MIAMI NEWS TIMES "Home-built aircraft are five times as likely to crash as professionally made planes. And if an accident does happen, their pilots are seven times as likely to die, according to federal investigators. In 2011, 212 home-built aircraft have crashed around the United States, killing 63 people."
Well...I see the source of a couple of aspects of those quotes, though the paper is just plain wrong on one statistic. Allow me to give you my perspective.
The "Seven Times as Likely" claim came from the Nall Report a few years back. Makes a great quote, but I'm not enamored of their process. The Nall Report computes its figures based on the estimated number of hours flown, by GA aircraft vs. homebuilt aircraft.
They don't make the annual hours estimate. The FAA does. The FAA holds the basis for that estimate very tightly; there's rarely any insight into the process. Several years back, I was in a telecon with the FAA, the Nall Report folks, and a couple of people from EAA. Didn't get much insight, but one aspect is stunning.
Take Geico, that man-about-town RV-10 driver. According to the FAA, he flies his RV 29 hours per year.
Say you convince Geico of the error of his ways. He sells the RV, and buys a Cessna 172.
The FAA now assumes he flies *two hundred* hours per year. Just because he's now flying a certified airplane.
How did the FAA come up with that 29 vs. 200 values? We don't really know. It's a combination of survey inputs and a string of assumptions as to how many homebuilts on the registry are still active. Low estimate for homebuilts, of course. No basis for that, it's just their estimate.
Divide that 200 by 29. You get just about 6.9...which is the amount the Miami paper said that homebuilts are *worse* than production airplanes.
Odd coincidence, that.
The basic problem is that the estimate is based on a stream of unpublished assumptions. You can't check their numbers because you don't know HOW they came up with their numbers. The Nall Report guys just shrug and say, "We used FAA estimates." The FAA guys aren't talking.
The basic problem is what I refer to as comparable use. The FAA probably estimates how many hours privately-owned aircraft fly, adds that to an estimate of how many hour hours charter aircraft flew, adds it to an estimate of how many hours the country's Gulfstreams and Beechjets flew, and divides that by the total number of GA aircraft.
But of course, those homebuilts aren't flying charters. They (mostly) aren't corporate-owned, they aren't flown by pilots specifically hired to fly them. They aren't required to undergo 100-hour inspections, and their upkeep is a matter of a private owner's pocketbook, not a corporation's tax deduction.
Frankly, if you do an accidents-per-100,000-flight-hour comparison between Learjets and Stinsons, the Stinsons will come out far worse.
The fair comparison would be to a Cessna 172 or Piper Warrior, to an estimate of how many hours a private owner flies per year for personal travel and recreational flying...which is all the homebuilt owner can legally do.
But of course, those numbers...other than the 200 hours per year overall average...aren't available.
So, what to do?
How about we *assume* the homebuilt owner and the Cessna owner fly the same number of hours per year?
To me, it's a beautiful solution. Anyone can then compute the "Fleet Accident Rate" (number of accidents per year divided by the number of airplanes of that type). There are no mysterious undocumented assumptions; no "we got the numbers from the FAA" sort of excuses. Number of accidents/number of aircraft. Simple.
The last time I ran this analysis, homebuilts had a 46% higher accident rate than the overall US fleet. They had a lower rate than several types, such as the PA-18. If you eliminated the aircraft still in their test periods, the homebuilt rate dropped to about 15% higher than the overall rate.
When you consider that the aircraft are built, flown, and often designed by amateurs, I don't think 15% is all that bad. Yes, it's higher when you include the airplanes in the test period...but consider: How many production-airplane buyers buy a plane with zero hours? They're all going to have at least one test flight by a company test pilot...and if something happens, a professional test pilot, with scads of experience in that model, is just who you want at the controls to keep from becoming a statistic.
As for the Miami paper's claim that in an accident, "[homebuilt] pilots are seven times as likely to die," that's bull.
For the 1998-2012 period, here are the fatality rates (percentage of accidents with at least one fatality) for the overall US fleet, for homebuilts, and a number of common GA aircraft.
Overall 21%
Homebuilts 27%
Cessna 18%
Cessna 172 13%
Cessna 182 21%
Cessna 210 25%
Cessna 185 15%
Piper 22%
Piper PA-28 20%
Beech 32%
Beech Bonanzas 34%
Mooney 25%
Cirrus 41%
Diamond 18%
You can argue speeds vs. configuration, but there's no way the homebuilt fatality rate is seven times higher.
Ron Wanttaja
....
