Using Fiscal Indicator Systems to Predict Municipal Bankruptcies

Abstract

U.S. state governments over the years have developed a number of methods for predicting fiscal distress and defaults by their local governments. These systems encompass a range of conceptual understanding and practical approaches, and involve varying data requirements and analytic demands. A widespread prediction strategy is to compute indexes of fiscal stress based on financial and other indicators. This chapter assesses the predictive accuracy of three well-documented and carefully argued indicator systems, by computing indicator scores for a sample of three bankrupt general-purpose local governments and six matched jurisdictions that did not declare bankruptcy, over a ten-year period that included the years leading up to the three bankruptcy proceedings. The goal was to answer two primary questions. First, are the indicators all measuring the same thing? How closely do different indicator systems agree in their scoring of a given jurisdiction? Second, to what extent are the indicators predictively valid? How good a job do they do of predicting severe fiscal distress and of distinguishing between cities that did and did not enter bankruptcy? The system developed and currently used by Ohio’s Auditor of State performed well. The system developed in 2002 for Michigan’s State Treasurer (no longer used) and a system subsequently proposed by academic researchers as a potential improvement to it (not implemented) both performed less well.

Appendix

This appendix provides more information about each monitoring system’s specific indicators and index computation.

Table 14.6 lists the indicators used in Michigan’s system (as described in Kleine et al. 2002, 2003; Kloha et al. 2005a). All three of the environmental indicators (numbers 1–3) make explicit comparisons over time. The system’s six financial indicators (numbers 4–9 in the table) include three (4, 7, and 9) that are peer-referenced, and one of the three constant-dollar interperiod comparisons also uses a peer benchmark (number 3). We calculated those indicators for the California sample by using the sample’s mean values as the benchmarks. The index is computed by summing the individual indicator scores. An index score of 8–10 points indicates a “fiscal emergency”; 6 and 7, a “fiscal warning”; 5, a “fiscal watch”; and 0–4, “fiscally healthy” or “neutral.”

Table 14.6 Component indicators of the Michigan Fiscal Stress Index

The system of indicators proposed by Crosby and Robbins (2010, 2013) as an improvement over the Michigan system uses ten indicators to generate an additive index than can range from 0 to 10 points. Table 14.7 lists the indicators. Of the ten indicators, nine draw information from the government-wide statements, and one (number 6) is for the general fund only. None makes comparisons across time. Two (numbers 7 and 8) are implicitly peer referenced (D. Robbins, personal communication) by setting their benchmarks at approximately the means for all Michigan jurisdictions. For the California application we modified the benchmark values to the means for our sample cities: $1071 for number 7, and $1715 for number 8. An index score of 8–10 points indicates “fiscal stress”; 5–7, a “fiscal watch”; and 0–4, “neutral.”

Table 14.7 Component indicators of the Crosby and Robbins Fiscal Stress Index

Ohio’s indicator system, summarized in Table 14.8, includes 16 financial indicators and one behavioral indicator. Of the financial indicators, ten are drawn from funds statements (we used government-wide information as a substitute for two of these, because of data-availability limitations), five from the government-wide statements, and one from the notes (at least it was for our project; Ohio appears to maintain a database of financial information inputted by local governments). Eight of the indicators make interperiod comparisons, of which three assess short-term trends (three fiscal periods). Ohio’s indicator system in its native form does not generate an index score. If six indicators are at “critical” values or eight indicators have “cautionary” or “critical” values, the subject local government is judged to have an “elevated risk” of fiscal distress. For purposes of comparing systems, we rescaled the Ohio system to an index score ranging from 0 to 10 points, with five point representing the threshold for “elevated risk.”

Table 14.8 Component indicators of the Ohio Auditor of State Index