Working Papers

“What Drives Demand for Fake Reviews? Evidence from Amazon Product Reviews”

Abstract: This paper investigates key drivers influencing online sellers' decisions to invest in fake reviews. I first construct a model in which sellers strategically invest in fake reviews to influence consumers, who form beliefs about product quality based on observed reviews. I then empirically test the model's predictions on the main determinants of fake review prevalence using Amazon product review data. To address the challenge of unobservable fake reviews, I use a machine learning model trained on verified fake Amazon reviews sourced from Facebook groups to estimate the likelihood of fake reviews for Amazon products. I find that products are more likely to have fake reviews when they are durable, are new, have more online shoppers with experience, or (in certain circumstances) face more substitutes. These findings help platforms and regulators better understand the economic environment to combat fake reviews.

Works in Progress

“Estimating the Informativeness of Rating Systems” (with Babur De los Santos and Chungsang Lam)

Abstract: We study how the coarseness of rating scales shapes the informativeness of online ratings about product quality. We combine three approaches. First, a simulation shows that a six point scale is more informative under moderate taste dispersion, while a two point scale can dominate when dispersion is very high. Second, a controlled Bookworm experiment that manipulates vertical and horizontal differentiation finds that the two star scale better separates objective quality in a complex game, whereas the six star scale is more discriminating in a simple game. Third, using Yahoo Movie and EachMovie, we construct a benchmark for average perceived quality with film and viewer fixed effects and compare platform's raw rating rankings to this benchmark. The six point scale aligns more closely with the benchmark than the thirteen point scale, consistent with greater informativeness when tastes are dispersed. The findings imply that scale design should match consumers' preference heterogeneity for products to improve information transmission and welfare.