Assessing Municipal Bond Default Probabilities

Abstract

In response to a request from the California Debt and Investment Advisory Commission, we propose a model to estimate default probabilities for bonds issued by cities. The model can be used with financial data available in Comprehensive Annual Financial Reports that cities are required to publish. The study includes modeled default probability estimates for 261 California cities with population over 25,000. Our model relies on case study evidence, logistic regression analysis of major city financial statistics from the Great Depression -the last time a large number of cities defaulted -as well as logistic regression analysis of more recent city financial statistics.Independent variables in our model include (1) the ratio of interest and pension expenses to total revenue, (2) the annual change in total revenue, (3) the ratio of general fund surplus (or deficit) to general fund revenues and (4) the ratio of general fund balance to general fund expenditures.

Local Public Finance Dynamics and Hurricane Shocks

“…When considering the magnitude of this change in default risk, it is important to consider that municipal bond ratings change minimally on average. Approximately 90% of municipal bond ratings remain unchanged over a two-year period (Holian and Joffe, 2013). Furthermore, ratings agencies assess bond risk only for municipalities that pay them to do so.…”

“…Furthermore, ratings agencies assess bond risk only for municipalities that pay them to do so. In their literature review on the determinants of municipal bond default, Holian and Joffe (2013) find that cities that choose to be rated are more likely to have a lower default risk that cities that do not choose to be rated. This implies that hurricanes likely increase the 10-year default risk of un-rated municipal debt more than 0.1 percentage points documented here.…”

Is Good Credit Good? State Credit Ratings and Economic Insecurity, 1996–20121

“…We use the general obligation credit ratings for state governments as our focus, consistent with previous studies analyzing government credit ratings (Holian and Joffe 2013). Though there are three primary credit‐rating agencies, we use credit ratings issued by S&P because S&P is the largest agency and has the most extensive coverage of state governments during our period.…”

“…S&P credit ratings are based on an alpha-symbolic ordinal scale. We convert S&P's ratings into a numerical ordinal scale consistent methodologically with other studies of credit ratings (Holian and Joffe 2013) by assigning values from 1 to 22 to reflect the number of steps in S&P's rating scale. Our numerical scale is linear with one-notch differences in S&P's scale corresponding to a numerical difference of one.…”

Local public finance dynamics and hurricane shocks

“…Approximately 90% of municipal bond ratings remain unchanged over a two-year period (Holian and Joffe, 2013). Furthermore, ratings agencies assess bond risk only for municipalities that pay them to do so.…”