Fudging the nudge: information disclosure and restaurant grading.

• Author: Ho, Daniel E.
• Position: Introduction through III. Fudging by Inflation: San Diego, p. 574-616

INTRODUCTION I. THE LANDSCAPE OF GRADE REFORM A. Historical Antecedents B. Public Health Doubts C. Los Angeles Faith II. EMPIRICALLY ASSESSING GRADING A. The Confounding Nature of Grade Reform B. Jurisdictional Variation C. Our Approach III. FUDGING BY INFLATION: SAN DIEGO A. Regulatory Background B. Inspections Data C. Empirical Results 1. Uniform Grade Inflation 2. Scoring Consistency IV. FUDGING BY NOISE: NEW YORK A. Regulatory Background 1. Inspections 2005 to 2010 2. The 2009 Comptroller Audit 3. Letter Grading 4. Internal Assessment B. Inspections Data C. Empirical Results 1. Grade Variation 2. Grading Changes Scoring 3. Scoring Inconsistency V. EXPLAINING THE SCORING DIVERGENCE VI. INTENDED AND UNINTENDED EFFECTS A. Health Outcomes B. Perverse Resource Allocation VII. IMPLICATIONS A. Design Matters B. Retargeting Transparency CONCLUSION APPENDIX A. DOHMH Data Integrity 1. December Version 2. December Version and DOHMH Site 3. Comparison of Five Versions B. Classification Algorithm C. Robustness 1. Types of Violations 2. Consistency in Other Types of Inspections 3. Administrative Hearings 4. Random Sample from the DOHMH Website D. Corroborating Evidence from Eight Other Jurisdictions E. Sources for Tables 1 & 2 INTRODUCTION

When does disclosure work? Mandated disclosure to solve informational failures--and to empower parties to make informed decisions--has long been recognized as a theoretical matter. (1) Examples of mandated disclosure abound across regulatory areas as diverse as securities regulation, (2) campaign finance, (3) product safety, (4) energy regulation, (5) employment law, (6) environmental law, (7) and health law. (8) Yet despite the fact that disclosure is a mainstay of the regulatory toolkit, a fierce debate persists about the conditions under which disclosure works. (9)

Over the past few years, one of the most important regulatory developments has been the emerging focus on "targeted transparency." The chief insight, based on behavioral research, is that the public faces significant cognitive limitations in processing information. (10) More information is not always better. (11) Instead, effective forms of regulatory disclosure are "targeted": simplified disclosures embedded at the point of decisionmaking to "nudge" parties along. (12) In the terms of Cass Sunstein and Richard Thaler, interventions should focus on structuring choices to nudge parties toward decisions that are less prone to heuristics and biases of decisionmaking. (13) Age grading of children's toys, (14) star ratings for SUV rollover risk, (15) and smart energy meters (16) arguably embody such prescriptions.

The Obama Administration has embraced targeted transparency and behavioral insights in its regulatory approach, (17) most notably in the appointment of Cass Sunstein as Administrator of the Office of Information and Regulatory Affairs (a.k.a. "nudger in chief" (18)). Executive Order 13,563, which reaffirms cost-benefit analysis of proposed regulations and mandates retrospective review of existing regulations, champions "provision of information to the public in a form that is clear and intelligible." (19) In a series of memoranda to agency heads, Sunstein further refined the Administration's approach: "Agencies should consider how best to eliminate unnecessary complexity and to simplify people's choices." (20) Information technology and intermediaries should serve that end, with agencies encouraged to release "complex information and data in standardized, machine readable formats [to] enable consumers to make informed decisions." (21) In 2011, the Administration convened a National Science and Technology Council Task Force on "Smart Disclosure." (22)

Agencies, in turn, have developed a host of proposals in line with targeted transparency. The Securities and Exchange Commission promulgated a rule requiring standardized (machine-readable) risk-return summary disclosures for mutual funds. (23) The Environmental Protection Agency and National Highway Traffic Safety Administration proposed motor-vehicle letter grading for fuel economy and greenhouse gas emissions. (24) The Food and Drug Administration simplified sunscreen labels to minimize consumer confusion. (25) The Department of Housing and Urban Development issued a grant to the Center for Neighborhood Technology to create a national Housing and Transportation Affordability Index. (26) And the Consumer Financial Protection Bureau is currently experimenting with simplified mortgage disclosure forms. (27) The central ideas of targeted transparency continue to inspire scores of normative proposals. (28)

Targeted transparency's poster child is restaurant sanitation grading. The central idea is to summarize sanitation inspections with letter grades ('A,' 'B,' or 'C') and post these in entryways of restaurants to succinctly and intuitively inform consumers. In theory, the disclosure helps consumers select restaurants based on health risk, which in turn incentivizes restaurants to clean up. In the seminal synthetic work on targeted transparency, Archon Fung, Mary Graham, and David Weil systematically review disclosure policies and associated empirical research across regulatory areas, finding restaurant hygiene disclosure to be one of two "highly effective" regimes because of its simplicity and comprehensibility. (29) Indeed, Fung, Graham, and Weil use restaurant grading as the motivating example of how to "embed" disclosures with individual decisionmaking in an informative and comprehensive fashion. (30) Restaurant grading, according to them, exhibits congruence between policy and consumer goals to reduce food-poisoning risk, with only a moderate chance of misinterpretation. (31) Similarly, Omri Ben-Shahar and Carl Schneider argue that mandated disclosure has generally been a failure across policy areas, but they point to restaurant grading as a salutary exception and as the prototype for promising regulatory alternatives. (32)

Restaurant grading is widely considered a paragon of disclosure regulation. (33) In a landmark study, Ginger Jin and Phillip Leslie reported that the adoption of grading in Los Angeles in 1997 caused a 20% reduction in hospitalizations for foodborne illness. (34) Mayor Michael R. Bloomberg, who introduced restaurant grading to New York City in July 2010, called grades "wildly popular" (35) and concluded that as a result "the City made restaurants cleaner, safer and more transparent." (36) Forbes magazine described New York's system as "The Most Effective Regulatory Disclosure Ever." (37)

Other jurisdictions, in turn, have jumped on the bandwagon. Over the past ten years, in addition to Los Angeles and New York, Georgia, (38) Hartford, (39) Louisville, (40) Mississippi, (41) Toronto, (42) Albany County (NY), (43) Cuyahoga County (OH), (44) Kern County (CA), (45) Maricopa County (AZ), (46) and San Bernadino County (CA) (47) have implemented grading. Around this time, grading was proposed in Florida, (48) New York State, (49) Washington, D.C., (50) Albuquerque, (51) Chicago, (52) El Paso, (53) Kaufman (TX), (54) New Haven, (55) Pasadena, (56) Pittsburgh, (57) San Francisco, (58) Alameda County (CA), (59) Ashland/Jackson County (OR), (60) Huron County (OH), (61) Kanawha County (WV), (62) Santa Clara County (CA), (63) and Ventura County (CA). (64) The Center for Science in the Public Interest advocates that "[s]tate and local governments should pass laws requiring the posting of inspection grade cards in the windows of all food establishments." (65)

Restaurant grading has its critics, however. New York City Council Speaker Christine Quinn called the City's grading system "inconsistent" (66) and "borderline harassment." (67) In March 2012, Speaker Quinn convened a raucous, six-hour oversight hearing and reported from a convenience sample that 66% of all restaurateurs (and 59% of restaurateurs who received 'A's) found the system "poor." (68) Time magazine called the system "arbitrary and imperious." (69) The Wall Street Journal and New York Times documented suggestive evidence of gaming of the grading thresholds, which we use as a starting point of our analysis of New York below. (70)

Despite the pivotal role that grading commands in the debate over information disclosure (and the exhaustive review of the literature by Fung, Graham, and Weil), restaurant grading's merits turn out to rest on remarkably fragile empirical grounds. The only large-scale empirical study of grading examines Los Angeles around 1997. (71) To cure this empirical gap, this Article amasses large-scale microdata from over 700,000 restaurant inspections in ten other jurisdictions to evaluate the efficacy of restaurant grading. (72) For expositional simplicity, the analysis of our research team focuses on San Diego and New York, but the findings generalize to the other jurisdictions. (73) We show that the benefits of grading are vastly overstated, and costs vastly understated. The regulatory design, implementation, and practice in these jurisdictions are flawed at their core. As practiced, regulators fudge the nudge.

The findings, in brief, are fourfold. First, nearly every restaurant in San Diego receives an 'A,' limiting the meaningfulness of grades. Second, New York grades vary widely, but, unlike San Diego's underlying numerical scores, New York scores exhibit little substantive consistency. A score (or grade) in one year predicts little about the restaurant's cleanliness down the road. Third, differing inspection criteria provide one compelling explanation for the difference in consistency between San Diego and New York. The relative complexity of New York's inspection criteria appears to impede uniform scoring across inspectors and restaurants. Fourth, grading in New York has had no discernible health benefits but may come at a large, previously unrecognized cost in administrative resources. Specifically, grading reallocates inspection resources away from...