A nonlinear probability analysis of the decision to insure municipal bonds

This dissertation research investigates the occurrence of insured municipal debt, and it attempts to model the insurance decision. The sample used is Louisiana municipal bonds issued in 1990 and 1991. In addition to using a different sample of bonds than has been done, more recent data is studied, a wider variety of security types are included, a different method of studying net benefits of insurance is used, and probability modelling is utilized.;The probability analysis results found several significant variables that help to explain the insurance purchase. These are the dollar amount of the issue, the term of the bond, and per capita levels of issuer debt, revenue, and expenditure. Additional determinants of the insurance decision were sought by including political variables. Since none of the variables tested were significant, more research needs to be done to ascertain whether other factors serve to affect the issuance of insured debt.;Finally, a comparison of results of the two analyses is made to gain additional insight into the portion of bond issues for which the insurance decision cannot be explained with the existing methodology and variables. Three factors seem to be present in issues that were misassigned in both exercises. These factors are relatively large premiums, use of net interest costs, and decreasing coupon payments.;The results of an analysis of the net benefits of insurance for the Louisiana sample are similar to those conducted in the past, but with additional insight gained. Scrutiny of the specific bonds for which negative net benefits are estimated indicates the size of the premium and use of net interest costs as possible influences. This finding deserves further investigation.