When Cirrus announced more than a decade ago that it would produce an airplane with a built-in parachute and that it would become the best-selling airplane of the next century, the doubters said the FAA would never certify it. And even if it did, Cirrus would never be able to sell such a crackpot idea. Of course, the FAA did certify the airplane and Cirrus briefly overtookCessna last year as the leading builder of single-engine piston airplanes.
But what about that other promise Cirrus implied, the one that this new design would be the safest GA airplane ever? A series of high-profile crashes-the latest in California in which the parachute may have ripped away from the airplane in a high-speed deployment-has dogged Cirrus attempt to position itself as a safety leader. Like it or not, every Cirrus crash-and there have been 29-has attracted the scrutiny of the aviation press simply because the airplane appears to not be delivering on its promise of being safer. At a casual glance, the reverse appears to be true; the model seems almost snakebit.
But is it really? Or is this a perception problem, one involving high expectations: We dont expect Cirrus airplanes to crash. When they do, we become both surprised and intrigued. Are we just obsessing on every Cirrus crash because it, well, isnt a Cessna or a Mooney? After all, Cirruses are designed to be safer than their peers. In addition to the Cirrus Airframe Parachute System (CAPS), the airplanes include other safety-enhancing details like high-G seats. This is in stark contrast to airplanes originally designed and certificated in the 1950s, but which can still be bought new.
To find out if the problem is perceived or real, we exhaustively reviewed the accident records of six other major airplane types that could be loosely considered to be in the Cirrus category. We wanted to find out how new and older Cessnas compared along with other models, including offerings from Diamond, Mooney and Lancair. We talked with several industry watchers-including Cirrus CEO Alan Klapmeier-and insurance executives. The results tell us that Cirrus may be the victim of its own success.
As the sidebar on the next page highlights, three Cirrus SR22s were involved in fatal crashes in as many weeks earlier this year. Although the NTSB has not yet determined a probable cause of those accidents, its unlikely theyre related.
To see if there were any common threads among other Cirrus accidents, we looked at the NTSBs data involving Cirrus airplanes beginning in 1999 through the end of February 2005. During that time, 29 Cirrus accidents or incidents made it into the NTSBs files. Of those 29, 10 involved SR20s.
Similarly, 13 of the 29 accidents/incidents resulted in a fatality, with five for the SR20. A cursory look at each fatal accident reveals only that pilots flying Cirrus airplanes get themselves into the same kinds of situations as pilots flying other airplanes.
As one example, the VFR-only Private pilot of an SR20 flew his perfectly good airplane into a mountain near Sierra Vista, Ariz., on April 10, 2001.According to the NTSB, Airmets for mountain obscurement, moderate turbulence and moderate rime/mixed icing conditions were in effect for the area. No anomalies were noted that would have prevented normal aircraft operation, the NTSB said.
On November 3, 2002, another VFR-only SR20 pilot managed to do pretty much the same thing. He came to grief near Las Vegas, N.M. The NTSB noted on the day of the accident, the pilot did not file a flight plan or receive a formal weather briefing from a FAA flight service station. The Boards probable cause determination inevitably references the pilots inadvertent flight into IMC and failure to maintain clearance with terrain.
None of the poor decision-making involved in these two and similar accidents-involving a Cirrus or any other aircraft-can realistically be blamed on the airplane.
The only credible way to judge the relative safety of the Cirrus line of aircraft is to measure its accident rate against airplanes of like type. But this is a task mired in a world of fuzzy data and subjective judgment. First of all, what constitutes a like type? And to calculate accurate accident rates, you need accurate hours-flown data, something that has proven elusive to obtain, at best. Lacking this kind of data, we can make reasonable estimates of hours flown and we can also measure accidents as a ratio of total aircraft produced or registered. Fair warning: Such analysis has to be taken in context. Its not the best way of comparing apples to apples, but its the best data available to us.
Starting with the wide view, we estimate that the Cirrus fleetwide total hours stood at about 550,000 hours as of late January 2005. This is an estimate based on information provided by Cirrus. Not counting the crash of a pre-production airplane on a test flight in 1999, the Cirrus line has suffered 29 accidents as of February 2005, 13 of them fatal. That works out to an overall accident rate of 5.3/100,000 flight hours and a fatal rate of 2.4/100,000.
The NTSB recently announced that GA total accidents in 2004 were the lowest they have ever been since records have been kept, but the most recent rate data we have is for 2003, which indicated a GA overall rate of 6.7/100,000 and a fatal rate of 1.3/100,000. Fleetwide, Cirrus airplanes appear to have a slightly lower overall accident rate than GA as a whole, but a measurably higher fatal rate. (Another aspect of the accident record supports this; more on that in a moment.)
When Cirrus airplanes are examined individually, the picture doesnt change much. Nineteen of all 29 accidents have involved the SR22, and these account for eight of the crashes involving fatalities. This gives the SR22 an overall accident rate of 5.4/100,000 hours and a fatal rate of 2.3/100,000. The SR20 has a similar record, although its fatal rate is .3 higher per 100,000 than its stablemate SR22.
Although the Cirrus fleet may appear, by perception, to have a problem accident rate-judging by the high premiums the insurance industry sets for the Cirrus, insurers seems to think so-in reality, the Cirrus lines overall rate may actually be better than average. We say may because, as noted, the hours-flown numbers for these comparisons are difficult to confirm. The Cirrus fleetwide fatal rate appears somewhat higher than the GA average, especially for the SR20.
But what about compared to other like airplanes? If you define like airplane as a new-age composite model-say the Lancair or Diamond lines-Cirrus loses some of its luster. Lancair and Diamond have fewer airplanes flying-about 200 for the Lancair and about 240 in the U.S. for Diamonds DA-40 Star compared to nearly 1800 Cirrus airplanes-so one might argue that the comparison is unfair. But exposure is exposure, and both Lancair and Diamond have it.
The DA-40 only recently suffered its first fatal accident, and only three U.S. accidents of any kind have been reported to the NTSB. While the Star is a lower-performance airplane than either the Cirrus SR20 or SR22 and probably isnt used in the same way, the DA-40s record is, nonetheless, remarkably good.
The Lancair models certainly are used on the same kinds of missions as the Cirrus and, again, this line of aircraft has a better record thus far than the Cirrus airplanes do, at least in terms of fatal accidents.
We could find only three accidents of any kind for Lancair certified production airplanes and one of these concerned a factory test flight. All were minor and involved no fatalities. In our interviews with insurance executives, one said of Lancair, Give them time, theyll catch up to Cirrus. Whether this is true or not remains to be seen. But for now, the Cirrus accident history doesnt shine much against its composite rivals.
And what of the traditional manufacturers, like Cessna, Mooney, Beech and Piper? How do they compare with the Cirrus? Here, the analysis is murkier yet-what from those manufacturers is like a Cirrus? Its reasonable that a new Cessna 182 buyer might also consider an SR20 or SR22 or a retractable from Mooney or Beechcraft. So, take your pick. We examined accident patterns of older and newer models of Cessnas 182 and two Mooneys.
Our estimate of the overall accident rate for Cessna 182s-all models and some 13,500 airplanes registered in the U.S.-comes to 6.8/100,000 overall and 1.5/100,000 fatal. About 27 percent of all the 182 accidents we examined were fatal, versus 45 percent for the Cirrus. (For hours-flown for the other aircraft, we used the report-at-sale figures for each model from the Aircraft Bluebook Price Digest, as accurate an estimate as anyone has.)
When new Cessna Skylanes-those made since 1997-are examined, a different pattern emerges. Cessna has sold about 830 of the fixed-gear, non-turbocharged Skylanes since 1999, and when we searched for accidents during the same period we used for the Cirrus crashes, we found 32. Cessna wasnt able to provide an overall fleet-hour estimate but said 182s fly about 200 hours a year. The Bluebook estimate is less so we used 150 hours for our calculations and estimated fleet hours at 367,000, for an overall accident rate of 8.6/100,000 and a fatal rate of 2.4/100,000, the same as all Cirrus airplanes. (The SR20 has a worse fatal accident rate than the Cessna; the SR22 is better.) Further, 28 percent of Skylane accidents proved fatal. Of course, if the airplanes really fly 225 hours a year, those rate numbers would be quite different. We simply dont have the data to know that.
Finally, we examined Mooneys, a model we think is a contender among pilots who buy Cirrus airplanes. Our estimate of the accident rate for the Mooney Ovation, for instance, is 5.7/100,000 overall, with a fatal rate of 1.9/100,000. Of the accidents we reviewed, 33 percent were fatal. An older popular Mooney model, the M20J or 201, has, according to our estimates, an overall accident rate of 3.5/100,000 and a fatal rate of 1.27/100,000, well below the Cirrus on both counts. According to our review, 22 percent of Mooney M20J accidents proved fatal.
What it All Means
Are these numbers credible enough to put the Cirrus airplanes in a fair context? Well concede that because of the lack of accurate hours-flown data, the overall accident rate comparisons are iffy. But one comparison is telling: the fatal rate and the percentage of Cirrus accidents that have proven fatal. Nearly half of all Cirrus accidents-45 percent-have been fatal.No other production model we could find comes close to that record, including a couple of twins we reviewed with a reputation for being skittish.
Given this record, its tempting to question the Cirrus vaunted crashworthiness claims but, unfortunately, the types of accidents the Cirrus airplanes have gotten into wouldnt have been survivable in any airplane, given the pattern of controlled flight into terrain at high speeds, loss of control in weather or in upsets, and improper IFR operations. In other words, these are largely pilot-error and poor-pilot-judgment accidents and apparently not related to the design of the airplane itself.
What Do Insurers Say?
Generally, aviation insurance company executives echo the idea that theres no common thread to Cirrus accidents. One exec we spoke with summed up the industrys overall view by saying that many accidents simply resulted from an old problem: stupid pilot tricks.
That said, the aviation insurers we spoke with do think the Cirrus accident rate is higher than it should be. He said concerns about the airplanes safety record has the industry asking, Why do we seem to be losing so many?
In the beginning, aviation insurance companies really didnt know what to make of the Cirrus. The first models-SR20s-offered the simplicity of a Cherokee with the speed, capability and decision-making requirements of a more complex airplane, one exec told us.
To date, there is still very little insurance industry experience with either Cirrus model, even after almost six years of operation in private hands, he added. But what experience there is has resulted in higher premiums.
Ultimately, said one insurance exec, the Cirrus models tend to attract a different kind of pilot. This pilot is fairly new to aviation and relatively inexperienced. Which is why many insurance companies require initial and recurrent training in the airplane.
What Does Cirrus Say?
We asked Cirrus CEO Alan Klapmeier for comments on his airplanes safety record. In a wide-ranging conversation, he shared with us several thoughts about his airplanes and the pilots who fly them.
In a nutshell and unsurprisingly, Klapmeier believes theres nothing wrong with the airplane. Instead, he maintains Cirrus has produced safe, easy-to-fly airplanes featuring relatively low workloads when compared with their peers. Admitting frustration with Cirrus safety record, Klapmeier told us that pilot over confidence comes close to defining the problems inherent in many of the accidents drawing attention to his airplanes. In discussing the design and engineering decisions Cirrus made and implemented, he noted that, now that we made control manipulation easier, we see people using the airplane more because of its lower workload. In fact, Cirrus believes the airplanes are regularly used for cross-country work more than other types in their class. Unfortunately, theres no real data on that.
Klapmeier says the cross-country performance of a Cirrus is now available to pilots who havent put up with or been willing to gain the appropriate levels of experience necessary to safely and comfortably travel long distances, perhaps through multiple weather systems. In other words, pilots who might find themselves lacking confidence in their ability to safely fly a more-complicated Bonanza or Centurion have overblown confidence in their ability to safely get from Point A to Point B. The experience you acquired along the way to becoming comfortable flying a Bonanza along that same route, for example, would help you better make the go/no-go decision, or one to turn around, land and wait out some bad weather, according to Klapmeier.
Aviation is not good at teaching judgment, he told us. As a result, the company is looking at ways the Cirrus Standardized Instructor Program, CSIP, can improve this critical element of training. Yes, training needs to be done differently, according to Klapmeier, whose nightmare scenario involves a non-Instrument-rated pilot flying an airplane with all the toys.
Instilling good judgment is the final goal of the plans and improvements Cirrus has in mind for training pilots to fly its airplanes. In essence, Cirrus is looking at developing a mental proficiency program, one which gets pilots thinking about modifying their personal minimums based on their recent experience.
Finally, we asked Klapmeier if the higher rates insurance companies charge to cover Cirusses was fair. No, its not fair, he told us. He added that the composite airframes repairability is improving and that the markets reaction is drastic. Despite the number of mishaps that make the news, Klapmeier told us there are more [Cirrus] airplanes in the fleet than people think. And perhaps thats part of Cirrus safety-perception problem.
Is there a problem with the Cirrus airplanes? Probably not, but that doesnt automatically let the company off the hook. Viewed in context, the Cirrus CAPS, the glass panel, crashworthy seats and spin-resistant wing appear to have thus far produced an airplane with what may be an average safety record, but not an exceptional one. In that sense, Cirrus has fallen short of its goal to produce the safest GA aircraft ever, although longer-term experience could yet see it realized.
The high percentage of Cirrus accidents that are fatal appears to reflect how the airplanes are used-as all-weather, cross-country machines-by pilots not necessarily trained or experienced enough for that kind of operation. All the safety gadgets Cirrus has added, including the parachute, havent proved capable of overcoming a lack of skill or basic errors in judgment by pilots, just as with all the airplanes flown before.
As we see it, the problem for Cirrus is two-fold. One, it needs to improve the risk perception among insurers by demonstrating a declining accident rate. Second, training will have to keep pace to reduce the percentage of accidents that are fatal. Otherwise, we think Cirrus faces a real and serious risk of crippling litigation as its fleet size increases.
On the first count, the overall Cirrus accident rate appears to be declining as the fleet flies ever more hours. Thats good news. On the second, Cirrus needs to show that its efforts at training pilots-laudable though they may be-can actually deliver results in fewer fatal crashes.