Tilapia Farm Business Management & Economics
The same procedure will be used to
dene a normal distribution for the
price of feed (cell I14). In this instance,
a mean of 12 and a standard
deviation of 1.40 will be used. The
graph will indicate the prices KSh
7.80 and KSh 16.20 as boundary
values. Finally, a triangular distribution
will be developed for the FCR (cell I13).
Usually, triangular distributions
are selected when there is insufcient
background information to dene a
normal distribution; nevertheless,
the user might have a clear idea of
the value most likely to occur and the
values that would mark the boundaries
of the distribution (maximum and
minimum). For the FCR, it will be
assumed that the likeliest conversion
factor is 3.46, but could vary between
2 and 5. As it was explained before,
the cell I13 must be selected, and
then the icon “Define Assumption”
must be clicked on. Next, the
triangular distribution is selected
and the minimum, maximum, and
likeliest values are entered.
Up to this point, probability distributions
have been dened for three budget
items (price of tilapia, feed price, and
FCR). The goal of the exercise is to
measure the effect on net returns/ha
of the variability in the mentioned
items. To achieve this, the user must
select the cell S49, which contains the
formula that calculates net returns/ha,
and then click on the icon “Dene
Forecast” (the third icon from left on
the Crystal Ball bar). A new window
will pop up requesting information
on the cell. The name of the cell must
be specified in the first box (Net
returns/ha) and the units in the
second box (KSh). After clicking on
“OK,” the window will disappear
and the background of the cell S49
will change to blue, which indicates
that Crystal Ball has been instructed
to evaluate the effects of the variability
in tilapia price, feed price, and FCR
on the net returns/ha.
At this point, everything is ready to
run a simulation with Crystal Ball.
This program repeatedly recalculates
the enterprise budget until an
iteration limit is reached (the default
specication is 500 trials). Then, the
enterprise budget will be recalculated
a total of 500 times, but each iteration
will use different values for the
price of tilapia, feed price, and FCR.
Crystal Ball selects the values for
each iteration based on the defined
probability distributions through a
sampling procedure known as Monte
Carlo Sampling Method. As the
enterprise budget is recalculated 500
times, instead of obtaining a single
estimate of net returns/ha (as it is
typically done in a spreadsheet), a
whole range of possible outcomes
occurring with different levels of
probability will be shown.
To get started with the simulation,
the user must click the icon “Start
Simulation”—the eleventh icon from
the left on the Crystal Ball bar. Crystal
Ball will conduct the 500 iterations,
which may take a few seconds. At the
end, a graph will appear on screen
displaying the resulting probability
distribution for net returns/ha. Figure
9 shows that net returns/ha can
range from KSh – 100,000 to KSh
600,000 (the range is between
KSh -74,333 and KSh 586,000 to be
more exact). These results can be
interpreted as follows: net returns/ha
could be as low as KSh -74,333 if,
as an unfortunate event, low tilapia
prices coincide with high feed prices
and high FCR. Similarly, net returns/ha
could be as high as KSh 586,000 if tilapia
prices are high and feed cost is low.
Sometimes it may be useful to determine
what the probability is that net
returns exceed or do not reach a certain
arbitrary level. For instance, the user
may be interested in assessing the
likelihood of obtaining net returns above
300,000 KSh/ha. To do this, the
user must define the limits of the
relevant portion of the distribution.
These limits can be specied in the
boxes located at the left and right
sides of the graph. Enter 300,000 in
the left box. Crystal Ball automatically
calculates the certainty level
associated with the range KSh 300,000–
586,000, which is 32%. In other words,
there is a 32% likelihood of obtaining
net returns above KSh 300,000.
In summary, Crystal Ball is a very
useful tool to measure the amount of
uncertainty, or risk, of a specific
operation in a fast and efficient
fashion. In the example, the risk implicit
in the operation of the 1-ha tilapia
farm was described in terms of
variability in tilapia and feed prices.
However, this basic model can be
expanded to incorporate the effects
of variability in other budget items,
such as pond yields. The incorporation
of additional factors of uncertainty
in the model will result in a more
accurate measurement of the risk
levels associated with the activity.
3 3
. . .