Why are airplanes expensive relative to family income?

The latest issue of AOPA Pilot magazine contains the following data:

  • In 1967, median household income in the U.S. was $12,000 and a four-seat Piper Cherokee was $16,000.
  • Today, median income is roughly $48,000 and the same airplane (they still make it!) is $230,000.

Where it took 16 months of income to buy a basic four-seat airplane in 1967, it takes 57 months now. The question is “Why?”

Airplanes are made in small quantities and have not benefited from automation and capital investment in tooling the way that automobiles have, so you wouldn’t expect their price relative to incomes to have fallen. On the other hand, there have been some efficiencies introduced such as computer-controlled machining so the number of labor hours should have been slightly reduced.

Workers in airplane factories are not paid more than average and roughly the same number of working hours are required to build an airplane.

Airplane companies are not ridiculously profitable. In fact, many struggle to survive.

People often cite litigation as a reason for aircraft being expensive and say that one third of the price of a new airplane goes to liability insurance, but that still doesn’t account for most of the increase.

It can’t be regulation because the FAA was just as bureaucratic back then and most of the designs that are being produced today were certified in the 1950s and 1960s.

[Unrelated items from the same magazine: (1) We lost 8,314 bombers during World War II, counting only B-17s and B-24s and counting only those lost in the European theater. Each B-17 carried a crew of 10; each B-24 carried 7-10 men. (2) a Formula 1 driver survived a crash in which the G forces were estimated to be 178.]

14 thoughts on “Why are airplanes expensive relative to family income?

  1. I suspect if you factor in costs from all along the supply chain, one third going to pay off lawyers is an underestimate. Purely a subjective guess – not sure how you’d come at real numbers. Starting in the 1970’s we started to see insane and very expensive judgements land on small aircraft companies all along the supply chain. That was the time prices started to jump upwards.

    Which makes me wonder – will new aircraft manufacturers have to hike prices as they accumulate similar legal nonsense? Have they found a new way around the problem?

  2. One factor is that modern general aviation airplanes are equipped with tens of thousands of dollars of electronics (GPS, transponder, moving map, engine monitoring, intercom, etc.).

  3. Presenting the median income next to the aircraft price, as AOPA Pilot does, suggests that the price of the aircraft should naturally track the median income, but this is specious. We would be better served to first answer in isolation the underlying question of why the nominal price of the aircraft has increased so dramatically.

    In January of 1967, the M2 money supply stood at $486.4 billion. By November of this year, the M2 money supply had increased to $7,430.8 billion, a monetary expansion (inflation) of 1,528% [1].

    Since those new dollars have been bidding up the factors of aircraft production (land, materials, labor, etc.) [2], and the components required to build this aircraft have not changed substantially over time (the same amount of aluminum goes into the Cherokee now as did in 1967), it is not ridiculous to assume that the nominal price of the aircraft should, in the long run, follow the extent to which our Federal Reserve has diluted our dollars.

    Working forward from 1967, if the price of the Cherokee had followed the expansion of the M2 exactly, then it would cost $244,434 today, which is pretty close to the actual price of $230k.

    Since real wages have been on the decline, it would make sense if demand for Cherokees has diminished a bit over the years. Combined with the improvements in production efficiency that have been realized, it makes sense that today’s price should be somewhat lower than what extrapolation from monetary data would indicate.

    Looking at the question from another angle, in 1967, a Piper Cherokee cost about 450 ounces of gold. Today that airplane costs about 284 ounces of gold, representing a modest year over year decrease in the price level of Cherokee airplanes, which is what one would actually expect given technological improvements and the much greater availability of ready substitutes for new Cherokee aircraft (such as the large stock of used Cherokee aircraft).

    Over the last 40 years, the Federal Reserve has been sharply and steadily devaluing the dollar, which distorts the economy, effects a transfer of wealth to favored groups, and makes most of us poorer and less able to buy Piper Cherokee aircraft. This fact has been partially obscured by our fantastic advances in technology, communications, and productivity. Because of these advances, prices should fall year over year in most industries [3]. The extent to which prices have not fallen therefore represents a hidden ‘inflation’ that is never calculated when we look purely at the degree to which prices have steadily risen.


    [1] http://research.stlouisfed.org/fred2/series/M2NS

    [2] Boeing and Lockheed Martin, typically early recipients of newly printed dollars through lucrative defense contracts, need the same aluminum that Piper does, for example.

    [3] This seems radically contradictory to our lifetime of experience with rising prices, but with a moment’s reflection it should become clear why this is in fact the case (in a free market without a coercive and inflationary central bank). Through technology, society gets better at producing ever larger amounts of goods and services of value, i.e. wealth (both in aggregated and per capita). With more produced goods and services to go around, we should each have to give up less of the things we have a fixed amount of (our time, our money) to obtain some given amount of those goods and services.

  4. Publicly traded companies have a legal responsibility to their shareholders to make profit, and as I understand it the company directors can be prosecuted if they don’t make profit the number 1 objective of the business (even at the expense of other employees, the environment, etc.).

    So in 1967, if Piper were making 1 million profit a year, the next year the share holders would expect 1 million + inflation + a bit more or they’ll move their investment. And they expect the same every year, or the CEO (or another bigwig) will probably lose their job. So the prices go up.

    I’ve just noticed this is a Harvard Law blog…. I’d hope the reputation Harvard has mean you know more about economics than me, so the above might not be totally accurate….

  5. Phil –

    The point about the G-forces in an F1 crash might be a bit of a red herring if you don’t know how long those G-forces were sustained. A golf ball bouncing off the ground is subject to a few thousand G’s, and if it was to just instantly stick to the ground it would be half that. Of course, a human body is a bit more squishy, so it couldn’t stop as fast, but I would like to see the G-forces in faceplanting from 6 feet onto concrete, I bet they’re not insubstantial.

  6. Is this a phenomenon peculiar to airplanes? Have other niche luxury goods (e.g., yachts, expensive cars, high-end audio) similarly increased in price relative to normal income? I don’t know, since such things are well beyond my budget and taste. I do know that the cost of a home in some urban areas has increased far faster than wages, which threatens to make home ownership a high-end luxury for many.

  7. “air plane”: Piper is not a publicly traded company. In fact, they went bankrupt some years ago and their assets were handed, for almost nothing, to “New Piper”. Ergo this is not a story of increasing profits.

    Ted: If you got into an unairconditioned, noisy, slow Piper 4-seater, I don’t think you would call it a “luxury good”! True luxury goods have some cachet from being expensive, but there are usually cheaper substitutes from competitors (e.g., you can buy a Honda Accord if you don’t need the snob appeal of a BMW). Home ownership in Manhattan isn’t a good comparison because real estate there is scarce and there have been big reductions in crime and big increases in traffic congestion, both of which have made living in the city more desirable. An airplane is built in a factory. There is no limit to the number of airplane factories that can be built and you’d expect them to sell at the average cost of production plus a slight profit, which, to judge by the lackluster profits of the small airplane companies, is what they do sell for.

    Tweety: The latest airplanes do have some extra avionics, but I think the airplane manufacturers buy those for relatively cheap.

    Travis: Much of your analysis makes sense to me, but the same argument should apply to a lot of other simple manufactured goods that are still made in the same ways as they were in the 1960s. Yet we do not see huge increases in the prices of these other items compared to the income of an average household. It is true that some of these items may seem more expensive now that we are flooded with cheaper products from China, but in the photography world there are a lot of background supports, light stands, and other weird items that are still riveted together in the U.S. and Europe and cost about the same as they always did (relative to wages).

    Preston: I fear that you are right. I think the guys who make vacuum pumps, for example, are forever getting sued when an airplane crashes (on the theory that maybe the pump failed and shut down the attitude indicator (“artificial horizon”)). When you have a relatively small fleet of aircraft needing these pumps and they sell for a few hundred dollars each, there isn’t enough profit to pay for too many legal defenses. A friend had a Piper Malibu that needed a new hydraulic pump for his landing gear. The cost for the FAA-approved-for-airplane-use pump was $20,000. The identical pump was sold by the manufacturer for other applications for a few hundred dollars. What alternative do they have, though, if selling something for an airplane entails the risk of a massive lawsuit?

  8. You can get a Zenair STOL 801 with the exact same Lycoming engine as the current Piper and possibly with some of the fancy electronics for 64k.

    If you used a rebuit engine it can be about half that.

    Planes aren’t really commodity items like cars. The Piper has the brand recognition (since 1960) and I suppose that means something to some people. And there may be a market segment that won’t fly a home built plane.

  9. Philip:

    I should have noted in my original post that I certainly don’t dismiss the impact of tort actions and increased insurance costs, which inevitably propagate through the supply chain.

    I just don’t believe that they tell the whole story.

    The automobile industry provides a reasonable basis for comparison, as it has largely escaped massive liability judgments.

    In 1967, the average cost of a new car was $2,750 [1]. Today the average new car costs roughly $27,800 [2].

    So while a new car in 1967 cost the median earner 2.75 months of work, today it costs the median earner nearly 7 months of work (before taxes), an increase of 2.53x versus 3.59x for the Cherokee. Some of that discrepancy is undoubtedly related to liability. The point I wish to stress though is that both goods cost markedly more with respect to the median income, and this is true with a number of industrial products [3].

    It should immediately be pointed out that new cars today are clearly ‘better’ than cars in 1967, but also that they are largely composed of cheaper components. Cars today contain less steel and iron, and consume far fewer man-hours to build, than did automobiles of 1967. As advances in computer chips have amply demonstrated, the price of a commodity is dominated by the inescapable marginal costs (raw materials, labor, liability, etc.), not by the value delivered or even by the enormous engineering costs that are amortized.

    On a related note, this case study might be of interest to you. In 1967, tuition at MIT was $1,900 per year [4]; today it is $34,986 [5], an increase of 1,841%. Over that same period, the MZM [6] increased from $429.6B to $8,066.1B, an expansion of 1,878% [7]. If MIT tuition had exactly tracked MZM from 1967 forward, tuition today would be $35,674 – almost absurdly close to today’s number [8]. As with airplanes and cars, there are certainly other contributing factors here as well, but none I think that should represent more than a two-fold increase in price, whereas what we see here is a nearly twenty-fold increase in nominal price.

    And of course, as with cars and airplanes, the increase in the price of tuition at MIT has grossly outpaced income growth.


    [1] http://www.thepeoplehistory.com/1967.html

    [2] http://www.edmunds.com/advice/buying/articles/45310/article.html

    [3] Unfortunately, I was unable to find 1967 pricing for photography equipment in my brief search. I think though that we would probably notice it less if it had been creeping up relative to incomes, since the same multiple applied to a smaller nominal number tends to be less shocking to our sensibilities.

    A more significant factor to consider is that the price of photography equipment is much more strongly dominated by distribution and channel costs than is a Piper Cherokee, costs which have decreased dramatically in recent years.

    When comparing the price of goods made in China (and to some extent, anything that has to compete with them), it is necessary to keep in mind that the Chinese government has been buying up huge sums of dollars (and US government debt) to artificially hold down their currency. This represents, in effect, a subsidy from Chinese citizens to American consumers.

    [4] http://alum.mit.edu/docs/ne/reunions/1967_It_Happened_In.pdf

    [5] http://web.mit.edu/newsoffice/2007/tuition.html

    [6] I think that using a very slightly broader measure of money supply than M2 is justified in this case for the following reason:

    As new money works its way through the economy, it does not raise all prices at once. Rather, some groups naturally must get the advantage of spending the new money before it has bid up prices elsewhere. Depending on how the new money is introduced, it is even conceivable that some industries will not see price increases at all, or at least not for a long time.

    Because of the enormous amount of federal aid money, federal research grants, and other government subsidies that flow to universities, universities clearly experience the effects of newly created money earlier than most anyone else.

    [7] http://research.stlouisfed.org/fred2/series/MZMNS

    [8] Another plausible hypothesis for this almost uncanny correlation is that MIT administrators, through their exposure to the management of MIT’s enormous endowment, have learned that the true measure of the worth of a dollar is inextricably bound to the number of other dollars in circulation, and therefore aim to set the dollar price of tuition in such a way that it does not decrease in real terms.

  10. Travis: Thanks for the more detailed analysis. I hadn’t realized that cars had gotten more expensive relative to monthly income. Certainly there hasn’t been any significant substitution of cheaper materials and methods with Piper as there has been with cars. College tuition inflation is an interesting example. Since these are non-profit orgs, you’d expect them to charge what they pay for labor plus a small markup to pay for capital expenses and maintenance. The number of administrators at MIT per teacher has doubled since 1967, so that explains some of the price inflation. Then there is the fact that the facilities have been gold-plated, e.g., the $52 million new Z-Center gym (no soap provided!) augmenting the rotting old 25-yard pool (which subsequently was remodeled at what I’m sure was a staggering cost). Finally, I guess, we have to look at the $12 billion in cash that has been accumulated by MIT since 1967. The accountants say that the school loses money on education ever year; the bank statements keep getting fatter…

    N: If I can ever scratch up $1 million for a lake house here in New England, I want a Zenair CH 701 on floats for the dock. I’ve heard that they land so slowly that you can land in the water with the wheels down and it won’t flip over.

  11. At dinner tonight, my friend Richard brought up a good point. The production volume was much higher in the late 1960s, which reduced the per-plane cost of overhead. An airplane company needs marketing, accounting, sales, purchasing, engineering, quality control, FAA paperwork, etc. departments, whether it makes 1 plane or 1000 planes. At 1000 planes per year, the overhead cost per plane is considerably lower than at 1 plane per year.

    My friend Tom weighed in via email: “In most industries, being heavily regulated offers some protection against lawsuits. Not so in aviation: that’s really all you need to know to answer your question. Plane makers have been hit with a double whammy of regulation plus litigation. It’s a fatal combination.”

  12. I asked Richard at dinner how much he thought the aluminum in a light airplane cost and he said “no more than a few thousand dollars”. http://www.metalprices.com has aluminum at about $1 per lb., so that could mean between $1000 and $2000 for a typical four seater, though at least some aircraft aluminum parts are presumably some kind of higher cost alloy.

  13. To tweety’s point, I realize that there are big reliability issues between avionics and, say, bicycle electronics, but for under a grand I can put a GPS based navigation system with wind speed and barometric altitude on my bicycle, and get heart rate and a few other bits of data besides.

    I haven’t looked at the issues of how many extra nines avionics need versus automobile or bicycle telemetry systems, but I’d bet that the costs of producing modern avionics, even once you amortize out engineering, are less than the old mechanical systems, and the remainder is all in liability insurance.

  14. Safety and compliance with administrative requirements account for a lot of the costs here. A recent aircraft designed in Australia (the Boomerang) is a 2 seat light trainer built to FAR23 specifications. CASA (our aviation administration) put a great many roadblocks in the way and required some massive modifications to the prototype. This has reached a point where manufacturers like Cessna are not going for FAR23 any more in light trainers and are building to the recreational specification instead which is cheaper and less stringent.

    Then there is liability insurance. Recently a manufacturer of carbs for light aircraft announced that they can no longer produce float carbs due to the excessive liability insurance costs. Basically the costs of insuring a *part* were so high, that the company can no longer supply them to the market. Given they are one of the only suppliers of this part, general aviation could be in a spot of trouble here.

    To put things in perspective, for someone not wanting a career in aviation, the smart way to go is to fly recreational aircraft (under 650kg MTOW) which can be bought for 60k (australia) for an excellent little two seat Jabiru and have performance comparable if not better than a Piper Cherokee. Cheaper to learn, buy and maintain, and don’t look quite as dodgy as some of the older pipers and cessnas.

    BTW – I fly cherokees myself (Warrior II and III) and I’m not too phased about the cost of the airplane if it’s reliable and keeps me up in the air. The alternative isn’t worth thinking about.


Comments are closed.