The election of Donald Trump has revived hopes of air traffic privatization in the United States. The case for doing so goes something like this. The Federal Aviation Administration (FAA) - the governmental agency that manages the nation’s skies – operates using outdated technology. Although newer, more effective systems are available, government bureaucracy makes procuring them challenging. A non-governmental entity would be better positioned to address this issue, more easily cutting through bureaucratic red tape that has long hindered the FAA.

But in an industry where mistakes can prove deadly, any change in the status quo raises questions about safety. Hence, reform advocates have been quick to point out the safety benefits of privatization.

This idea is purportedly supported by numerous studies. The Government Accountability Office (GAO) for example, found the number of ‘air proximity incidents’ – instances where the safety of the aircraft is threatened – fell after air traffic service providers underwent some form of privatization. An analysis by the MITRE Corporation found that the safety record of specific ANPSs, “was equal to, or better than,” their record during government involvement. Similar results have been reported elsewhere.

Given these findings, the safety case for privatization is sound. At least that was what lawmakers were told during congressional hearings earlier this year. Yet closer scrutiny of this claim raises questions.

For one thing, air proximity incidents occur against the backdrop of millions of flights. For example, while the number of close calls in Germany went from 23 in 1995 to 8 in 2003, the number of flights handled during the same time period was over 2 million.  Similar, albeit smaller, figures were observed in the United Kingdom. From a scientific perspective, analyzing so many flights risks attaching significance to those events that result from chance alone. Simply put, the drop in close calls following privatization may well have little to do with privatization itself.

The MITRE analysis is also problematic. Its conclusions are based in part, on a 2005 study that relies heavily on descriptive reports from other aviation and trade publications. The author of that study states that in regards to privatization, “increased flight safety was indicated but could not be measured or forecasted.” In other words, even though verbal reports suggest that privatization-backed measures improved safety, there is no actual data to back up this claim.

Perhaps most important is the question of whether close calls, and more broadly fatal incidents, should be the sole criteria of interest when measuring safety. More than 3.5 billion passengers took to the skies in 2015 on some 37 million commercial flights. Yet the global jet accident rate (measured as the number of aircraft lost per 1 million flights) was amongst the lowest in history, As Frederic Lemieux notes, limiting our measurement of security and safety to fatal incidents alone in this environment narrows our appraisal of risk. Moving forward means rethinking how aviation safety is measured.

There are many reasons to favor privatization. Inadequate funding, flawed governance, and a, “status-quo-oriented organizational culture.”  But based on the available evidence, better safety isn’t one of them.