This discussion paper resulted in a publication in 'Transportation Science'
We derive the expected user costs of US domestic air travel delay variability taking into account scheduling behavior of travelers. Travelers do not only consider mean arrival delays, but also face scheduling costs because they arrive too early or too late at their destination. The model allows travelers to anticipate arrival delay variability by choosing an earlier flight. We show that the expected user costs of US air traffic delays are underestimated by 16% if arrival delay variability is ignored.