Nall Reports Show Drop In Accident Rates, Get New Platform


In addition to publishing the 29th and 30th editions of its Joseph T. Nall Report, the Aircraft Owners and Pilots Association (AOPA) Air Safety Institute (ASI) has launched a new platform designed to provide “access to more current accident data in near real-time.” According to AOPA, the new platform will allow accident data to be updated on a rolling 30-day cycle. It will also let users select and compare accident analysis graphs for multiple years from 2008 up through 2020.

“I am excited that this major effort has significantly accelerated the accident analysis process,” said ASI Senior Vice President Richard McSpadden. “This allows us to release the 29th and 30th Joseph T. Nall Reports, which provide a snapshot in time for 2017 and 2018 data, respectively.”

Digging into that data, the 29th Nall report found that the total number of general aviation accidents in 2017 decreased from 1,227 in 2016 to 1,204. The 2017 analysis also saw an improvement on 2016’s overall accident rate of 4.98 accidents per 100,000 hours, coming in at 4.81 per 100,000 hours. The 2017 fatal accident rate also dropped slightly from 0.78 per 100,000 hours the previous year to 0.76 per 100,000 hours. The 30th edition of the report, covering 2018 data, noted an increase in total accidents to 1,224. However, it found a further drop in the total accident rate to 4.56 per 100,000 hours as well as a decrease in the fatal accident rate to 0.74 per 100,000 hours.

Continuing a longstanding trend, personal flights made up to largest percentage of fatal non-commercial fixed-wing accidents in both years—82.7 percent in 2017 and 80.7 percent in 2018. Landing accidents, of which there were 335 in 2016, dropped to 314 in 2017 and rose to 322 in 2018. Weather-related accidents spiked from 23 (12 fatal) in 2016 to 42 (32 fatal) in 2017. 2018 saw a decrease in weather-related crashes to 23 (21 fatal). In all three years, the majority of those accidents involved VFR flight into IMC.

The Nall Report analyzes aviation accident data from years where probable causes have been determined for at least 80 percent of the accidents that occurred. Segments covered include non-commercial and commercial fixed-wing operations, non-commercial and commercial helicopter operations, and sport/experimental operations. The full reports can be viewed here.

Kate O'Connor
Kate O’Connor works as AVweb's Editor-in-Chief. She is a private pilot, certificated aircraft dispatcher, and graduate of Embry-Riddle Aeronautical University.

Other AVwebflash Articles


  1. Not many people are flying, due to the economic situation. Hence, the rates of accidents are correspondingly lower. General aviation accident rates are invidious. You can track them all you want, but you’ll discover that there’s no set value or relevance for these numbers. Depending on the health of the economy, we will see adjustments to those figures. You’ll also notice a commonality of probable causes for fatal mishaps. VFR flight into IMC being one of them.

    • David, your point that reduced flying is having an effect upon the number of GA accident is well taken, but as I read the article I gathered that the rate they note is accidents per 100,000 flight hours. This should in take into account reduced flying and still make an “apples and apples” measure(except any increased safety due to less crowded skies, i.e. lower controller workload enabling more assistance, etc). That said, I also noted that the numbers the article discusses are based upon 2017 and 2018 data, which should not be affected by the current flying slowdown.

      • The 100,000 rate is type specific data, to assess the safety aspects of a particular aircraft. When these rates are lower, it simply means the 100,000 marks are taking longer to reach. That correlates to lower aircraft usage fleet wide. Mr. Browne, of You Tube Blancolirio Channel fame, may have hit upon a pertinent fact: The people flying GA aircraft are wealthy, and they’re not exactly committed to staying either proficient or being trained to higher standards of aviation acumen. They simply hook up the automation and GPS their butts to Point B. This is not the first example of wealth precluding safety. Mr. Munson, a famous baseball player, met his Waterloo aboard his Cessna Citation, which he was demonstrating to his former flight instructor. Cessna had to bite the bullet after a huge lawsuit resulted in their being held liable for selling their aircraft to a less-than-qualified pilot/owner. There are dozens of other examples, but I’ll cite a particularly well-known one: The Bonanza was jocularly called the “Fork-Tailed Doctor Killer.”

      • The problem with comparing accidents on the basis of flight hours is that the process used in estimating flight hours has been hosed up by the FAA’s triannual re-registration program.

        The FAA provides these estimates as part of its annual GA Survey. The survey determines the average hours flown by each aircraft and estimates what percentage of the fleet is active. The number of hours flown is based on the total number of aircraft registered, multiplied by the estimated percentage that are active, and finally,multiplied by the survey’s estimate of the average flight hours flown by each aircraft.

        But when the FAA starting requiring owners renew their registrations every three years, and commenced removing aircraft from the registry if the owners didn’t comply, this messed up the process for estimating flight hours.

        The FAA started this policy in 2010. Here’s the net change in the number of aircraft registered in the US over the last ten years:

        2010: -477
        2011: -6039
        2012: -15659
        2013: -34205
        2014: -5407
        2015: 1818
        2016: 6279
        2017: -8339
        2018: -18123
        2019: -4141

        In 2018, for example, more than 18,000 aircraft were removed from the US registry. These aircraft didn’t suddenly quit flying that year. This was solely due to the cyclical nature of the re-registration process.

        But from 2016 to 2019, the estimation of total flight hours drops due the fact that the official size of the fleet dropped by more than 30,000 aircraft. A drop that is administrative in nature, and doesn’t truly reflect the number of aircraft flying nor the number of total flight hours of the fleet.

        This isn’t the fault of the FAA survey people, nor is a reflection on the Nall folks. They all do good work, but the FAA registry is the ONLY source for fleet-size information, and it varies for reasons that are unrelated to actual operations.

        The lesson is that one must consider studies that compare results year-to-year with a jaundiced eye. The TOTAL number of accidents each year is a good metric, but of course, will vary. Estimations of accidents per 100,000 flight hours are handy, but must be viewed in light of the effect of FAA administrative de-registrations.

        • The aviation industry shouldn’t dwell on statistics, they should focus on fatal accident causal factors, and gear training and proficiency standards toward reducing those mishaps. My research on this issue has revealed the Automation Factor. These new gizmos are tremendously helpful for flight crews, but when they’re pickled, or especially when they malfunction, some pilots are left with diminished flying skills to deal with demanding flight conditions: Low airspeed, low power, high drag configurations, task saturation, etc. The Houston Cirrus tragedy had me scratching my head. A 400 hour pilot couldn’t complete THREE approach attempts in VMC weather to a major facility, and wound up spinning into a parking lot. The Cessna 414 mishap at Santa Ana may have been a similar type of accident.

          Pilots aren’t flying their airplanes, and that aspect of their operational activity has a significant bearing on their ability to take command without automation during critical phases of flight.

          • My god the tenants of train train train are ingrained in this community. In aviation systems engineering training is only used when first Prevent, then Guard, then Warn and if the prior measures fail – TRAIN. I will give you a time honored challenge ….. Fuel accidents….. Nothing can be done according to the common feeling. You have to train, use a watch stick your tank. Strangely the statistic didn’t change and in a polar opposite of your theory, higher trained pilots ATP, Commercial, CFI are more susceptible. So we have delivered over 60,000 quality fuel quantity instruments to this community – yep high tech stuff. Our record after 10 years is No that is zero fuel related accidents. Knock on wood. It appears that just simply warning a pilot of a condition is surprisingly effective in addressing a common issue. If this is combined with or conjoined with a totalizer and comprehensive and progressive trustable warnings we now move this system to Guard. And again this statistic is now zero. We can’t PREVENT fuel accidents but simply applying appropriate technology provides relief….. dramatic relief. Now can this be applied to other systems – yes definitely. Training is important, but effective and trustable warnings and guards provide real benefit. The automotive community have caught on, where driver assist – assists the most mindless driver and accident and fatal accidents continue to drop. The future points to a more automated flying future- like it or not We will see if the old maxims fall to the wayside Just in the little area we participate in -the benefit is obvious. And the trend in this s tat is tic is falling

          • Scott P, you’re missing the point. It’s not training, it’s FLYING. Pilots these days use automation. Fine. But when those gizmos are uncoupled, or when they fail, pilots have to take over the chores of flying their plane. Some pilots apparently weren’t up to that challenge.

  2. Having read through several “NALL Reports”, I feel that the data is well presented.. I’m certain there will be a special mention (caveat), of the unique COVID economic effect on aviation or any other industry.. Not to draw a over weighted reference on a “year over year basis”. As well as the FAA A.T.C. publication of “By The Numbers” , and/or any other data being presented during the pandemic.

  3. Aviation, in particular General Aviation, is influenced heavily buy economic factors. No matter how you slice it, when the chips are down, fewer people fly planes. I was absolutely floored when I tried to arrange an Aztec Proficiency flight at our local airport. The cost was prohibitive, to say the least. I rented an Aztec for $104.00 per hour wet, and I thought that was pricey. I’m not cheap, but I’m not Bill Gates, either. Airplanes cost lots of money (I know, I owned one.) and they’re not a staple resource for our existence. These days, only very well-off people can afford to own–or even rent–planes.

  4. In may past experience as an airport manager, as well as operating a full service FBO, the 80/20 rule applied to aircraft flying is very accurate. Primarily, 80% of the flying is accomplished by 20% of the airplanes. To develop and maintain profitability as an FBO, one has to identify and go after the 20% who fly their airplanes regularly.

    If there are 100 airplanes based at the local airport, there are 65-75 of them in annual. 25-35 are not. Rounding out the number to 70 airplanes legally available for flight, 14 (20%) will be flown regularly accumulating on an average, 100-200 hours per year. But the remaining 56 will be lucky if they see 10-25 hours per year of wind underneath their wings. We maintained several airplanes whose only time they were started was during the following annual never flown in between. Plus, a significant portion of the GA fleet gets pencil whipped annuals, especially those flown sporadically.

    I am not an advocate of more regulations nor government meddling in my personal affairs. However, using the present method the FAA uses to determine fleet hour flown is highly flawed, and is a guesstimate at best. While applying the 100,000 hour formula to statistically attempt leveling the playing field of actual accident data, it does not accurately apply those stats as it relates to everyday pilot proficiency nor mechanical causes.

    For example, the airplanes flown 100-200+ hours per year are highly reliable. Pilots who fly them are highly proficient. Airplanes flown 10-25 hours per year, are maintenance headaches, with the constant debate between owner and maintenance facility regarding continually deferred maintenance. Throw in the pencil whipped annuals by the roving/mobile IA who performs 3-5 ” annuals” in an afternoon that the local FBO never has access to or an opportunity for. Now safety has a perfect storm for lack of use, deferred maintenance, and poor pilot proficiency. And the largest number of so-called active GA airplanes, new or old, fall into the 80% category. Owner flown LSA’s statistics gleaned from aircraft ads and used aircraft sales show most of the fleet is flying only 10-20 hours per year.

    I believe each airplane owner needs to report to the FAA, as recorded in the aircraft records, hours that particular airplane has flown annually. In addition, yearly pilot flight time needs to be reported to the FAA. I have seen pilots who flew regularly someone elses spam-can airplane, or commercial bizjet regularly and safely, but have not had time for a variety of reasons to fly their own. Many times these airplanes are high performance, experimentals, or antiques like a Staggerwing, Stearman, LSA, or modern tailwheel airplanes that normally require much higher stick and rudder skills. If they have an accident in one of those airplanes but are flying mostly common airplanes more regularly, proficiency numbers would not be accurate. Proficient in the boss’s CJ5 does not mean proficient in the family Luscombe or Swift.

    If the flying community supplied yearly accumulated flight times and each airplane reported hours flown on that particular airframe, accident investigations would be simplified and causes much more clearly defined. Insurance requires airframe time and yearly flying hours. We are giving that data to the insurance companies, why not the FAA?

    I believe those two stats would present much more accurate data regarding pilot errors, proficiency, and maintenance contributions to accidents. In addition, there would be accurate data reflecting the variety of economic drivers that increase or decrease flying activity.

    We lost 64% of our fuel sales in the last quarter of 2008’s economic meltdown. Every FBO in our region reported similar declines. Rental hours dropped similarly. Several regular customers did not have their airplane’s annualed when due. Many airplanes sat for an additional six months to a year, before the next annual was performed. Covid-19 is making a similar contribution. However, under current statistical guesstimates based on flawed and often optimistic flying estimates, we will not be able to learn the affects of Covid until 2022-23. 2017-18 Nall reports are helpful, but certainly very difficult to accurately spot trends related to a current, long lasting, worldwide pandemic. It’s sort of closing the proverbial barn door long after the cows have been gone.

    If GA aircraft ownership and flying was steady no matter what the economic and political landscape, then maybe we can get good stats using the present method. But my participation in GA has not shown any one year predictable of the next. But it is undeniable when it comes to aircraft maintenance or flying proficiency, use it or loose it.

    • I was president of a military Aero Club, and I was also in Airfield Management at military airfields. The one emphasis I conveyed to our club administration was maintenance. How many pilots have rented junk from FBOs that operate for profit at the expense of renters? Dirty, decrepit, barely airworthy planes are the offerings from many schools and clubs. At the very least, our club was non-profit, and our wrench benders were pretty good hands (Though some genius decided we would be better off with a Progressive Maintenance System. This was in place when I became president, and our planes were all grounded due to the lack of an AI to approve them for flight.) The margin of profit for aircraft rental agencies is thin. So, something has to give. Unfortunately, that something will materialize when a component falls from your rental aircraft in flight.

  5. Terrible shifts occur in how you view things as you approach the end of life, I am finding. For example, in reading all this I find myself thinking “In the case of personal GA, is it really all that important to agonize over the accident rate?”