Using internet search data to predict aggregate retail sales and enhance firm‐level revenue expectations
研究发现，公开上市零售企业的互联网搜索强度指数能预测分析师的预测误差，并带来约2%至3%的超常收益；该指数还能作为个人消费支出的领先指标，用于预测总体零售销售。
Abstract This study examines whether a simple measure of internet search intensity for publicly traded retail firms can enhance the capital market's firm‐level revenue expectations and provide insights into economy‐wide retail sales. At the firm level, the search index is predictive of analyst nowcast and forecast errors after controlling for past sales, deferred revenue, firm characteristics, and firm and time fixed effects. An implementable trading strategy generates abnormal returns of roughly 2% to 3% from the fiscal quarter end through the earnings announcement, well above transaction costs. We also find that approximately two‐thirds of the abnormal returns occur around earnings announcements, with an even greater fraction for firms with coarser information environments. At the macro level, we find that the permanent, seasonal, and transitory components of our search intensity index align with those of the Census Bureau's retail sales data and US real gross domestic product, suggesting our measure is a leading indicator of personal consumption expenditures, a key driver of aggregate output. The aggregated search index nowcasts aggregated publicly traded retail firm sales both within and out‐of‐sample after controlling for past sales.