The Effects of Protest Format and Agent Use on Residential Appeals Adjustments

This study examines how protest format and use of an agent affect the magnitude of the adjustments obtained in property tax appeals. The sample includes all single-family residential properties in Harris County, Texas for which a property tax protest was filed in 2010 or 2011 (n = 290,022). A protest can be settled in one of three formats: electronically (only when an agent is not used), an informal hearing, or a formal hearing. Because an owner's decision to use an agent is voluntary, I use a two-stage estimation process to help control for potential selection bias. The first-stage probit model is used to generate the predicted probability that an owner will use an agent, while the second-stage Tobit regression models the appeals adjustment as a function of protest format, agent use, control variables, and the selection-bias variable. Empirical results show that when an owner does not use an agent, appeals adjustments for electronic settlements are smaller than for informal settlements, and appeals adjustments for informal settlements are smaller than for formal settlements. When an agent is used, results show the opposite: appeals adjustments for informal settlements are greater than for formal settlements. Results also show that appeals adjustments are smaller when an agent is used compared with owners representing themselves, regardless of whether the hearing is formal or informal. Results are generally consistent with agency problems whereby property tax agents accept suboptimal appeals adjustments. Results from the first-stage probit model provide evidence on factors that affect the likelihood that an owner will use an agent.