Does Getting Forecasts Earlier Matter? Evidence from Winter Advisories and Vehicle Crashes
中文摘要
大量投资被用于提高预报的准确性和提前发布能力。然而，延长预报提前期的价值尚不明确。利用美国冬季天气预警和车辆碰撞事故数据，本文表明，即使提前期较长的预警不如提前期较短的预警准确，它们仍能减少碰撞事故。此外，边际收益并不随提前期延长而下降。这些收益源于个人和机构的应对。当预警更早发布时，人们前往的地点更少，扫雪作业人员也会加强道路维护作业。这些结果对如何提供有效预报具有政策意义。（JEL Q54, Q58, R41）
Abstract
Significant investments are directed toward improving the accuracy and early availability of forecasts. However, the value of longer lead times on forecasts is unclear. Using data on winter weather advisories and vehicle crashes in the United States, I show that advisories with longer lead times reduce crashes, even when they are less accurate than advisories with shorter lead times. Further, marginal benefits do not decrease with lead time. The benefits come from individual and institutional responses. When advisories arrive earlier, people visit fewer places, and snowplow crews intensify the road maintenance operations. These results have policy implications for providing effective forecasts. (JEL Q54, Q58, R41)