Association for Public Policy Analysis & Management

Scholars have applied regression discontinuity designs in the study of local bond elections to examine the impacts of bond election passage on such outcomes as property values (Cellini, Ferreira, and Rothstein, 2010) and student achievements (Hong and Zimmer, 2016). When a bond proposal passes voter referendum, issuing a bond and building infrastructure are the two direct consequences. While issuing a bond changes the amount of debt outstanding, infrastructure investments may strengthen the fiscal or tax bases of a jurisdiction. Both mechanisms can change credit risks of bond issuers. While debt outstanding may mechanically increase credit risks, capital investments may have a nonlinear impact. The effects of capital investments on credit risks can have an inverted-U shape. Credit risks reach the lowest when capital investments are neither too “low” nor too “high”. Changes in credit risks can be measured with bond yields in secondary bond market and credit ratings of bond issuers.

This paper attempts to estimate the causal impacts of debt outstanding and capital investments on municipal credit risks with a regression discontinuity approach. The bond election data of Texas are available for school districts, cities and counties between 2003 and 2016, with a total number of bond elections of 3,546. Localities passing bond elections see an increase in their amounts of debt outstanding and levels of capital investments, while those failing the election would not. Using a regression discontinuity design, I compare the bond prices and credit ratings of local governments in the vicinity of passing a bond election. The running variable is the percentage of votes for the bond proposal in each bond election. The outcome variables are municipal bond prices measured as bond yields in the secondary market and credit ratings of general obligation bonds for each issuer. Viewing credit risk as a key indicator of fiscal health, this study can help understand whether local governments have borrowed too much or too little for capital investments.