There was a lengthy discussion about this in the subreddit. https://www.reddit.com/r/spacex/comments/61walr/why_are_f9_boosters_not_tested_to_full_capacity/ ht

There was a lengthy discussion about this in the subreddit. https://www.reddit.com/r/spacex/comments/61walr/why_are_f9_boosters_not_tested_to_full_capacity/ https://www.reddit.com/r/spacex/comments/61walr/why_are_f9_b... I didn't find many of the comments enlightening, but the general argument seems to be this: expert consensus (by both SpaceX and the insurance company) is that the risk of failure is simply not that much higher than a normal launch. So it doesn't make sense to waste a probably-safe trip to LEO; you might as well get paid. Even in the extreme limit where there was a 50-50 chance of failure, it's clear that there would be some payload that would be worth attaching (e.g., a standard satellite design with a low marginal construction cost, or even a high school satellite project).