PLAN PUEBLA AFTER SIX YEARS* /
By Don Winkelmann
International Maize and Wheat Improvement Center
Mexico City, Mexico
With policy makers giving ever more emphasis to agricultural devel-
opment, interest in special programs aimed at disadvantaged farmers is
increasing sharply. Some are convinced that such programs represent the
only hope for-traditional farmers. Others, arguing that costs are ex-
cessive when compared with results or when compared with more general
programs, are skeptical about the results from or the need for such ef-
This paper treats one such program, Plan Puebla in Mexico. The dis-
cussion is organized around the following themes: the goals and organi-
zation of the program; achievements; factors playing a role in impeding
its progress; and implications for policy.
GOALS AND ORGANIZATION
Plan Puebla is a development effort aimed at the traditional far-
mer. Launched in 1967, the project ". .is an attempt to tackle simul-
taneously. .two development problems -- food shortage and low income
in agriculture. .by obtaining a massive increase in the yields of a
basic crop among smallholders. ." 1/ More specifically, the project's
objectives have been: 1) to develop, field test, and refine a strategy
for increasing yields of a basic crop among smallholders; 2) to train
technicians in the use of this strategy; and, added later, 3) to double
the yields of the basic crop within five years.
The project is set in the State of Puebla, less than two hours drive
from Mexico City. This area was chosen for a variety of reasons. Its
roughly 50,000 farms were small, averaging about 2.5 hectares according
to a 1967 survey. Farmers accented one basic crop, with over 70 percent
of the land in corn according to a 1967 survey. Actual yields have
tended to be far below potential yields. The probability of drought is
not great, less than one year in ten for severe drought and less than
three in ten for moderate drought. From the project's inception,
*The author gratefully acknowledges the help of the staff of Plan Puebla
and of Drs. R.J. Laird and R.D. Osler of the Centro Internacional de
Mejoramiento de Maiz y Trigo (CIMMYT). The views expressed are those
of the author and are not necessarily shared by CIMMYT or by the Rocke-
government infrastructure was in place in the form of credit agencies,
market roads, and crop insurance; and further, markets with stable prices
were readily available.
While Plan Puebla has always emphasized maize, this crop has been
regarded as a vehicle for catalyzing continuous change among the project
region's farmers. Several major lines of work are identified below.
The research program emphasizes intensive experimentation on far-
mers' fields, a tack now regarded as one of the hallmarks of Plan Puebla.
After five years of experimentation, a number of crop recommendations
are available. These depend on the soil conditions of the Plan area,
the seeding date, and the restriction imposed by limited financing. Com-
mon differences between Plan recommendations and traditional practices
are presented in Table 1.
Representative Practices Recommended by Plan Puebla
and a Representative Traditional Strategy a
Zone 1 Zone 2 Zone 3 Zone 4 Zone 5 Traditional
Seeding Rate 50,000 50,000 50,000 50,000 50,000 30,000
Nitrogen c 110 130 130 110 100 30
Phosphorus 50 60 40 50 0 24
Applications 2 2 2 2 2 1
Data provided by Dr. Antonio Turrent, CIMMYT.
Plants per hectare.
Kilograms of nutrient per hectare.
In dealing with local institutions, especially the three government
banks and the major fertilizer dealer, the project staff identifies work-
ing rules which inhibit the adoption of the recommended strategies and
seeks to have these rules changed. 2/ While a substantial effort is
made to influence the relationship between local institutions and far-
mers, there is little emphasis given to relating the Plan to federal re-
search and extension agencies focusing on agriculture.
The technical staff works with farmers through groups. The groups
are organized to facilitate teaching the new technology to farmers, to
expedite the acquisition of credit and fertilizers, and to smooth repay-
ments of loans. Groups are spontaneously organized and elect a leader
who works closely with an extension technician. The staff includes five
such technicians, one for each of five different extension zones into
which the area has been divided.
During the project's initial stage emphasis was put on identifying
production strategies. Later, the accent was shifted to extension.
More recently, emphasis has been placed on identifying impediments to
farmer adoption of recommendations. Nevertheless, the essential func-
tions and the organizational format of the Plan have changed little
since the second year, continuing to feature on-farm experiments on
maize with diffusion of the new technology through farmer groups.
The achievements of the Plan can be assessed in several dimensions.
Some of these can be measured, albeit with difficulty, e.g., the in-
crease in maize production attributable to the Plan. Some are measurable
only at substantial cost, e.g., changes in the production of other crops
which have occurred because of the lessons learned in maize. Yet other
achievements are intangible and not subject to measurement, e.g., those
occurring in individual communities because of the farmer groups orga-
nized by Plan Puebla or those arising from the training program.
All of this implies that evaluating program benefits is difficult.
Even simplifying the evaluation problem by looking only at changes in
maize technology and yields is not totally satisfactory. The difficulty
of estimating whose maize has been influenced by the Plan remains. One
of the procedures adopted below assumes that only those farmers receiving
credit through the Plan have been influenced. This, of course, under-
states the impact of the Plan to the extent that others follow the re-
commendations. Another alternative examines the area's increase in ni-
trogen use on non-irrigated maize. Here too, however, one can only
guess at what would have happened to nitrogen use in the absence of the
MEASURES OF ACHIEVEMENT
Several measures of achievement under Plan Puebla are presented in
Tables 2 and 3. In Table 2, it is assumed that only those farmers who
have received credit under Plan auspices have been influenced. Given
that assumption, it would appear that approximately 12 percent of the
area's farmers were associated with the Plan in 1972, its sixth year of
operation; alternatively, roughly 20 percent of the area's non-irrigated
maize was in the program.
The implications of the restrictive definition of Plan participation
adopted in Table 2 are clearly seen in Table 3, in which data are pre-
sented related to nitrogen use on non-irrigated maize in the area. Ni-
trogen application as a measure of Plan participation is easily justi-
fied on the basis of Plan recommendations. Three substantive changes in
traditional practices have been advocated: the use of larger quantities
of seed; more fertilizer; and split fertilizer applications. The recom-
mended use of additional fertilizer is easily the most expensive part of
the new package, and program activities revolve around getting ferti-
lizer into farmers' hands.
Although Table 3 would seem to suggest that nitrogen use among area
farmers doubled between 1967 and 1970, two caveats are in order. The
data may not be entirely representative of all non-irrigated maize pro-
duction. The influence of trend and/or weather vis-a-vis Plan Puebla
Farmer Participation and Maize Production Parameters
in the Plan Puebla Area
1968 1969 1970 1971 1972a
Field Experiments 60 63 72 57 58
Communities with Groups 31 60 94 101 103
Farmer Groups 3 128 218 183 312
Credit (1000 pesos) 75 4,900 9,600 7,600 11,101
Hectares in Plan 76 5,743 12,661 14,438 17,581
Participation Farmers b 103 2,561 4,833 5,259 6,202
Yield of Participants b 3,590 2,809 2,670 2,618 2,920
Yield of Non-Participants 2,144 1,631 1,775 1,783 2,378
Sources: Cano, Jairo and Don Winkelmann, "Plan Puebla: Analisis de
Beneficios y Costos", El Trimestre Economico, Vol. 39: Octo-
ber/December 1972, pp. 783-796, and Gomez, Mauro, "V Reunion,
Plan Puebla", February 1972 (mimeograph).
aBased on Plan Puebla staff estimates.
bTons per hectare.
is unknown, although rainfall data would suggest no appreciable weather
Nonetheless, the influence of Plan Puebla on non-irrigated maize
production may have been appreciably greater than is reflected in Table
2. For example, if Plan participation were measured by the use of over
one-half of the recommended level of nitrogen on at least one field,
participation rates would have been 47 percent in 1970. 3/ Hence, while
it is highly unlikely that Plan Puebla has resulted in a doubling of
yields, the influence of the Plan has been undeniably significant. The
approximate magnitude of that significance is examined in greater de-
tail in the following paragraphs.
Costs and Benefits
The estimated costs of operating the Plan through 1971 appear in
Table 4. Exclusive of those involved in project training activities,
eight staff members are required to operate the program. However, all
training costs, training of staff and of others, are included in Table
4, thereby overstating the costs related to increasing maize production.
Concurrently, as implied by the proceeding discussion, total benefits
tend to be underestimated.
At least two attempts to compare project costs and benefits have
been undertaken. One, by Cano and Myren, 4/ arrived at a benefit-cost
ratio in excess of seven. This study projected future expansions in
Frequency Distribution of Average Nitrogen Use Per Hectare on All
Non-Irrigated Maize for a Sample of Farmers
from the Plan Peb
from the Plan Puebla,
Sample Average Kg/Ha
Farmers in Sample
Note: Frequency distributions and calculations by D. Winkelmann based
on data taken in Plan Puebla surveys of 1967 and 1971.
aSome farmers, those with non-irrigated maize and those for whom separa-
tion of fertilizer use on irrigated and non-irrigated maize was impos-
sible, were eliminated from random samples representing all of the area's
farmers. There is no reason to believe that their elimination has in-
troduced bias between frequency distributions.
Estimated Cost of Operating Plan Puebla by Year
Year Total Costs a
Source: Centro Internacional de Mejoramiento de Maiz y Trigo, CIMMYT
Program Review, 1972. Centro Internacional de Mejoramiento de
Maiz y Trigo, Mexico, D.F., 1972 (mimeograph).
aTotals include staff time needed to provide 62 man-years of training
through June, 1972. They do not include the costs of maintaining the
production which can already be seen to be over optimistic. Moreover,
it utilized a model for estimating yields in the absence of Plan Puebla
which could lead to errors of undetermined sign in estimating project
A second attempt to compare benefits and costs employed conserva-
tive assumptions and almost certainly underestimated benefits relative
to costs. In order to reduce the possibility of attributing to the Plan
what was really a consequence of weather or trends, Cano and Winkelmann 5/
assumed that only those farmers formally associated with the program ex-
perienced increases in production, that the increases were limited to
maize, and that increases are maintained for only ten years. This is
like assuming no one learns from his neighbor, fathers do not teach sons,
all knowledge of improved practices will be forgotten after 1977, and
there is no spill-over to other crops. Given these exceedingly conser-
vative assumptions, an interest rate varying from 12 to 18 percent, the
price of corn varying from $49.50 to $75.20 per metric ton, and a charge
for family labor equal to half of the local wage, the benefit-cost ra-
tio ranged from 1.1 to 2.9 (Table 5). 6/
Benefit-Cost Ratios for Plan Puebla Under Alternative Assumptions
Corn Prices a
$49.50 $60.20 $75.20
I. All fertilizers
and seed obtained
through Plan Puebla
r=.12 b 1.18 1.45 1.84
r=.18 1.13 1.40 1.78
II. Only first fertilizer
and seed obtained
through Plan Puebla
r=.12 b 1.40 2.01 2.88
r=.18 1.26 1.82 2.60
Source: Cano and Winkelmann, op. cit.
aMaize prices are in dollars per metric ton. The first two assume that
Puebla is either exporting to or importing from the world market. The
final price is the government support price.
Rates of interest, r, of .12 and .18 are respectively the subsidized
rate for certain classes of farmers and an estimate of the marginal
efficiency of capital.
It has sometimes been said that Plan Puebla cannot be duplicated
in other countries, that its resource requirements are such that the
project must be regarded as a special case. The most direct response
to this challenge is to point out that ten variants of the program are
now operating in Latin America -- six in Colombia, two in Mexico, one
in Honduras, and one in Peru. In each of these, alumni of the Puebla
training program are playing an important role.
FACTORS INFLUENCING PROGRAM ACHIEVEMENTS
Critical interest excited by projects like Puebla tends to focus on
how quickly farmers accept new technology, and on identifying and alle-
viating impediments to more widespread adoption.
Agreement is lacking as to what constitutes a reasonable standard
of progress for small-farm programs in LDC's. Puebla's farmers have not
adopted new technology as quickly as the Punjab irrigated farmers adop-
ted new wheat technology. On the other hand, the Puebla experience has
been more favorable than that of most earlier efforts to introduce new
technology to Latin American farmers. Indeed, Puebla itself is a res-
ponse to the experiences of those earlier efforts in Latin America.
Whether or not Puebla has made "reasonable" progress, it is useful
to speculate on what factors may have limited that progress. A myriad
of such factors have been advanced. The treatment here accents those
which have received most attention with no attempt to be exhaustive.
Discussion is organized around the factors commonly held to be most in-
fluential in the diffusion of new technology: adequacy of technology
with respect to profits and risk; adequacy of information; and availa-
bility of credit and inputs.
Adequacy of the Recommended Technology
Some have said that farmers do not adopt Plan recommendations be-
cause the "package" is not good enough. This challenge is usually
based on the profitability of the Plan as compared with traditional tech-
Comparison of Puebla and wheat data from Pakistan's Punjab does not
support this argument (Table 6). These data show the new technologies
have comparable yield advantages over the old technology in farmers'
fields as well as in experimental plots. While such comparisons are not
conclusive -- issues such as representativeness and costs are obvious
considerations -- they are suggestive. The Puebla "package" compares
favorably with the Punjab "package" in terms of production and, as
changes in costs appear to be similar, in terms of profitability.
Others, pointing to the Punjab's irrigation and Puebla's reliance
on rainfall, assert that risk has played a prominent role in influencing
adoption in Puebla. Proponents of this view hold that the Punjab's ir-
rigated farmers saw little additional risk and larger incomes associated
with the new technology, and hence, moved quickly to adopt it. Puebla's
farmers, on the other hand, are said to have been attracted by the
Various Yields (Tons/Hectare) of Maize in Puebla
and Wheat in Pakistan's Punjab
Plan Puebla Punjab
Experimental Yields a 6.64 4.50
Adopters' Yields 2.80 2.50
Traditional Yields b 1.89 1.48
aThe Puebla yield is a 1968/69 average calculated from estimated res-
ponse surfaces based on experimental data and using profit maximizing
levels of fertilizer. See Turrent, Antonio, "Value of Agronomic Re-
search in a Project to Rapidly Increase Crop Production", in Delbert
Myren, ed., Strategies for Increasing Crop Production on Small Holdings,
Centro Internacional de Mejoramiento de Maiz y Trigo, Mexico, D.F.,
1970. The Pakistan yield is the average of nine experiments seeded in
November, 1968, at Lyallpur Station, Fourth Annual Technical Report,
Accelerated Wheat Improvement Program, West Pakistan, 1968/69, Lahore,
September 1969. All plots were seeded in Mexipax, the then dominant
variety of dwarf wheats. Experimental yields were reduced by hot winds
at a critical time.
Since the 1968 Puebla yield was based on a biased sample of 103 farmers,
only 1969 data are reported in CIMMYT, El Plan Puebla, Discussion Paper
No. 8, Table 4. Pakistan averages are for 1968 fall wheat seedings.
Eckert, Jerry B., "The Economics of Fertilizing Dwarf Wheats in Pakis-
tan's Punjab", July 1971 (mimeograph), p. 11.
prospect of higher incomes represented by the potential yields promised
by Plan recommendations, but were deterred from following the recommenda-
tions by their perceptions of accompanying risk.
This argument is consistent with the impressions gathered from in-
terviews of a small sample of farmers. 7/ Nineteen farmers not following
Plan recommendations were asked why they did not associate themselves
with Plan Puebla and follow Plan recommendations. Seventeen responded
with answers indicating an unwillingness to go into debt when the wea-
ther, and hence the resulting production, is uncertain.
Discussion with production scientists validates the conviction of
the farmers. In years with severe moisture stress, more densely planted
fields can give lower yields than fields less densely planted. This oc-
curs because available moisture is consumed in the formation of stalk
and leaves, with little left over for grain formation. With yields lower
and costs higher, net incomes for these circumstances would be decidedly
When such results are likely, the argument featuring risk contin-
ues, the farmer prefers the lower expected income and risk of the tra-
ditional strategy to the higher expected income and risk of the recom-
mended strategy. Indirect tests of this view of farmer behavior were
made by Edgardo Moscardi 8/ and by Antonio Martin del Campo. 9/ In
each study a behavioral model based on income and risk aversion was used
to derive relationships among empirical variables.
The Moscardi study suffers from data limitations. Done in 1971,
the dependent variable was measured by appearance on Plan Puebla spon-
sored credit lists in 1970. Two critical problems emerge from the use
of this dependent variable. First, the adoption of new technology by
the farmers of the sample is almost certainly understated. Second, its
dichotomous nature makes standard regression analysis mute with respect
to statistical significance. While the model says little of the role
of risk in decision making, it was found that the variance-covariance
matrix of independent variables for participants was significantly dif-
ferent from that for non-participants. 10/ This suggests that the two
classes of farmers were drawn ffom different populations.
In Martin del Campo's study, the principal model utilized the aver-
age level of nitrogen application on non-irrigated maize as the depen-
dent variable. On the basis of 1971 survey data it was concluded that
the relationship between the dependent and independent variables was
neither strong enough nor sufficiently consistent with the hypotheses to
support the assertion that risk has played a dominant role in farmer
acceptance of Plan recommendations.
Two variants of the Campos models are found in Tables 7 and 8. In
Table 7, six of the estimated regression coefficients have high statis-
tical significance. Two of these, family size and off-farm work, are
not consistent with hypotheses based on risk averting behavior. The
latter is consistent with high opportunity costs for farmer labor, but
this explanation is not convincing given the existence of well-developed
labor markets. The four remaining coefficients, all related to areas,
are essentially in proper order with respect to risk according to Plan
One interesting result is that farm size and nitrogen use are not
related in a significant way. The model of Table 8 reinforces this
finding. Here the data of Table 7 were partitioned according to loca-
tion. Two essentially homogeneous agro-climatic regions were formed.
In neither region is fertilizer use related in a significant way to farm
In any case, the coefficients of the models are too low to warrant
further discussion of the regression results. As in studies by Moscardi
and Campo, the results themselves do not give strong support to the ar-
gument that risk has played a prominent role in farmer decision making.
This is not to suggest that the adoption of new technology is indepen-
dent of risk, but rather, that the models reported here do not offer
strong support for theargument. Even in the face of these empirical
results, the author remains convinced that risk is a dominant force in-
fluencing farmer response to new technology, especially that of small
farmers in rainfed regions.
Results of a Multiple Regression Relating Average Nitrogen Use per
Hectare of Non-Irrigated Maize to Several Independent Variables a,
Mean Regression "t"
Variables Values Coefficients Values
X1 Family Size 5.7 2.70 2.27
X2 Age of Farmer 51.8 .039 .19
X3 Farmer Schooling 2.2 3.13 1.80
X4 Non-Irrigated Maize, Hectares 2.3 2.05 .76
X5 Cultivated Land, Hectares 3.3 -2.00 .85
X6 Sale of Animal Products 1618.8 .0008 -1.41
X7 Sale of Alfalfa 82.0 .0003 .92
X8 Sale of Fruit 1717.8 .0001 .49
X9 Sale of Small Grains 74.1 .006 1.56
X10 Sale of Vegetables 3478.4 .0002 .81
Xl1 Farmer's Off-Farm Income 2376.0 .008 -2.50
X12 (X11) .0000 .93
X13 Other Off-Farm Income 1770.6 .0004 .51
X14 Equipment 12057. .0001 -1.21
X15 Zone 1 -30.9 -2.86
X16 Zone 2 -25.0 -2.23
X17 Zone 3 -43.9 -3.90
X18 Zone 4 -41.3 -3.72
R2 = .26
Source: Martin del Campo, op. cit.
aBased on data from farmers with some non-irrigated maize in the 1971
Simple correlations among variables were low, exceeding .5 in only two
Adequacy of Information
Puebla Plan staff disseminate information to farmers in a variety
of ways. Some of these aim at transmitting information about recom-
mendations to farmers, especially to participating farmers. Others have
the function of inducing farmers to participate. It is in the latter
dimension, attracting participants, that the information system is held
to be faulty.
Criticism focuses on the high-yielding plots. All formal partici-
pants in the Plan, i.e., all who receive credit through the Plan, are
said to have high-yielding plots. These plots or fields are one of the
principal devices for demonstrating the new technology. It is known
than significant proportions of the participants have not adhered close-
ly to Plan recommendations, even though those recommendations were un-
Results of-Two Regressions Relating Average Nitrogen
Use Per Hectare of Non-Irrigated Maize
with Several Independent Variables a
('t' values in parenthesis)
Variables Zone A Zone B
Family Size 2.12 5.34
Age -.26 .45
(-1.19) ( .83)
Years of Schooling 3.94 2.65
(2.23) ( .70)
Off-Farm Income -.06 -.14
Proportion of Land in Maize 14.50 94.67
b (-1.17) (1.49)
Farm Size 1.98 10.31
( .20) ( .43)
R2 .35 .61
Source: Moscardi, op. cit.
aBased on data from farmers with some non-irrigated maize in the 1971
Measured in logarithms.
Why participants have diverged from Plan recommendations is not
clear. A variety of possible explanations can be offered: planting
densities were held down and fertilizer applications were made late so
as to reduce risk; fertilizer has been received late because banks have
had difficulty in processing requests; credit limits fertilizer pur-
chases even among formal participants; farmers have difficulty in judg-
ing seeding densities properly; and some participants were not fully
informed about recommendations. Each of these probably explains a part
of the divergence. In any case, in years with adequate rainfall, de-
partures from recommendations led to lower yields than if recommenda-
tions were exactly followed. Thus, it may be reasonable to assume that
project area farmers, hearing the recommendations, see so-called high-/
yielding plots which promise less than double their corn yields (see
Table 2) and decline to participate.
Availability of Credit and Inputs
Several factors related to infrastructure may have influenced adop-
tion of the Plan's technology. The first is that credit has not been
in amounts sufficient to meet farmer demand, or if available, much time
and trouble has been entailed in its acquisition. The high probability
that fertilizer will be delivered late may also have impeded the rate
of adoption of new technology. Conversations with farmers suggest that
this consideration has been important in the timing of fertilizer appli-
cation since, if fertilizers arrive after seeding, they cannot be applied
at seeding. Seeding densities also have been influenced as farmers re-
duced seeding rates because of apprehension over the availability of
fertilizers. Late deliveries of fertilizers may have been a consequence
of tardy applications for loans, time involved in processing loans, re-
luctance of government banks to carry large inventories of fertilizers,
and general shortages arising at certain times of the year.
These discussions are related to earlier remarks on risk, in the
present case the risk that inputs actually solicited would not arrive
on time. There is, as well, an indirect relation which suggests, for
some risk averters, a modest refinement to the argument. Risk averters
may be disposed to follow a new technology on at least some part of
their farms unless other factors intervene. The time and psychic costs
associated with acquiring credit tend to be fixed and will impose a
lower limit on the amount of credit a farmer will seek. If that lower
limit exceeds the upper limit he is willing to accept because of the
risk involved in using the new technology, the farmer will not seek
credit for employing new technology on any part of his farm.
Other Factors Peculiar to the Plan
Plan Puebla has relied heavily on the use of farmer groups. New
technology is taught through groups, credit is arranged through groups,
and repayment is expedited through groups. Proponents of this format
argue that it is a convenient way to deal with large numbers of farmers
and that in fact, banks and technicians could not or would not deal with
individual small farmers. 12/ Others counter that the functions should
be separated, so that groups are employed in teaching the new technology,
while all credit relationships should be carried on individually.
An opposing viewpoint maintains that farmers resist affiliation
with groups, and will not do so even if it means foregoing credit. Al-
ternatively, groups may limit the entry of prospective members; and
hence, the reliance on groups for diffusing new technology tends to be
self defeating. This last argument tends to be easily rationalized.
The entire group can find itself without credit if a single individual
defaults on a previous loan. This induces the group to take a conserva-
tive attitude toward prospective members, refusing entry to those who
might not meet payments.
A second impediment peculiar to Plan Puebla is said to be the loan
insurance required by one of the principal sources of credit. In order
to qualify for a loan from the Ejido Bank, farmers must take out certain
kinds of insurance. This insurance is designed to protect the lender in
the event of crop failure or the death of the borrower. It adds consi-
derably to the cost of the loan, and consequently, farmers may be reluc-
tant to accept the added cost because of the conviction that the insur-
ance is not effective.
While other impediments have been alleged to influence Plan pro-
gress, e.g., social structure or unwillingness to change, the factors
given above are those most commonly mentioned. Their implications for
policy are the subject of the following paragraphs.
IMPLICATIONS FOR POLICY
To review, Plan Puebla has several distinguishing characteristics:
1) it is a pilot project aimed at increasing maize production, develop-
ing a strategy for diffusing new technology among smallholders, and
teaching that strategy to others; 2) the strategy developed emphasizes
research in farmers' fields, groups for instruction and credit, and the
use of farmer-monitored demonstration plots; 3) the project features
close coordination with credit agencies but little connection with
federal research and extension agencies; and 4) evaluation of the pro-
gram's success is quite sensitive to the way in which farmer participa-
tion is defined.
While a large number of unanswered questions remain, witness the
earlier discussion, some lessons have been learned. It is generally
agreed that, although a pilot project, such efforts should have close
ties with national agencies. Moreover, while pilot projects are useful
ways to acquire information, train staff, and work out modus operandi,
few deny that they should include plans for expansion into national pro-
grams. Finally, workaday experience provides ample evidence that care-
ful efforts must be made to distinguish between official rules and the
way such rules are translated into practice.
Narrowing the focus to the tactical questions opened above, three
have implications for operating policies in Puebla-like programs. First,
to the extent that risk is influential in farmer decision making, profit
maximizing recommendations might be inefficient. What this suggests is
that agronomists devote more attention to establishing and demonstrating
the risk associated with various agronomic strategies. Recommendations
to risk averters will tend to call for fewer inputs and be less inten-
sive than recommendations aimed at maximizing expected profits. Farmers
concerned with risk will find such recommendations more consistent with
their goals, and may be expected to adopt such recommendations more
readily, achieving higher yields and greater incomes than with tradition-
Adoption could be expanded by reshaping the existing insurance pro-
gram, by making it one of crop insurance, protecting the farmer from
lower net incomes in the event of bad weather rather than loan insurance,
risk would be further reduced and Plan recommendations made more attrac-
Looking to the information system, it would seem advantageous to
label as high-yielding plots only those in which Plan technicians insure
that recommendations are fully implemented by the farmer. In this way,
potential participants would have more accurate impressions of the trade-
offs between yields and risks.
Finally, there is the issue of the groups and the circumstances
which favor conservatism on their part, i.e., no financial loss occurs
to the groups from rejecting a potential member but loss can result from
accepting a member. Recognition of this asymmetry could be an argu-
ment for opening up an alternative avenue of financing, one featuring
merchants, while maintaining the groups for distributing information.
In rejecting a would-be buyer who requires credit, a fertilizer
merchant gives up the potential profit from the sale and the loan.
Applying other terminology, groups lose only from type II error, re-
jecting when they should accept, while merchants lose from both type I
and type II error. Merchants, then, should be less conservative in fi-
nancing fertilizer purchases than are groups. The merchants, in turn,
would be financed by banks whose aim would be to increase fertilizer
use and production, while avoiding the burden of individual financing
for large numbers of small borrowers. Local merchants also have the ob-
vious advantage of knowing their clientele. If financing could be ar-
ranged through several merchants, so as to insure reasonable competition,
this could be an effective supplement to, and perhaps substitute for,
financing through groups.
Beyond the question of tactics, there are broader questions of
strategy. Some argue that it is not necessary to mount special pro-
grams for small farmers, that they too are best served by national pro-
grams aimed at all farmers producing a given crop. The experience in
El Salvador, where average maize yields increased by some 60 percent
from 1966 to 1971, is an apparent example of a national program with
Related to this, but so important as to be a separate issue, is the
question of what really restricts the adoption of new technology. The
conventional wisdom of agricultural development no longer accents tradi-
tionalism and unwillingness to change, but rather, recognizing the far-
mer's ability to manage limited resources in a precarious environment,
looks to other restrictions -- risk, management, the information system,
credit, and ready access to inputs. Because of the impact of risk, Plan
recommendations probably have been inconsistent with the farmer's objec-
tives. More efficient recommendations for risk averters should more
closely approximate customary practices. It is argued that a host of
other forces impede change, including the pattern of production, and the
fusion of production practices and ritual. 13/ By knowing how all of
these elements interact, policy can be more effective in alleviating
In still another dimension, some believe that Puebla is too narrow
in its goals; that it is sufficient neither to think in terms of one
crop nor simply in terms of production. Examples of more broadly fo-
cused projects are found in Colombia and Kenya. Still another format
emphasizes first one crop and then a series of new activities as the
first goals are satisfied. The relevant questions here are the extent
to which: 1) program resources are available; 2) complementarities ex-
ist between production and other activities; 3) a broad thrust is so
blunted as to lose its catalytic effect; and 4) farmers can be counted
on to take a longer view as their incomes increase.
Finally, even if Puebla were totally successful in raising maize
production, is this in itself enough to promote shifts into other acti-
vities? These activities should permit the area's farmers to market
more of those resources in relative abundance. Obviously, maize must
be regarded as a vehicle for impelling still further change. This can
occur as the new skills acquired in producing high-yielding maize and
the accompanying higher incomes operate to relax restrictions on pro-
duction, most notably those stemming from risk, management, and liqui-
But where does the farmer turn for relevant information on alter-
natives? This question leads some to argue that projects like Puebla
must incorporate succeeding crops as well as the basic crop. Others
use this issue to reinforce their argument for ties with national pro-
grams. With close-working relationships, a more responsive national re-
search and extension force merges as a result of the experiences gained
in promoting cereals production.
More research on Plan Puebla is underway, and more is in the plan-
ning stage. The hope is that factors influencing adoption will be iden-
tified more precisely. Beyond this, given the importance attached to
diffusing the new technology among disadvantaged farmers, other Puebla-
like ventures should place more emphasis on experiments aimed at re-
solving questions on the adoption of new technology.
1/ Centro Internacional de Mejoramiento de Maiz e Trigo, El Plan Pue-
bla, 1967-69. (Centro Internacional de Mejoramiento de Maiz y
Trigo, Mexico, D.F., 1969), p. 2.
2/ A formal study of the Plan's impact on institutional behavior was
launched in 1972 by Heliodoro Diaz, formerly the Plan's coordinator
and currently a Ph.D. student in the Land Tenure Center, University
3/ Martin del Campo, Antonio, "La Capacidad del Agricultor para Adop-
tar Technologia: Un Caso de Estudio Productores de Maiz de Tem-
poral en el Area del Plan Puebla", Colegio de Postgraduados, Cha-
pingo, Mexico, 1972 (unpublished M.S. thesis).
4/ Cano, Jairo, and Delbert Myren, "Benefit Cost Analysis of the Pue-
bla Project", in Delbert Myren, ed., Strategies for Increasing
Agricultural Production on Small Holdings. Centro Internacional
de Mejoramiento de Maiz y Trigo, Mexico, D.F., 1970.
5/ Cano, Jairo and Don Winkelmann, "Plan Puebla: Analise de Benefi-
cios e Costos", El Trimestre Economico, Vol. 39, October/December
1972, pp. 783-792.
6/ It should be noted that the costs reported by Cano and Winkelmann
were lower than those given in Table 4. The principal difference
occurs in the first year, when indirect costs as reported by CIMMYT
were higher than those in Cano and Myren. Adjusting for this dif-
ference, the ratios seen in Table 5 would be revised downward, but
would remain greater than one in every case.
7/ Moscardi, Edgardo, "Riesgo y Transferencia de Techologia Estudio
para el Caso del Plan Puebla", Colegio de Postgraduados, Chapingo,
Mexico, 1972 (unpublished M.S. thesis).
9/ Martin del Campo, Antonio, op. cit.
10/ Moscardi, Edgardo, op. cit.
11/ In a random sample of 60 Plan participants, only 16 were found to
completely follow Plan recommendations. The standard used was to
be within 10 percent of the amount of fertilizer recommended, to
apply the fertilizer at the recommended times, and to be within 10
percent of the recommended seeding rate. Virtually all of the de-
partures occurred in the timing of the fertilizer applications and
in the seeding rate.
Dorantes, Jose Antonio Avila, "Evaluacion de Factores que Actuan
Como Restricciones en la Aplicacion de las Recomendaciones del Plan
Puebla", Colegio de Postgraduados, Chapingo, Mexico, Primavera de
1973 (unpublished M.S. thesis).
12/ Group formation may also be argued to lead to developments beyond
maize. By 1973, however, less than 15 percent of the groups were
engaging in formal activities outside of those focused on the pro-
duction of maize.
13/ Mbithi, P.M., "Innovation in Rural Development", Institute for
Development Studies, University of Nairobi (mimeograph).