Thanks @joshua-goodman.com for highlighting the paper.

Here is the ungated version papers.ssrn.com/abstract=420...

9/ Ungated link to the paper: papers.ssrn.com/abstract=420...

8/ A big thanks to my advisors, friends, family, reviewers, and fellow researchers who provided valuable feedback and tremendous support without which this study wouldn't have been possible.

7/ Bottom line: Weather forecasts are valuable! Improving forecast lead times has material economic benefits. People and institutions pay attention to forecasts and advisories.

6/ My estimates suggest longer lead times on winter advisories reduce approximately 13.2 crashes per 100,000 people each year, resulting in an annual savings of approximately USD 190 million.

5/ I provide evidence of two mechanisms at work here:
• Individual behavior: People visit fewer places when they get advisories earlier
• Institutional response: Snowplow crews intensify road maintenance operations when advisories arrive earlier

4/ In fact, at the current timescale of forecast horizon, the marginal benefits of additional lead time don't decrease--longer the lead time, fewer the crashes.

3/ Winter advisories with longer lead times DO reduce vehicle crashes, even when they're less accurate than advisories with shorter lead times.

Research Question: Significant resources go into improving #forecast accuracy, but we know little about the value of longer lead times. Does getting warnings earlier actually provide economic benefits? I examined this question in the context of winter weather and crashes.

Agree about the name--it is confusing. Though I didn't think about it earlier, sampling within a pre-defined insurance unit may reduce basis risk even more. e.g., in home ins, a unit could be homes that share similar characteristics (age, roof type, elevation). Thanks for highlighting this.

Crop cutting experiments (CCE) based crop ins in India might be an example. I believe the sample farms for CCE are randomly chosen within an insurance unit (a village). CCEs determine the actual yield, which determines the payout. Although, not sure if the CCE farms are themselves insured or not.