We appreciate the engagement by Jin and Leslie (2019) (J&L) with our study (Ho, Ashwood, and Handan-Nader, 2019) (HAH). Understanding the causal effect of restaurant grading on health outcomes remains a pivotal case in scholarship and policy debate about information disclosure (see, e.g., Ben-Shahar and Schneider, 2011, 2014; Fung et al., 2007; Loewenstein et al., 2014; Weil et al., 2006), and we are grateful for their attention to this issue.
We write to clarify three points in response to the new analyses reported by J&L. First, J&L’s response insufficiently engages with the role of the Southern CA salmonella outbreak. We show that because their analyses are only conducted within the 1995-99 window, they are unable to account for the peak of the salmonella outbreak in 1994. Extending their analysis with only two more years of pre-treatment data confirms that there is no evidence that grading affected hospitalizations and invalidates CA (or Northern CA) as a control group. Second, we discuss the potential for spillover effects, which would also threaten the original design in Jin and Leslie (2003). The dramatic and nearly identical drops in hospitalizations across all of Southern CA, beginning from the salmonella crisis in 1994 through 2001, are inconsistent with spillover effects from a local, retail-level disclosure policy. Third, we consider evidence based on non-salmonella pathogens, which show substantial violations of parallel treatment trends and reveal no evidence of any decrease in hospitalizations or illnesses around 1998.