U.S. Flight Delay in the 2000s: An Econometric Analysis

Many researchers have undergone to better understand, quantify, and improve operations of the National Airspace System (NAS). In this paper, a series of successively more complex econometric models relating average delay against schedule in the NAS to key causal factors including airport congestion, total traffic, and en route weather. The estimation results suggest that airport congestion, measured by arrival queuing delay, has been a major contributor to average delay (about 32%), but a model with one explanatory variable is inadequate to describe the reality of a system. Thus, along with traffic and weather conditions, the models also take into account the 10 airports with the worst on-time statistics in 2007. Results indicate that Newark International Airport (EWR), John F Kennedy International (JFK), O'Hare International Airport (ORD), Philadelphia International Airport (PHL), and Charlotte Airport (CLT) are airports with a disproportionate large effect on system wide performance. LaGuardia Airport (LGA), interestingly, has a disproportionately small effect. Even with the above factors accounted for, seasonal and secular influences are significant. December, January, and February have the worst delays, while the several years after the 9/11 attacks decreased delay substantially. Such comparisons highlight the fact that statistically modeling of NAS performance is still limited both in terms of understanding and data.