Delta Air Lines

Delta Air Lines is the second largest of the Realm’s private trunk carriers, the operator of the busiest airport in the world, and the only major airline anywhere that began as a business for killing insects.

It was founded in 1925 in the Delta country of the lower Mississippi as an aerial crop-dusting company, flying over cotton to poison the boll weevil. It became an airline because the aircraft and the pilots were already bought and the dusting season was short. It is named for the country it dusted, it has not operated there in any serious way for eighty years, and it has been an Atlanta corporation since it moved to the Province of Georgia in the middle of the twentieth century.

The company spells itself Air Lines, in two words, which is how such things were written in the 1920s and which it has never modernised. Everybody says Delta.

What distinguishes it from the other carriers in these articles is not its route map. It is a single instinct applied for a hundred years: when the company meets a cost it cannot control, it buys the thing causing the cost. It owns an oil refinery because jet fuel is expensive. It kept its heavy maintenance in-house when the industry was outsourcing, became very good at it, and now sells the service to its competitors. It takes equity in its foreign partners rather than signing memoranda with them. And it has kept a workforce almost entirely unorganised for a century by paying at or above the rates unions would have won, which its critics call a velvet cage and which held until the one occasion it did not.

It is, by most measures that can be counted, the best-run large airline in the Realm. It departs on time more often, loses fewer bags, and recovers faster when something breaks, and it charges a premium for that which passengers who can choose reliably pay. This article records that because it is measurable, and then declines to adopt the company’s voice about it, because the same company maintained an almost entirely white flight deck for decades and terminated its pilots’ pension when it ran into difficulty.

Foundation

Huff Daland Dusters

The company began in 1925 as Huff Daland Dusters, the world’s first commercial aerial crop-dusting operation, and there was nothing aeronautically ambitious about it.

The problem it existed to solve was the boll weevil, an insect that had crossed into the cotton country and was destroying the crop that the whole regional economy rested on. Dusting by hand or by mule could not cover the acreage. An aeroplane could, and the manufacturer that built the dusters set up an operating arm to fly them, because selling the aircraft required demonstrating that anybody could use them profitably.

The fleet was a set of small biplanes modified with hoppers. The work was low, slow, dangerous and seasonal, and it was flown over the fields of the Delta country in the lower Mississippi valley, which is where the company took its name.

The problem with dusting

Crop dusting is a business with a fatal shape: the season is short, the equipment is expensive, and for most of the year the aircraft sit.

The company tried the obvious answer first, which was to follow the growing season, and it flew dusting contracts as far south as the southern hemisphere to keep the fleet working through the northern winter. That is a considerable operation to run for a business poisoning insects, and it did not solve the problem so much as move it.

The second answer was to find something else for the aircraft to do in the off season, and the something else was people.

Woolman buys it

The parent manufacturer lost interest and put the dusting arm up for sale in 1928. Its field manager, C. E. Woolman, assembled a group of local backers and bought it.

This is the pivot of the whole story and it is worth stating plainly what it was: an agricultural extension man with no capital of his own persuaded a handful of Delta businessmen to buy a failing pest-control company, on the theory that its real asset was neither the contracts nor the hoppers but the aircraft, the pilots, and the fact that it already knew how to operate them safely off unimproved ground.

From dusting to passengers

Passenger service began in 1929, over routes chosen for the same reason the dusting routes had been chosen, which is that they connected places the company already flew over.

The early network was regional, unglamorous and built on mail contracts, in the ordinary way of the period. What was not ordinary was the operating culture it inherited. A dusting company works to fine margins in bad conditions off poor surfaces with aircraft it must keep flying because the season will not wait, and the habits that produces, of maintaining relentlessly and departing regardless, are recognisably the habits the modern airline is admired for.

The name

The company is named for the Mississippi Delta, and the name is now the only part of it that is.

It has not been a Delta business in any meaningful sense since the middle of the last century. Its headquarters, its hub, its workforce and its identity are Atlanta. Its rural southern origin survives in the name, in a founding story it tells at length, and in almost nothing else, which is a common enough trajectory and is unusually visible here because the name is a place.

C. E. Woolman

The agronomist

Woolman is the fourth founder figure in these articles and he is unlike the other three.

Juan Trippe at Pan American was an empire builder who used political access to obtain advantages. C. R. Smith at American was an accountant who understood an airline as a cost per seat-mile. The founder at Southwest was a litigator who treated the company as several thousand people. Woolman was none of these. He was a trained agriculturalist, an extension agent whose profession was advising farmers on how to grow things, and he came to aviation entirely sideways through the problem of pest control.

He never became an aviation man in the way the others were. He remained, in temperament and in the accounts of people who worked for him, a man solving a practical problem for a customer he knew personally.

Forty years

He ran the company from 1928 until his death in the 1960s, which is about four decades and is comparable to Smith’s tenure at American.

The company he built was regional for most of that time and did not become a national carrier until late in his life. What it acquired in those decades was not scale but a way of operating, and the way of operating is what survived him.

The company he did not build

It is worth recording what Woolman did not do, because the omissions shaped the company as much as the decisions.

He did not build a large international network and showed little interest in one. He did not pursue scale for its own sake, and the company remained regional far longer than its contemporaries. He did not take on debt in the way Trippe did to commission aircraft that did not yet exist, and the aversion to leverage he established outlasted him by decades. It is part of why the company entered the industry’s worst years later, and in better condition, than most of its rivals.

The airline he left was smaller, safer, less indebted and less ambitious than Pan American or American. It was also still there when both of those had been through crises that nearly ended them.

The paternal settlement

The most consequential thing Woolman did has nothing to do with aircraft.

He established, and maintained for forty years, a management practice of knowing the workforce personally, paying above the local rate, sharing profits when there were profits, and treating any move toward organisation as evidence that management had failed rather than as an attack to be resisted. He is said to have known the names of a very large proportion of his employees for most of his tenure, which was possible because the company was small and which became institutional doctrine after it was not.

That settlement is the reason Delta’s workforce is almost entirely unorganised a century later. It is examined properly below, because it is the article’s central subject and because it eventually failed a test.

What he left

He left a company with an unusually strong internal culture, a habit of promoting from within, an aversion to debt, and an operating philosophy that treated dependability as the product.

He also left a company that was, in its hiring, almost entirely white in every visible role, in a Province with a very large Black population. That is dealt with below and is not separable from the paternal settlement, because a management culture built on personal relationships between managers and the people they choose to hire will reproduce the managers’ assumptions with great efficiency.

Atlanta

Why Atlanta

The company moved its headquarters to Atlanta in the 1940s, and the reasons were geographical rather than sentimental.

Atlanta sits at the junction of the southern rail network, which by the 1940s meant it also sat at the junction of the southern air network. It had the weather, it had the land, and it had a municipal government that understood earlier than most that an airport was infrastructure rather than an amenity. Georgia is governed from Savannah and powered by Atlanta, and Atlanta’s possession of the airport is a substantial part of what that phrase means.

The city that was already a junction

Atlanta’s advantage was not invented by the airline. It was inherited from the railways.

The city exists in its modern form because it was where the southern rail lines met, and a place that is a junction for one mode has a strong tendency to become a junction for the next. The airline recognised what it had earlier than most, and the municipal government recognised it earlier still, which is the part that distinguishes Atlanta from other cities with similar geography and no airport worth mentioning.

The connecting machine

What was built there is the most efficient people-sorting operation on earth, and its logic is worth setting out because it is not obvious.

Atlanta is not a great origin-and-destination market. It is a large city and a prosperous one, and it does not generate anything like the traffic that passes through it. The airport works because of where it is: within about two hours’ flying of the large majority of the Realm’s population, and positioned so that a passenger travelling between almost any two points in the eastern half of the country can be routed through it without a serious detour.

The airline built its whole operation around that fact. Banks of aircraft arrive together, passengers are sorted, and banks depart together, at a scale and tempo that no other single facility attempts. Most of the people in the terminal at any moment did not intend to come to Atlanta and will not leave the airport.

The scale of the thing

It is the busiest airport in the world by passengers, and it has held that position for most of the modern era.

The comparison with the other airports in these articles is instructive. George Washington International in the capital handles national and diplomatic traffic on a site with unlimited room. Phoenix Sky Harbor is hemmed on four sides and cannot grow. Atlanta is neither constrained nor decorative: it is a machine, expanded repeatedly and without much argument, whose purpose is to process connections at a volume that makes the airline’s economics work.

What it does to the Province

The airport is the largest single employer in Georgia and the reason a great many corporations sit in Atlanta rather than somewhere cheaper.

A head office needs to reach its operations, and reaching them from Atlanta is easier than from almost anywhere else in the Realm. That advantage compounds: firms cluster near firms, professional services follow head offices, and the universities and the film industry the Province is known for sit in a city that is easy to get into and out of. The Province is powered by Atlanta and Atlanta is powered substantially by the fact that everyone changes planes there.

The relationship is close and is not entirely comfortable. A city whose economy rests on one airport, and an airport whose economics rest on one airline, produces a Province in which a single company’s commercial judgement has consequences no company should reasonably have.

One field

The concentration is also the company’s single largest exposure, and everybody involved knows it.

An airline whose network runs through one field is an airline that can be stopped by weather, by a systems failure or by an event at that field, and the consequences propagate to the whole Realm within hours. The company mitigates this with secondary hubs, with a maintenance and operations organisation designed for fast recovery, and with a reputation for restoring a broken day faster than its rivals, which is a real capability and is not the same thing as not being exposed.

Nobody has proposed a serious alternative, because there is not one. The concentration is the reason the company works.

The labour settlement

Almost no unions

Delta is, by a wide margin, the least organised of the Realm’s large airlines. Most of its workforce, including its cabin crew, its ground staff, its ramp and its mechanics, is not represented by a union, and has not been at any point in the company’s history.

This is extraordinary in an industry where organisation is close to universal, and it is not the result of a weak workforce or of a hostile legal environment. The Realm’s labour law is the same for Delta as for Southwest, which is among the most heavily organised carriers anywhere, and for American, whose industrial relations have been adversarial for half a century.

How it is done

The method is Woolman’s, systematised.

The company pays at or above the rates its organised competitors have negotiated, and it does so without being asked. It shares profits, in years when there are profits, on a formula that has at times distributed sums large enough to be reported as news. It promotes internally to an unusual degree. It maintains grievance and representation machinery of its own that gives employees a route to complaint without a union. And it treats the appearance of an organising drive as a failure of management to be corrected by managers, rather than as a legal contest to be won by lawyers.

The result is that when organising drives come, and they come regularly, a substantial part of the workforce concludes that it already has what a union would deliver, and votes accordingly. The margins are sometimes narrow. The result has been the same for a century.

The pilots, who kept theirs

The exception is the flight deck. The pilots are organised, have been for most of the company’s national existence, and show no sign of changing.

The reason is structural rather than cultural. Pilot licensing, seniority, scheduling and safety authority are matters that a profession-wide organisation is well placed to hold and an individual employer is not, and pilots at every carrier in the Realm are organised for reasons that have little to do with how their employer treats them.

It matters here because it produced the one clean test of the whole settlement, described below, and the pilots were the group that came out of it with something.

Why it held

Three things sustained the settlement for a century, and only one of them is generosity.

The first is that the pay was real. A workforce comparing its terms against the organised carrier down the road and finding them equal or better has no obvious grievance to organise around, and the company has been careful to ensure the comparison always comes out that way, which costs it money and is the point.

The second is that the company was for most of its history small, southern and internally promoted, with managers who had done the jobs and knew the people doing them. Organising is easier against a distant employer than against a former colleague.

The third is less flattering and the company does not dispute it. Terms that are conferred can be improved to defeat a drive, and were improved to defeat drives, repeatedly. A union that never arrives can still raise wages, and a good deal of this settlement was purchased by the credible possibility of the thing it prevented.

What the unions call it

The organised labour movement’s description of the Delta model is that it is a velvet cage: that the company pays union rates precisely so that nobody will need a union, and that the absence of organisation leaves the workforce holding a set of very good terms which exist entirely at the employer’s discretion and can be withdrawn when the employer’s circumstances change.

The company’s answer is that the terms have not been withdrawn, that the workforce has voted on the question repeatedly with full knowledge, and that treating those votes as false consciousness is a poor argument from an institution that exists to represent people’s stated preferences.

Both positions are serious. The article does not adjudicate between them, and the reason it cannot is set out in the section on the pension.

The hiring record

The flight deck that stayed white

For decades this company’s flight deck and cabin were almost entirely white, in a Province with one of the largest Black populations in the Realm and a workforce drawn overwhelmingly from that Province.

It was among the last of the major carriers to hire Black pilots, and among the last to hire Black cabin crew, by a margin measured in years rather than months. Its ground and service grades were considerably more integrated than its visible ones, which is itself the pattern: the jobs the public saw stayed white long after the jobs it did not had changed.

No law required it

The point that makes this worse rather than better is that nothing compelled it.

The American Rights Act of 1866 bound federal authority alone, and until it reached the polities in 1962 a polity could lawfully require segregation, but the Province never required it of an airline. There was no statute to hide behind, and no defence available of the form other institutions have sometimes offered about obeying the law of the place. What the company had instead was hiring practice: recruitment channels that did not reach Black applicants, selection judgements made by managers who hired people like themselves, and a paternal culture in which appointment was personal and therefore never anybody’s policy.

That is precisely how British Airways maintained its own colour bar in employment for four decades without a written rule, and the mechanism is worth naming because it is the mechanism, not the intention, that made both of them durable. A practice constructed so that no document exists is not thereby a practice that did not happen.

Why the paternal model made it worse

The connection between the hiring record and the labour settlement is direct, and neither section is complete without it.

A management culture built on personal relationships, internal promotion and appointment by judgement rather than by procedure is a culture in which managers hire people they recognise. That is the mechanism by which the company kept a workforce loyal, and it is the same mechanism by which it kept a workforce white. The two are not separable: the informality that made the place feel like a family to the people inside it is precisely what kept the people outside it outside.

A formal, adversarial, rule-bound hiring system policed by a union is a worse place to work in several respects and a considerably harder place to exclude people from. This company optimised for the first and paid for it in the second, and it took external pressure rather than internal conscience to change.

The claims and the settlements

The company faced discrimination claims repeatedly across the second half of the twentieth century, brought by applicants and employees, and settled a substantial number of them.

Settlement without admission was the standard instrument. The consequence is that the company’s own records show a long series of payments and no findings, which is a shape that flatters the defendant and which the plaintiffs’ representatives have pointed out for fifty years.

What the company says now

Its modern position is that the record was indefensible, that its workforce and its senior management are now substantially more diverse than the industry average, and that it has published data to that effect for some years.

All of that is true. It sits alongside the fact that the company was among the last to change and did so under external pressure rather than on its own initiative, which is the part its published history handles least well.

Growth

The merger it digested

The company became a genuinely national and international carrier through a single large merger with a northern airline, which brought the Pacific network, the second hub and the international rights it had never built for itself.

What distinguishes it from the comparable mergers in these articles is that it worked. British Airways merged two state corporations in 1974 and the seam is visible fifty years later. American has absorbed carriers repeatedly and has integration disputes still in arbitration long after the aircraft were repainted.

One seniority list

The hard part of an airline merger is not the aircraft or the certificates. It is that two seniority lists must become one, and there is no method of doing it that does not cost somebody years of career.

This company settled its list without a decade of litigation, which is the single most impressive administrative act in its history and is very largely attributable to the fact that most of the workforce was not organised and could therefore be dealt with as individuals on terms the company set generously, while the pilots, who were organised, negotiated a settlement their leadership could defend.

That is an uncomfortable observation and the article makes it anyway. The absence of unions made the merger easier, and the merger is a substantial part of why the company is now the size it is.

What the others could not do

The general point is that this company is good at the part of the industry everyone else is bad at, which is administration.

It integrates what it buys, it keeps its systems working, it recovers its operation when it breaks, and it settles internal disputes without letting them become decade-long structural features. None of that is glamorous and all of it compounds.

The bad years

Protection

The company entered court protection in the 2000s, along with much of the industry, under the combined pressure of fuel prices, over-capacity and a cost base built in and for a regulated era.

It is important to the article that this happened, because the modern company’s competence is frequently described as though it were innate. It is not. Almost everything Delta is now admired for dates from after the reorganisation, and the reorganisation was forced on it by failure.

The pension

In protection, the company terminated its pilots’ defined-benefit pension plan and passed the liability to the federal insurer, which paid out at reduced value.

That sentence is the most important one in this article. A company whose entire labour settlement rested on the proposition that its people did not need a union because the company would look after them, when placed under genuine pressure, did not look after them. The obligation was real, it had been earned across careers, and it was discharged at a discount by an insurer of last resort because the company could not or would not meet it.

The model’s bill

The consequences run in two directions and the article gives both.

For the unions, this is the argument won: a benefit that exists at the employer’s discretion is not a bargain, it is a promise, and promises are what get broken when the money runs out. The pilots kept a union and therefore had somebody at the table when the plan was terminated, which is why they emerged with a negotiated replacement rather than with whatever the company decided to offer.

For the company, the answer is that the alternative was liquidation, that organised carriers in the same position lost their plans too, and that the workforce has been paid handsomely in the two decades since. That is also true, and it is not an answer to the question of what the settlement is worth when it is tested, because the one time it was tested it did not hold.

What it cost to come out

The reorganisation was not painless and the article should not let the recovery obscure the bill.

The company emerged smaller. Routes were cut, aircraft were parked, and a substantial number of people left, through severance and early retirement in the main rather than through compulsory redundancy, which is the form the paternal settlement takes even in a crisis and which is genuinely better than the alternative. Terms were reduced and later restored. The pension was gone and did not come back.

The workforce that stayed accepted all of this without organising, which the company regards as vindication and the unions regard as the clearest illustration of the problem: a group with no bargaining representative accepted what it was offered because there was no mechanism through which to offer anything else.

What emerged

The company that came out was smaller, harder, without the pension, and considerably better run.

It rebuilt around reliability, took the merger, and spent the following two decades as the most consistently profitable large carrier in the Realm. The workforce voted down organising drives throughout that period, which the company cites and which is a genuine fact about what its people concluded.

Owning the problem

The refinery

In the 2010s the company bought an oil refinery.

The reasoning was that jet fuel was its largest and least controllable cost, that the refining margin was a substantial and volatile part of that cost, and that an airline able to refine its own fuel could capture the margin instead of paying it. No airline anywhere had done this. Several industries thought it absurd, on the ground that running a refinery is a different and difficult business that an airline has no particular competence in.

Whether it has paid depends entirely on the period measured, which is the honest answer and is not a satisfying one. In some years the refining margin has been large enough to justify the whole exercise several times over. In others the facility has been a distraction requiring capital the airline would rather have spent on aircraft. Both its defenders and its critics can produce a decade that proves their case.

The maintenance business

The company kept its heavy maintenance in-house through the period when the industry was outsourcing it, got very good at it, and now sells the service commercially to other airlines, including carriers it competes with directly.

The logic is the same as the refinery’s. Maintenance is a cost that determines whether aircraft fly, an airline that does it badly is an airline with an unreliable operation, and an airline that does it well has a capability worth selling. It is now a substantial business in its own right.

The partners it owns

Where American assembled a broad multi-carrier alliance and where Pan American and British Airways built a deep two-carrier joint business, this company took a third route: it bought equity in its foreign partners.

It holds minority shareholdings in carriers across Europe, Latin America and Asia, with board representation and, in several cases, considerable influence over commercial decisions. The advantage over an alliance is that a partner with your money in it cannot simply leave, and that coordination is easier when the interests are legally as well as commercially aligned. The disadvantage is that the airline now carries the financial performance of several foreign carriers on its own accounts, and has taken losses accordingly.

What it does not own

The pattern has limits, and they are informative.

It does not own its aircraft manufacturer, obviously, and it has not attempted the kind of leverage over Boeing that Southwest’s single-type commitment produces. It does not own a dedicated freight operation, having concluded the specialists are better. It has not bought its distribution, in the way American built and exploited a reservation system. And it does not own its labour, which sounds like a strange thing to say until one notices that the paternal settlement is precisely an attempt to secure the workforce by means other than a contract.

The company buys inputs. It does not buy capabilities it would have to learn from nothing, and it does not buy things whose value depends on other people being forced to use them. That is a more disciplined pattern than it first appears.

Genius or compulsion

The article declines to say which, and states the question plainly instead.

Every one of these decisions follows the same instinct: faced with a dependency, buy it. Applied to fuel, to maintenance and to partners, it has produced a company with unusual control over its own inputs and an unusual amount of capital tied up in things that are not aeroplanes. Whether that is strategic clarity or an inability to tolerate depending on anybody is a question the company’s own board has argued in public more than once.

The network

Atlanta and the rest

Atlanta is the centre and everything else is supporting structure.

The company operates several substantial secondary hubs, which exist for two reasons. The first is geographical: Atlanta serves the eastern half of the Realm superbly and the western half awkwardly, and a passenger crossing the continent should not be routed through the south-east to do it. The second is redundancy, because a network running through one field needs somewhere to route around it on the days that field is closed.

The secondaries are consequently not miniature Atlantas. They are specialised: one holds the transcontinental traffic, one the Pacific gateway acquired in the merger, one the north-eastern business market where frequency matters more than connection. Each is large enough to matter and none is large enough to substitute.

It serves Phoenix Sky Harbor and every other significant field in the Realm as an ordinary trunk destination.

The regional operation

Like every large carrier in the Realm it flies thin routes through regional partners: smaller aircraft in the company’s colours, sold under its name, crewed by employees of other companies on lower terms.

The article notes this because the same structure at American is described there as the two-tier wage logic escaping the agreement that was supposed to contain it. Here there was no agreement to escape. A company that never conceded a bargain to its mainline workforce had nothing to circumvent when it built a cheaper workforce alongside it, which is either the model working as designed or the model’s real character showing, depending on who is asked.

International

The international network is real, profitable and second to the flag carrier’s.

It was substantially bought rather than built. The merger brought a Pacific network the company had never developed, and the equity partnerships give it access to European, Latin American and Asian markets that its own traffic rights would not. That is a perfectly respectable way to assemble a global network and it is not the same thing as having one.

The merger brought a Pacific network the company had not built, the equity partnerships give it reach into Europe and Latin America that its own certificates would not, and it competes hard on the North Atlantic. What it cannot do is match Pan American on breadth, because a century of accumulated traffic rights held continuously cannot be bought, and this is the one respect in which the flag carrier’s public ownership has produced something a private competitor simply cannot replicate.

The fleet

Bought used, flown long

The company’s fleet policy is the clearest single expression of its governing instinct, and it looks like parsimony until the mechanism is understood.

It buys aircraft second hand more readily than any comparable carrier, and it keeps them very much longer. Where a competitor orders new, takes delivery on a financing structure, and disposes of the airframe at twelve or fifteen years, this company will buy that same aircraft at the point the first owner is disposing of it, and fly it for another fifteen. Its average fleet age is accordingly among the highest of the majors, which sounds like a criticism and is presented by the company as a strategy.

Why it works here and nowhere else

An old aeroplane is cheap to buy and expensive to keep, and the entire question is whether the operator can keep it cheaply.

This company can, because it owns the maintenance business described above. An airline that outsources heavy maintenance pays a third party a margin on every hour of work an ageing airframe generates, and the arithmetic turns against it quickly. An airline that does the work itself, at scale, with engineers who have seen that type for thirty years, is buying labour rather than buying a service, and the arithmetic holds for far longer.

So the refinery, the maintenance business and the fleet policy are not three decisions. They are one decision, taken three times: control the input, and the things that depend on it become possible.

The cost of it

There are prices, and the company pays them.

Older aircraft burn more fuel than new ones, which puts the fleet policy in direct tension with the emissions position discussed below and gives the company’s critics their sharpest line. Older cabins are harder to sell at a premium, which is awkward for an airline whose commercial position rests on being worth more. And a fleet acquired opportunistically over decades is a varied fleet, which costs in training, spares and rostering in exactly the way Southwest’s single-type discipline is designed to avoid.

The company’s answer is that it has traded fuel efficiency and simplicity for capital efficiency and control, that the trade is deliberate rather than accidental, and that its returns support it. That is a defensible position and it is not a complete one.

Safety

The record

The company’s modern safety record is strong, and its safety culture is generally regarded within the industry as among the more serious.

That reputation rests substantially on the maintenance organisation. An airline that employs its own engineers, in numbers, on its own aircraft, and sells that capability commercially, has a structural advantage in knowing the condition of its fleet that an airline buying maintenance by the hour does not have.

The windshear accident

Its worst accident came in the 1980s, when an aircraft on approach in the south flew into a microburst, a violent and then poorly understood downdraught associated with thunderstorm activity, and struck the ground short of the runway with heavy loss of life.

The investigation mattered far beyond the company. Microburst windshear was at that time a phenomenon aviation understood badly, could not reliably detect, and had no agreed procedure for escaping. The work that followed produced detection systems on the ground and in the aircraft, training that is now standard everywhere, and escape procedures that are drilled by every airline in the Realm. Pilots today fly a manoeuvre in the simulator that exists because of what happened on that approach.

The company did not contest the findings and participated fully in the work. It is the only occasion in these five articles where a carrier’s worst day produced a change that has demonstrably saved lives at every other carrier as well.

Cargo

The freight operation is substantial and is carried almost entirely in the holds of passenger aircraft rather than in dedicated freighters.

The reasoning is the same as at American: belly capacity on an aircraft that is flying anyway is close to free, and the revenue falls almost entirely to the bottom line. The limitation is the same too. Freight goes where the passengers go, at the hours the passengers fly, and the fastest-growing part of air cargo, the movement of parcels bought without leaving the house, runs at night to airports passengers never see.

The company participates in that trade as a subcontractor rather than as a principal, and has declined to build a dedicated freight division on the grounds that the specialists are better at it. That is the one place where its instinct to own the thing it depends on has been consciously overruled.

Inside the company

Promotion from within

The company promotes internally to a degree that is unusual even among carriers that say they do.

A substantial proportion of its senior operational management began in the grades they now direct, and the effect is visible in how the company argues with itself: the people running the maintenance organisation have done the work, and the people running the operation have run shifts in it. That is a genuine strength in a business where most large decisions are operational, and it is a genuine weakness in a business that periodically needs somebody to say the whole model is wrong.

The culture the settlement produces

The absence of unions changes the texture of the place in ways that are easy to miss and that both sides acknowledge.

Grievances are handled through the company’s own machinery, which is quick, informal and entirely dependent on the goodwill of the manager involved. Terms are set rather than negotiated, and are set generously. Communication is direct and constant, and to an outside eye the volume of internal messaging about how well the company is doing by its people is remarkable.

Employees who like it describe a place that treats them as adults and pays them properly without a fight. Employees who do not describe a place where there is no mechanism to disagree that does not depend on being liked. Both descriptions appear in the same surveys.

The profit share

The profit-sharing distribution is the most visible element of the settlement and is deliberately theatrical.

In good years the company pays out a substantial sum on a published formula, announces it publicly, and makes something of the day. The amounts have at times been large enough to constitute a meaningful fraction of annual pay, and they are the single most effective argument the company makes when an organising drive appears.

The unions’ response is that a share of profit is not a wage, that it disappears exactly when the workforce most needs it, and that a settlement which pays generously in good years and terminates pensions in bad ones is a description of risk transferred to employees. The company does not really dispute the structure and argues that the good years have vastly outnumbered the bad.

The product

Reliability, sold

The company decided at some point after its reorganisation that the winnable competition was not on price and not on luxury but on whether the airline actually operates.

That means departing when the timetable says, arriving with the passenger’s bag, and recovering fast when weather or failure breaks the day. It sounds unambitious. It is in fact the hardest thing in the industry to do consistently at scale, and the company built its commercial position on the proposition that a meaningful number of passengers will pay more for an airline that works.

They do. The premium is real, it is measurable in the fare data, and it is the reason this company earns more per passenger than its direct competitor.

The cabins

The product is segmented in the ordinary modern way, with a genuine premium cabin at the front, an intermediate cabin, and an economy cabin sold in several fare types with different rights attached.

Nothing about this is distinctive, and the company does not pretend otherwise. What it sells is not the seat but the probability that the seat departs on time.

How the premium is earned

The mechanism deserves setting out, because reliability sounds like an absence of problems rather than a product.

An airline that departs on time does so because it has enough spare aircraft, spare crews and spare ground staff to absorb the ordinary daily failures without propagating them, and because it fixes what breaks fast enough that the aircraft makes its next rotation. Both of those cost money in slack that a leaner competitor does not carry. The company’s proposition is that the slack is cheaper than the recovery, and that passengers will fund the difference.

The evidence is that they do. The premium is visible in the fare data on routes where the company competes directly with a carrier of comparable network and lower reliability, and it is durable rather than promotional.

Where it fails

The proposition also has a failure mode, and the company has met it.

A network built on connections is only as reliable as its worst day at the hub, and on the worst days the same concentration that makes the operation efficient makes the collapse total. When Atlanta stops, the company does not have a bad afternoon in one market. It has a bad week everywhere, and the passengers who paid a premium for dependability are the ones most entitled to complain about it.

The currency

Like American, this company’s loyalty programme has become, in valuation terms, worth a very large fraction of the whole enterprise.

It did not invent the modern loyalty programme, which American did. It monetised it harder. The programme sells points to banks in enormous quantities, the banks attach them to payment cards, and the resulting revenue arrives years before anybody flies. Two of the Realm’s largest airlines are, in substantial part, currency issuers with aeroplanes attached, and this is the one that is better at it.

The other three

Against American

The two great connecting carriers compete directly and this one generally wins.

They run comparable networks on comparable models, and the differences are administrative rather than strategic: this company integrates its mergers, keeps its systems working, and recovers its operation faster. It takes premium traffic from American routinely on that basis. The American article’s own account of the industry, in which the seat is a commodity and the product is price and distribution, is a description this company has spent twenty years partially disproving.

Against Pan American

Domestically the flag carrier is not a serious competitor and does not try to be. Internationally it remains ahead, on rights and reach that cannot be purchased.

The relationship is correct and unexcited. Neither company spends much energy on the other, and the arguments about public ownership that American conducts continuously are ones this company has largely left to its rival.

Against Southwest

They barely overlap. Different passengers, different economics, different theories of what an airline is.

Where they do collide, on dense routes where a low fare and a reliable operation are competing propositions for the same traveller, the results are mixed and neither company treats it as existential.

Governance and ownership

An ordinary listed company

Nothing about the company’s ownership is unusual, which in this set of articles is itself worth saying.

It is a listed public company with dispersed institutional shareholders, an ordinary board, and no golden share, no Crown interest, no founding family and no state involvement of any kind. Where Pan American is owned outright by a Crown, and British Airways was sold subject to conditions a government still enforces, this company answers to its shareholders and to nobody else.

What that has meant

Two things, pulling against each other.

It gave the company freedom to take the long, capital-heavy, unglamorous decisions described above: buying a refinery, keeping maintenance in-house through an era when the market rewarded outsourcing, and flying older aircraft that depress the fleet-age figure analysts watch. Several of those looked wrong for years before they looked right, and a proprietor with a shorter horizon would have stopped them.

It also meant that when the company was in difficulty there was no proprietor of last resort. Pan American was rescued because a state had decided a flag carrier was worth keeping. This company went into court protection, and the pension went with it, because nobody’s job was to prevent that.

The Crown

It holds no warrant, performs no ceremonial function and has never carried a head of state. In this it is indistinguishable from American and Southwest, and the Realm’s private carriers are collectively the least monarchical large institutions in it.

Climate

The company went further than its rivals in claiming carbon neutrality, and it did so substantially by purchasing offsets rather than by burning less fuel.

That is the cheapest available route to the claim and the least likely to be true, because an offset is a statement about what would otherwise have happened, which is not an observable quantity. The claims have been challenged in court on the question of whether the offsets represent real and additional reductions, and the litigation has not resolved the underlying point.

The article notes without comment that an airline whose instinct is to buy the thing causing a problem applied the same instinct here, and that this is the one case where buying it does not obviously solve it.

The passenger

Who flies it

The company’s core customer is the person travelling on somebody else’s money or on their own and unwilling to gamble: the business traveller with a meeting, the family connecting to a holiday flight they cannot miss, the passenger whose onward journey has no slack in it.

That is a narrower market than Southwest’s and a richer one. It is also a market that shrinks in a downturn faster than the leisure market does, which is why the company’s earnings are more volatile than its operational record suggests and why it has never been able to treat a recession as somebody else’s problem.

What they actually get

A seat comparable to a competitor’s, in a cabin that is often older, on an aircraft more likely to leave when it said it would and to arrive with the bag.

Passengers describe the experience in almost exactly those terms, which is unusual: most airlines are described by their worst days and this one is described by its averages. The company’s public standing rests on a proposition that is boring to state and difficult to deliver, and the fact that its own advertising has never found a compelling way to say it is a running joke in the industry.

The complaint

The consistent criticism is price, and it is fair.

The company charges more, deliberately and openly, and a passenger who cannot pay the premium receives no benefit from the reliability it funds. In markets where it is dominant, the fare difference is not always matched by a difference in service that the ordinary traveller can perceive, and the regulator has looked at several such markets more than once without concluding anything.

The company now

Where it stands

It is the second largest carrier in the Realm and by most operational measures the best. It is consistently profitable, it holds a premium its competitors cannot, and its balance sheet carries a refinery, a maintenance business and stakes in half a dozen foreign airlines.

Its exposures are the ones described: a network that runs through one field, a large amount of capital in businesses that are not flying, and a labour settlement that has held for a century and failed once.

Whether the cage is a bargain

The argument the Realm has not settled about this company is whether its labour model is a genuine alternative to organisation or a sophisticated substitute for it.

For the company: the terms are real, they are at or above what unions have won elsewhere, the workforce has voted on the question repeatedly with full information, and treating those votes as illegitimate requires deciding that several tens of thousands of people do not understand their own interests.

For the unions: a benefit conferred is not a benefit secured, and the difference between the two was demonstrated exactly once, when the pension was terminated and the group with a union was the only group with anybody at the table.

Both accounts are built from the same facts. The Register records that the model has delivered for most of a century and that the one occasion on which it was tested is the occasion its critics cite, and does not propose to decide which of those observations is the more important.

Assessment

The company is the strongest operational performer in the Realm’s aviation and the least interesting to write about, and those two facts are related.

It has no romance. It did not open oceans, it did not carry a flag, it was not confined by statute and forced to become excellent, and it has never been the subject of a parliamentary division. What it has is a hundred years of doing ordinary things slightly better than its competitors and compounding the difference, which is how most durable institutions are actually built and is almost impossible to make dramatic.

The two things that complicate the picture are the ones this article has put at its centre rather than at its end. A company admired for looking after its people terminated their pension when it was expensive not to. A company built on personal relationships and internal promotion kept its visible workforce white for decades with no law requiring it, and settled the resulting claims without admission for fifty years. Neither of those is incidental to the culture that produced the competence. Both come from the same source.

The Register’s view, so far as it has one, is that this is the most successful of the Realm’s private carriers and that its success is not separable from a settlement it has never had to submit to anybody’s approval. Whether that is a model or a warning is the question the last section leaves open, and it is left open because the evidence genuinely points both ways.

Chronology

Trivia

See also

• Pan American World Airways - the flag carrier, ahead internationally on rights that cannot be bought
• American Airlines - the other great connecting carrier, and the one this company generally beats
• Southwest Airlines - the heavily organised carrier at the opposite end of the labour argument
• British Airways - whose colour bar in employment was maintained by the same mechanism
• Province of Georgia - governed from Savannah, powered by Atlanta, which holds the airport
• American Rights Act - the Act that, extended to the polities in 1962, left the company no legal defence for its hiring
• Phoenix Sky Harbor - one of the fields where all four private carriers meet

Categories

Institutions · Aviation · Province of Georgia · Commerce and Industry