MSc

Cocoa (Theobroma cacao L.) is a very important crop as it provides food, income,
employment and resources for poverty reduction. It ensures livelihood for millions of
small holder farmers and offers raw material for the multibillion global chocolate
industries. Despite the fact that Kerala has enormous potential in terms of suitable
agricultural land, cocoa has failed to become a significant crop. As its domestic
production is not sufficient to meet the increased demand, the industry has to resort to
substantial imports. So, a comprehensive study titled “Multiphase analysis of cocoa
production in Kerala” has been made on different aspects of cocoa cultivation,
management practices, production and the constraints faced by actual growers.
The trend analysis and forecasting of yearly area, production and productivity of cocoa
in Kerala using advanced time series models employed on the data for the period from
1980-2017 revealed a distinct quadratic trend for the area under cocoa, having an
increasing trend now and more or less linear stochastic trends for production and
productivity. The Holt’s exponential smoothing model was identified as the best to
predict yearly area under cocoa with an adjusted R2 equal to 0.94. The yearly production
of cocoa could be well modelled by ARIMA (0,1,1) with an adjusted R2 = 0.72. By
incorporating area under cocoa as an independent variable, ARIMAX (0,1,0) model
could improve the R2 to 0.84 to predict the yearly production of cocoa. The productivity
of cocoa seemed to be constant for several years (0.45tonnes/ha) which was well
predicted through the simple exponential smoothing model with an adjusted R2 = 0.84.
Evaluation of the performance of 100 selected cocoa hybrids in the Cocoa Research
Centre, College of Horticulture, KAU, Vellanikkara showed that the peak average
monthly yield was in the month of November (18.14pods) followed by the yield in
October (18.04) and December (14.56). A pattern of biennial tendency persisted for the
yearly yields of the hybrids. The results of General linear model repeated measures
ANOVA highlighted the existence of a significant Time x Factor interaction with a
partial eta squared equal to 0.14 where factor denotes different subgroups of cocoa
hybrids with homogeneous yield. After the first harvest, the peak average yield was
noticed during the fifth year irrespective of different low and high yielding groups.
The income from cocoa farming depends on healthy pods harvested. So, an attempt was
also made to account for the frequency of number of infected pods from each tree and
it could be well demonstrated by geometric distribution which is a special case of
Negative binomial distribution. Owing to the fact that the infected pods might be the
outcome of external factors like weather variables, the influence of those factors with
cocoa yield was also investigated. A stepwise regression of yield on previous five
month’s accumulated weather variables resulted in a parsimonious prediction equation
with total number of rainy days as the single regressor which could explain 66% of the
variation in yield. The adjusted R2 could be enhanced to 69% by incorporating
maximum temperature as the second most important regressor. The vide variation
realised in the average monthly yield of cocoa hybrids could be well captured through
SARIMA (1,0,0) (1,1,0)12 model with an adjusted R2 = 0.92.
An empirical analysis to identify the factors perceived by farmers to influence their
cocoa production and ultimately their income was performed taking a total sample of
100 farmers from Veliyamattom Panchayat of Idukky district and Iritty Panchayat of
Kannur district. From a path analysis through structural equation modelling several
linear regression equations could be generated simultaneously leading to prediction
equations for cocoa yield and income. The final model iterated resulted in goodness of
fit measures viz; comparative fit index = 0.96 and Tucker Lewis index = 0.94. Price of
cocoa turned out to be the most prominent factor which contributed to the income of a
cocoa farmer highlighting the importance of fixing the marketing price of cocoa.
Second factor was yield per tree which was the outcome of good quality seedlings,
efficient cultivation practices, plant protection and disease management measures,
protection from rodent attacks etc. Importance of access to credit which would help to
overcome the problems of lack of capital was emphasised. Financial problems such as
inability to get assistance from financial institutions, lack of proper marketing facilities
including drying and fermentation facilities of cocoa beans also were noticed.
Probit analysis identified the factors viz; age of the farmers, land holding size,
experience in cocoa cultivation, membership in organisations like Krishibhavan,
farmer’s club, Cooperative society, Banks, SHGs etc. and frequency of contact with
extension personnel to be significant for decision making to implement plant protection
measures which were inevitable for successful crop management and ultimately leading
to the net income of farmers.
The yield gap analysis revealed that as against the potential yield (dry bean weight) of
4kg/tree/year, the national average yield from cocoa farmers was only 2.5 kg/tree/year
resulting in a yield gap of 37.5% which need adequate attention.