21.8 41.8 19.6 45.1 23.3 54.9 225 42.5 27.3 52.8 30.2 665

2-3 12.0 53.8 14.2 59.3 12.4 67.3 16.8 59.3 17.8 70.6 16.4 82.9

3-4 7.7 61.5 9.8 69.1 8.4 75.6 10.5 69.8 12.7 83.3 3.3 86.2

4-5 8.0 695 6.5 75.6 55 81.1 10.6 80.4 2.9 86.2 4.7 90.9

5-6 5.0 74.5 4.4 80.0 2.2 833 5.4 85.8 2.2 88.4 0.7 91.6

6-7 4.8 79.3 1.8 81.8 4.7 88.0 1.5 87.3 25 90.9 1.8 93.5

7-8 2.2 81.5 3.6 85.4 0.7 88.7 1.8 89.1 1.5 92.4 1.8 95.3

8-9 2.1 83.6 2.2 87.6 1.8 905 2.5 91.6 0.7 93.1 0.4 95.6

9-10 3.3 86.9 1.1 88.7 1.5 92.0 1.1 92.7 1.8 94.9 0.7 96.4

10-11 1.5 88.4 1.8 90.5 1.1 93.1 0.8 93.5 1.1 96.0 0.0 96.4





Table 24-Continued

Scheme Operating as a Compensatory Scheme Operating as a Combination of a Compensatory
Financing Mechanism Only Financing Mechanism and 20 MT of Grain Reserve

Uninsured Import Bill Level Uninsured Import Bill Level

110 Percent b/ 120 Percent 130 Percent 110 Percent 120 Percent 130 Percent
Payments a/ Relative Cumulative Rel. Cum. Rel. Cum. Rel. Cum. Rel. Cum. Rel. Cum.
Frequency Frequency Freq. Freq. Freq. Freq. Freq. Freq. Freq. Freq. Freq. Freq.


($ billions) (percent)

11-12 1.4 89.8 1.1 91.6 1.1 94.2 1.4 949 0.0 96.0 0.7 97.1
12-13 1.5 91.3 1.1 92.7 1.5 95.6 1.1 96.0 0.4 96.4 0.4 97.5
13-14 1.4 92.7 1.1 93.8 0.4 96.0 0.4 96.4 0.4 96.8 0.0 97.5
14-15 0.4 93.1 2.2 96.0 0.4 96.4 0.3 96.7 0.7 97.5 0.0 97.5
15-16 1.1 94.2 0.0 96.0 0.0 96.4 0.0 96.7 0.0 97.5 0.0 97.5
16-17 1.1 95.3 0.4 96.4 1.1 97.5 0.4 97.1 0.0 97.5 0.4 97.9
17-18 1.1 96.4 0.0 96.4 0.0 97.5 0.4 97.5 0.0 97.5 0.4 98.3
Greater than 18 3.6 ... 3.6 ... 2.5 ... 2.5 ... 25 ... 1.7 ...
Expected Present
Value of Total
Payments
($ billion) 5.108 4.348 3.690 4.006 3.247 2.603


a/ In computing present values, a discount rate of 8 percent has been assumed.
b/ The figures under this alternative have been duplicated from Table 22 for comparison.







S Table 25-Probability distribution of present value of payments made through the compensatory financing mechanism
operating in conjunction with a 20 million metric ton grain reserve, under alternative grain release rules a/

Release Rules

Price Greater than $200/MT Price Greater than $170/MT Price Greater than $200/MT
Shortfall Greater than 5% c/ Shortfall Greater than 5% Shortfall Greater than 3%


Payments b/ Relative Cumulative Rel. Cum. Rel. Cum.
Frequency Frequency Freq. Freq. Freq. Freq.


(percent)


($ billion)

0-1

1-2

2-3

3-4

4-5

5-6

6-7

7-8

8-9

9-10

10-11


20.0

42.5

59.3

69.8

80.4

85.8

87.3

89.1

91.6

92.7

93.5






Table 25-Continued

Release Rules

Price Greater than $200/MT Price Greater than $170/MT Price Greater than $200/MT
Shortfall Greater than 5% c/ Shortfall Greater than 5% Shortfall Greater than 3%

Payments b/ Relative Cumulative Rel. Cum. Rel. Cum.
Frequency Frequency Freq. Freq. Freq. Freq.


($ billion) (percent)

11-12 1.4 94.9 0.7 94.2 0.7 94.
12-13 1.1 96.0 1.5 95.6 1.1 95.6
13-14 0.4 96.4 0.7 96.4 0.7 96.4
14-15 0.3 86.7 0.0 96.4 0.0 96.4
15-16 0.0 96.7 0.0 96.4 0.0 96.4
16-17 0.4 97.1 0.7 97.1 0.7 97.1
17-18 0.4 97.5 0.4 975 0.4 975
Greater than 18 2.5 ... 2.5 ... 25 ...
Expected Present
Value of Total
Payments
($ billion) 4.006 3548 3.874


a/ The excess over 110 percent of trend in each country's current import bill is covered.
b/ In computing present values, a discount rate of 8 percent has been assumed.
c/ This column has been duplicated from Table 22 for comparison.








Appendix 5


COUNTRY LIST



A. Developing Countries with Foreign Exchange

1. Asia Group
a. Malaysia
b. Republic of Korea

2. North Africa/Middle East (OPEC Group)
a. Algeria
b. Iran
c. Iraq
d. Libya

3. Latin America: Venezuela

B. Developing Countries with Foreign Exchange Constraints

1. Asia Market Economies
a. Bangladesh
b. Burma
c. India
d. Indonesia
e. The Philippines
f. Sri Lanka

2. North Africa/Middle East (Non-OPEC)
a. Middle Income
(1) Morocco
(2) Turkey
(3) Other Middle-Income: Cyprus, Jordan, Lebanon, Syria, Tunisia
b. Low Income
(1) Afghanistan
(2) Egypt
(3) Sudan
4Yemen Arab Republic







3. Sub-Sahara Africa


West
a. Middle Income
(1) Ghana
(2) Other Middle Income: Angola, Ivory Coast, Liberia
b. Low Income
(1) Nigeria
(2) Sahel Countries: Chad, Mali, Niger, Senegal, Upper Volta
(3) Other Low Income: Benin, Cameroon, Gambia, Guinea, Sierra Leone

East
a. Middle Income: Zambia
b. Low Income
(1) Tanzania
(2) Uganda
(3) Zaire
(4) Other Low Income: Malagasy, Malawi, Rwanda

4. Latin America

a. Middle Income

(1) Brazil
(2) Chile
(3) Colombia
(4) Ecuador
(5) Mexico
(6) Peru
(7) Other Latin America: Costa Rica, Cuba, Dominican Republic, El Salvador,
Guatemala, Honduras, Jamaica, Nicaragua, Panama, Paraguay, Trinidad

b. Low Income: Bolivia, Haiti










The sixty-five countries are grouped according to the categories established for IFPRI's
Research Report No. 3, Food Needs of Developing Countries: Projections of Production and
Consumption to 1990. They are grouped by income based on the average 1973 GNP per
capital: middle income, US$ 300 or more; and low income, less than US$ 300. Asian countries
all fall under the low income group.








REFERENCES TO APPENDIX 1




OVERVIEW OF PAST APPROACHES TO

THE INSTABILITY OF GRAIN MARKETS



Bailey, W. R.; Kurish, F. A. ;and Rojko, A. S. "Grain Stocks, Issues and Alternatives-A Progress Report."
U.S. and World Food Security. U.S. Congress. Senate, Committee on Agriculture and Forestry, 93rd
Congress, 2nd Session, March 15, 1974.

Cochrane, Willard, and Danin, Yigal. "Reserve Stock Grain Models and the World, 1975-1985." Analysis of
Grain Reserves, a Proceedings. ERS-634. Washington, D.C.: U.S. Department of Agriculture and the
National Science Foundation, 1976.

Danin, Yagil; Summer, Daniel; and Johnson, D. Gale. Determination of Optimal Grain Carryovers. Office
of Agricultural Economic Research, Paper No. 74:12. Chicago: University of Chicago, 1975.

Eaton, D.; Steel, W. S. ; Coho, J. L. ; and Revelle, C. S. "The Joseph Problem: How Large a Grain Re-
serve?" Mimeograph, Department of Geography and Environmental Engineering. Baltimore: Johns
Hopkins University, March 1976.

Gray, Roger. "Grain Reserves Issues." Speech delivered at the 1974 National Agricultural Outlook Con-
ference, December 9, 1974, in Washington, D.C. Mimeographed.

Gustafson, Robert. Carryover Levels for Grains. Technical Bulletin No. 1178. Washington, D.C. : U.S.
Department of Agriculture, 1958.

Hazell, Peter B. R., and Scandizzo, Pasquale L. "Market Intervention Policies When Production is Risky."
American Journal of Agricultural Economics 57 (1975):641-649.

Heuth, Darrell, and Schmitz, Andrew "International Trade in Intermediate and Final Foods: Some Welfare
Implications of Established Prices." Quarterly Journal of Economics 86 (1972):351-365.

Johnson, D. Gale. "Increased Stability of Grain Supplies in Developing Countries: Optimal Carryovers and
Insurance." World Development 4 (1976): 977-987.

Just, Richard E.; Lutz, Ernst; Schmitz, Andrew; and Turnvosky, Stephen. "The Distribution of Welfare
Grains from Price Stabilization Under Distortions." American Journal of Agricultural Economics
59 (1977):652-661.

"The Distribution of Welfare Gains from Price Stabilization: An International Perspec-
tive." Journal of International Economics. Forthcoming.

Konandreas, Panos, and Schmitz, Andrew, "Welfare Implications of Grain Price Stabilization: Some
Empirical Evidence for the United States," American Journal of Agricultural Economics, 60, (1978):
74-84.

Massel, Benton F. "Price Stabilization and Welfare." Quarterly Journal of Economics 83 (1969): 285-297.







Newbery, D. M. G. "Price Stabilization with Risky Production. Institute for Mathematical Studies in the
Social Sciences, Working Paper No. 69. Stanford, California: Stanford University, 1976.


Oi, W. "The Desirability of Price Instability Under Perfect Competition." Econometrica 29 (1961): 58-64.

Reutlinger, Shlomo. "A Simulation Model for Evaluating Worldwide Buffer Stocks of Wheat." American
Journal of Agricultural Economics 58 (1976): 1-12.

Samuelson, P. A. "The Consumer Does Benefit from Feasible Price Stability." Quarterly Journal of
Economics 86 (1972): 476-493.

Sarris, A. H.; Abbot, P. C. ; and Taylor, L. "Grain Reserves, Emergency Relief, and Food Aid." Mimeo-
graph. Washington, D.C. : Overseas Development Council, March 1977.

Trezise, P. International Grain Reserves. Washington, D.C. : Brookings Institution, 1976.

Turnovsky, S. J. "Price Expectations and the Welfare Gains from Price Stabilization." American Journal
of Agricultural Economics 56 (1974): 706-716.

U.S. Department of Agriculture. The World Food Situation and Prospects to 1985. Foreign Agricultural
Economic Report No. 98, Washington, D.C. : U.S. Government Printing Office, December 1974.

Walker, R. L.; and Sharpies, J.S. Reserve Stocks of Grain: A Review of Research. Agricultural Economic
Report No. 304. Washington, D.C.: U.S. Department of Agriculture, August 1975.

Waugh, Frederick V. "Does the Consumer Benefit from Price Instability?" Quarterly Journal of Economics
58 (1944): 602-614.


























































Panos Konandreas joined IFPRI as a Research Associate in August, 1976. Previously, he was a Post-graduate
Research Agricultural Economist at the University of Calfornia where he completed his Ph.D. studies. He
has published articles on the welfare implications of grain price stabilization and the international grain
market.

Barbara Huddleston, a Research Fellow, joined IFPRI in January, 1976. She was previously Director of the
Trade Negotiations Division, Foreign Agricultural Service, U.S. Department of Agriculture and has also
served as a trade specialist in the Africa Division, U.S. Department of Commerce. Her publications deal with
agricultural trade issues, including grain reserves and food aid.

Virabongsa Ramangkura, joined IFPRI as a Research Associate in December, 1976. He is on leave from the
Faculty of Economics, Chulalongkorn University, Bangkok, Thailand.



96







International Food Policy
Research Institute



Board of Trustees


Sir John Crawford
Chairman; Australia

Ralph Kirby Davidson
Vice Chairman; U.S.A.

Ojetunji Aboyade
Nigeria
David E, Bell
U.S.A.

Norman E. Borlaug
Mexico

Mohamed EI-Khash
Syria

Nurul Islam
Bangladesh

Affonso Pastore
Brazil

Puey Ungphakom
Thailand

Lucia G. Reca
Argentina
Roger Savary
France

Andrew Shonfield
United Kingdom

V. S. Vyas
India

Ruth Zagorin
U.S.A.

John W. Mellor, Director
Ex Officio, U.S.A.





IFPRI PUBLICATIONS


MEETING FOOD NEEDS IN THE DEVELOPING WORLD: LOCATION
AND MAGNITUDE OF THE TASK IN THE NEXT DECADE

COMMODITY TRADE ISSUES IN INTERNATIONAL NEGOTIATIONS,
by Barbara Huddleston

RECENT AND PROSPECTIVE DEVELOPMENTS IN FOOD CONSUMP-
TION: SOME POLICY ISSUES

POTENTIAL OF AGRICULTURAL EXPORTS TO FINANCE INCREASED
FOOD IMPORTS IN SELECTED DEVELOPING COUNTRIES, by Alberto
Valdes and Barbara Huddleston

FOOD NEEDS OF DEVELOPING COUNTRIES: PROJECTIONS OF
PRODUCTION AND CONSUMPTION TO 1990


SW wr w WV




University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - - mvs