MOTAD and Focus Loss - University of Florida
Download
Report
Transcript MOTAD and Focus Loss - University of Florida
Farm Portfolio Problem: Part II
Lecture XIII
MOTAD
Hazell,
P.B.R. “A Linear Alternative to
Quadratic and Semivariance Programming
for Farm Planning Under Uncertainty.”
American Journal of Agricultural
Economics 53(1971):53-62.
Fall 2004
Farm Portfolio Problem II
2
Hazell’s approach is two fold. He first sets out
to develop review expected value/variance as a
good methodology under certain assumptions.
Then he raises two difficulties.
The
first difficulty is the availability of code to solve
the quadratic programming problem implied by EV.
The second problem is the estimation problem.
Specifically, the data required for EV are the mean
and the variance matrix.
Fall 2004
Farm Portfolio Problem II
3
The variance of a particular farming plan can be
expressed as
1 s
x j xk
(chj g j )(chk gk )
j 1 k 1
s 1 h 1
n
n
1
chj x j g j x j
s 1 h 1 j 1
j 1
s
n
n
2
2
Fall 2004
Farm Portfolio Problem II
4
Hazell
suggests replacing this objective
function with the mean absolute deviation
s
n
1
A chj g j x j
s h 1 j 1
Fall 2004
Farm Portfolio Problem II
5
Thus, instead of minimizing the variance of the
farm plan subject to an income constraint, you
can minimize the absolute deviation subject to
an income constraint. Another formulation for
this objective function is to let each observation
h be represented by a single row
n
yh chj g j x j
j 1
n
yh yh chj g j x j
j 1
Fall 2004
Farm Portfolio Problem II
6
min sA yh yh
s
x
n
h 1
st chj g j x j yh yh 0
j 1
n
g x
j 1
j
j
j
bi
n
a x
j 1
Fall 2004
ij
Farm Portfolio Problem II
7
s
min sA yh
x
n
st
j 1
h 1
chj g j x j yh 0
n
g x
j 1
j
j
j
bi
n
a x
j 1
Fall 2004
ij
Farm Portfolio Problem II
8
Table 1. Hazell’s Florida Farm
Obs.
Carrots
Celery
Cucumbers
Peppers
1
292
-128
420
579
2
179
560
187
639
3
114
648
366
379
4
247
544
249
924
5
426
182
322
5
6
259
850
159
569
Average
253
443
284
516
Fall 2004
Farm Portfolio Problem II
9
Obs.
Carrots
Celery
Cucumbers
Peppers
1
39
-571
136
63
2
-74
117
-97
123
3
-139
205
82
-137
4
-6
101
-35
408
5
173
-261
38
-511
6
6
407
-125
53
Fall 2004
Farm Portfolio Problem II
10
y1 y2 y3 y4 y5 y6
min
x
x1
x2
x3
x4
25 x1 36 x2 27 x3 87 x4
x1
x2
x3
10000
x4
39 x1 571 x2 136 x3 63 x4 y1
74 x1 117 x2 97 x3 123 x4
139 x1 205x2 82 x3 137 x4
0
y2
y3
6 x1 101 x2 35 x3 408 x4
173 x1 261 x2 38 x3 511 x4
6 x1 407 x2 125 x3 53 x4
253 x1 443 x2 284 x3 516 x4
Fall 2004
200
Farm Portfolio Problem II
y4
y5
y6
0
0
0
0
0
0
11
Focus-Loss
Two
factors make Focus-Loss acceptable
First, like Hazell’s MOTAD, the Focus-Loss
problem is solvable using linear programming.
Second, Focus-Loss has a direct appeal in that
it focuses attention on survivability
The
first step in the Focus-Loss
methodology is to define the maximum
allowable loss
Fall 2004
Farm Portfolio Problem II
12
n
L E ( z) zc E (c j ) x j E ( F ) zc
j 1
L
- Maximum allowable loss
E(z) - Expected income for the firm
zc - Required cash income
E(cj) - Expected income from each crop, j
xj
- Level of the jth crop (activity)
E(F) - Expected level of fixed cost
Fall 2004
Farm Portfolio Problem II
13
Given this definition, the next step is to define
the maximum deficiencies or loss arising from
activity j.
rj E (c j ) r
*
j
where rj* is the worst expected outcome. For
example, a crop failure may give an rj of -$100
which would represent your planting cost
Fall 2004
Farm Portfolio Problem II
14
Given this potential loss, the Focus-Loss
scenario is based on restricting the largest
expected loss to be above some stated level
L
rr x j
k
Fall 2004
Farm Portfolio Problem II
15
max 72 x1 53.4 x2 88.8 x3 200
x
x1
x2
x3
x1
x3
30 x1 20 x2 40 x3
5 x1 5 x2 8 x3
60 x1
44.5 x2
74 x3
Fall 2004
Farm Portfolio Problem II
12
8
400
80
1
L 0
3
1
L 0
3
1
L0
3
16
The
choice of k = 3 is somewhat arbitrary.
Two points about the Focus-Loss
Fall 2004
Allowing L - the Focus-Loss solution is the
profit maximizing solution.
L can become large enough to make the linear
programming problem infeasible.
Farm Portfolio Problem II
17
A Better Justification for k
One
alternative for setting k results from the notion
that
r j t p j
*
j
Thus,
if we let tp be -1.96, the maximum loss would
be 1.96 j
L
j t p x j k
Fall 2004
Farm Portfolio Problem II
18
Direct Expected Utility
We
have been discussing several
alternatives to utility maximization based on
efficiency criteria or ad hoc specifications
of risk aversion as in the case of focus-loss.
One alternative is direct use of expected
utility.
Fall 2004
Farm Portfolio Problem II
19
Table 2. Data for Direct Utility
Maximization
Corn
Soybeans
Wheat
Observation 1
176.24
94.81
97.09
Observation 2
232.93
114.39
120.18
Observation 3
273.01
144.50
108.75
Observation 4
221.59
114.32
87.48
Observation 5
-7.87
97.22
100.46
Observation 6
247.59
126.41
108.34
Observation 7
226.79
113.49
98.16
Observation 8
250.11
123.27
107.60
Observation 9
255.99
136.15
102.81
Observation 10
246.91
131.04
104.68
Average
212.33
119.56
103.56
Fall 2004
Farm Portfolio Problem II
20
Parameterization of the
Expected Utility Model
Total
acres do not exceed 1280.
Annual profit of $271,782.
Amortizing this amount into perpetuity using a
discount rate of 15% yields a total value of
$1,811,880.
Assuming
the debt-to-asset position of the
farm is 60%, the value of the asset
represents equity of $724,752 and debt of
$1,087,130.
Fall 2004
Farm Portfolio Problem II
21
Assuming
an interest rate of 12.5% yields
an annual cash flow requirement of
$135,891 to cover the interest payments.
Assuming a family living requirement of
$50,000 yields a minimum cash
requirement of $185,891.
Wi 176.24 x1 94.81 x 2 97.09 x 3 538861
Fall 2004
Farm Portfolio Problem II
22
1 W1b 1 W2b
1 W10b
10 b 10 b
10 b
max
x
x1
x2
1280
x3
176.2 x1 938
. x2 97.1 x3
232.9 x1 114.4 x2 120.2 x3
W1
W2
538,861
538,861
246.9 x1 1310
. x2 104.7 x3
Fall 2004
Farm Portfolio Problem II
W10 538,861
23
Table 4. Portfolio from
Expected Utility
r
x1
x2
x3
-0.001
929.12 350.88
0.00
239,231.70
75,923.56
-0.1
882.59 397.41
0.00
234,915.10
72,865.46
-1.0
532.69 747.31
0.00
202,455.70
50,130.77
-10.0
Fall 2004
4.93
0.00
284.59
Farm Portfolio Problem II
24