AND EFFECTS OF ALTERNATIVE POLICIES
ON COSTA RICAN COFFEE FARMS
JOHN LEWIS BIEBER
A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
The author wishes to express his sincere appreciation to Dr.
W. W. McPherson, Chairman of the Supervisory Committee, for his
guidance and supervision throughout all phases of this research and
for his valuable suggestions and criticisms in preparing this manu-
Thanks are also due Dr. C. E. Murphrec, Dr. C. W. Fristce, Dr.
K. C. Gibbs and Dr. L. H. Myers for reviewing the manuscript and
offering assistance. The author also wishes to thank Dr. H. L.
Popenoe, Director of the University of Florida's Center for Tropical
Agriculture,for his assistance in obtaining funds for support of the
project. The assistance of the University of Florida's Computing
Center is recognized and appreciated. The author also wishes to
express thanks to Mr. D. W. Parvin for his aid in interpreting the
Special thanks are also due numerous people in Costa Rica who
graciously supplied information and insights necessary to the comple-
tion of this work. The Oficina del Cafe under the direction of Sr.
Alvaro Castro Jimenez supplied transportation, office facilities, and
technical assistance. The author depended heavily on the advice of
Ing. Hugo Castro, Ing. Rogello Acosta and !ig. Edwin Marin in the
initial phases of the study.
The cooperation of the Extension Service of the Costa itica
Minist-y of Ayricult ,re is also appreciated. Ac!now\Jledgimnt and
thanks are due the ,nny people who supplied input-uutput information.
Additional assistance was given by others including Carlos
Arroyo, Hester Barres, James Ross, Russel Desrosiers, Oscar Benavides,
Robert F. Voertman, J. Robert Hunter, and Ernesto Sanarrusia.
Special thanks are also due Miss Linda DiDuonni and Mrs. Sandi
Davis for typing the first draft and to Mrs. Lillian Ingenlath for
typing the final manuscript.
TABLE OF CONTENTS
ACKNOWLEDGMENTS . . . . . . . . . .
LIST OF TABLES . . . . . . . . . . . .
INTRODUCTION . . . . . . . . . . . .
An Economic Background . . . . .
The Importance of Coffee to Costa Rica
The Marketing Situation for Coffee . .
The Pros and Cons of Diversification .
Approaches and Attitudes toward
The Problem and Objectives .
SCOPE AND METHOD OF STUDY . . .
Selection of Areas and Farms
Description of Farms Studied
Palmares-San Ramon . .
Alajuela . . . . .
Acosta . . . . .
The Linear Programming Model
The Enterprises and Restraints
The Sources of the Budgets .
Assumptions of the Model . .
RESULTS . . . . . . .
Optimal Cropping Plans
Farm I .
Farm 4 .
Farm 8 .
. . . . . . 20
. . . . . . 20
. . . . . . 22
. . . . 57
. . . . . . 57
Farm 12 . . . . . . . . . .... 64
Farm 13 . . . . . . . . 65
Farm 14 . . . . . . . . . .... 66
Farm 15 . . . . . . . . ... . 67
Farm 16 . . . . . . . . ... . 68
Policy Analysis . . . . . . . .. .. . .69
Education for Better Farm Management ..... .69
Taxation or Price Reduction . . . . . 75
Payments for Coffee Removal . . . . .. 81
Price Differentiation . . . . . ... .87
Reduction of Credit . . . . . . ... .90
Movement of Labor . . . . . . ... .93
Extra Credit . . . . . . . . ... .98
Subsidies for Alternatives . . . . . . 101
Stability of Alternative Crops . . . . . . 103
Comparative Costs of Coffee Removal . . . . .. 110
SUMMARY AND CONCLUSIONS .. . . . . . . . 134
Effects of Improved Resource Allocation . . ... .134
Effects of Taxes or Price Declines . . . ... .136
Effects of Credit Reduction . . . . . ... .137
Effects of Price Differentiation . . . . . 138
Effects of Increased Credit . . . . . . . 139
The Choices of Alternative Crops . . . . .. 140
Comparative Costs of Coffee Removal . . . . .. .43
Potential Effects of Technological Advances
in Coffee Production . . . . . . . .. 144
The Qualifications of an Acceptable Alternative . .. 145
Diversification Versus Production Control . . 146
GLOSSARY . . . . . . . . . . . . 148
APPENDIX . . . . . . . . . . . 149
LITERATURE CITED . . . . . . . . . . . 196
BIOCRAPHI CAL SKETCH . . . . . . . .... . 202
LIST OF TABLES
Partial matrix of coffee selling activities . 37
Comparisons between reported and optimal
incomes and coffee output . . . .. . 71
The effects of improving coffee production
technology on coffee production and
income . . . . . . . . . . 74
Comparisons of optimal incomes and coffee
outputs with traditional versus new enter-
prises on farms in Alajuela and Acosta . . 76
Optimal coffee production with specified
price declines per fanega of coffee . . . 77
Optimal manzanas in coffee with specified price
declines per fanega of coffee . . . . 79
Optimal net farm income with specified price
declines per fanega of coffee . . . .. 80
Alternative crops increased first by declines
in coffee prices . . . . . . . 82
Coffee production with annual payments for
coffee removal . . . . .. . 83
Net farm income with annual payments for
coffee removal . . . . . . . .. 85
The alternative crops increased first by
annual payments for coffee removal . . .. 88
Comparisons of income and coffee production
for differentiated prices versus single
prices . . . . . . . . . . 89
Optimal coffee production with various
levels of reduced credit . . . . . . 91
Optimal net income with various levels of
reduced credit . . . . . . . 92
15. Optimal manzanas in coffee with various levels
of reduced credit . . . . . . . 94
16. Marginal returns to credit with various levels
of reduced credit . . . . . . ... .95
17. Optimal family income with and without off-
farm employment opportunities . . . ... .96
18. Optimal farm income and the reduction of
permanent labor . . . . . . ... .97
19. Effect of additional credit on income and
coffee production . . . . . . ... .99
20. Optimal farm outputs of various crops on
Alajuela farms given base capital and
credit constraints and extra credit
equaling (400 and (2,000 per manzana .... 100
21. Optimal output of various crops on Acosta
farms with high fruit prices given base capital
and credit and extra credit equaling (400 and
12,000 per manzana . . . . . . . 102
22. The effect of a blackberry subsidy on optimal
coffee output . . . . . . . ... 1.04
23. The effect of a strawberry subsidy on optimal
coffee output . . . . . . . . 104
24. Relationships between long-term interest
rates and manzanas planted to limes . . .. 105
25. Lime prices and lime production . . . ... .106
26. Prices of oranges necessary to initiate
orange production on farms in Alajuela
and Acosta . . . . . . . ... .109
27. Alternative methods of reducing coffee
output on farm 1 . . . . . . ... .111
28. Comparative costs of various methods of
coffee output reduction on farm 2 . . .. 112
29. Alternative methods of reducing coffee
output on farm 3 .... . . . . . . 113
30. Alternative methods of reducing coffee
output on farm 4 . . . . . . .'. 114
Comparative costs of various methods of
reducing coffee output on farm 5 .
Comparative costs of various methods of
reducing coffee output on farm 6 .
Comparative costs of various methods of
reducing coffee output on farm 7 .
Comparative costs of various methods of
reducing coffee output on farm 8 .
The comparative costs of various methods
of reducing coffee output on farm 9 .
The comparative costs of various methods
of reducing coffee output on farm 10
The comparative costs of various methods
of reducing coffee output on farm 11
The comparative costs of various methods
of reducing coffee output on farm 12
The comparative costs of various methods
of reducing coffee output on farm 13,
given high fruit prices . . . .
The comparative costs of various methods
of reducing coffee output on farm 14,
given high fruit prices . . . .
The comparative costs of various methods
of reducing coffee output on farm 15,
given high fruit prices . . . .
The comparative costs of various methods
of reducing coffee output on farm 16,
given high fruit prices . . .
The comparative costs of various methods
of reducing coffee output on farm 13,
given low fruit prices . . . .
The comparative costs of various methods
of reducing coffee output on farm 14,
given low fruit prices . . . .
The comparative costs of various methods
of reducing coffee output on farm 15,
given low fruit prices . . . .
. . . 116
. . 118
. . . 120
. . . 121
. . . 122
. . . 123
. . . 125
. . 126
. . . 127
. . . 128
. . . 129
146. The comparative costs of various methods
of reducing coffee output on farm 16,
given low fruit prices . . . . . . 132
An Economic Background
Costa Rica is a small country in Central America with a popula-
tion of 1.6 million people living on an area of 19,700 square miles
(46, p. 11). The economy is dominated by agriculture which employed
over 56 percent of the active work force in 1963 (45, p. 13). In
addition, agricultural products accounted for approximately 30 per-
cent of the gross national product in 1966 and about 80 percent of
the value of Costa Rica's exports (17, P. 3). Alleger described
Costa Rica as a nation of small farmers (3, p. 33). This character-
ization does not mean that land is evenly distributed among the pop-
ulation, since farms of over 100 manzanas represent only 10.5 percent
of all Farms and cover 70 percent of the total farmland area (23,
p. 46). Nevertheless, the land distribution figures must be inter-
preted with a realization that many large farms are in remote areas
and that, except for sugarcane and bananas, the intensively grown
crops are dominated by small and middle-sized farms.
In recent years, Costa Rica enjoyed the highest per capital in-
come in Central America. The estimated grc-s national product per
person in 1965 was $415 compared to an average of $303 for all of
Central America (55, p. 3).
In recent years Costa Rica has had a balance of payments problem.
Foreign debts have had to be continually r-mclnocot.ated. From 1950 to
1967 the value of exports of coffee, bananas, cocao, cotton, beef,
and sugar increased from 592.1 to 702.8 million colones (56, p. 34).
While the value of exports has increased despite unfavorable price
trends, the value of imports has increased at a much faster rate.
The value of exports exceeded the value of imports by 24.8 percent
in 1967 and by 12.2 percent in 1968 (24).
Production of the basic food crops has exhibited erratic growth
since 1950. Corn imports exceeded exports in 1951, 1954, 1956, 1959,
1964 and 1966. Per capital consumption of corn in 1966 was below the
1950 level (56, p. 42). Rice production has almost tripled since
1950, with a stronger upward trend occurring after 1958. While the
change is largely due to an increased acreage, the average yields
have increased since 1955 (56, p. 5). Per capital consumption of rice
has increased. Bean production has also increased significantly,
doubling the 1950 output in 1967. However, net imports of beans
occurred in 1951, 1956, 1957, 1959, 1961, 1962 and 1964. Per capital
consumption has increased 25 percent since 1950 (56, p. 28).
The Costa Rican population increased from 1,028,175 in 1955 to
1,648,815 in 1960 (15, p. 2), an annual average increase of more than
4 percent. While population growth increases the demand for food,
this demand is also increased by a rise in per capital incomes. If
food output is to keep pace with the demands of both population and
income growth, further changes in inputs and technology will be re-
quired (13). The allocation of scarce Foreign exchange to food
imports limits the ability of a developing nation to purchase capital
goods needed for economic growth (48, p. 5).
Another problem facing Costa Rican planners is public finances.
The costs of governmental services are increasing. Education is an
example (55, P. 33). Schools are being expanded but more and more
children crowd them. With over 35 percent of the population under
ten years old, education is costly (19, p. 47). Rapid population
growth slows occupational changes (1).
Thus while the Costa Rican economy has given the nation the
highest per capital income in Central America and supports the most
advanced social and educational programs in the region, it faces
demanding requirements if it is to continue to grow.
The Importance of Coffee to Costa Rica
Coffee is Costa Rica's leading agricultural commodity, the chief
export crop, and a major user of agricultural credit and labor.
Coffee has historically accounted for 17 to 26 percent of the value
of agricultural output (54, p. 37). Despite a long history of pro-
duction, coffee output has increased markedly in recent years. From
1955 to 1963 coffee output was doubled. Yields increased 27 percent
while acreage increased over 58 percent (53, p. 10). Considerable
effort was put into research and extension work which emphasized the
use of modern inputs in coffee production. Higher coffee prices in
the 1950's encouraged the expansion of coffee onto new lands. The
modernization of coffee production is indicated by the fact that in
1963, 27 percent of the farms reported fertilizer use on 53 percent
of the land planted to coffee (23, p. 176). Fertilizer responses
gave 18 to 233 percent increases in coffee yields in trials running
from 1952 to 1957 (69, p. 60). Since 1963. production increases in-
dicate an even larger use of off-farm ir--ts. In 1967, coffee output
was more than three times the average output in the 1948 to 1952
period (29, p. 63). Thus coffee production has led in a change from
traditional toward modern farming methods.
Coffee is the major source of foreign exchange for Costa Rica.
It accounted for 38.9 percent of total value of exports in 1966, 41.2
percent in 1965 and 42.2 percent in 1964 (15, p. 13). Thirty percent
of the economically active population is associated with the coffee
industry (54, p. 37). This underestimates coffee's importance as a
source of wage earnings since many children harvest coffee and are
not considered part of the economically active population. The
harvest season usually coincides with school vacations. Also, earn-
ings from coffee picking greatly exceed other farm labor earnings.
Another benefit that comes from coffee production is soil con-
servation. A well-cared-for coffee planting protects the soil from
driving rain and contour ridges that slow runoff are permanent and
reinforced with woody root systems. Much of the land utilized by
coffee is unsuitable for annual cropping unless very elaborate ter-
racing is used (67).
Lastly, coffee is a source of tax revenue. Taxes of $0.45 per
quintal of exported coffee, 2.00 per quintal of coffee consumed
internally, and 0.20 per fanega of coffee fruit processed at the
beneficios support the operations of the Oficina del Cafe. In addi-
tion, an advalorum tax contributes to the national treasury. This
advalorun tax is graduated in the following manner: 10 percent if
the average price exceeds $42.50 per quintal, 72 percent if the
average price falls between $40.00 and $42.50 per qu;ntal, 5 percent
if the average price falls between $37.50 and $40.00 per quintal, and
2-- percent iF Lhe average price falls between $35.00 and $37.50 per
quintal. No advalorum tax is paid if the price falls below $35.00
per quintal (54, p. 56).
Declines in coffee prices are burdensome to governments which
depend upon coffee earnings for foreign exchange and tax revenue
(30, p. 15). This is particularly true in Costa Rica where a 14
percent decline in price resulted in a 57 percent loss in tax revenue
per quintal and a 3 percent decline in price resulted in a 53 percent
loss in tax revenue per quintal in the marketing years from 1965-66
to 1966-67 to 1967-68. The price decline from 1966 to 1968 brought
about a drop in tax revenue estimated at 24 million colones (55,
The Marketing Situation for Coffee
Historically, coffee production has gone through highly cyclical
price periods. Prices fluctuated in a cyclical pattern accentuated
by periodic unplanned changes in supply caused by unfavorable weather
conditions in Brazil (32, p. 454). When drought and frost cut
Brazil's output the price would rise. This high price encouraged
renovation of old coffee orchards and the planting of new ones.
Later, when recovery of damaged groves occurred, output surpassed
the earlier level. Prices then were pushed to new lows causing
abandonment or neglect of coffee farms until unfavorable weather
again stimulated high prices. Over the years, there have been short
periods of shortage and high prices followed by long periods of sur-
plus and low prices (59, p. 8). These drastic price fluctuations
stimulated coffee producers to seek remedial schemes. The first of
these was a Brazilian law blocking new plantings in 1902. Politically
unpopular, this law was repealed and followed by valorization schemes
with which large quantities of coffee were purchased and held off
the market (32, p. 456). Brazil borrowed from British banks to make
coffee purchases in 1906, 1917, 1921 and 1927. Larger and larger
crops were encouraged which became more and more difficult to store.
Brazil burned over 78 million bags of surplus coffee from 1931 to
1944 (73, p. 423).
Coffee production in other Latin American countries expanded in
response to Brazil's price supporting activities. Growth elsewhere
caused Brazil to market a smaller percentage of the world coffee
(73, p. 425).
International agreement to control coffee output was first
attempted in Bogata in 1936. Later, the Inter-American Coffee
Agreement was signed by 14 producing countries and the United States.
Selling quotas were established and the agreement lasted from 1941
to 1948 (73, p. 423). In 1957, coffee producing nations agreed upon
a voluntary system of export regulation. This agreement failed to
check the downward price movement (59, p. 13). Finally, the Inter-
national Coffee Agreement received the support of 46 producing and
consuming countries in 1963. By chance, a severe frost cut the
Brazilian output that year and, as a result, quotas were increased
to give help to the coffee consumers as prices rose (59, p. 14). The
demand For coffee is believed to be inelastic with respect to price.
If this is true, a frec-:narket solution would result in lower gross
income to producers than would occur if output were limited by some
kind of a cartel arrangement. Economists generally view trade agree-
ments with a skeptic eye. Both cxperiece and theory show that, with
price maintenance at or above a free-trade level, pressures and temp-
tations arise to break or by-pass the agreement since high prices
tend to encourage more production (37, p. 108).
Nevertheless, the International Coffee Agreement has reduced
price fluctuations and is considered to be effective enough to justify
extension for five more years from 1968 (8, p. 188). However, the
reduction in world output from 1965-66 to 1966-67 occurred mostly in
Brazil where weather has historically caused wide production varia-
tion. The 1963 Agreement was concerned with regulating sales and
made no attempt to regulate production.
In 1968, a new article was written into the International Coffee
Agreement setting up the Diversification Fund (34). This change
provides for compulsory payments (U. S. $0.60 per bag in excess of
100,000 bags) into the Fund. This acts as a tax on coffee. Extra
incentive is given to coffee producing nations to devise crop diver-
sification projects. Eighty percent of the compulsory payments can
be used within the producing country on approved projects. If not
used domestically, the unused payment must be paid to the Fund in
freely convertible currency. Thus each country will be motivated to
develop local diversification projects by the desire to conserve
scarce foreign exchange.
Producing countries have used different programs to check coffee
output. As mentioned earlier, Brazil first attempted to control out-
put in 1902 with a law prohibiting new plantings. More recently pay-
ments have been made for coffee removal (bO). Participation in the
program has been voluntary,with 648 million coffee trees pulled out
from June, 1962 to December, 1965. Plans had called for the removal
of two billion trees and part of the lack of response was blamed on
inflation which lowered the value of the fixed payments from $0.04
to $0.01 per tree. The effect on coffee output was small. The
program included no restriction on planting. In Colombia, Inter-
national Development Bank (IDB) loans have been used to further
credit, infrastructure, education, commercialization and industrializa-
tion of alternative products. In Mexico, rubber, citrus and avocado
plantings are being promoted in a program directed at the small
coffee producer (27). In El Salvador, sugar, corn and rice produc-
tion has been expanded. Modern corn and rice production gave returns
reported to range between $3.00 and $7.50 per $1.00 spent on new in-
puts. Nevertheless, the expansion of cereals was accompanied by a
reduction in cotton and beans, rather than coffee (74). In Guatemala,
pilot tests of tea, citrus, dairy and oil palm have been initiated.
Over $1.9 million in foreign money has been invested in the program.
Total costs surpassed $5.8 million invested on 282 farms covering
3,900 hectares (6). Guatemala also is developing a rubber industry.
In 1965, rubber was planted on 26,000 acres. Projections estimate
gross returns of $24 million from 80,000 acres (75).
In Costa Rica, credit for new coffee plantings has been restrict-
ed. In addition, the Universidad de Costa Rica, the U. S. AID Mission,
the University of Florida, and the Centro para la Promocion de
Exportaciones and Inversiones cooperated in a series of observation
trials in six locations scattered throughout the western part of the
coffee growing region. The municipality of Turrialba has initiated a
regional diversification program with the Instituto Interamericano de
Ciencas Agricolas. Financial support from the Oficina del Cafe and
the Agency for International Development (AID), plus technical
assistance from the Instituto Interamericano de Ciencas Agricolas,
the Peace Corps, and the Ministerio de Agricultura,makes this a truly
cooperative venture. The emphasis is primarily on research with a
rapid follow-up of pilot commercial plantings. The project began
with basic studies of fast growing trees, Tilapia fish ponds, and
macadamia nut production. The stated main purpose of the project
was to institutionalize an attitude of dynamic change (9). While it
is still too early to evaluate results, leaders in another municipal-
ity have talked of imitating Turrialba with a diversification project
of their own.
In summary, a review of the literature on Latin American coffee
diversification showed more discussion and hypothesizing than prac-
The Pros and Cons of Diversification
There has been much discussion of diversification in recent
years as a method to foster economic development. In a comprehensive
study, Dalrymple (22) has compared monoculture and diversification.
The advantages of rroncculture include the following:
1. In some cases the monoculture crop has a clear comparative
advantage both at domestic and international levels. The financial
gap between the monoculture crop and next best alternative has been
found to be too wide to permit rational change,
2. It may be easier to raise yields to give higher returns from
an established crop than to press for :ore complex cropping systems.
The short--run returns to increased spacializaticn with economies of
scale may be quite high. The new knowledge and skills required to
improve production of an existing crop may be easier to learn than
the technical requirements of a totally new crop.
3. Monoculture is generally focused on export crops which provide
a developing country with needed foreign exchange and an easily
administered tax system.
4. Certain crops have more prestige and social status than
The disadvantages of monoculture bring out accompanying economic
1. Producers under monoculture are subject to high risk induced
by technical change and insect and disease problems. This is often
labeled "putting a lot of eggs in one basket."
2. Because many of the mcnoculture crops are perennials, a
production lag may follow a decision to increase output. By the
time the crop comes into production considerable investment has
already been made. Excessive reaction to favorable price situations
may occur when this lag follows a major change in resource allocation.
Readjustment of supply to face a lowered price will be sluggish even
in the face of losses since marginal costs may be easily covered.
3. The low price elasticity of demand for coffee results in
sharp, short-term price fluctuations caused by weather and biological
factors. The resulting high prices in the short run may trigger ir-
reversible investments. These investments plus technological advances
can be expected to increase supply while demand is not likely to
increase faster thon population growth. As a result, prices are
exc:< c i:L to we;-i n over time.
4. Trade agreements limit sales to key markets and therefore
increase both price and gross returns for the commodity with an
inelastic demand. Unless each producing country takes actions to
correct the internal distortions of high incentives for the commodity
covered by trade agreements, an imbalance encouraging overproduction
of that commodity will result (53, p. 9).
The advantages and disadvantages of diversification are roughly
the inverse of those of monoculture. Diversification may be advanta-
geous if it more fully utilizes labor and reduces economic risk (48,
p. 24). A number of different sources of income gives protection
against severe loss caused by insects, disease or bad weather condi-
tions which may affect one particular crop but not others. Labor
requirements may be spaced in such a way that one crop uses labor
when another has a slack work period. Shifting from an export crop
to food crops can lead to improved nutritional levels, especially if
more fruits and vegetables are introduced into the diet. Also,
import substitution may save scarce foreign exchange (22, p. 27).
On the other hand diversification Faces certain limitations.
Research has been focused on a few major export crops (22, p. 39).
Without much experience or local scer.tific investigation to support
a new enterprise, the innovating producer faces -igher uncertainties
with respect to the crop response t-o .innvorEbl e factor and condi-
tions (41, p. 1). The market for th.e ~lternat i ve crop may not be
sufficient to absorb expand nr product n at profita :bl prices (22,
p. 41). Even i a potential demand exists the marketing facilities
may not be ade-:uate to rnovi the new output to consumers. The econ-
omies of scale may resul t in poor efficiency as a greater number of
crops are produced and volume of some crops is reduced. This is
particularly a hazard for a new crop introduced without sufficient
volume to utilize efficient processing machinery.
In cases where the established crop is a perennial, a high per-
centage of the costs are fixed. Therefore, replacement by an alter-
native requires that total costs, since all costs are variable, be
considered in comparison with the variable costs in the case of the
established crop. Also, costs associated with removing the old crop
must be added into the cost of establishing the new enterprise.
Another problem arises if a new crop uses either more or less
labor than the established crop (22, p. 38). If the labor require-
ment is much higher, labor scarcity may prevent adequate handling of
the new crop. If much less labor is used, unemployment has social
ramifications that may be prejudicial to the establishment of a new
The quantity of research, extension work and information services
will have to be expanded if changes require more complex agricultural
systems (48, p. 24). Diversification projects are likely to fail on
farms where administration and management inadequacies greatly limit
the returns to coffee because the new enterprises are likely to be
even more difficult to manage (27).
A!2roachjz and Attitudes toward Diversification
Crop diversification is defined as a movement away from mono-
culture with the growing of new or additional crops (22, p. i). iore
detailed descriptions may be conflicting ad the evaluation of diver-
sification as a policy measure depenJs greatly upon just what meaning
is used. A most restrictive definition, and one often thought of, is
the transFer of land from r:cnoculture to alternative uses. However,
other resources besides land may be shifted from one use to another.
Thus, diversification occurs if operating capital or labor is put to
alternative use. For instance if labor or fertilizer is applied to
strawberries rather than to coffee, diversification occurs. Diver-
sification may even occur without a shift in resources or a reduction
in primary crop output. This is possible if unused resources are
associated with monocultural production. Therefore, if a farm were
to begin to grow a crop of dasheen on swampy ground formerly unused,
using surplus family labor and operating capital, this would be an
example of diversification. Thus for the purposes of this study,
diversification is defined as a positive action which reduces the
relative importance of the primary crop.
Crop diversification may occur at either the farm level or the
national level. A recent advisory group proposed that Costa Rica
should concentrate diversification efforts on areas unsuited for
coffee, where mechanization was feasible (68, p. 3). The program in
this case would be to expand output using resources not now used to
Crop diversification may either expand domestically consumed
crops or promote new export crops. Although import substitution is
reccgnied as beneficial, planners seek to increase earnings of foreign
exchange with new exports. The export dei.rr,nd for a product facing a
small country is often highly price elastic (57, p. 1). Thus, gener-
ally the new export crop has an advantage over domestic crops in that
price will remaiir more stable as outout is expanded. The small size
of the domestic market is coupled with a shortage of capital funds and
modern know-how to limit diversified economic potential (16, p. 33).
Furthermore, overproduction of basic food crops may result in govern-
ment losses if high support prices are coupled with export subsidies
Others suggest that domestic crops offer better diversification
possibilities because benefits of technological change are passed on
to the consuming countries (66, p. 432). Since a particular good from
one country has perfect or close substitutes produced in other coun-
tries, the demand curve facing each country is elastic. However, the
common agricultural export crops are inelastic when the entire world
market is considered since they do not have close substitutes, do not
have many uses, and do not take a large share of the consumer's in-
come (44, p. 41). After a technological change is widely adopted,
the result of increased output is often lower gross revenue. In
this situation, the consumer benefits from lower prices and a higher
The innovator may initially supply the domestic market when low
early yields are compensated by high prices. Later, costs may be
reduced to permit export or industrial use at much lower prices. In
addition, the scale of operations needed for new export ventures
exceeds the capacity of existing producing units and marketing facil-
ities (17, p. 12). The development of a new export crop requires
efficiency if it is to nmet the estabished comupet Ltion.
Another area of debate rceners around the question of who should
diversify. Two nearly opposite views have developed: an efficiency
criterion of marginality proposes to re;ove the low profit producers.
In opposition, a criterion of welfare seeks the removal of those least
hurt by shifting to alteri.n tive production.
The removal of marginal producers is deemed desirable by both
the Oficina del Cafe in Costa Rica and the Asociacion National del
Cafe in Guatemala (54, p. 32; 27). Fernandez (27) defined marginal
farms as those where costs exceed returns and also small farms where
returns are low. One may note that this approach avoids antagonizing
the politically influential in Guatemala as marginal farms are
"liberated" for other uses. In Costa Rica, selective credit restric-
tions were used to limit the expansion of coffee, particularly into
areas producing low quality-low yield crops. The Oficina del Cafe
favors this policy because it helps to maintain a higher average of
quality as well as to restrain production.
Newman (53, p. 14) has criticized the marginal producer definition
for being concerned with absolute rather than comparative advantages.
Small farms are unsuited for such alternatives as dairying and fruit
production because they cannot take advantage of economies of scale
open to large operators. Small farms also lack reserve capital or
credit availability to enable them to invest in the more productive
alternatives. They are less able to withstand possible loss of a new
venture. The opportunity cost associated with removing coffee includes
interest charges on foregone income, which may be limiting for farms
near the subsistence level of income when perennial alternatives are
Welfare considerations cannot be quantified for interpersonal
comparisons. However, value judgments need not be made if the level
of alternative output is used to evaluate different policies of diver-
sification. The problem then becomes the calculation of the net costs
of removal of a quantity of coffee from production on different farms.
If the means of production control is alternative use of land,
this may be stated algebraically as follows:
C = P C V + C + C
r c c a a s
C = net cost of removing one fanega of coffee,
P = price of coffee,
C = cost of producing one fanega of coffee
V = value of alternative product produced with resources made
available as coffee is reduced by one fanega,
Ca = cost associated with producing V ,
Cs = cost of removing trees producing one fanega of coffee.
Thus, the comparative advantage in coffee production may differ from
the absolute advantage where C = P C If coffee is purchased and
destroyed, the cost is higher as C = P .
A more general formula can be stated for calculating the unit
cost of coffee removal.
C = II I
C = net cost of coffee removal
I, = net income before change
I = net income after change
Ql = coffee production before change
Q := coffee production after change
IdJealy, diversification would seek to reduce coffee without re--
ducing income However, given a price sitLation which sti rulates
excess use of resources in coffee production, a more practical policy
would attempt to minimize the costs of controlling output.
The Problem and Obiectives
Although coffee is Costa Rica's most important commodity, con-
tinued dependence on that one crop is considered detrimental to
prospects for economic growth. This harsh statement is supported by
a political-economic situation in which the sales of coffee to the
high consumption markets is now limited by international agreement.
Thus if substantial growth is to be achieved, it must occur in some
other segment of the national economy.
Technological changes are occurring in coffee production. These
changes enable coffee to be produced at lower unit costs as modern
inputs are added to traditional land and labor. The result is to
increase yields. If resources are not shifted away from coffee, this
increases production. The problem then arises as to which resources
should be shifted to what alternative uses. Also, what policy mea-
sures will facilitate changes which are both efficacious and equitable?
One may also ask, are Costa Rica's farmers functioning as profit
maximizers? Theodore Schultz has claimed that in traditional agri-
culture farmers are not only profit maximizers but that they also are
quite efficient profit maximizers (62. r. 44). llo,ever, most of the
farms studied do not truly fit the definition of traditional agricul-
ture used by Schultz (63, p. 30). Cultural techniques for coffee
have not remained unchanged for generations and further changes are
occurring (32, p. 432). Costa Rica's emphasis on primary education
already has accomplished much to-ward the investment in human resources
necessary to change traditional attitudes (63, p. 201). Schultz has
designated the human agent as the key variable in explaining differ-
ences in agricultural productivity (63, p. 17). The existence of
experimental farms, scattered research plots, extension agents and
agricultural schools also takes Costa Rica out of the category of
Costa Rica admittedly does have many traditional farmers. How-
ever, the threat of overproduction of coffee does not come from that
direction. Recent yield increases indicate non-traditional behavior,
while increased coffee acreage has been relatively unimportant (15,
Schultz describes a transitional classification of agriculture
between the traditional and the modern (63, p. 107). Vast.disequilib-
rium is said to exist with differences in marginal productivity and
overuse or underuse of factors. The expanding use of fertilizers and
pest control chemicals (15, p. 7) indicates that transitional changes
toward modernization are occurring in Costa Rica.
Nevertheless evaluating the Farm case studies with respect to
expectations given by Schultz's theories may be interesting. Tradi-
tional farm situations could be expected to give marginal value
products near or at current prices of resources. On the other hand,
transitional farm situations may put extremely high values on certain
resources. Linear programming solutions may be compared with actual
farm operations to judge the efficiency of farm decision makers.
This may be of particular interest in explaining the continuance of
traditional meLhods with some crops while changes occur with others.
Specifically the objectives of this study were to evaluate farm
income opportunities from producing coffee and from alternatives and
to determine the effects that selected programs would have on coffee
production, resource use and incomes. At the same time, comparisons
can be made of current resource use and optimal resource use to test
the hypothesis that the farmers are income maximizers and that tradi-
tional economic behavior has economic motivations. In addition, the
comparisons may be used to determine what changes or adjustments in
farm operations would be profitable.
SCOPE AND METHOD OF STUDY
Selection of Areas and Farms
Coffee is grown in Costa Rica under a wide range of climatic
and ecological conditions. Three geographic areas were included in
this analysis. The data were taken from a study carried cut under a
University of Florida AID Contract in Costa Rica, No. la-261.
Budgeted comparisons between coffee enterprises and leading alter-
natives in 12 areas are given in An Economic Analysis of Coffee Pro-
ducing Areas, Costa Rica (11).
Palmares-San Ramon, Alajuela and Acosta were the areas selected
For this intensive study. This selection was based on the following
1. Coffee and alternative crops should be found growing under
similar ecological conditions.
2. Enough coffee should be produced to make changes important
to national totals.
3. Different areas should represent a wide range of coffee
productivity and alternative choices.
Some consideration was given to including the Turrialba area in
this study. HnwevPer, the diversification project in that area was
just beginning to generate completely new data when this work was
Within the three areas s'lectcd, the local extension agents
selected farms they regarded as typical of size and class categories
common in each particular region. Data were collected by the agents
listing the resources available and the resources used on each farm.
A monthly breakdown of labor was supplied for each enterprise. In
addition, budgets for several crops grown in the areas were made
available from the Banco Nacional de Costa Rica and the Banco de
Credito Agricola de Cartago. The list of enterprises was further
supplemented by crop cost study reports of the Ministerio de Agri-
It is admitted that the farm case study approach cannot be sta-
tistically supported. However, by drawing upon the prior knowledge
of the local agricultural scientists, costs could be held to a
fraction of the costs of working with a large random sample. Because
it is likely to be the better farmers who contact and work with the
extension agents, one may expect that the "typical farms" of the
extension agents may be above average. Nevertheless, this direction
of bias need not be undesirable since it is this group of farmers
who are most likely to first respond to economic incentives with
either increased coffee output or diversification to alternative
Furthermore, the farms selected are in no way expected to be
averaged to give a quantitative measure of policy response for more
than each farm itself. The generalizations possible must be limited
to direction and nature of change which may in turn iead to specula-
tions concerning the response of the total coffee industry.
Although quantitative analysis is made at the farm level, extrap-
olation to larger areas must be with descriptive or qualitative anal-
ysis. In this manner, associations can be made between policy re-
sponses and various resource situations or enterprise possibilities.
The uniformity, irregularity or lack of response can be noted for
various policy alternatives. Thus the sample may indicate the kind
of farm likely to support or oppose a particular political measure.
Description of Farms Studied
The farms selected for analysis are all located on the Pacific
side of the Continental Divide. Farms were selected to represent the
most common size-type categories found in the three areas.
This area is made up of the intensive coffee-growing districts
of the cantons of the same names. The Canton of Palmares, except for
the districts of Candelaria and Esquipulas, which are not well suited
for coffee, reported coffee on 94 percent of the farms and 37 percent
of the land in 1963 (25).
The soils are fluvio-lacustrine groups containing diatomite
(iO, p. 3). Internal drainage may be a problem on level areas. A
distinct dry season extends from December into April and Good Hard
Bean (2) type of coffee is grown at elevations beLween 900 and 1,200
meters. Practically no rain falls from December to February (64).
In the districts of San Isidro and San Ramor of the canton of San
Ramon the Good Hard Bean type of coffee is produced at elevations of
1,000 to 1,200 meters. In these districts, coffee utilizes 12 percent
of the land and it found on 83 percent of the farms. Tobacco, corn
and beans are the major crops after coffee in the combined area.
According to .:ih. 1563) census, coffee was gro.n on 2,783 manzanas,
corn on 417 manzanas, tobacco on 314 manzanas, beans on 226 manzanas,
and sugarcane on 106 manzanas. Coffee output of the districts within
the area made up over 3.6 percent of the national total and yields
were roughly 37 percent above the national average in 1963 (25).
Eight farms were used to cover a range of small, medium, and large
units growing coffee alone and in combinations of coffee and tobacco.
Farm 1 is a small coffee-tobacco unit in San Ramon. It is situ-
ated at an elevation of 1,080 meters on land described as moderately
rough. The farm owner applied modern technology in the forms of
fertilizer, insecticide, herbicides, and foliage fertilizer. The
farm contains five manzanas with three manzanas planted to coffee and
two manzaras planted to tobacco and corn. The labor force was made
up of two hired male employees plus the family labor of a man, a
boy, and two girls. Fixed expenses, including permanent labor, taxes,
depreciation and maintenance, and interest on the land investment
totaled (6,186. Short-term operating capital including credit was
estimated at (4,400.
Net farm income was calculated as gross returns less reported
annual expenses and estimated fixed costs including rent, taxes and
hired permanent labor. Net farm income was estimated at (22,500.
The farm was chosen as an example of mixed croppingr usinc modern
technology and hand iabor.
Farm 2 in San Ramon is a small farm specializing in coffee.
Coffee was grown on all its seven :manzanas, Labor was supplied by
five male and two female employees plus the family labor of two men
and two boys. The fixed expenses, including permarent labor, totaled
(16,690. Operating capital and credit available for variable costs
totaled 14,000. Net farm income was estimated at <15,250. This
farm was chosen as an example of a specialized modern coffee producer.
Farm 3 in San Ramon is a larger farm producing coffee, corn and
tobacco. The farm covers fifty-six manzanas of which fifty manzanas
were planted to coffee. The farm has a flue-cured tobacco contract
for three manzanas. There were nine permanent employees and an
administrator. The fixed expenses totaled t60,100 and combined
operating capital and credit totaled 071,100. Net farm income was
estimated at 134,130. This farm was chosen as an example of a larger
farm in the process of undergoing technological change.
Farm 4 is located in Palmares on nearly level land at an eleva-
tion of 980 meters. It is a middle-sized farm growing both coffee
and tobacco. The farm reported fifteen manzanas planted to coffee
and five manzanas in flue-cured tobacco. The farm hired two permanent
employees and family labor consisted of three men, three girls, and
two boys (part-time). Fixed expenses were calculated at 1!4,840.
Annual operating expenses were (16,900. The net farm income was
estimated at t14,700. This unit was selected as an example of a farm
specializing in tobacco with low yielding coffee grovn on other land.
Farm 5 is a medium-sized coffee farm in Palmares. The land is
nearly level at an elevation of 1,020 meters. All ten manzanas were
planted to coffee. The farm work was done by four men; two were hired
and two were membe-rs of the owning family. The fixed costs were
estimated at 010,732 and the sum of the operating capital and short-
term credit amounted to ,6,400. The net farm income was estimated
at (15,520. This farm was representative of a level of technology
commonly used o:i iQ.ms slowly adopting changes.
Farm 6, also located in Palmares, is found on gently sloping
land at 1,030 meters' elevation. It is a small farm with four manzanas
of land of which 3.75 manzanas were planted to coffee. The farm hired
four permanent employees. It had no administrator nor family workers.
The (12,480 estimate of fixed expenses included very high permanent
labor costs. There were (4,000 available for annual expenses, in-
cluding credit and operating capital. The farm operated at a loss
estimated at 6,760. This farm was chosen as an example of a small
property owned by an absentee owner.
Farm 7 is located on moderately rugged land at 1,025 meters'
elevation in Palmares. It is a small farm of six manzanas with three
manzanas of coffee and one manzana of sun-cured tobacco reported.
Pasture was grown on two manzanas. The farm employed one man and
extra work was supplied by two members of the owning family in times
of emergency. Fixed expenses were calculated at (6,000 and (4,300 were
reportedly available for variable expenditures. The estimated net
returns were (3,345. This unit was selected to represent a small
farm with mixed production.
Farm 8 is also located on moderately rugged land in Palmares.
It covers two manzanas of land of which one manzana was planted to
coffee and one-half manzana is planted to sun-cured tobacco. One
permanent employee was hired and two mer. and t:.o boys (part.-time) of
the owning fami ly worked on the farm. Fixed e.ixpenses wiere estimated
at (2,882. Operating capital and short-term credit was limited to
(900. The coffee planted was not yet in production but anticipated
yields gave expected annual returns of (2,653. This Farm was selected
as an example of a very small farm using multiple cropping.
The Alajuela area is located around the town of the same name.
The land is made up of rolling hills with occasional areas of nearly
level topography. In general, the soil is rich, being influenced
by reoccurring ash fall, which results in andosols with high organic
matter content, although considerable variation is found in both top-
soils and subsoils (71, p. 26). The elevation ranges between 700
and 1,100 meters. The districts of Alajuela, San Jose and Desamparados
were used to represent the region in the 1963 census data. In 1963,
average coffee yields for these three Alajuela districts were 27 per-
cent above the national average and coffee occupied 1,980 manzanas.
In the same year, there were 82 manzanas of pineapples, 137 manzanas
of corn, 186 manzanas of beans, 17 manzanas of tomatoes, and 14
manzanas of cassava reported (25). Coffee was reported on 76 percent
of the farms and occupied 26 percent of the land area in 1963.
Although rains may occur throughout the year, a dry season ex-
tends from December into April (64). The coffee produced is the Hard
Bean type (2). The representative districts selected from the 1963
census produced 2.43 percent of the national output of coffee. How-
ever, the ecological conditions of these districts extend into parts
of adjacent districts.
Four farms were selected for study from this region. Farm 9 is
located near an elevation of 1,100 meters on gently rolling land in
Alajuela, It is a large family farii with 40 m nanzanas, all in coffee.
IPolitical subdivisions do not coincide well with ecological
areas; th;'lrefo"re, the census data must be interpreted cautiously.
Labor was supplied with ten permanent employees and the family labor
of two men and two boys (part-time). The fixed expenses were estimated
at 48,575 and short-run credit arid operating capital totaled 22,200.
The estimated net returns were 90,000. This unit was chosen to
represent large coffee farms with both good coffee and horticultural
Farm 10 is located on rolling land above 1,000 meters. All of
its 15 manzanas were planted to coffee. Labor was supplied by five
permanent employees and two men and a boy of the family. The fixed
expenses were estimated at 21,875. The operating capital and annual
credit totaled 15,750. Estimated net returns were 17,160 but young
plants raised expected future returns to 30,000. This farm was
selected to represent the medium-sized specialized farm with good
Farm 11 covers 10.5 manzanas of nearly level land at 700 meters
elevation in Alajuela. Coffee was planted on five manzanas. Pine-
apple was r-own on the remaining land. The farm utilized the labor
of one hired employee and one man, two boys, one woman, and one girl
of the family labor force. The fixed expenses were estimated at
9,670. A sum of 8,850 was available for annual operating expenses.
Farm income was estimated at 42,980. The horticultural crop was
sold into the high-priced fresh fruit market. If the farmer received
processing prices for the pineapple, farm income would fall to 9,330.
Pineapples would be discontinued since returns at processing prices
would fall to about one-half of the cost of production. This farm
was chosen to represent the middle-sized producer of coffee and
fresh market fruit.
Farm 12 lies on rolling land at 1,100 meters near Alajuela. The
farm contains 5.5 manzanas of land of which 2.5 manzanas are planted
to coffee and the remainder is planted to sugarcane. Five men made
up the farm's work force; two were hired and three were members of
the owning family. The fixed expenses were estimated at 7,905 and
credit and capital for annual expenses summed to 1,115. The annual
net return to the reported sugar and coffee enterprises was estimated
to be 6,122. This farm was chosen to represent small mixed-crop
This area is around San Ignacio de Acosta. This zone lies to
the south of the Central Valley on rugged, eroded latosols of an
intermountain valley. The coffee of San Ignacio is mostly grown at
elevations between 900 and 1,200 meters. In the 1963 census, 98 per-
cent of the farms reported growing coffee on 15 percent of the area.
In 1963, coffee was reported on 1,098 manzanas which greatly sur-
passed 317 manzanas of beans, 174 manzanas of corn, 65 manzanas of
sugarcane, and 24 manzanas of cassava (25). Most of the land was in
pasture or forest.
The dry season is less pronounced in Acosta than in Alajuela.
Only January averaged less than 50 mm of rainfall during the period
from 1961 to 1965 (64). The heavier and more uniform rainfall brings
about a slight reduction in quality from that which the elevation of
the area would suggest. Quality Falls into the Hard 3ean category
(2). Leaching and erosion have reduced The natural fertility of the
Traditional practices h\/e persisted in the Acosta area. The
district coffee yields were 64 percent of the national average. San
Ignacio produces only 0.67 percent of the Costa Rican coffee; however,
conditions are similar in other districts along and beyond the southern
rim of the Central Valley. The area was included in the study to
represent a poorer coffee-growing region which fits a definition of
"marginal land" based on low output per unit of land and labor (27).
Farm 13 is a large coffee producer in Acosta. The farm covers
101.5 manzanas of rugged land averaging 1,100 meters elevation.
Coffee occupied 90 manzanas and 1.5 manzanas were planted to oranges.
The remaining ten manzanas were used to produce a joint crop of corn
and beans. Labor was supplied by four male family members and
thirteen hired men and one hired woman. Fixed costs were estimated
at 47,496 and 51,500 were available for annual expenses. The net
farm income was estimated at 66,878. This farm was chosen to rep-
resent large farms with poor soils.
Farm 14 is a medium-sized unit in Acosta located on rouah land
at 900 meters elevation. The size is 16 manzanas of which 7 manzanas
were planted to coffee. There were 4 manzanas of oranges and 5
manzanas of corn and beans reported. The farm had 3 adult male family
workers. The fixed expenses are 01,184 and 5,945 were available for
annual operating expenses. The estimated net returns were l l,4L46
on this middle-sized farm with mixed cropping.
-Fa-m 15 is a small Acosta unit on nearly level ground at 1,100
meters. It had 2 manzanas planted to coffee with interplanted orange
trees supplementing farm income. The farm family supplied the labor
of one man and three women. The fixed costs were estimated at 424.
Credit and operating capital totaled 825 for short-run expenses.
The farm reported a net income of (1,003. However, the costs included
an abnormally high number of new plants which suggested that an
investment was being made. The net income estimate was therefore
adjusted to l1,403.
Farm 16 is a second small farm in Acosta. The farm contains
6.5 manzanas of which 1.5 are planted to coffee with scattered orange
trees. The labor was supplied by one man and one woman of the owning
family and also a hired man and a hired woman. The fixed expenses
totaled 2,988 and 675 were available for variable costs. The farm
income was estimated at 0l,339. This farm was considered typical of
the small, mixed-crop farms of the area.
The Linear Programming Model
It would be presumptuous to claim that a study using one year's
data and many imported or estimated production coefficients would be
sufficiently accurate to permit the calculation of optimal management
plans to the five decimal places provided by the University of Florida
computer. However, the lack of a high degree of accuracy of data
should in no way be a deterrent to the use of the computer to solve
problems of a practical nature.
Linear programming is now widely used as a farm management anal-
ysis techniqjqe. However, in this particular study the emphasis is
put upon policy rather than production analysis. The input and output
coefficients may or may not represent true possibilities for the farm
groups studied. However, they do represent the expected possibilities
open to th-ose farmers based upon limited experimental data, reported
experience. and the opinions of the extension agents who advise them.
Therefore, the model attempts to predict how it would pay the farm
decision makers to react to a series of policy measures given certain
expected cost-output relationships for coffee and various alterna-
tives. The results were expected to provide implications with respect
to actions that farmers would be likely to take in response to program
The linear programming model is a computational method used to
minimize or maximize a linear function given a series of linear in-
equalities as restraints (33, p. 7). Net returns or profits were
maximized for each farm studied. The matrix in each application of
the model was comprised of the following sub-matrices: A production
sub-matrix was made up of input-output coefficients for the monthly
land and labor requirements, capital requirement, rotation require-
ments and coffee land requirements for the various crop production
activities. Costs were included as negative entries into the profit
row and yields were included as negative entries in transfer rows.
A selling sub-matrix consisted of price coefficients in the profit
row and unitary entries in the product transfer rows. A transfer
sub-matrix was comprised of coefficients of columns representing
labor purchasing, coffee planting, coffee destruction, borrowing,
long-term credit and fixed cost transfer.
The matrix included input coefficients for the monthly land and
labor requirements of the various enterprises. Additional restraints
included operating capital, rotational limitations, and coffee
plantings. Production coefficients were entered in transfer rows
to be sold via selling enterprises. Mornthly labor purchasing was
limited to a 2 to 1 ratio of adult male labor which allowed two
temporary workers to be hired for each permanent employee a a labor
supervision restraint. Planting and destroying coffee trees were
entered with cost and investment coefficients. New investments were
not limited but were given a cost through a borrowing column. Coffee
harvesting was handled by using both contract and family labor harvest-
ing. A special family labor restriction was used during the harvest
months. Excess labor was allowed to be sold during the coffee harvest
but not at other times.
?arametric programming was used to estimate the effect on opti-
mization when changes in policy and in production levels occurred.
Coffee yields were programmed downward 50 percent. Coffee price was
reduced (100.00 and printouts were made at 5.00 intervals. A con-
tinued annual payment for coffee removal was considered. Operating
capital was reduced 70 percent. A special credit row allowed the
operating capital requirements of expandable alternative crops to be
reduced. Credit was increased. A two-price system allowed differ-
entiated coffee prices to reflect returns to coffee sold in tradi-
tional and in new markets. Prices and yields of certain alternatives
were moved upward and downward. All new crops ,were excluded by down-
ward price manipulation. Family labor and hired labor were permitted
to be shifted off the farm. Lastly, higher yielding coffee aiter-
natives were allowed to enter with additional investments.
The farms selected were analyzed using linear programming to
maximize profits for a one-year period. Annual crops were considered
using land, labor, and operating capital coefficients taker directly
from enterprise budgets.
Enterprises requiring new long-term investments were considered
using mnir ntenance costs and return estimates for an annual period
after commercial production would become stabilized. Establishment
costs of permanent crops were calculated with opportunity costs
included as part of the investment to be considered using interest
charges. Thus, loss of income during the establishment period was
estimated and added to material and labor costs making up the in-
vestment calculation. The cost of making this investment was computed
using a low interest rate of 6 percent which assumed special long-
term credit subsidies for permanent crops.
The coefficients for intermediate length enterprises were calcu-
lated by summing the budget entries for the years of duration of the
crop. Thus the unit used was a "planting unit" or that area planted
each year and assumed an averaging of resource use over time. This
admittedly limited the strategy of cropping plans to be considered,
but this simplification greatly reduced costs of analysis while
coinciding with the generally followed procedure of "evening out"
inputs over time.
Presently established long-term coffee enterprises were entered
into the matrix without calculating investment costs. Additional
plantings were allowed but were associated with long-term interest
charges on the investment.
Depreciation of the investment was not considered for the per-
manent crops. If the annual expected returns surpassed returns of
other alternatives after interest costs were paid on the investment,
then it was assumed that the investment was an addition to net worth
which offset the original expenditure.
To further clarify the model, the mechanics of the less conven-
tional manipulations are given in more detail. Operating capital was
taken as a short-term credit restraint. All annual costs except in-
terest, transportation, and contract harvesting were used to determine
the requirement for operating capital of each enterprise. Reported
expenditures were taken as the row constraint for each farm.
Rotational restraints for annual crops were handled in the
following manner: Total farm land was used as the constraint on the
right-hand side. Each manzana of permanent crop used one unit of the
rotational limitation. Then the crops requiring rotation received
a coefficient equal to the minimum number of years during which only
one crop would be permitted. A coefficient of one allowed an average
of no more than one crop of beans per year. A coefficient of five
allowed only one crop of cucurbits in five years. This restraining
row cannot be used to set up the sequence of a rotation but does
insure that the optimal solution does not include a degree of crop
specialization contrary to required pest control practices.
Coffee plantings were used in three different constraining rows.
To insure that new coffee land was charged the cost of planting
coffee, maximum coffee land was set equal to or less than reported
coffee land plus land newly planted to coffee. To account for tree
removal costs of shifting land from coffee to other uses, another
row set optimal coffee land equal to or greater than reported coffee
land plus the land from which coffee was removed. A third constrain-
ing row set coffee removal at no more than reported coffee land to
block irregular possibilities as payments were iiade for coffee
removal in parametric operations.
Coffee harvesting requirements were handled in the following
manner: Coffee selling enterprises were given a coffee harvesting
requirement. An option was given for ha i esti n, allowing either
contracted labor or family labor to be used. The first means of
harvest used a cost of i40.00 per fanega. The utilization of family
labor used monthly labor resources and family labor resources at a
level of four or five hours per fanega each month during the harvest
Parametric programming was used to anticipate the effect of
various policy manipulations and to test the stability of the optimal
plan in the face of certain price and yield changes.
The use of selling enterprises made the programming of price
changes straightforward. A change row in the matrix contained coef-
ficients of change. Programming cards determined the magnitude of
change and the frequency of printout. Thus, the solution of the
program was continually re-evaluated as additions or substractions
were made to coefficients in the objective function. Cj' = C. + X.
(N) where C.' is the new price, C. is the old price, X. is the change
J J J
row entry and N Is the parametric multiplier ranging from 0 to a
given maximum. Prime decline used a negative X. value.
Yield changes were programmed using "PARAROW" parametric addi-
tions or subtractions. The solution of the program was re-evaluated
as a changing multiple of change row coefficients was added to the
coefficients of a designated transfer row. P.' = P. X. (N) where
J J J
P.' is the changed yield, P. is the old yield, X. is the chance row
J J J
coefficient and N is the parametric multiple e. Proportional changes
of several enterprises producing the same product was allowed by
setting X.'s = P.'s.
The effect of special credit facilities to finance the production
of the non-t-rditional cash crops was handled by a parametric reduction
of the use of normal operating capital of the favored enterprises.
This assumes a selective policy of credit expansion for crop diver-
sification. The "PARAROW" operation was mechanically like the yield
change procedure. K.' = K. + X. (N) where K.' is the new credit co-
J J J J
efficient, K. is the old credit coefficient, X. is the exchange row
coefficient and N is the parametric multiplier. With (X.) equal to
K. and N equal to one, K.' equals zero and the credit needs of
favored crops are all supplied by the new unrestricted source of
The effects of a reduction of credit were analyzed by using
"PARARHS" procedures. The parametric programming of the credit
constraint, a right-hand side value, used an exchange column coeffi-
cient set equal to the credit constraint. The parametric multiplier
ranged from 0 to .70. This assumed 30 percent farm supplied operating
capital as the lower limit of practical credit reduction. K = K +
(-Xr) N where Kr' is the modified credit constraint, Kr is the old
credit constraint and -Xr is the coefficient of the exchange column.
In a similar manner, an increase in credit was programmed with avail-
able credit being increased up to c2,000 per manzana.
A programming procedure was devised to evaluate the effect of
differentiating farm prices between sales into the new and into the
tradition! market. Since the new market price was roughly 25 per-
cent below the traditional market price and the new market took
-oughly 25 percent of total output, the differentiated new market
price was set at 80 percent of the current farm price and the tradi-
tional market price set at 106 2/3 percent of the current farm price.
The coefficient of the new market coffee selling enterprise was not
changed by the paral etric additions. The traditional market selling
enterprise was given a functional coefficient of zero. The exchange
row included coefficients for both the traditional market selling
enterprise and the current coffee selling enterprise. The former was
the estimated price that coffee would receive were it sold only in
the traditional market. The latter was a negative price coefficient
devised to exclude the average or current coffee selling activity as
the sale via the higher priced selling enterprise was permitted.
Table 1. Partial matrix of coffee selling activities
Averaged New Traditional
S market market market
Profit 200 1.60 0
Coffee transfer row I 1 1
Limit to traditional market 0 0 1
Change row -200 0 213 1/3
By referring to Table 1, it may be seen that parametric changes
will first block selling in the averaged market and then permit
selling in the traditional market. Sales in the traditional coffee
market were limited to 75 percent of the reported output of each
farm. The single printout of the solution was called when the values
of the change row were added to the functional. The price coefficient
for the averaged market was reduced to zero. That of the new market
was unchanged and that of the traditional market was increased to
106 2/3 percent of the averaged price. In this manner -he marginal
return for additional coffee on each farm was set equal to the
corresponding marginal returns to the country and each farm would
receive a price based upon new market returns for production in
excess of the quota. In this manner, the benefits of higher prices
in the traditional markets could be passed on to the producers with-
out increasing incentives for overproduction.
The payment for the removal of coffee trees was programmed as
a continuing payment made after the removal of coffee trees. An
activity for coffee destruction was given an annual cost to force
the payoff for coffee removal to be made in five years. Subsidizing
coffee tree removal was programmed with a change row entry in the
coffee tree destruction column. Payments from 0 to 1,800 per manzana
were covered with printouts on 200 intervals. The calculation of
payments equal to 20 percent of coffee's gross returns per manzana
was made using interpolation where there was no change in resource
use between the 20 percent payment and one of the printed outputs.
When the straight-line interpolation could not be made, the problem
was re-run with output demanded where the payment was equal to 20
percent of the gross returns to coffee.
Algebraically, Cd' = Cd +Xd (N); where, Cd' is the return for
destroying one manzana of coffee, Cd is the cost of destroying one
manzana of coffee, Xd is the unitary exchange row entry and N is the
parametric multiplier which represents varying levels of subsidy
payment. In addition, the exchange row contained a large negative
entry in the coffee planting column to block ne;w (offee planitings if
payments were made for coffee removal.
The effect of outside employr'ent on family incomee was studied
using param'-tric changes of the objective function. In an original
matrix, a column represented the reduction in monthly labor supply as
a man left the farm for other employment. Additional entries rep-
resented the effect on the monthly hiring of temporary labor and
monthly family labor for coffee harvesting. Also, a constrained row
restricted movement to adult male family members. Since this use of
resources received no returns it would not come into the initial
optimal solution. Then returns were entered with parametric changes
to record the response of income and output as outside opportunities
increased to the level established by minimum legal wage laws. There-
fore, CW = X (N); where, C equals outside wage returns per man, X
equals the exchange row entry of the legal minimum wage, and N equals
the parametric multiplier ranging from zero to one.
The effect of moving permanent employees to other jobs was
handled in a similar manner. The only differences were that the
exchange row coefficient was smaller, reflecting the part of the
calculated permanent labor costs that are not cash expenditures,
and entries for family coffee harvesting were not applicable since
permanent employees are paid by the fanega for harvesting coffee.
Movement was restricted to adult male employees.
Parametric changes in the interest rate for long-term investments
were programmed with the same procedure as price changes. The cost
of borrowing was increased from 6 to 20 percent of the investment.
An investment rcw requires that money be supplied to meet investment
requirements via a borrowing activity. An exchange row entry is
multiplied by an increasing number so that 1' = 1 + Xi (N); where,
1' equals the new interest rate, 1 equals the old interest rate, X.
equals the change row element in the borrowing column and N equals
the parametric multiplier. This procedure was used to evaluate the
effect of interest rate changes on the stability of resource use.
The fact that economic development is a long-term phenomenon
may stimulate objections to the use of a model maximizing returns to
a single year. With more work and more computer expenditures, it
would have been possible to build a growth model maximizing returns
over a period of five, ten, or even twenty years. Such a study could
quite dramatically illustrate the gains to be derived from more in-
vestment capital, cheaper interest rates, longer term loans and the
cumulative effects of modern inputs.
However, there is a danger in trying to extract too much informa-
tion from limited data. The data were collected with interviews in a
single year. The effects of one crop on successive crops are not
yet known. Any errors in reporting, interpretating, and evaluating
the data will be multiplied not only by the simple coefficient of
time but also by complex coefficients reflecting the fact that each
year's output becomes part of the next year's inputs.
To be accurate, the long-term growth model also requires correct
predictions of the interest rates and future credit availability.
An even more difficult prediction centers on technology. '.ill yields
remain constant? Will they remain proportional if they change? Will
some resources become outdated? If changes occur, when will they
occur? Will technological changes be accompanied by changes in factor
prices and product prices? The common economic practice is to avoid
these questions by assuming constant technology. This excludes a
chief source of growth from the growth model.
Those crops of greater economic importance are likely to receive
the greater agronomic research. This means that specialization or
monoculture is likely to be encouraged by technological change.
Also, larger commercial farms can be expected to change faster than
small, near subsistence units. Thus, while it is possible to predict
the direction or tendencies of certain differences in crop technology,
the magnitude of those differences will be difficult to estimate.
These statements should not be taken as a general argument
against long-term planning or programming. Such an attack on problems
is worthwhile despite the inherent difficulties. However, with the
data and prior knowledge available, it was decided that more practical
information could be gleaned from a single-year model having more
The Enterprises and Restraints
One of the basic assumptions of the linear programming model is
that a finite number of alternatives and resource restrictions exist.
The number of combinations of factors must be limited but the degree
of limitation is arbitrary and depends upon the use that is to be
made of the model.
In this study, the objectives were to compare coffee with alter-
native enterprises. A number of different processes or methods of
growing coffee and different processes for scrTe of the alternative
enterprises were considered because, with differ-nt farm resources
and parametric changes, one particular technology was not obviously
superior to the others. Also, an averaging of inputs and yields is
not particularly meaningful since distinctly different technologies
related to differences in yields are known to exist. In the decision
as to the number of processes for an enterprise the availability and
the accuracy of the data were taken into account. Activities were
chosen to represent different intensities of the use of labor and
operating capital as well as different levels of modernization.
Activities found on the poorer farms were included in the better
farms' alternatives. However, the activities requiring high levels
of technical skills were blocked for the poorer farms in the initial
solution and were considered only in special parametric procedures.
The activities were coded from budget data supplied by the ex-
tension agents and also from the budgets of the agricultural credit
reports of the Banco Nacional de Costa Rica and the Banco de Credito
Agricola de Cartago and the Ministerio de Agricultura y Ganaderia.
These sources supplied data concerning material inputs, labor hours
and timing of various work operations and expected yields of selected
enterprises in particular areas. In addition, activities were syn-
thesized from foreign input-output data which were modified to
anticipate Costa Rican conditions by adapting labor requirements for
particular work operations from currently grown crops. Research
studies of the Universidad de Costa Rica were used to supply data
for certain horticultural activities.
Activities based on foreign or small research plot d-ta were
entered with what was believed to be "conservative" yield estimates.
The results of plot trial yields were estimated by using the lowest
variety yield which was not significantly different (at a 5 percent
ievel) from the highest variety yield.
One problem that crcse was the fitting of the labor requirements
within the monthly labor con:i:r a ints. In most cases the farm enter-
prise reports placed the labor for each work operation within a
given month; however, in a few instances the reports spread work
over a two- or three-month period. Overlapping time periods would
have made the computations much more costly. Instead, two processes
were sometimes used for timing the input use of labor, and the labor
uses in other cases were arbitrarily placed in months so as not to
compete with coffee harvesting.
The use of operating capital and short-term credit was handled
together to avoid antagonizing the cooperating farmers. The extension
agents felt that loan information was personal and requested that
this section be deleted from the original forms. Therefore, total
short-term expenditures were used as the right-hand-side constraint
for the operating capital and credit row for each farm. This assumed
that the farmers were using as much bank credit as they could get.
As a result, the credit situation is oversimplified in the model, but
the complications of overborrowing for consumption or non-agricultural
uses are thus avoided. These complications would be difficult to
identify using interviews since some common practices are of question-
For the Palmares and San Ramon areas, the programming matrix
was made up of 81 rows and 94 columns. The rows included 12 for
monthly labor use, 12 for monthly land use, 12 for temporary monthly
labor use, 4 for monthly family harvesting labor, I for operating
capital and credit use, 3 for tobacco contract; I for investment, I
for fixed expenses transfer, 14 for product transfer, 3 for coffee
land, I for coffee harvesting, 1 for corn shelling, I for land
transfer, I limiting row for traditional coffee market sales, 2 crop
rotation requirement rows, 2 labor movement rows, 9 exchange rows
for parametric changes, and I profit row.
The columns included 10 coffee growing activities, 9 tobacco
growing activities, 8 corn growing activities, 4 joint corn and bean
growing activities, 4 bean growing activities, 2 sesame growing activ-
ities, 2 castorbean growing activities, 5 dairy activities, 2 beef
activities, 1 peanut growing activity, 2 buckwheat growing activities,
I chickpea growing activity, 1 pigeon pea growing activity, I mixed
crop producing activity, 1 annual to monthly land use transfer activ-
ity, 3 coffee selling activities, 3 tobacco selling activities, 8
grain selling activities, I calf selling activity, 1 milk selling
activity, 4 monthly labor selling activities, 2 yearly labor selling
activities, 12 temporary labor hiring activities, 1 borrowing activ-
ity, 1 coffee planting activity, 1 coffee-destroying activity and one
fixed cost transfer column. A listing of the row entries of each of
the activities programmed is given in the Appendix.
Only farms I and 2 were programmed to allow the use of all the
activities coded for the area. Farms 5, 6, 7 and 8 were not using
the same high level of technology as that found on farms 1 and 2.
The two most productive coffee-growing activities were blocked by
removing the coffee production transfer card. A change row entry was
substituted and an investment entry added so that a parametric pro-
cedure would evaluate the acceptability of the change if education
were to permit its occurrence.
Farms 3 and 4 were below-average coffee producers. They were
permitted to use only the least profit table coffee-growing activities
of the area. Again a parametric procedure allowed higher production,
assuming that education and long-term investment could make the higher
In the Alajuela area, the matrix was composed of 94 rows and 87
columns. The rows included 12 for monthly labor, 12 for monthly land,
12 for temporary labor, 4 for monthly family labor for coffee harvest-
ing, I for investment, 1 for operating capital, 1 for coffee harvest,
19 for product transfers, 4 for rotation limits, 1 for yearly land,
1 for fixed expenditures, 3 for coffee land, 2 for off-farm labor
movement supply, I for limiting sales in the traditional coffee
market, 12 for changes in parametric modifications and 1 for profit.
The columns included 9 coffee growing activities, I lime growing
activity, 1 orange growing activity, 3 corn growing activities, 2
corn-bean growing activities, 3 bean growing activities, 1 pineapple
growing activity, 1 strawberry growing activity, 6 sugarcane growing
activities, 3 cassava growing activities, 2 cucumber growing activ-
ities, 2 sweetpotato growing activities, 2 tomato growing activities,
2 sweet pepper growing activities, 1 peanut growing activity, I
chickpea growing activity, 1 buckwheat growing activity, 1 pigeon pea
growing activity, 1 dairying activity, 3 coffee selling activities,
16 alternative product selling alternatives, 2 cofFee harvesting
activities, 2 all-year labor selling activities, 4 monthly labor
selling activities, i2 monthly labor hiring activities, 1 fixed cost
transfer column, 1 land transfer column, I coffee planting activity,
I coffee destroying activity, and 1 borrowing activity For long-term
With the exception of farm 9, the two highest-yielding coffee
activities were blocked for the original solutions and allowed to
enter with parametric changes. Lime and strawberry selling activities
reflected prices estimated for processing use. These prices were
013.00 per quintal for limes and (l.00 per pound for strawberries.
These prices were conservative estimations for the Central American
Common Market and were programmed both upward and downward to fit
conditions of the fresh market and world market, respectively.
Actually, average current prices are much higher in the local fresh
fruit markets; however, these high prices would be unstable in the
face of any sizable change in quantity.
Tomatoes and sweet peppers were also priced for processing use.
The price used for sweet peppers was lower than the quoted contracting
price because pepper contracts were tied to tomato contracts.
Cassava, cucumbers, and sweetpotatoes were priced at the reported
market lows of the two years prior to the survey. Parametric changes
lowered vegetable prices to levels competitive in the world market.
While substantial changes in technology would be required if
the fruit and vegetable activities replaced coffee, these changes
were permitted in the model because the extension and research facil-
ities seem capable in the zone. The experiment station of the
Universidad de Costa Rica is located in the zone and specializes in
horticultural crops. It is easier to sell new ideas when they have
been tested under local conditions and the results are being applied
by the agronomists on their private commercial farms.
In the study of the Acosta area, the matrix contained 65 rows
and 60 columns. The rows included 12 for monthly labor, 12 for
monthly temporary labor SL'p!)v, 1 for investment, 1 for operating
capital, 3 coffee 1nd limiting rcws, 5 for monthly family labor for
harvesting, I fixed expenditure transfer row, 1 row limiting coffee
sales to the traditional markets, I coffee harvesting row, I yearly
land supply row, 9 for product transfers, 2 for off-farm labor move-
ment and I profit row.
The columns included 7 coffee growing activities, 2 joint coffee-
orange growing activities, 3 joint corn-bean growing activities, I
blackberry growing activity, 2 orange growing activities, 2 beef
producing activities, 5 dairy activities, I lime growing activity,
3 coffee selling enterprises, 8 selling enterprises for other farm
products, 2 coffee harvesting activities, 5 monthly labor selling
activities, 12 monthly labor hiring activities, I fixed cost transfer
column, 2 yearly labor selling activities, 1 coffee planting activity,
I coffee destroying activity, and I borrowing activity for long-term
Since the growing periods for all the crop activities programmed
for Acosta overlapped, land was programmed as a single resource
instead of being divided into monthly intervals of use.
Ecologically, Acosta is poorly suited to annual cropping. Corn
and beans were included because they are traditionally grown. Other
annuals were excluded from the area's model in order to conform with
The coffee yields in Acosta are less than the yields of the other
two areas. Poorer technology may have resulted from relative isolation
in past years. However, lower fertility is chiefly responsible for
lower yields. Two higher-yielding coffee growing activities were
blocked in the initial solution but were allowed to enter in a
parametric procedure representing technological change.
Two price leve!s were used for fruit selling in Acosta since
selling opportunities could be greatly affected by the nearness to
market outlets. Risks and transportation costs would be reduced if
a processing plant were built in the area. Prices were discounted
30 percent for limes, 25 percent for oranges and 50 percent for black-
berries when local outlets were not anticipated.
In calculating production costs for the production activities,
short-run interest charges of 8 percent were added to the costs of
The Sources of the Budgets
It was necessary to use agronomic data from several different
sources to construct the matrix of input and output data used in the
study. Farm resource information and input-output data were provided
by the extension agents in each area analyzed. The extension agents
collected data from the farms they considered typical of the various
farm size and type classifications found in their particular region.
Host of the budgets for coffee, corn, beans and sugarcane were
provided by the cooperating extension agents. These were supplemented
by data provided by the Banco Nacional de Costa Rica (18) and the
Banco de Credito Agricola de Cartago (53).
San Ramon and Palmares farms were grouped together in the
Analysis. !ng. Efrain Abarca collected data from San Ramon including
budgets used in the most productive coffee activities yielding 27.6,
25.7 and 20.0 fanegas per manzana. Ing. Danilo Zamora collected data
from Palmares farms which reported coffee yields of 19.0, 15.0,
and 5.3 fanegas per manzana. The best yields included herbicide use,
three applications oF fertilizers, insecticide use and moderate
pruning and weeding labor. Common practices included the uLe of
fertilizer and insecticides and gave yields above the national average.
Coffee activities using traditional methods were programmed from data
of the Banco Central de Costa Rica (7). Moderately heavy labor with
few purchased inputs produced a yield of 9.0 fanegas per manzana.
An activity of semi-abandoned coffee was based upon conversations
with Ing. Hugo Castro. Yields up to 4.0 fanegas per manzana were
obtainable without purchased inputs other than sacks ard without labor
except harvesting and enough weed cutting to allow the pickers to walk.
The extension agents' farm budgets also included corn and bean
activities. Common corn yields ranged from 13.3 to 40.0 quintales
per manzana. Higher-yielding activities were programmed from data
furnished by Ing. Walter Villalobos from 4-S Club plots at Santa Ana.
Yields were modified to 70.0 quintales per manzana maximum to corre-
spond with the reportedly poorer growing conditions. The moderate
use of fertilizer and insecticide, as reported in the worksheets of
the Banco Nacional, yields 48.0 quintales (18).
Bean activities were based on budgets from the following sources.
Modern technology yielded !8 and 20 quintales per manzana according
to budgets derived from a ministry of agriculture publication (51).
The Banco Nacional supplied budgets of low-yielding bean crops from
broadcast planting that yielded only 4.2 quintales per manzana and
tradition! methods that yielded 9.6 quintales per manzana (18).
Joint corn and bean production activities were programmed allow-
ing combinations of the average and poorer yielding corn and bean
activities commonly grown together.
Tobacco growing activities were programmed from budgets of the
Junta de DeFensa del Tobaco (36) and the Banco Wacional (18). Yields
ranged from 18 to 20 quintales per manzana but the budgets from the
Banco Nacional used lower levels of inputs.
Sesame was programmed using a budget of traditional methods
supplied in Ospino's work (58). A budget of modern practices for
growing sesame were synthesized using data from the United States
(21, 38, 39). An estimate of yield expectations was placed at 20
quintales per manzana despite reported yields up to 35 quintales per
The castorbean production activities were based on synthesized
budgets based upon foreign agronomic data (20, 72). Yields were
estimated at 34 quintales per manzana.
The traditional activities for producing dairy were based upon
reported budgets from Atenas by Ing. Adrian Prado. More intensive
dairy production activities were based on budgets from Heredia
supplied by Ing. Carlos Norza. Production ranged from 400 to 1,200
bottles of milk per manzana with extensive land use and from 1,000
to 2,000 bottles with more intensive operations.
An extensive beef calf producing activity and a moderately inten-
sive beef producing activity were budgeted by Ing. Ramon Castro in
San Carlos with one cow per 5.0 manzanas in the first case and one
cow per 1.5 manzanas in the second. Much higher range productivity
was reported in studies made in Puerto Rico (14). However, the cost
of production w-ould not be covered by Costa Rican prices.
Pigeon pea production activities were based on budgets synthesized
from agronomnic data chiefly from Hawaii (31, 40) modified by Costa
Rican recommendatiur;s (50). Reported yields reached 20 quintales per
Buckwheat was included as a catch crop. Yields were programmed
at 8 and 20 quintales depending upon the time of planting. Reported
production in Mexico (70) and Ceylon (4) showed that this crop could
be grown in tropical countries.
A peanut production activity was included although part of the
soils may not be well adapted. Yield was programmed at 20 quintales
per manzana as Banco Nacional data from Alajuela were used to synthe-
size a budget.
Chickpeas were included as a dry season catch crop. Cultivation
is similar to beans (47) and the yield expectations are 5 quintales
A mixed crop enterprise was synthesized from other budgets
combining corn and legumes with high labor inputs.
Coffee price was determined by an unweighted average of the
prices paid by the beneficios in Palmares in the 1966-67 crop year.
Corn and bean prices were reported by the extension agents. Peanut
and sesame prices were included in the credit worksheets of the Banco
Nacional (18). Castorbean price was computed from the world dollar
price. Buckwheat was priced arbitrarily low to reflect probably
limited acceptance as a feed grain. The chickpea and pigeon pea
prices reflected estimated wholesale prices based upon retail prices
in Scn Jose as compared with beans.
In the study of the Alajuela farms,budget data on coffee produc-
tion were supplied by Ing. Guillermo lontenegro. Yields ranged from
25 fanegas per manzana on the best farm to 16 fanegas per manzana on
the poorest farm. The maximum yield permitted without technological
change modifications was 20 fanegas per ,:aiizana except on farm 9 which
had already adopted modern production techniques. Traditional and
semi-abandoned coffee production activities were included with the
same coefficients used in the Palmares-San Ramon matrix.
Lime production was programmed with input-output data synthesized
from Florida sources (42). Yields were 400 quintales per manzana.
This approximates a U. S. yield of 462 bushels per acre. Costs of
establishment used U. S. costs but annual labor costs were modified
by data from Costa Rican orange production budgets.
The orange production activity was programmed from data of
modern orange production in Guatemala (49) and Florida (28). Yields
vary with the age of the trees but an estimated yield of a mature
grove was taken at 1,020 cien (hundred fruit). This approximates
300 boxes of fruit per acre.
Tomato production activities covered common and modern producing
techniques. The common yield of 11.25 tons per manzana was reported
in a budget worksheet of the Banco Nacional (18). Experimental
results of agronomic trials show yields that surpass 20 tons per
Corn production activities were based upon 4-S Club budgets
which reported yields of 90 quintales per manzana and upon extension
agent reports of corn yielding 60 quintales per nanzana. Corn produc-
tion was also programmed in joint activities with beans where the
output of 48 fanegas and 20 fanegas of corn was produced jointly
with 18 and 4 fanegas of beans in budgets supplied by Ing. Guillermo
Bean production was programmed with three distinct levels of
technology. Modern inputs fielded 20 cuintales per manzana (51),
broadcast beans yiclded 5 qui;tales per ilanzana and traditionally
planted hl ans yielded 10 qi.-itta es per narlnzra (18).
The strawberry producing activity used Florida production data
(12) modified by incomplete data from Alajuela and Heredia farms.
Yields were programmed at 250 quintales per manzana. This approxi-
mates 14,600 pounds per acre. Good California yields, for comparison,
ranged from 48,000 to 60,000 pounds per acre (43) and Israel increased
its average yields from 3,000 to 10,350 pounds per acre in six years
Sugarcane production activities were programmed from budgets
supplied by the Banco Nacional (18) and the extension agency. Produc-
tion ranged from 60 tons per manzana to 100 tons per manzana per
harvest. In a period ranging from 38 to 48 months,three harvests were
made. Production and inputs were totaled for the entire periods.
Three cassava producing activities were programmed. Traditional
methods were represented by a budget which yielded 150 quintales in
a 21-i:month growing period (18). Intermediate yields were received
by a budgeted production lasting two years (58). Higher yields, 275
quintales per manzana, were programmed with more modern inputs and
longer growing period of 26 months (26). Total production per month
increased with age but quality declined.
Cucumber production activities represented different timing of
modern techniques based upon agronomirc Jta furnished by the Alajuela
experiment farm (52). Yields were 140 quinteles per manzana.
The sweetpotato production activity bas-d on modern inputs (5)
yielded 300 quintales per manzana which tripled the yield of tradi-
tional methods reported by the Banco Nlacional (18).
A peanut producing activity utilized a budget reported by the
Cartago bank (jS). Yieids of 25 'quintales per manzana -were expected.
Pigeon pea, buckwheat and chickpea production activities were
included, based upon the same sources as used for the Palmares-San
Prices were based on an unweighted average for coffee, reported
lows of vegetable prices, and estimates of potential industrial
prices for tomatoes, peppers, strawberries and citrus. Since the
projected horticultural marketing assumed much greater volume than
current sales, price predictions were lower than average prices but
subject to considerable error.
In the study of the four farms in Acosta, coffee production
budgets were supplied by Ing. Rodrigo Cavallini of the San Ignacio
Extension Agency. The better methods used fertilizer or other
purchased inputs and yielded 10 fanegas per manzana which was con-
sidered high for this region. Traditional methods yielded six
fanegas and used small quantities of fertilizer and heavy labor in-
puts. Coffee was grown with oranges and bananas on some farms with
poor yields of both coffee and fruit. Coffee yielded five and six
fanegas and oranges yielded 1,000 to 3,000 fruit per manzana on
An activity used to program the possibilities of technological
change was supplied by Edwin Marin of the Oficina del Cafe. He
budgeted production yielding 14 fanegas per manzana in an adjoining
district. An activity was also included representing semi-abandonment
and yielding only ,two fanegas per manza.ra,
A high yielding orange producing activity was included and based
upon data taken From foreign sources (28, 19). This activity yielded
102,000 oranges ccnpared to 53,300 oranges per manzana produced on
a farm in Acosta.
Corn and bean production was reported with low to very low
yields. Joint cropping produced 29, 6.4 and 16 fanegas of corn com-
bined with 11, 9.6 and 6.6 fanegas of beans.
The same beef and dairy production coefficients used in the
Palmares study were included in the matrix.
Lime production was included with a yield of 400 quintales per
manzana based upon Florida data (42).
The inclusion of a blackberry producing activity also was based
upon a composite of information from Florida (65) and Costa Rica.
Yields of 16,000 pounds per manzana were anticipated.
Assumptions of the Model
There are certain assumptions and limitations of the model which
should be clarified before the results are interpreted. Linear
programming uses profit maximization within a set of constraints as
a single criterion for allocating resources. This would deviate
from actual practice especially in those cases where the magnitude
of gain is so slight as to not make a more complicated program worth
the trouble when compared to a simpler, more easily managed plan of
operations. Also, the model does not take uncertainties and risks
into account. Risks, that may be either real or imagined, enter into
farm decision making. Yields and prices vary from year to year.
The farm operator will actually be interested in maximizing profits
only within soine range of acceptable risk.
The model forces all decisions to be made at once. Since it is
a static model, growth possibilities are not taken into account. This
is particularly troubiecome in the case of short-term credit and
operating capital restriction. Because of this feature, the model is
conservative insofar as new resources are not permitted to be
generated over time.
All units are considered divisible. This does not cause a
problem except in the case of cattle and labor movement. Theoreti-
cally, part-time employment could explain fractional units of labor
Optimal Cropping Plans
In general, cropping patterns determined with the initial
linear programming model did not greatly differ from the reported
practices. With one exception, land was fully utilized during at
least part of the year. Unused permanent labor was often available
except during coffee harvest. The restrictions on hiring temporary
day-wage labor were generally not constraining. Where horticultural
crops were considered, operating capital was restrictive. This
restriction occurred also on the smaller, poorer coffee farms.
Marginal returns to short-term credit, calculated by using reported
expenditures as a base, were either zero or well over the established
interest rate. Parametric changes found coffee production to be
stable in the face of moderate coffee price and yield decreases but
responsive to technological improvements and credit manipulations.
Changes in the availability cf credit, the interest rate, the labor
supply, and product prices had a much greater effect on the alloca-
tion of resources among the various alternative crops than they had
on coffee production. A brief summary of optimal resource use will
be giver for each of the 16 farns.
The optimal solution of this coffee-tobacco farm coincided with
the reported production. Two manzOn.ss were planted to thinly spaced
corn and tobacco and three manzanas were used to produce coffee. In
addition, two manzanas of buckwheat were planted in the dry season.
The net returns (W19,781) fell slightly below the farmer-estimated
returns (22,500) because the model included extra labor costs and
lower corn prices. Land and burley tobacco contracts were limiting
rows with marginal values estimated at 4,069 and 215, respectively.
Temporary labor was hired in October and January. Permanent labor
was fully utilized in May and October as well as from November through
February where the model permitted coffee harvesting to exhaust the
labor supply. A slight excess (81) in operating capital occurred.
The second farm from San Ramon was a small specialized coffee
producer. The optional solution for the linear programming model
derived for this farm situation also produced only coffee. Net
returns in the model were ;18,024 compared to 15,250 calculated
from the farmer report. The model permitted the use of a higher-
yielding coffee activity which proved slightly more profitable than
the technology actually used. With tobacco activities blocked, the
only effective constraint was land. Permanent labor was exhausted
only from November through February when coffee harvesting utilized
all the labor. The farm optimal solution used 3,157 of 4,000 avail-
able capital and credit. There were 193.2 fanegas of coffee produced.
This farm produced good coffee yields with very high costs.
The model permitted traditional coffee activities on this farm but
the five best coffee-producing activities of the region were blocked.
The optimal solution contained less-intensive coffee production,
more-intensive use of the non-coffee land, and no reduction in coffee
acreage. Coffee output was reduced 25 percent from the reported
output. This change was accompanied by an increase in net income
from (34,130 to 35,543. The optimal plan made heavier demands on
management with seven producing activities instead of three. Sales
included 750 fanegas of coffee, 52.8 quintales of tobacco, 108.4
quintales of corn, 49.1 quintales of beans, and 16.3 quintales of
The flue-cured tobacco contract row was exhausted as all three
maiizanas permitted for tobacco were planted in the farm model.
Permanent labor resources were exhausted in all months except May
and October. Coffee harvesting exhausted the labor supply during
November, December, January and February. Temporary labor was hired
in March, June, July, August and September.
The optimal plan used less than i45,666 of a 71,100 operating
capital and credit constraint.
The fourth farm studied produced coffee and flue-cured tobacco.
Tobacco was the chief money earner with coffee grown to supplement
income with very low labor inputs. The linear programming model of
this farm situation resulted in considerable changes in resource use.
Coffee output was increased from 80 to 225 fanegas and tobacco
production was reduced from 90 to 46.8 quintales. In addition, corn,
beans, and mixed crops were substituted for tobacco. Sales also
included 36.7 quintales of corn, 59.9 quintales of beans, 2.6
quintales of pigeon peas, and 81 pounds of chickpeas. Net returns in
the model situation were 331,667 which greatly exceeded thZ 14,700
estimate of income under reported resource use.
The model used all available land from July through January,
all available operating capital and credit, and all permanent labor
in April, June, July, August and the harvest months from November
Temporary labor was hired in January and August. Family coffee
harvesting was limited by the family labor supply in December. Hiring
labor was not constrained by monthly supervisory limitations which
allowed two temporary workers for each family or permanent employee.
The tobacco contract allotment was not exhausted.
The fifth farm specialized in coffee. The maximization of the
linear programming model gave results similar to reported resource
use. The net income for the programmed model was 16,876. The in-
crease over the reported income of 15,520 was explained by small
savings in the accounting of harvesting costs and the fact that some
of the reported expenditures were long-term investments.
In the model, permanent labor was fully employed during the
harvest period from November through February. All 10 manzanas of
land were used and family coffee harvesting was limited by the avail-
ability of family labor in December. An excess in the operating
capital and credit row occurred in the mode! because the reported
annual expenditures included some long-term investments.
The Farm produced 190 fanegas of coffee using the highest yield-
ing coffee activity permitted in the model.
The sixth farm studied was a small coffee farm wi th an absentee
owner. The farm .-odel allowed coffee yields slightly above those
reported by the farm. This reduced the net loss to 760. This
loss occurred with an increase in both yield and coffee acreage
above the reported numbers.
In the model, all four manzanas of land were planted to coffee
for profit maximization. This occurred with 0.25 manzanas of new
The optimal solution maximized returns to land. Credit and
operating capital were not limiting in the model, since reported
preharvest expenditures were 4,000 compared to a 2,565 optimal
preharvest expenditure. A large surplus of permanent labor occurred
in all months except during the coffee harvest period.
The new coffee planting was not stable in the face of interest
changes on the required investment. The rate initially used was 6
percent representing a minimum charge which government banks have
used in a policy to subsidize agricultural investment. The new
coffee was not planted if interest charges rose 0.8 percent. Beans
and sesame were planted instead of coffee. The farm loss increased
from 760 to 780.
The results of programming the seventh farm more than doubled
reported net farm income from 3,345 to 7.147. In the optimal
solution, coffee production was expanded with 0.26 manzanas planted
to new coffee. Other production activities included were sun-cured
tobacco, corn and beans. Optimal output, before parametric changes,
was maximized with the production of 62.0 fanegas of coffee, 19
quintaies of tobacco, 72.8 quintales of corn, and 18.9 quintales of
The optimal production exhausted the credit and operating capital
row, land rows from September through January, and the sun-cured
tobacco contract. Permanent labor was exhausted in all months except
July. Temporary labor was hired in all months except July and August.
Labor supervision did not limit hiring temporary labor in any month.
The newly planted coffee was not stable when interest charges
on investments were increased. Interest charges would block new
plantings when the rate was increased 0.1 percent.
The eighth farm model results optimized resource use with the
production of 19 fanegas of coffee, 7.8 quintales of sun-cured
tobacco, 24.0 quintales of corn, 5.4 quintales of beans, and 6.5
quintales of sesame. Optimal net returns were 4,005 compared with
a reported income of 2,653. Coffee acreage in the optimum program
was the same as the farmer's reported acreage.
Credit was severely restricted on this small farm. The shadow
price indicated that an additional colon of credit or operating
capital could return d2.96 additional net income.
Land was fully utilized except in the month of February. Extra
labor was available in all months except during the coffee harvest
season. No temporary labor was hired. The restriction that allowed
0,5 manzanas of sun-cured tobacco w'as not used up.
Fa rm _
The ninth farm produced coffee with modern inputs. The model
included activit!os for horticultural crop prodjct-on. The farm
maximized income with coffee monoculture which yielded 86,761 net
returns. The Farmer's estimate of net returns was 90,00. Expendi-
tures for temporary labor were underestimated which limited the
operating capital and credit row. The monthly labor supplies were
fully used in January, February, June, July, September, November and
December. Temporary labor was hired in June, July and September.
Supervision limitations did not restrict temporary labor hiring.
Analysis of the tenth farm showed that an increase in farm in-
come could accompany crop diversification. Net returns were increased
from 17,160 to (29,489 when optimal use of other crops replaced
coffee. However, this farm was reportedly upgrading coffee technol-
ogy in order to receive (30,000 expected net returns with coffee
monoculture. The optimal cropping pattern included five different
crop-growing activities producing 213.3 fanegas of coffee, 48.0 tons
of tomatoes, 148.0 tons of sugarcane, 708.0 quintales of sweetpotatoes
and 3.14 quintales of chickpeas. Coffee trees were removed from 4.3
manzanas of land. The permanent labor supply was used up in January,
February, April, August, September, November and December. Temporary
workers were hired in January, February, September and December. The
operating capital and credit constraint was limiting in the model.
Supervisory capacity did not limit hiring temporary workers.
On the eleventh farm, programmed optimal solutions did not
change coffee acreage from the reported land use. Other crops replaced
pineapples. The reported net returns of Cl2,980 exceeded the optimal
income of (24,536 of the model. This difference occurred because the
model contained fruit prices for industrial use while tha Farmer
produced 5.5 manzanas of pir.eapple for the fresh fruit market in
San Jose. The high domestic market prices of fruits and vegetables
were not used in the model, since these prices could not be received
if any significant portion of coffee resources were shifted into
Sales to maximize net revenue included 100.0 fanegas of coffee,
2.0 quintales of beans, 710.9 quintales of limes, 13.0 tons of tomatoes,
0.8 tons of green peppers, 236.0 tons of sugarcane and 254.5 quintales
of sweetpotatoes. Land was fully utilized from November through April.
Permanent labor was fully utilized in all months except October.
Temporary labor was hired in January, February, April, June, August,
September and December. The capacity to supervise labor was not an
effective restriction in any month; however, slack supervisory capac-
ity was reduced to 29.4 hours in January. The operating capital and
credit row limitation was exhausted. Roeational limitations were not
The twelfth farm was programmed with only the poorer coffee-
producing activities of the Alajuela region. The farm reported poor
coffee yields averaging 10 fanegas per manzana. The programmed
optimal solution contained the traditional coffee activity which
yielded nine fanegas per manzana. The activity based on the reported
resource use was dominated by another coffee-producing activity. Net
returns were maximized with the production of coffee, corn and beans
ard limes which netted 07,078 in the model. These returns were
higher than the 6,122 calculated from the Farmer report. Farm out-
put included 22.5 fanegas of coffee, 24.1 quintales of corn, 4.8
quintales of beans and 717.2 quintales of limes.
Excess permanent labor occurred in all months except the coffee
harvest period. No temporary labor was hired. All coffee was
harvested with Family labor. Credit was severely restricted.
The thirteenth farm was programmed using two different sets of
fruit prices representing industrial prices with and without a
processing plant located in the Acosta region. Two completely dif-
ferent diversification pictures are presented since it is questionable
whether or not the area could support a processing plant. First,
higher fruit prices were used to evaluate diversification alternatives.
Limes were priced at 0i3.00 per quintal, blackberries at (1.00 per
pound and oranges at (4.00 per hundred. The results of profit
maximization in the model showed a sizable departure from the reported
resource use. This does not dispute a theory of farmer profit
motivation since completely different horticultural alternatives were
placed in the model. Nevertheless, the model shows that if fruit
prices were moderately high and stable, considerable changes would
occur and coffee output would be reduced on farm 13.
Profits in the model were maximized with 89,948 netted from
mixed coffee and lime production. Coffee tree destruction was
programmed for 32.94 manzanas. Credit was limited, causing both semi-
abandoned methods and traditional low-yield methods of coffee to be
used in coffee production. Permanent labor was fully utilized in
January, February, May, July, August, September, October, November
and December. Temporary labor was hired in January, February, flay,
July, August and September. The limit on supervision for temporary
workers was not effective.
Farm production was comprised of 195.6 fanegas of coffee and
17,780 quintales of limes. This coffee output was reduced from the
present output of 900 fanegas. However, without favorable fruit
prices, the maximum income for farm 13 was 67,694. Therefore, the
reported income of 66,878 closely approximated the maximum of the
model when only traditional crops were grown and coffee output was
900 fanegas. Other sales included 333.5 quintales of corn and 126.5
quintales of beans. In the model, a higher-yielding corn and bean
activity substituted for the corn and bean activity actually reported
In the second analysis, low fruit prices were programmed at
9.00 per quintal of limes, 0.50 per pound of blackberries and 3.00
per hundred oranges. With the lower fruit prices, the available
credit and operating capital row was not a limiting factor. All
coffee land remained in coffee but no new coffee was planted. The
most advanced technology permitted by the model was used to produce
corn and beans. Permanent labor was exhausted in all months except
April and June; however, temporary labor was hired only in January,
March, July, August, September and October. All but 96 fanegas of
coffee were harvested with contract labor.
As with the preceding farm, the fourteenth farm was programmed
using high and low fruit prices. With the higher fruit prices,
optimization of the model resulted in considerable departure from
reported practices. Although no coffee trees were removed, output
of coffee was reduced from 70.0 to !8.4 fanegas with coffee produc-
tion activities using low-yield and semi -abandoned methods. Nine
manzanas of limes utilized the non-coffee land and 3,600 quintales of
limes were produced. Net returns in the model were (24,222. The
farm reported income was much lower with l 1,446 netted from tradi-
tional grain crops and common coffee practices.
The operating capital and credit row were severely limited in
the model with a marginal return of i.51 per 1.00 of credit. Per-
manent labor was exhausted except in March and April. The model
resulted in hired labor in January, February, May, June, July, August
and September. Labor supervision was not a limiting factor.
The optimal solution in the model for farm 14 was completely
different when lower fruit prices were used to represent alternatives
without a nearby processing plant. The maximization of net returns
with traditional crop alternatives resulted in 70.0 fanegas of coffee
output. Maximum net returns in the model were higher than reported
returns, 16,718 compared to ll ,446. The increase was accomplished
with higher returns to corn and beans which were programmed with
moderately poor yields that greatly exceeded reported farm production.
Nevertheless, insofar as coffee production was concerned, the model
results coincided with the actual reported production.
Maximum net returns occurred with the production of 70.0 fanegas
of coffee, 261.0 quintales of corn and 99.0 quintales of beans.
Operating capital was limiting and the family labor supply was ex-
hausted in all months except March, April and June. Labor was hired
in May, August, September, October, NIovember and December.
The fifteenth farm also was given two different fruit price
situations. With high fruit prices used first to represent production
potentials given a processing plant in the region, the farm model
maximized net returns at i5,459. In comparison, reported production
netted only 1,403. Profits were maximized by shifting 0.26 manzanas
of coffee into lime production. Farm output included 17.4 fanegas of
coffee and 105.9 quintales of limes.
The farm model used all credit and operating capital available.
Labor was fully used only from October through February during the
coffee harvest. Land was fully utilized and no temporary labor was
hired. Of the 5,459 netted on the farm, 2,353 were received from
off-farm coffee harvesting.
When lower fruit prices were used in the linear programming
analysis of farm 15, all land was planted to coffee. Farm production
in the model slightly exceeded reported coffee output. Income was
above that reported because the model included temporary outside
income from coffee harvesting as part of farm income. Of the 5,209
net returns, 2,308 was income from coffee harvesting work on other
farms. Commonly grown coffee replaced the reported poorer yielding
Coffee production maximized the returns to land and harvest time
labor. Credit w:as not a limiting resource when low fruit prices were
used in the model. The family labor supply exceeded all monthly
demands for labor and no labor was hired. The excess family labor
was sold during the coffee harvest period ard was unused the remainder
of the year. Twenty fanegas of coffee were produced with the optimal
use of resources.
The sixteenth farm was also programnied with high and low fruit
prices. The farm had reported 7.5 fanegas of coffee production from
1.5 manzanas of coffee. When high fruit prices were used, the
maximization of profits shifted resources into lime production and
coffee land was utilized in semi-abandoned coffee production. Net
returns reached 4,307 when maximized which surpassed the 01,339 cal-
culated as expected income with reported resource use. However,
t2,273 out of the 4,307 represented harvest labor sales.
Credit was severely restricted when high fruit prices reflected
a strong local market. Over 5.30 marginal returns were estimated
per colon of additional credit. Almost 1.2 manzanas of land were idle
in the optimal solution. Output included 3.0 fanegas of coffee, 382.4
quintales of limes, 18.2 quintales of corn and 27.3 quintales of beans.
Permanent labor was fully employed only during the coffee harvest and
no temporary workers were hired.
When low fruit prices were used in the programming model, limes
were excluded and resources were allocated to traditionally grown
coffee and low-yielding corn and bean activities. Credit was severely
restricted with a marginal return for operating expenditures estimated
at A3.22 per 1.00. Profits were maximized with 0.77 manzana of land
idle. Excess permanent labor occurred in all months except during
the coffee harvest. No temporary workers were hired. Optimal sales
included 9.0 fanegas of coffee, 54.0 quintales of corn and 32.2
quintales of beans. Income was maximized at 3,612, of which 2,145
cane from labor selling for coffee harvest.
Po i y Ana 1 ysi s
Education for Better Farm Manacement
The first section of this chapter compared reported and optimal
programmed resource use given the price and credit relationships faced
by the farmers without a positive policy to foster crop diversifica-
tion. Nevertheless, the changes in resource use may be attributed to
a general policy of better farmer education and expanded extension
work that would be required before part of the alternatives programmed
could be put into practice.
Therefore, the first policy consideration to be considered repre-
sents education for better farm management. It is often heard that
farmers produce coffee because of nor-econcmic motivations, that they
are reluctant to change, or that they maintain traditional cropping
patterns out of ignorance. Table 2 shows comparisons between the
reported and optimal values for net returns and for coffee outputs.
The results do not indicate that irrational overproduction of coffee
was prevalent. Of the 16 farms studied, only two farms demonstrated
overproduction of coffee. Overproduction on farm 3 could be attrib-
uted to overintensification where returns could be increased by
reducing inputs and yields. The huge quantities of fertilizer re-
ported represented either mismanagement or hidden investment.
Overproduction on farm 10 occurred because either sugarcane or horti-
culture gave returns high enough to replace part of the moderately
high-yielding coffee. The lower programmed coffee output on these
two farms was countered by higher output on other farms.
While the sample is too small to support broad generalizations,
some overproduction of coffee beyond that quantity dictated by strict
1Hidden investment in this case may arise when increased annual
expenditures occur for a short period before yields are increased to
Table 2. Comparisons between reported and optimal incomes and coffee
Net returns Coffee output
Farma Reported Optimal Reported Optimal
coloness) coloness) (fanegas) (fanegas)
1 22,500 19,781 82.8 82.8
2 15,250 18,025 179.9 193.2
3 34,130 35,543 1,000.0 750.0
4 14,700 31,667 80.0 225.0
5 15,520 16,877 190.0 190.0
6 6,760 760 56.3 76.0
7 3,345 7,147 45.0 62.0
8 2,653 4,005 15.0 19.0
9 90,000 86,761 1,000.0 997.2
10 17,160 29,489 300.0 213.3
11 42,980 24,536 100.0 100.0
12 6,122 7,078 25.0 22.5
13A 66,878 89,948 900.0 195.6
13B 66,878 67,694 900.0 900.0
14A 11,446 24,222 70.0 18.4
14B 11,446 16,718 70.0 70.0
15A 1,403 5,459 10.0 17.4
15B 1,403 5,209 10.0 20.0
16A 1,339 4,307 7.5 3.0
16B 1,339 3,612 7.5 9.0
aThe letters A and B are used to designate different fruit pricing
used in Acosta. A's are used to indicate the situations using high
fruit prices and B's are used to indicate the situations using low
profit maximization is indicated. Most of the potential for income
improvement occurred with the land not planted to coffee. Labor
selling activities for the coffee harvest made sizable contributions
to the higher optimal incomes of the smaller farms.
Of the 20 farm'situations studied, three reported higher incomes
than the optimal permitted in the model. In farm I and farm 9 this
difference was due to a lower wage cost estimation in the farmer
report than was permitted in the model. More significantly, the re-
duced optimal income on farm 11 came about because the model only
allowed pineapple sales for export or industrial use while the farm
produced for a much higher domestic fresh-fruit market.
Markedly increased incomes were accompanied by reductions in
coffee output only when high return fruit or vegetable crops were
considered as alternatives. Thus the higher incomes are associated
with higher risks. Optimal allocation of resources reduced coffee
output 3.06 percent on the eight farms of the Palmares-San Ramon area,
and 6.46 percent on the four farms of the Alajuela area. For Acosta,
optimal resource allocation increased coffee output 1.16 percent
when low fruit prices were used but decreased coffee output 76.26
percent when high fruit prices were used, A sizable reduction occurred
in the Acosta output only when high yielding alternatives were compared
with low yielding coffee.
Changes which reduced coffee output brought activities into
production which utilized more modern inputs than the alternative
crops usually receive in Costa Rica. This gives rise to a question
of whether or not the same relationships would exist if new coffee
growing activities represented higher levels of technology in coffee
production. In order to keep the model representative of current
technological proficiency, certain high-yielding coffee activities were
flagged from use in 17 of the 20 farm situations. Then parametric
programming was used to increase coffee transfers until reported yields
of the better farms were equaled. The results showed that on some
farms modernization would be justified if over 60 percent of antici-
pated yields were obtainable. In the Alajuela and Palmares-San Ramon
areas, it would pay most farms to adopt modern inputs if over 80 per-
cent of the recorded yields were obtainable. In Alajuela technological
change permitted yield increases from 20 to 25 fanegas per manzana.
In Palmares-San Ramon the yields were increased from 19 to 27.6
fanegas per manzana. Yields were increased from 10 to 14 fanegas per
manzana in Acosta.
The results of the programmed technological advances are given
in Table 3. In the Alajuela and Palmares-San Ramon farm situations,
optimal coffee output was increased 72.4 percent and farm income was
increased 110.23 percent when higher-yielding coffee activities
modified the output of nine farms. In the Acosta programs, coffee
output was increased 198.81 percent and net income was increased
9.38 percent when technological changes for coffee were permitted in
the four farm situations when high fruit prices were used. Given
low fruit prices, coffee output only increased 15.93 percent but in-
come increased 26.98 percent in the Acosta farm situations.
Higher levels of technology brought snmll increases in coffee
acreage on a number of the farms. The planting of new coffee was
limited by credit restrictions. The gains iI programrrd income were
sufficiently high to motivate change if the improved technology can
The effects of improving coffee production technology on
coffee production and income
Coffee output Incomes
Current Improved Current Improved
Farm technology technology technology technology
1 ,032. 1
aThe letters A and B are used to designate different fruit pricing
used in Acosta. A's are used to indicate the situations using high
fruit prices and B's are used to indicate the situations using low
increase production to the levels reported on the better farms. The
higher yields obtainable in Alajuela and Acosta resulted in greater
increases in income in comparison with the Acosta farms. Coffee
production, therefore, is expected to increase unless positive aids to
diversification or production controls are put into effect.
Comparisons between optimal coffee production and maximum net
returns for traditional versus new enterprises are shown in Table 4.
On the Alajuela farms, returns from sugarcane approached those of
horticultural crops and optimal coffee output was not reduced when
the new crops were included among the alternatives. However, incomes
were increased where coffee yields were limited to 20 fanegas per
manzana. Actually, the higher credit requirements of the horticultural
alternatives caused the optimal coffee tree destruction to be lower
when horticultural crops were included on Farm 10.
In the Acosta farm situations the exclusion of fruit production
alternatives lowered optimal income and caused a sizable increase in
coffee output. Farm 13 and farm 14 had sufficient capital and credit
to respond to fruit production opportunities with sizable increases
in income. The additional income potential of fruits and vegetables
was limited on the smaller Farms by a shortage of operating capital
Taxation or Price Reduction
The stability of coffee production in farm management plans was
further examined with price declines parometrically programmed into
each fari situation. Table 5 shows the effect of price declines on
coffee output. Reduced output with ,5.00 per fanega price declines
was limited to the blocking of new plantings of coffee trees and the
Table 4. Comparisons of optimal incomes and coffee outputs with
traditional versus new enterprises on farms in Alajuela
__I~ ~ __ I_ __ __
CM- L! r~-
OO LiO 00 OlO (D r- N COO0
l-k- 0 -C*- lA
C N Co '-9 OO 0 Lf\ -0 CM irc a \^OOl- 00 r-
N c- c\, O 0 N N r-- 0 m c 0 C- AZ
co a~o) \,o o) m Ln a r o-:1- 0 c -o N)-
\ 1 - a) o- -- u-
--\o-o- NN NN-m n
0n C- 0 r- 0 ~
0 ~ Co 0'i0C)-
ON O0- nr-(o
00r\0 L 0 ND
L~ NC-M Cr- ir 0
Lr\cM cN~ni-r-(- o
r~~CM- r- M
-L CN 0 OO
x0N 00 -N 0 0
N M r CO 0 c 0
cNl rm 0- ) \o CM
LrN 00-t N 0 0
LM \kO of Lr~ c000o
N Lfl\ 0CD o0) r- CM
N LM 0 CO CN r 0 c Q
N- a0-'- N
C 0 0 0 0 0 0 NC c. 0 tr\ k 0 0 C 0-- 0
oo -- CM ru 0 nN or. C0 N C-\ o o -- r cM
-- r) .-> -r u\ \n r 0 0 C7, C)i fi -c. L-\ LA V3 V
substitution of corn and beans for low-yielding coffee on farm 14.
Taxes up to $10.00 per fanega would have little effect on coffee out-
put of the farms programmed. Except for the newly planted coffee,
coffee production was not responsive to price declines up to 20.00
per fanega. High-yielding coffee and semi-abandoned coffee were the
most stable. With high coffee yields, the alternatives are poorly
competitive and with very low-yields, stability is assured by severe
credit limitations. Resources were shifted away from coffee when the
coffee price fell to new market price levels.
Table 6 shows the effect of price declines on coffee acreage.
Coffee acreage was more stable than coffee output. Coffee land was
notably more stable on the smaller farms as price declines were pro-
The effect of price declines on farm incomes is shown in Table 7.
Price declines reduced income more rapidly than output and had the
strongest effect on those farms highly specialized in coffee produc-
tion. For example, given a (40 per fanega price decline, income fell
54 percent on the three largest coffee producing farms and only 22
percent on the farms with optimal production less than 100 ganegas.
Net returns on farm 2, farm 5, farm 9 and farm 13B, all of which
specialized in coffee production, were reduced over 50 percent by a
50 per fanega price decline. Of those farms specializing in coffee
production, only on Farm 10 did diversification possibilities hold
income above 50 percent of initial levels when price was reduced 25
percent. In the Acosta farm situations, high fruit prices resulted
in both higher incomes and greater income stability in the face of
coffee price declines. The effect of a price decline on income was
o C000-- Lt c 00 C 0mcO L O C) 0 -- -- Lr LnL
N C LP I
OO04mmOOBC OO O NJ
* * * * * * *
o o c 0o00 00 Ln Co
(-1 U 0 1;~i 1; -* u-\ m-
CM r0-.000 Co 0 N-- C0 C.- .-
_Ln cMmco 0
- 0.0 0 0- 0 0 0 -)- o 0 -Ln \' O0 N-- C0 M l-r
U-'-- * * *
00 0 0C)C)O
t rN- 6 Lr\ C)
r-O C _r O--.T- o0 r.O L-\ Lu-
r0- LC) C) C:) 0 r-C) U Lru
0 C rL C 0 rC) f- r- M- -
- LA en*
Cr)-- o oC)" O C) C,-- 0C ; CM rL-o r-C r- CM -- -
u 0-- -C 0 0A ,
r *-- Co L-\ Co -.f c* 0 C \ C'D! C 0 - C '- U-
N- M r't-..z- u'. N-CO 0n Co '- ** "n i- d -t Lr\ Ui\'. 'D 0
a L n-M U r C14 00\O -l-: N UN' CO G- N r-%-C
0I rC \.O -c 0 o oL -- -0 CO0 O lo M-
C' I c~ r' -O3 N C'c .N i C' C iL Lra N4 r ( M r'> j- N"
C- i- N c M -) C4 M INn-
I I I
I \ oro0 m m 0WM
0 n- c c ---t r---O LrM r --\.o-o c \o CD \- mCr'. N
CD I C"O CD ,7 r ~ r 00 C14 Loo T'O "CD of\LD o o 0 c'4 Ln
-- ".-O C4 O 0 0"\--.-i r",%\D r- \o0 in-, -- CV _0 cL -
I N -N -C C' C4ON- cN-
I I IL
C- \. r-4- r-- r-' C 0 -- C-4 M C-4 00 Cy) \ O C" (D re-
CL 0 C U -\ ._C C4 CD 0 ,0 N0 r,,"'QD r m -- L-\ ',, 00
, -- U I- O4 ci C 4 C-4 00 o4,
I Cr,- Oc ',- .-" N rr- cO ur--l.CO ir n .,i o - O C r-,O CM
-- I o 4 (y. -T In r-,- c,-j \- o L o o cIM 3 c4 U--
I) nI I
r-I C - -eri --L a -' C- D) N o --o r \ CO M-\ -\ N-cM
u- -1 0- -- L-4 o. c.- cu N C O -c \ N
rn Q) I >c r-~' rf~\ rnL Cl L- U\ mo \o co Co Cc M O~3 I' -
W 1 00 C1l 4' \D WO ,o -I- CD Co (n r----o) U CO- i X (D -C
--N N U- r~~ I-- o I-r N u c-D CoL, cN r- 0 o C4
o a C : c
CL I o 01\ CO L ,% i r oo r>Lroo r O O '3 cM O cM Lr\
o I I
ao N4rnop-tN N-O\ c0 0 o -LMNoCo Ot Nc CO N (j-(M r-
F-O Lo0- N c t-- cY- C -- cO N-) ONC
(0 I o a ri I -n -C 0 rLT i 4
I -- N\ COlNN CO 0N-
j ( : L1m > N-N- C' o L oa mm CO co oo N CM o o N c i Q CM
44a Oa .o V .-..o N--T"J mj',-4 mN-.~L u-\I
0 | CO N O'C CN -
C N N c'rc.- Lr un<0 oQ
I -- M - -. -
greater on the farms from the Palmares-San Ramon area where expandable
alternatives fell further behind coffee in terms of net returns.
The alternative crops that increased first as coffee prices were
programmed downward are shown in Table 8. In the Palmares-San Ramon
area the farms first responding to a price decline were those with
unused tobacco allotment. Among the Alajuela farms studied the larger
farms responded first to a price decline. This also held true among
the farms from Acosta although irregularities occurred with respect
to which set of fruit prices were used.
Payments for Coffee Removal
The programmed effect upon coffee output of annual payments for
coffee tree removal is shown in Table 9. Farm 1, farm 2 and farm 9
were not responsive to removal subsidies because of high returns to
coffee. Farm 16 was not responsive because of severe credit limita-
Annual payments equal to 20 percent of gross coffee earnings per
manzana were effective in changing optimal resource allocation in five
of the 20 farm situations studied. Given a 25 percent discounted
price for 25 percent of the output of coffee, the payment of 20 per-
cent of base gross returns could be made for the withdrawal of coffee
production without extra taxes or loss to coffee producers. The
following formula can be used for calculating a self-paying subsidy
for coffee removal. Let the coffee price be unchanged as coffee in
excess of the quota is taken out of production.
PT + (P DP)N PT
T+N T + SN
Table 8. Alternative crops increased first by declines in coffee
Farm per faneqa Crop increased
1 100 None
2 91 Beans, sesame
3 40 Beans, corn
4 15 Tobacco, beans
5 57 Beans, sesame
6 58 Beans, sesame
7 90 Castorbeans, milk
8 15 Tobacco
9 21 Sugarcane, beans
10 22 Tomatoes, sugarcane
11 43 Limes, peppers
12 60 Limes
13A 14 Limes
13B 14 Corn, beans
14A 25 Limes
1IB 55 Corn, beans
15A 29 Limes
15B 58 Corn, beans
16A 100 None
16B 32 Corn, beans
00 NOC4 rr- CD 0 0 00 00C 0 CD0
0c 0 c U0 -. m~ I' ( 0 D 0 '
00 r-.cM Lr\ 0 0 ) CO Lt 7 \0 \ C7.C)Ct 0o
N cc oD 0 C r. 0 N. ,; c N L; u. c- 0c
00 0-) CD \c Lf'--Q-NN--m -
-r -s -m
co (D N n 0 C) CD -t L0 0 0 Lr\ C CC )0 0
(7-,'~ N 7. 0 Nj\-. C7, r-.CJ CN C" Co \,v C. o 0r
M -T 0cnOD 0 0 Ln-r- \,DCD "0~0 0
00 a,- \, -" 0-% r-. o, N- mN-1 0 N Co4 Cr\ c-0C '
Co0-. c-t --- C;"%- -NN U-t .
CCo N N0 (Y CQ N N 0. C\ N N 4 4 CO 00
CN Or ) G Lr\ 0 N. M N. N- 0 "n -t Co 0 N-'\0 cn 7\
Co N ~00 ~0 c~~0 00 Nc f ~0- 0
0N C7) \0o C\ -D-\ 0L\ M 0 Cl M Co 0 r-C 0 -
CoO' u~ Ln or N- U\ L- 0-' -- 0 c.- .-- -- N- Nr
r -4 crJ.-- L-Co
Co -C 0 00 cn C) 0 N, cn 0 Lr\' D -D 0 0 00
C4 l;) c L ( -* r o r- 14 c 1; 1 c c; c:; -*\ c
N C-4 0 [f0 0 C 0 N 0 N L\ \ Co 0T CN 0 C. C>
00 () U\ 04(J"r-,Ln-C)--- 0 C:C710 r--N-.-N
-NN-- r, '4N4- I-Q-m
<~ m n M 'r, < O
- N1 m -r Lr \. 0 N-Co0 cD- CN m~ crZi in. 'ikn '
S = TD
T + N ND
P = price in traditional market,
D = discount for sale in new market,
T = traditional market quota,
N = surplus over quota,
S = share of payment for exiting firms.
This indicated that coffee could be removed from farm 3, farm 4, farm
10, farm 13A and 13B with a scheme to pay for coffee removal out of
total sales revenue without lowering the average price to producers.
However, such a scheme would reduce coffee output over 10 percent
only on farm 3, farm 10 and farm 13A. A shortage of operating capital
blocked coffee removal since available credit was reduced by the cost
of coffee tree removal.
Annual payments, as shown in Table 10, either increased or did
not affect Farm income. One would expect such a program of self-
financing allotment payments to be politically acceptable since it
would not lower the income for any farm. However, the effect upon
coffee production appears to be relatively low with respect to the
quantities of coffee produced above traditional market quotas. Addi-
tional money could be made available from the Diversification Fund of
the International Coffee Agreement but even with Lhis extr ; subsidy
the annual payments are unlikely to motivate much change. IF $0.30
per quintal of sales were made available for this subsidy, annual
payments could be increased less than 20 per Fancga removed, assuming
that payments were spread over iO percent of the ba:e production.
- Ln 0 cnLrL--\ -- co CO INr--\ -t r-o "M
co" 0 ,O, Lt- N-I _r-- 0 \.o --d- r- o r o "--. o" 0 -
or D0 0 0 r, (n -.. 0 CD -- f Lt.\ (D -3- n \.,
00 -ZtN C nOCN-
i-" roo oo N ".- o r'a,\ n O N- L n0 O0 o 00 Nr-
c0O 0\ C -0 r C- 0\ 0 r-_ -.D- \D0 r'- L\ oC
r-O O Co 0 -o '--4 O 'J C1 i-o o
NCoo Lr' -_ 0 C 0o N r OO r-rO c0-r "U\ o'
M0cO oC i \- r-.-- \o '.0 n N--- r-: \- LPn LU\-\ r
_- n- 00l r- C0 sN MoC4N -
Lr 0r r--0 -r-.' U- --Lr .- \ co, r-- C. -,\- r--,N
rco r-o O0 0 r'-... LnO Drr0. oo oL;
-0- CO CN N CO N-
Lrn o' 10 r--NC) L \Lr -r\.OCO M -.N C\00 Lr\ -- fN
aD N ) r-- N 0 CO -1-0 \,O -3- r N- -- CO 00Ci C -
coN- N0- \ 0o N-_ .0 r- N- N -z\ 'D 0 L L\o (nr \o
-- CCO mCN r--C4 -
LC L\r rLP r- C L- Lr\ \o0 CO ) C C O O \ 0 r- N N
NO- Ltl- \N N-- '. O- o- -, N u\ -:
u~ COO 0 h cMn C14 0- d L\l M
m M- 03 N -N-N-
- Ln (r r- o Lr\L \ N -\. CO Cm c- CO m\ oN- s N
C 0 N '. CO CO O'O c00 N '. N L\ o C -
NOCm f'\ ) rN-- o CN-Nm- aL T d L'-r\-r-d rN
-o o4CO- \CO.N C0LON-
N-C Ln.'.r03 C -L C --Z LCaC O Gm' CNO N c-- O N ^'
O~on LCr '.0 N-~- '.0 -^ N- i r N-dZ '.0 L\ Lr\-3t r'
- ^- c' CO NM N CO M. N -
< on < on -l < C a: <
- N rCn -t Ln \' CO c Cmo N cM (M -T -d- L Lr\ 0 \'.
This could increase total payments to"approximately (60 per fanega.
Therefore, before annual payments may effectively remove coffee, the
alternative use of resources must give returns competitive with coffee
and ample credit must be available to finance such alternative activ-
Payment for coffee removal could be more efficacious if made in
a lump sum provided that the marginal interest rate facing the farm
operate or is well above the rate at which money is available from
bank sources. A lump sum payment equal to the earning differential
between coffee and its next best alternative divided by the marginal
interest rate for the farmer would be necessary to motivate rational
change. Poorer farmers with scarce capital and poor credit standings
would be most responsive to such payments if they were given the
knowledge required to change traditional cropping patterns.
The value of a lump sum payment is shown in the following example.
Assume that the government can borrow money at 10 percent annually
for a 10-year period and that a farmer's marginal interest earning
rate is 30 percent. Then the cost of generating a perpetual psychic
income flow of (100 per year is 50 for each of 10 years. Higher
farmer interest rates or extended pay-off periods would further reduce
the annual cost of generating a given psychic income flow.
Because of the difference between bank rates and marginal interest
rates for near subsistence farmers, lump sum allotment payments would
cheapen the payments required to motivate change. installment pay-
ments computed as a share payment for destroyed coffee could be paid
to a government fund out of coffee export sales without lowering the
average price paid to remaining producers. The government fund then
could borrow to pay lump sum allotments for coffee removal. If the
above example of 30 percent farmer marginal interest holds true, then
the cost of generating an income flow is halved and changes in output
can be motivated through a program of allotment payments in 10-year
installments from coffee sales without lowering average price. Limi-
tations of credit affecting the outputs and incomes would be removed
as the subsidy payments would be a source of operating capital.
The cropping pattern encouraged by subsidy payments for coffee
removal is similar to the changed output programmed with coffee price
declines. Favored crops are listed in Table 11.
Price differentiation is theoretically efficient in that it
permits the national marginal returns for coffee to be passed back to
the farmer. Output of coffee should be reduced or held constant while
income is either increased or unchanged. One difficulty in applying
this measure is the establishment of the quota for the traditional,
higher-priced market. If optimal coffee production is above the his-
torical base used to calculate the traditional market quota, coffee
price averages will tend to be lowered. If optimal coffee production
falls below historical production, the price averages will tend to be
increased. If new crops shift resources away from coffee in the
initial optimal solution, higher marginal returns to coffee may cause
an increase in coffee production as some resources are shifted back
into coffee production.
Table 12 shows optimal incomes and coffee outputs with single and
differentiated coffee prices. Of those farms with increased incomes,
coffee output was unchanged on five farms, increased on two farms,
The alternative crops increased first by annual payments
for coffee removal
aWhen changes were not initiated by the levels of payment first
programmed, additional runs extended the levels of payment.
bLand was idled as credit became more limited due to coffee removal.
Table 1 .
Table 12. Comparisons of income and coffee production for differen-
tiated prices versus single prices
Single price Two prices
ce Two prices
-------------- -- --
and reduced on three farms. Of the eight farms with reduced coffee
output, five had income reduced below the single price optimum.
When differentiated coffee prices were used to separate sales
into new and traditional markets, optimal output was reduced or un-
changed on farms where the base used for establishing the traditional
market quota was equal to or less than the initial optimal coffee out-
put. However, in those cases where new activities replaced coffee
and optimal output fell well below the historical base, then the
optimal coffee output increased or was unchanged when prices were
differentiated. Income increases were associated with above optimal
production in the base period. Increased shifts toward lime or
vegetable production did not occur when prices were differentiated to
reflect different export earnings in new markets and traditional
Reduction of Credit
The reduction of credit has been suggested as an effective means
of controlling coffee output. As Table 13 shows, coffee output is
reducible if this measure can be imposed. However, there are effects
on farm incomes also (Table 14). In 10 of the 20 farm situations,
income was first reduced by credit restrictions without affecting
coffee production. In four situations where horticultural crops re-
placed coffee in the initial solution a reduction in operating credit
actually brought increased coffee production within certain levels of
constraint. The magnitude of the income reduction is greater where
the optimal solution contained alternative crops. Since credit avail-
ability was tied to coffee land in the model, a credit restriction
cut coffee production without encouraging diversification. Coffee land