Professional Documents
Culture Documents
A
s a consequence of increased Florida growers is pitaya, also known as 20–30 years, ensuring that with proper
foreign competition and declin- dragon fruit, a climbing vine of cactus care, the crop can provide a steady
ing returns to traditional agri- species native to the tropical forest re- stream of income (Crane and Balerdi,
cultural commodities, many growers in gions of Mexico, Central America, and 2005; Gunasena et al., 2006).
southern Florida have embarked on a
search for viable alternative agricultural
commodities. One commodity that Units
has gained the attention of southern To convert U.S. to SI, To convert SI to U.S.,
multiply by U.S. unit SI unit multiply by
Tropical Research and Education Center, University 0.4047 acre(s) ha 2.4711
of Florida, 18905 SW 280 Street, Homestead, FL 0.3048 ft m 3.2808
33031 0.0283 ft3 m3 35.3147
1
Associate Professor 3.7854 gal L 0.2642
2
2.54 inch(es) cm 0.3937
Intern 0.4536 lb kg 2.2046
3
Corresponding author. E-mail: eaevans@ufl.edu. 1.1209 lb/acre kgha–1 0.8922
of commercial viability over the as- a certain outcome. Accordingly, the taking into consideration the likely
sumed time period. As pointed out by stochastic variables chosen in this impact that increased imports could
Lien (2003), the traditional (deter- model were yield and price. In view have on the prices, the minimum, mid-
ministic) approach to budgeting, of the fact that historical data were dle, and maximum prices of $0.50/lb,
based on fixed-point estimates of pro- not available to empirically determine $0.85/lb, and $2/lb, respectively, were
duction and prices to predict point the distribution functions for either used to generate the price variable. The
estimates of financial results, seldom of these variables, it was necessary to stochastic yield multiplied by the pack-
holds in reality. While sensitivity anal- assume such distributions. Accord- out rate times the price equals the
ysis is a common response to this ingly, the uniform distribution and stochastic gross receipts.
problem, it too is limited in that the the GRKS distributions were chosen The costs of all recurrent inputs,
analysis does not give any indication to model the yield and price variables, with the exception of labor, were
of the likelihood of a particular result respectively. Among other things, assumed to increase by 1% per year.
being achieved and invariably involves these distributions were chosen, Many of the growers indicated that
considering changes in only one or a given the limited amount of informa- the cost of labor was increasing at
couple of variables at a time. A better tion needed to generate the random a faster pace compared with the cost
approach is to use stochastic budget- variable. The uniform distribution is of other inputs. Among other things,
ing, which accounts for some of the one for which the probability of oc- this was due to the increase in the
main uncertainties in the evaluation currence is the same for all possible minimum wage and to policies to
and provides indications of the distri- outcomes. The population of a contin- crack down on undocumented workers.
bution of the outcomes. uous uniform distribution is defined As such, the cost of labor was allowed
The stochastic budget used by a minimum and a maximum value. to grow at an average annual rate of
follows the approach outlined by The GRKS distribution is a two- 2% over the period. It was also assumed
Richardson (2006) and involves the piece normal distribution with 50% that the five acres of land could be
following steps. First, probability dis- of the weight below the middle value purchased for $200,000. The residual
tributions were assigned to the vari- and 2% less than the minimum, and value of the land was kept at $200,000
ables affected by the risk factors 50% above the middle value and 2% and the residual value of irrigation
outlined before. Second, the stochas- above the maximum. The population system was estimated at $2650 or
tic values sampled from the probabil- can therefore be defined given the 20% of the initial value. The annual
ity distributions were used in the minimum, middle, and maximum net cash flow was calculated as the
accounting equations to calculate pro- values. The distribution is used in stochastic gross receipts plus residuals
duction, receipts, and the KOVs. Third, place of a triangular distribution less the total cost (investment plus
the completed stochastic budget was when one knows only minimum in- recurrent costs).
simulated 500 times using the random formation about the random variable To compare the stochastic re-
values for the risky variables. The re- (Richardson, 2006). sults with the deterministic results,
sults of the 500 samples provided the For the yield variable, given that the average values for the period
information to estimate the empirical the fruit of the pitaya can be harvested 2009–2010 were computed based on
probability distribution for the unob- beginning in year 2 of cultivation, the information obtained from growers
served KOVs. This information can be with production stabilizing by year 4, and packinghouses. The average price
further analyzed using a cumulative and based on discussions with growers was $1.25/lb and average marketable
distribution function (CDF) of the and extension specialists, it was de- yields by year were 9375 lb/acre (year
KOVs, such as NPV and IRR, which cided to model per area yield (lb/ 2), 14,875 lb/acre (year 3), 18,450
shows the probability of these vari- acre) as follows: U2(10,000, 15,000), lb/acre (years 4–9), and 19,475 lb/
ables being less than a given value. U3(15,000, 20,000), and U4+(20,000, acre (years 10 onward). All other vari-
The probability distribution of the 22,500), where Ui(a, b) represents ables remained the same.
CDFs for NPV and IRR generated by the uniform distribution in year i and
the stochastic model provides a great a and b represent the minimum and Results and discussion
deal more information about the eco- maximum values (lb/acre), respec- The results of the deterministic
nomic viability of the proposed business tively. Hence, in year 2, it is assumed financial analysis (Table 1) indicate
than does the deterministic analysis. that yield will vary from a minimum of that the operation was financially fea-
Thus, for any chosen value of NPV or 10,000 lb/acre to a maximum of sible (profitable) on the basis of the
IRR, the respective CDFs would in- 15,000 lb/acre. Moreover, it was assumptions and using the average
dicate the probability of being below further assumed that the pack-out price and yields. Table 1 shows an
that value. The model was programmed rates (the percentage of crop har- NPV of $505,255, indicating that the
in Excel (Microsoft, Redmond, WA) vested that is marketable) would vary discounted benefits (present value of
and simulated using the Excel Add-In, as follows: year 2 (75%), year 3 (80%), the returns) over the 20-year period
Simulation & Econometrics to Ana- years 4–9 (90%), and years 10 onward far exceeded the discounted costs
lyze RiskÓ (Simetar, College Station, (95%). Achieving a 100% pack-out rate (present value of the costs) over the
TX) is difficult. same period. The IRR was calculated
As pointed out by Hardaker et al. The price variable was assumed at 16.2%, implying a profitable return
(2004), the variables chosen to be to follow a GRKS distribution. On on investment. Among financial ex-
stochastic should be those that are the basis of information received from perts, the rule of thumb is that a return
likely to have the biggest impact the local packinghouses that pur- of more than 15% to investment can
on the level of risk associated with chased the majority of the crop and be considered as quite favorable. The
248 • April 2011 21(2)
Table 1. A comparison of profitability indicators associated with establishing obtained in the case of changes in
and operating a 5-acre (2.0 ha) pitaya orchard in southern Florida using investment expenditures and seasonal
deterministic and stochastic budget models. inputs, although the impact on the
Modelx IRR was of lesser magnitude. For
Item z
Deterministic Stochastic instance, a 10% increase in either
variable would cause the IRR to de-
Main assumptions: cline from 16.2% to 14.7% and 15.2%,
Enterprise scale (acres) 5 5 respectively.
Initial investment ($) 393,744 393,685 A threshold analysis, or stress
Average recurrent input costs ($/acre) 7,896 9,094 test, was also conducted to determine
Average (expected) price ($/lb) 1.25 1.00 the extent of changes in any of the
Average (expected) marketable yield in (lb/acre) variables (yield, prices, investment
Year 2 9,375 9,375 expenditures, or seasonal inputs) that
Year 3 14,875 14,875 would be necessary to make the NPV
Years 4–9 18,450 18,225 equal to zero. As alluded to earlier, in
Years 10–20 19,475 19,238 light of the inherent risks involved in
agriculture, an NPV of zero at a given
Cash flow analysis: interest rate implies that the opera-
Net present value (NPV) at 5% ($) 505,255 246,227 tion is not profitable given that similar
Internal rate of return (IRR) (%) 16.2 10.8 returns could be obtained elsewhere
Benefit cost ratio (BCR) at 5% 1.93 1.54 from a less risky alternative. The re-
sults of the analysis as shown in Table
Sensitivity (scenario) analysis: 1 indicate that yield or prices would
NPV have to decline by 39.1%, whereas in-
at 3% 707,871 391,804 vestment expenditure or seasonal in-
at 7% 353,319 138,542 puts would have to increase by 136.8%
IRR if or 106.5%, respectively.
Yield or prices decreased by 10% (%) 13.6 — Table 1, in addition to showing
Investment expenditure increased by 10% (%) 14.7 9.7 the results of the deterministic model,
Inputs increased by 10% (%) 15.2 9.8 also shows the results for the stochastic
values based on the expected (average)
Threshold analysis: values for the stochastic variables,
Operation become unprofitable if: whereas Fig. 1 shows the results of
Yield or prices decreased by (%) 39.1 — the CDF of the stochastic NPV and
Investment expenditure increased by (%) 136.8 66.7 the cut-off point for the NPV ob-
Inputs increased by (%) 106.5 52.0 tained from the deterministic model
(indicated in the chart by the vertical
Breakeven analysis: line). Compared with the determinis-
Average payback period (years) 9 12 tic NPV value of $505,255, the sto-
z
Enterprise scale, size of the pitaya orchard; initial investment, amount of money required to start a business or farm chastic average NPV was $246,227,
operation, which includes the costs of land and capital equipment; expected value, the long-run average or the
weighted average of all possible values a random variable can take; NPV, a measure used to determine if a business with a standard deviation of $93,438
that results in a stream of benefits and costs in the future is worth undertaking, if positive the investment should be and a coefficient of variation (CV) of
made otherwise it should not; IRR, a measure of the rate of return used to evaluate the desirability of investing in
a business, the higher the value the more likely the business will be profitable; BCR, a measure used to compare the
37.9%. The minimum and maximum
value of benefits to the costs, a ratio greater than 1 is desirable; payback period, a measure use to indicate the length NPVs were –$19,032 and $541,286,
of time required for the accumulated receipts to cover completely the initial investment, a shorter period is respectively (Table 2). In general, the
preferred to a longer period.
y
1 acre = 0.4047 ha, $1/acre = $2.4711/ha, $1/lb = $2.2046/kg, 1 lb/acre = 1.1209 kgha–1. stochastic budget supports the deter-
x
Deterministic, model in which all variable are fixed or known with certainty, for example the price is known with ministic budget, namely that the in-
certainty; stochastic, model in which some of the variables are random, that is, they are determined by chance and vestment is economically viable, but it
hence are not fixed or predictable.
provides additional information on
the distributions of the net returns.
For instance, it shows that the NPV
benefit–cost ratio was calculated at To test the sensitivity of the dis- from the deterministic model is to the
1.93, implying that for each dollar count rate used (5%), calculations were extreme right of the distribution (Fig.
invested, the operator would earn also done using rates of 3% and 7%. As 1). It also indicates that even though
almost twice that amount. The break- can be seen in Table 1, the operation highly unlikely, there is a small chance
even analysis indicates that it would remained profitable, with NPVs of (probability of less than 1%) that the
take 9 years on average for the accu- $707,871 and $353,319, respectively. investment could be unprofitable
mulated revenues to equal the accu- Table 1 also shows that if yield or over the period. This is reflected by
mulated cost (i.e., for the accumulated prices were to fall by 10%, the IRR the small portion of the CDF that
net cash flow to equal zero). In other would change somewhat, falling from lies to the left (negative NPV) of the
words, it would require 9 years for 16.2% to about 13.6%. This suggests y-axis.
the investor to recoup all of his/her that the results were sensitive to changes A similar analysis was performed
investment. in price or yield. Similar results were for the IRR (Fig. 2). The data used
• April 2011 21(2) 249
PRODUCTION AND MARKETING REPORTS