Category: aviation


I recently read William Langewiesche’s Aloft (Penguin Modern Classics), his collected essays on flight.

One of these, justly regarded as a classic, deals with the loss of Valujet 592 near Miami in 1996, an accident which bears a strong resemblance, in his telling, to the parallel experience of rail privatisation in the UK. Deregulation permitted much of the business of running an airline to be reduced to contracts, in the “firm as nexus of contracts” model so beloved of business schools, and put out to tender. This resulted in ignorance, embitterment, ugliness, and failure. Go read the book.

But it struck me that a lot of Langewiesche’s work has a distinct message about the necessity of politics in the broadest sense. In American Ground he, among much else, says that the United States has the trick of drawing strength from argument, and a lot of his writing makes the case again and again for thrashing things out, for the danger of pensée unique and the more subtle danger of the consensus that substitutes for thought and conceals important prejudices and assumptions.

This requires, among other things, a degree of egalitarianism. You can’t have a useful argument with a steamroller that will simply crush you. It also requires a certain respect for the anomalous, the specific, the subculture, the historical quirk.

Another of the essays in Aloft is on air traffic control as a profession; Langewiesche makes the interesting point that the system has a lot of inefficient historical quirks and redundancies, which seem to be ripe for disruption and in need of modernising reform and you get the picture already, but which contribute in subtle ways to its resilience and its ability to recover from crises, which happen all the time.

His account of this recovery reminded me strongly of the Internet and especially of the NANOG mailing list, an institution which emerged, interestingly, around the same time as the US ATC system in its modern form, and which outsiders always want to clean up and straighten out.

He also discusses the notion of the “normal accident”, and makes a great point: couldn’t the idea that some things are just too complex and system-inherent accidents are inevitable be used as an excuse for big business to get out of its responsibilities? (Interestingly, the guy who invented it seems never to have thought of it that way.)

This is of course what happened with much of the broader social critique of planning, technocracy, modernism, and the like; although the critics hoped for anarchy or at least liberty, what they got when their ideas were implemented was neo-liberalism.

In the end, it strikes me that the meta-narrative here is that the institutions of anti-planning aren’t necessarily stable at the next level of abstraction up. They need constant policy inputs to function without ending up in one pathological mode or the other.

If you think that, you will probably end up in some sort of “varieties of capitalism” position. And you know? You can see market socialism from there with a good pair of binoculars. The rarity of this position explains why some of his readers probably find him conservative.

Experience the immense monotony

So far, the most embarrassing effort by a journalist to fill space on the MH370 story was certainly the ITN correspondent who opened their piece-to-camera with the words “Tuesday. Kuala Lumpur. The rain still falling.” You might hope that this was a deliberate allusion to “Tuesday. Africa. The hour of the lion”, supposedly the worst possible line of English prose, but this strikes me as optimistic.

Some people have more class. The Wall Street Journal, for example, has chosen to publish this short video about staring out of a small window in an aircraft at an unchanging sea, with the fine title Experience the Immense Monotony of an Ocean Search. Go on. Experience it.

There is an interesting rhyme here in the way that any hope is now symbolised by the image of Australian P-3C pilots staring impassively at the vast spaces of the Indian Ocean, presumably rather like the captain (or hijacker) of MH370.

Real air forces own boring stuff

David Axe’s piece on the Brazilian air force did the rounds a while ago. The obvious points: they do their stuff, and for cheap! Less obvious: there’s something disturbing about a state that regularly puts 8x Mk82 Snake-eyes into targets on its own territory as a police measure, even if the Super Tucanos are an economical platform that usually shacks the target and they have to deal with criminals who build their own airfields.

The implications in terms of the politics of state-building are discussed here, and you could be snarky and ask what the country behind this story is doing practising 1920s style air control.

But the point that struck me was that they chose to spend their money on the infrastructure of air power, building an airborne early warning, an intelligence-reconnaissance, and a maritime patrol type on the basis of the same aeroplane, one of their own Embraer regional jets. Now, when they’re looking at replacing the fleet of much-upgraded F-5s, they’ve selected the Saab Gripen NG, probably because sensor-fusion and communications are its specialist subject in the way the Eurofighter’s is so much power that the main thing to learn is:

managing the awesomeness

Compare the Royal Malaysian Air Force, which has managed to accumulate a fleet of F/A-18D Hornets, another fleet of Sukhoi-30MKI, but not to get enough radar to find a missing 777 an hour from their capital’s international hub airport. Some nations have air forces that choose useful; others, that make salesmen and a few fighter pilots happy.

This is relevant and depressing for the UK, as after all, between us and the Aussies, the RMAF is basically our doing, and more usefully, we decided we didn’t need our own maritime or overland ISTAR aircraft and the remaining one is theoretically only safe until next year. Another notch on the belt towards low-trust status? Think Defence has sensible thoughts on how to fix that bit.

Boris Island: Shag, Marry, Avoid?

The Airports Commission is eventually going to opt for Gatwick. Prediction. How did I come to this conclusion? And how representative is the commission as an institution?

Well, we can understand the decision-making process here with a simple model. There are various actors involved, who have preferences. To illuminate this, let’s play a game of Shag, Marry, Avoid. We could call these “Tolerate, Demand, Veto” instead, but I think it’s much more fun this way. Basically, any actor has a choice they would prefer, one they would accept failing that, and one that they would reject entirely.

Once we’ve specified the actors, and coded them as Shag/Marry/Avoid for each airport option, we can then look at a rule to model the decision process. This is the table I came up with.

Under my scoring, first-past-the-post would return the Rt. Hon. Heathrow Airport MP on the strength of its 8 marriage proposals. However, this isn’t a parliamentary election, and second choices and veto power are both important. Under an alternative vote-like system, I get 8 votes to marry Heathrow and 8 for all other options. Typically, AV requires 50%+1 for a first round win, so we reallocate the shags – i.e. the second preferences. This gives us Heathrow 11, Gatwick 13, and (Boris Island 3, Stansted 3, Somewhere Else 2) drop out. If we then subtract the avoids, i.e vetoes, we get Heathrow 7, Gatwick 12 and (Stansted 3, SE 2, Boris Island -8).

Another way of doing this would be something like approval voting, where voters cross off all the candidates they reject, and the least rejected candidates win. This gets at the minimum consensus aspect of the whole thing. SE and Stansted both have 0 vetoes, Gatwick 1, Heathrow 4, and Boris Island 11.

This of course raises the problem with approval voting – unless you do something else, it’s possible to have a winner nobody positively wants. In real institutions that use it, this is often dealt with by requiring a given number of electors to second you before you get to take part.

If we require more than one “marry”, Stansted and Boris Island are both eliminated, leaving Somewhere Else, Heathrow, and Gatwick. SE has no vetoes, Gatwick 1, and Heathrow 4. SE wins, but if there has to be a winner, then it’s Gatwick. Alternatively, as I’ve included lobbies for the North and Midlands in the scoring, we might read that as being “Gatwick, plus expansion outside the South-East”.

Yet a third option would be a balance of opinion. We might count a Marry as one point, a Shag as 0.5 points, and an Avoid as -1, implicitly coding a Shrug as 0. In that case you get net scores of Heathrow 5.5, Gatwick 8, Boris Island -9, Somewhere Else 2, Stansted 1.5.

The upshot, then. All three rules tried converge on the same result. Hardly anyone actually wants Boris Island and a lot of people hate it. Quite a lot of people support Heathrow, but a significant group of actors also hate it. Gatwick is everyone’s second choice – quite an advertising slogan, no? – and only one group wants to veto it.

And further, the airports commission does seem to be surprisingly democratic, at least in that you can simulate it as a plebiscitary process with the same inputs and it gives the same outputs.

No surprise blogging

From Crooked Timber:

Speaking of hypocrisy, what of the revelations (big surprise) that CIA had been briefing Pakistan govt on the drone strikes, while the latter was simultaneously denouncing them?

From the blog, December 2012, I point out that the CIA and the Pakistanis were evidently cooperating in so far as they coordinated their use of airspace over the NWFP, keeping the drones, army helicopters, and Pakistani F-16s out of each other’s way.

They weren’t just advised of the flying programme, but also of the targets and the intelligence files.

The awkward stretching movement

He’s referring to Peter Day of BBC News’s piece here. But the best bit is this bit:

Immediately, zonk! Back comes the seat as the person in front assumes relax mode as a default position. The mean battery hen space already allotted to an economy class passenger is clipped even further. You are suddenly presented with eight or 12 hours of prison.

So, faced with the anonymous unpleasantness of the passenger in front of me, I made a public vow. Never again – I promised listeners – would I press that tempting shiny button on the right hand side of the armrest. And ever since, I have not

I do this, or rather don’t, too, for precisely the same reasons. I may not be able to stop you flopping into my personal space but at least I can avoid inflicting it on the person behind me, and who knows, perhaps it’ll catch on.

I mean, first there was just me and now there’s this BBC guy, and pretty soon they’ll call it a movement. A sort of cramped, sleepless movement with a back like a greasy, Moldovan tugboat, but a movement none the less. An awkward stretching movement.

He’s also very much right about the dullards watching The Fast and the Furious 7: Deepening Righteousness when they could be looking at huge cracks in Hudson Bay or whatever.

Railtrack in the sky, at war!

To expand on something I wrote as a comment, one of the things I hate about wizard privatisation schemes and especially “total outsourcing” or whatever is as follows. The FSTA deal for the RAF’s new jet tankers is especially awful and exemplary.

Here’s the idea. Rather than buy the aeroplanes and do the job, we’ll buy “what we need” – a guarantee of however-many flying hours and tonnes of offloaded fuel where required. This will be awesome, because the availability is guaranteed. At the same time, the PFI contractors get to (supposedly) charter out the planes when not needed, which makes them money. This pays some of the costs and therefore it’s meant to be cheaper.

Guarantee, you say? What guarantee? What are these planes that never ever break down? Why can’t the RAF just buy some of them? Well, of course, they will break down but there’s a contract that says Airtanker will give the MOD a refund if they miss a slot. What good the refund is if you crash in the sea is left as an exercise to the reader. Obviously, they won’t just accept it, they’ll argue bitterly and sue everyone, and if Airtanker can’t airtank, does anyone imagine that the Treasury solicitor will find any money there in receivership?

You’ve seen this story, you’ve lived it – it’s Railtrack. Everything was ponies because contracts. It didn’t work, it cost more to litigate the disputes than it cost to build actual trains, and when it fucked up, there was no money to pay out the compo. It’s the Treasury’s attachment to the myth of immaculate compensation that’s doing the work here. As Bob Dylan sang, they say everything can be replaced

And this thing is costing £14bn for 11 planes, that sell for vastly less to the airlines. Well, say the PFI panjandrums, that does include every sheet of toilet paper used by the ground engineers, and all the fuel. But this is precisely the problem. How confident are you that their cost accounting is honest? You can’t really control it. How do you audit the price of jet fuel 20 years hence? There’s a reason why the Americans have the phrase, to sell someone a bill of goods.

Closer to home, think of one of those guys in the Tottenham Court Road flogging computers. If you take the pricier option, of course, I’ll throw in a bag and some USB sticks and an airbomb and that neat little fart-robot. The margin is on that stuff. If you take it you’re a sucker. It’s a similar process to the one I described here. Ryanair works by disaggregating its price, Airtanker and friends by aggregating it, but the point is the same – moving economic activity from the domain of competition and of legibility into that of monopoly and of mystery.

Both have the effect of destroying information and therefore efficiency. This is why taking railway maintenance back in-house saved money at Network Rail.

Update: I forgot to include this. US healthcare is basically completely think-of-a-number, nothing has a single price.

Don’t read this, read them!

Good NYT piece is good – working through the effort to arm Syrian rebels, with Saudi and Qatari money, Croatian surplus warstocks, and Jordanian airlift. Inevitably, the kit is moving aboard Il-76s. This time out, though, the aircraft are Jordanian (and occasionally other) military aircraft that sometimes operate as “Jordan International Air Cargo”, a nationalised freight line.

I don’t have much to offer except that it’s great to see the method applied, and it’s fairly common in that part of the world for an air force to have a semidetached heavy lift operation. Libyan Air Cargo operated their air force’s An124, Maximus is the UAE’s version of the same idea.

Awesomely, someone’s virtual-radar box on Cyprus collected the data. This seems to be well reported, so I’m not going to make a special effort. Hugh Griffiths of SIPRI is anyway involved.

The drone that didn’t crash

David Axe has a good piece on how vulnerable drones are to any kind of organised air defence.

“The predominant use of RPAs [Remotely Piloted Aircraft] over the past decade has been passive [intelligence] collection coupled with air-to-ground strikes in permissive airspace,” Air Force spokesman Lt. Col. Tadd Sholtis tells Danger Room. “There’s very little about their current capabilities — from their speed to their maneuverability to the range of visibility afforded to operators — that encourages operators to arm or employ them for air-to-air engagements in defended airspace.”

I thought I’d use this as a hanger to discuss a point I wanted to make in a row on Crooked Timber a little while back. Basically, it’s possible to deduce how much the Pakistani government really supports the drone program by looking for clues. Supposedly, the Americans just fax them a list and presume that they agree if no objection comes back.

However, we very rarely hear of a drone colliding with another aircraft over Pakistan. (In fact, the only cases I can think of were in Afghanistan, in airspace the US Air Force controlled.) Neither, so far as anyone knows, does the Pakistan Air Force (which is quite large, has modern aircraft, radars, and command-and-control systems, and did well against India and also in engagements with the Soviets during the 80s) treat them as unidentified contacts and intercept them for identification, still less shoot them down. Pakistan has carried out some very large military operations in the same places where the drones operate, using their air force, their army’s helicopters, which would tend to work in exactly the same air space as the drones, and indeed their own surveillance drones.

How did they manage? The answer is one word: deconfliction. Deconfliction is the process of planning air operations (peaceful or warlike) so that they do not conflict with each other. The fast jets have to keep out of area Z below 15,000 feet before 0650 so that the army helicopters can transit through it safely, while the drone has to stay above the close air support planes’ killbox and below the refuelling tanker’s high altitude towline pattern. It’s basically what air traffic controllers do, but completely vital if you want to use air power without this happening.

Now, as far as anyone knows, the US doesn’t object to Pakistani operations on the NW Frontier. Quite the opposite. This implies that the CIA is also deconflicting its drone flying programme to facilitate Pakistani operations.

It’s possible that both parties fax (fax!) each other tomorrow’s flypro, they never reply or acknowledge it, but the air-tasking cell on both sides acts on the information. It is certain, however, that they are cooperating, because of the lack of regular drone-Mi8 helicopter collisions, drone/fighter intercepts, etc.

A brief inquiry into the nature and consequences of think-of-a-number pricing

Inspired by this weekend’s story that you now pay a transaction fee to fly Ryanair even if you use a Ryanair credit card, which is apparently a thing that exists even if it sounds like it shouldn’t, I have been thinking about their business model. It is not what you may think it is.

This is one of those stories where there’s a business school whiteboard version, and then there is reality. The whiteboard version goes like this: they disaggregated all the service elements formerly bundled into the airline ticket, and priced them. As a result, through the wonder of revealed-preference, we know that people value some of them enough to pay for them separately. You then repeat the words “Southwest Airlines” several times, and ponies!

The cynical version starts off like this: ok sunshine, you expect me to believe a single write operation on their OLTP system costs £6? Seriously. If FR’s IT was that dreadful, they wouldn’t be able to run an airline.

Now, here’s the clever bit. Fares are prices; people mostly pick their carrier on fares. This is of course the point of running a low-cost airline. In case you doubt that low headline fares are the attribute of selection here, look at their website or adverts or airport presence. It says “The LOW FARES airline” everywhere.

You can’t choose to fly FR and be handled by some other company, so once passengers pick them, there is no competitive pressure on the fees. By transferring things from the headline fare into non-fare fees, part of the business has been removed from the domain of competition and moved into the domain of monopoly.

Now, as a rule, if something is more monopoly-like, do we expect its price to contain more or less margin? I think we know the answer to that one.

This may also create an opportunity for tax-dodging. Imagine a company that puts the IT systems that do its check-ins and credit card processing into a subsidiary that lives in, say, Luxembourg. You could move income from the operating business into the subsidiary. I have no idea if FR does this, and finding out would involve reading their report and accounts. Reading financial filings is something I do if people pay me, like enduring carols, riding, making PowerPoint slides, and going to Dubai.

However, they’re the airline that charged all its flying training candidates £50 to read their CVs, and then charged everyone who paid it by credit card another £50 three months later to “reconsider”, so…

Upshot! Disaggregating bundled products does not necessarily increase competition. In fact, it can actually transfer economic activity from the competitive sector to the monopoly sector. This creates opportunities for think-of-a-number pricing.

Think-of-a-number pricing is pricing that has no necessary link to economic reality, and is only constrained by a vague sense of plausibility. It is what happens when your electricity company decides to “estimate” your usage between two actual readings and sends you a gigantic bill, in the hope you’ll just pay it rather than calculating how much power over how much time that means and explaining to their call centre that this would have required quite an impressive industrial-grade circuit, and that a standard residential supply would have burned out in a giant blue flash, and as you aren’t speaking from beyond the grave, repeatedly threatening to change provider…until the enemy cracks and settles for a modest increase in the monthly payment.

Unlike normal prices, think-of-a-number prices convey no information whatsoever, other than the fact you’re being ripped off, which is only of use if you have alternatives. The informational function of prices is kinda important for the whole edifice of economics, and is the entire basis for the notion that tax-funded spending is by definition inefficient, so this is a non-trivial point.

From a business, rather than economic, point of view, this of course means that disaggregation creates margin.

In the last analysis, though, something very interesting is happening here. Weirdly, if you bundle all the services and sell them for a single sticker price, like a proper airline, they have a real price! Because, after all, you still compete on fares, so you need to keep costs down. The IT department has to contribute to the competitive effort. But if you break them out, they don’t, and because of think-of-a-number pricing, economic information is destroyed! And, as no competitive pressure is involved once you’ve sold the ticket, it is inevitable in a profit-maximising firm that an element of think-of-a-number pricing will happen!

And we have an existence-proof that people do actually behave in the way this requires, buying on headline price and then submitting to think-of-a-number pricing: it’s called Ryanair. Now, we could also go on to discuss their user experience design and the notion of “cooling out the mark”, but this blog post is long enough.