THE BOND MARKET

Early in 1982 Judy Johnson, Vice-President of Finance of a large, private, investor-owned utility in the Northwest, was faced with a financing problem. The company needed to issue bonds both to pay off short-term debts coming due and to continue construction of a coal-fired plant. Judy's main concern was estimating the bond market. The utility needed a reliable forecast of the interest rate it would pay on the bonds at the time of bond issuance.

Judy called a meeting of her financial staff to discuss the bond market problem. One member of her staff, Ron Peterson, a recent Cal State EastBay graduate, said he thought a simple linear regression model could be developed to forecast the bond rates. Judy then asked Ron to have a report on her desk the following Monday.

Ron knew that the key to the development of a good forecasting model is the identification of independent or predictor variables that logically relate to the interest paid by utilities at the time of bond issuance. After discussing the problem with various people at the utility, Ron decided to investigate the following candidate predictor variables: a utility ratio of earnings to fixed charges, Treasury bond rates, the prime lending rate, the Dow Jones index of 50 utility stocks, and the number of days of sick leave used by the employees of the utility.

Ron gathered data he believed might correlate with bond interest rates for utility bond issuances during 1980 and 1981. This cross-sectional data was gathered for various different utilities in the US who issued bonds during those years and was not listed in any special order .

The next step was for Ron to run MINITAB using the input data. To summarize:

Response Variable: Y = interest rate paid by utility at the time of bond issuance

Predictor Variable Candidates ( Only one should be chosen )

Variable 1: X ₁ = U tility's ratio of earnings to fixed charges

Variable 2: X ₂ = U.S. Treasury bond rates for at the time of bond issuance

Variable 3: X ₃ = P rime lending rate at the time of issuance

Variable 4 : X ₄ = Dow Jones index of 50 utility stocks at time of issuance

Variable 5 : X ₅ = the number of days of sick leave used by the employees

The data file is in BONDF0 9 .mtw . Note that the data variables Y, X ₂ and X ₃ which represent percents are entered with the decimal point moved two places to the left e.g. 15.3% is written 15.3, not 0.153 .

Judy seeks a simple linear regression model to forecast the projected interest rate on bonds to be issued shortly. The utility’s current ratio of earnings to fixed charges is 2.95; U.S. Treasury bond rates are expected to be 13% while the prime-lending rate is expected to be 19% on the date of issue. The Dow Jones utility index is expected to be 210 on the date of issue and 300 sick leave days will have been used by employees. P lease note that the above information becomes irrelevant for those variables not chosen ; data is given for all variables so as not to influence your choice . .

Write a report (do not simply write down computer output) to Judy that provides her with:

1. T he bond market interest rates forecast with an 8 5 % prediction interval .
2. The model equation on which the forecast is based and a brief evaluation of its statistical performance.

3 An explanation of why you chose the predictor variable that was used in your model rather than one of the other five , based on logic and /or a statistics. Devote one sentence to each variable.

1. Determine if the regression model behaves in a way that agrees with financial logic, e.g. what is the effect of a one unit change in the relevant range in your predictor variable.
2. A comment on the appropriateness of using the exact same model equation ( same predictor variable with same regression coefficient and same constant) to forecast bond interest rates today in the year 2009 ! ( answer yes/no and why-please do not hedge by using words such as "could")

The report should not be too technical but it should explain clearly the basis for your c hoice of predictor variable. It should be at most 1.5 pages long , double spaced, font Times New Roman 14 point; in addition , attach as an Appendix all relevant computer printouts from the model that you chose.

Note: Judy generally does not look at the Appendix so be sure to include in the report all information that you want to convey to Judy, including any limitations of your analysis.

Price: $7.07

Solution: The downloadable solution consists of 4 pages, 307 words and 4 charts.
Deliverable: Word Document