Notes

Outline

The Farm Portfolio Problem:  Part II Lecture VI

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.

"Hazell’s approach is two fold" 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.

"The variance of a particular..." The variance of a particular farming plan can be expressed as

"Hazell suggests replacing this objective..." Hazell suggests replacing this objective function with the mean absolute deviation

"Thus," 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

Slide 7

Slide 8

Slide 9

Slide 10

Slide 11

Slide 12

"Focus-Loss" 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

"L" 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

"Given this definition," Given this definition, the next step is to define the maximum deficiencies or loss arising from activity j. where rj* is the worst expected outcome.  Forexample, a crop failure may give an rj of -$100which would represent your planting cost

"Given this potential loss," Given this potential loss, the Focus-Loss scenario is based on restricting the largest expected loss to be above some stated level

"In the Anderson," In the Anderson, Dillon and Hardaker example:

"The choice of k=3..." The choice of k=3 is somewhat arbitrary. Two points about the Focus-Loss Allowing L -> -Infinity is the profit maximizing solution L can become large enough to make the linear programming problem infeasible. A Better Justification for k One alternative for setting k results from the notion that

"Thus," Thus, if we let tp be -1.96, the maximum loss would be 1.96 sj

"Direct Utility Maximization" Direct Utility Maximization To this point, we have discussed several alternatives to direct utility maximization which were based on efficiency criteria (as in the case of expected value-variance) or some ad hoc specification of risk aversion as in the case of focus-loss. Next, I want to discuss the application of a discrete form of expected utility analysis

Slide 21

"Parameterization of the Direct Expected..." Parameterization of the Direct Expected Utility Model First, we start by maximizing the expected profit given that the total acres do not exceed 1280.  This yields an 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.

"Assuming an interest rate of..." Assuming an interest rate of 12.5% yields an annual cash flow requirement of $135,891 to cover the interest payments. Further assuming a family living requirement of $50,000 yields a minimum cash requirement of $185,891.  Subtracting this from the equity yields a wealth constraint of

"A simple model for portfolio..." A simple model for portfolio choice can then be derived as

"In addition to rearranging the..." In addition to rearranging the sign on the objective function, we also must scale the problem.  Specifically, assuming an average payoff, the terminal wealth will be:

"In the original column," In the original column, the objective number is much smaller, which will lead to difficulties in the numerical optimization.  As a result, I have substituted wealth defined in thousands of dollars in the second column which helps scale the optimization problem.  In addition, I also scale the objective function directly.  In doing so, I try to get the optimal value of the objective function to be around -1 to -10.

"The optimal solutions are then..." The optimal solutions are then given by the crop rotations