Card for episode 4 of ten of the series « Politics makes the product », carrying the title « Leading products write their own laws » over a dark photograph
The Bonn political declaration, signed on 19 May 1987 on behalf of the French, German, Italian and British telecommunications ministers. Official UK copy, inventory 2013-149. Science Museum Group, CC BY-NC-SA 4.0.

Fourth episode of a ten-episode series on the weight of politics in the success of a product. The first three showed that no market-leading product escapes a public envelope.

A line drawing. A large circle, the government, contains the circle of the value network, which spills over into another government drawn in dotted lines. Inside the value network, the circle of the company, which also spills over into another value network. At the center of the company, the product team. Five numbered points of view: 1, the government seen from outside; 2, the government seen from inside; 3, the value network seen from the company; 4, the company seen from its own corridor, linked by a dotted red arrow coming down towards the company; 5, the company seen from its capacity plan, at the center of the product team, from which a green arrow rises beyond the government, the only one looking upwards. Point 2 is highlighted in yellow with the words you are here.
The plan of the series. This fourth episode stands at point 2, the government seen from inside.

The back room

Monday 16 February 1987, Funchal, island of Madeira.

The port of Funchal in Madeira, with a white ferry at the quay.
The port of Funchal, in Madeira, where the February 1987 meeting took place. Photograph Dietmar Rabich, license CC BY-SA 4.0, via Wikimedia Commons.

Fifteen European telecommunications administrations open a week of meetings to choose the radio technique for the future common mobile telephone. Two solutions are in contention: the first, wideband, is Alcatel’s, and France and Germany both back it. The second, narrowband, is carried by the Nordic countries and by the British.

The same slice of radio spectrum, cut two ways. On top, wideband: one very wide carrier filling the whole slice, Alcatel's solution backed by France and Germany. Below, narrowband: nine narrow carriers separated by gaps, carried by the Nordic countries and the United Kingdom. A frequency axis runs under both rows.
The choice is about the width of the radio channel: the same slice of spectrum, filled by a single carrier or cut into several. Under each solution, those who defend it on Monday morning.

Each country has sent a delegation, and each delegation holds two kinds of people. Engineers, split into small groups, who discuss the parameters one at a time; and a head of delegation, who sits in the plenary, the only meeting where every country sits together, where the specification is adopted and where each country has one vote. The ministers have not come, although they are the ones who decide their country’s solution. The group meeting here is called Groupe Spécial Mobile, and it will give its initials to the standard under discussion: GSM.

A three-column diagram. On the left, a national delegation: the administration's engineers, who take the parameters one at a time in the subgroups; the head of delegation, who carries his country's single voice and votes in the plenary; and two industrial experts, industry not being a member of CEPT but each delegation being allowed to bring two of them to every meeting. In the middle, the bodies: the services, radio and network subgroups, with no vote; the plenary, one voice per country, fifteen in Madeira, adoption by unanimity, chaired since 1982 by Thomas Haug; the Permanent Nucleus, CEPT's technical secretariat in Paris, which drafts the text and keeps it up to date. What leaves the room is a CEPT recommendation nobody is obliged to apply. On the right, outside the room: the ministers, not one of whom came and whose instruction comes down on the head of delegation, and the operators, who sign the Copenhagen memorandum on 7 September 1987 and commit to opening the service in 1991.
Three profiles in a delegation, three bodies where they sit, and two powers that decide outside the room.

On the Monday, Philippe Dupuis, the head of the French delegation, takes his British, German and Italian partners aside. Late on the previous Friday, Alcatel had got wind of what was being prepared in Madeira and had gone to the French government at a very high level; Temple later learned it was the prime minister’s office. Dupuis has lost his freedom to negotiate, and France will defend Alcatel’s solution, whatever its own experts think. Since the plenary decides only by unanimity, his single vote is now enough to prevent any decision.

A map of the north-east Atlantic and western Europe, with no borders at all, only coastlines. Funchal is marked in red off the coast of Morocco, and four lines link it to London at 2500 kilometres, Paris at 2410, Bonn at 2810 and Rome at 2780. A 500 kilometre scale bar sits at the bottom right.
Between two thousand four hundred and two thousand eight hundred kilometres of Atlantic separate the meeting room from the capitals issuing the instructions. Coastlines from Natural Earth, public domain.

The plenary can therefore no longer settle the choice between the two solutions. Its chairman, the Swede Thomas Haug, runs through the minor items to fill the time. On Wednesday, a tour de table gives thirteen delegations for narrowband and two against. Politics is blocking.

On Friday, the last day, Stephen Temple, the British delegate who tells the story of this week in his book on the history of GSM, settles into a back room Haug has given him. What happens in that room has no agenda, no minutes and no witness. Since one vote is enough to block everything in the plenary, the delegations come one at a time to negotiate here what they will not get in there, and what is settled here goes back to the plenary to be adopted. Temple compares the place to a doctor’s surgery.

A plan of the room. A long table surrounded by fifteen seats: thirteen green for narrowband, two red for wideband. Five seats are named: Sweden's Thomas Haug in the chair, the United Kingdom's Stephen Temple, Italy's Renzo Failli, and in red France's Philippe Dupuis and Germany's Armin Silberhorn. A corridor leads to a separate room, the Friday back room, with no agenda, no minutes and no witness, entered one delegation at a time.
The five delegations the story follows, and the split of the Wednesday. France and Germany stand on the wideband side because their instructions hold them there, not because their experts prefer it.

The Madeira package, or how a standard is negotiated like a treaty

Back to Wednesday, the evening of the tour de table. The partners meet in the hotel lobby at eight. Armin Silberhorn, the German delegate, and Philippe Dupuis arrive last, grim-faced: their governments have just hardened the instructions, wideband or nothing. Haug throws up his hands. There will be no agreement, then, he says.

Temple pushes the narrowband solution, which his experts judge better for handheld phones, one of the evaluation criteria the United Kingdom is defending. The German expert wants fourteen telephone channels per radio frequency, the British eight, because a smaller capacitor is enough to feed the burst of transmission once the channels thin out on a single frequency.

The Swiss want a system that works in valleys where the radio echo comes back late, and the Swedes defend a variant they alone support, hoping to trade it against the one their Norwegian ally is proposing. Everyone reasons correctly, defends a national industry, protects the investment their laboratories have already made, and behaves exactly as theory predicts. The whole thing seizes up.

The choice between the two solutions is out of the hands of the fifteen delegations, because it rests on a unanimity rule that can only lead to deadlock with so many players. Temple finds a lever: the specification of the standard, that is, the document listing the requirements the solution has to meet.

He brings in two things that week that will weigh more than the vote: two rules so that everyone leaves with something rather than with a complete disagreement. What is being decided that week is the system on which European mobile telephony, and then the devices that came out of it, would rest. Nobody in the room knows it yet.

The working assumption

The first rule is the working assumption. It governs the passage from the small groups to the plenary. Until then, each value was voted on as a final choice that steered the solution directly, and a delegate whose value lost had every reason to block, since he lost for good. The working assumption takes that reason away. The group stops asking which solution to pick and works parameter by parameter: for each one it adopts the best value known that day and records it as provisional. Anyone who wants to replace it later may do so, on three conditions: bring proof, have established it outside the GSM group, and show that their value does better, not merely as well. Temple compares the result to slow-setting cement: the shape has taken, it can still be corrected, but the effort needed to do so grows by the day.

A black Nokia 1011 mobile phone with its antenna.
The Nokia 1011, one of the first GSM mobile phones, in the early 1990s. Photograph Jkbw, license CC BY 3.0, via Wikimedia Commons.
Two frames compared. On the left, voting for a solution: two whole blocks, wideband and narrowband, and a red box saying that the loser loses for good and therefore blocks. On the right, agreeing parameter by parameter: a grid of eight green squares, each a provisional value, and a box saying that every value stays replaceable on proof.
The vote changes object: you no longer choose between two systems, you fix values that anyone can still beat on proof.

Take an example. Each radio frequency has to be shared between several conversations, and the number has to be fixed. The German expert wants fourteen, the British eight, because a pocket telephone gains battery life that way. Neither of them loses for good: the value recorded that day stays provisional, and whoever comes back with measurements made in his laboratory showing that another number does better replaces it. Every parameter of the standard is handled this way, in parallel, and no delegation has the best answer on all of them: each loses here and keeps a chance there, which is what makes the rule acceptable.

The loser of the day accepts the decision, since he keeps a chance of coming back with his trials, and the design moves on without waiting for him to be convinced. A mistake can be corrected later without reopening the whole agreement. Alain Maloberti, who has chaired the group in charge of defining the radio system since 1985, would say years later that without this rule the GSM specification could not have been written.

Two frames. Under the usual rule, a contested parameter blocks everything. Under the working assumption, every value is adopted provisionally and stays replaceable on proof established outside the GSM group.
Under the usual rule, a contested parameter blocks everything. Under the working assumption, everything is adopted provisionally and stays replaceable on proof brought from outside.

Package bargaining

The second rule is package bargaining. Temple proposes it just before the morning break and calls it the package, explaining to the room that everyone must leave with something. The Germans take the speech coder, the software that compresses the voice before sending it, the one from the German equipment maker PKI; the coder finally standardized would also take features from IBM France’s. The French take frequency hopping, which they attach importance to, the Nordics take the modulation method that had come out ahead in the Paris trials, and the Swiss get the parameter that handles the echo in their valleys.

A Motorola International 3200 phone with its manual, battery and charger.
A Motorola International 3200, a GSM phone from the early 1990s, with its manual and charger. Photograph Volker Artmann, license CC BY-SA 2.5, via Wikimedia Commons.

The United Kingdom, for its part, takes and gives up. Temple, with Ted Beddoes of Vodafone and Robin Potter of British Telecom’s research laboratories, cares about two things and they decide to keep only one. They keep the possibility of doubling the capacity of the system later on, by reserving room for a coder that compresses the voice twice as hard and fits two conversations where there is only one, which ticks one of the British evaluation criteria. They give up the speech coder from British Telecom’s laboratories, a credible candidate, but one whose German rival from PKI had wider support among the experts. Temple recounts that Potter withdraws his own proposal so that the package holds. Back in London, British Telecom reproaches him for having abandoned an in-house technique in favor of the Germans’. But he saw what Temple saw, a fleeting chance of an agreement on narrowband. The compromise is the price of the package. Whoever proposes it has to lose something visible, otherwise nobody else agrees to lose anything.

Eight channels, at two in the morning

That left the number of channels.

Two mobile phone antennas disguised as trees in Dublin.
Two GSM base stations disguised as trees, in Dublin. Photograph No machine-readable author provided. Blorg assumed (based on copyright claims)., license CC BY-SA 3.0, via Wikimedia Commons.

The experts shut themselves in that evening and agree on nothing. Temple goes down to see them around eleven at night. Three hours of argument, nothing gained. He tries a negotiator’s move: on a parameter with three options, the Swedes, alone on the third, have hinted they would drop it if their Norwegian allies’ option went through; so why not narrow the choice to two. The experts all turn on him and accuse him of bringing politics into a technical discussion. He would recount that in half an hour he had achieved unanimity on exactly one point: the problem was him. He goes back at it anyway, question by question, until he isolates the German expert on the only point that still interests him at that hour, the number of channels per radio frequency. He gets eight.

The experts’ meeting ends at half past two in the morning, and the famished delegates go down to the kitchens to find nothing but tins of sardines.

The whole world got eight channels per carrier frequency, and every GSM telephone ever built carries that number, settled in a discussion between two men, one of whom had just cornered the other in a back room of the hotel in the middle of the night.

Another witness tells it differently. Mats Nilsson, of the Swedish public operator Televerket, who ran the comparative trials in Paris, places the move from ten or twelve channels down to eight at an experts’ meeting held in The Hague in February 1987, shortly before Madeira, and describes it as a farmyard auction: any support for five channels, none; six perhaps, not really; eight then. Temple too writes that the experts were due to meet in The Hague before Funchal. Neither points to a document. The number is settled; the scene that produced it is not.

A band representing the GSM specification, with five blocks attached to it. Germany takes the speech coder from PKI and IBM France, France frequency hopping, the Nordic countries the modulation, Switzerland the echo setting for their valleys. The British block is double: it keeps room for a coder that compresses the voice twice as hard, and it gives up the speech coder from British Telecom.
The Madeira package: each delegation leaves with one parameter, and the United Kingdom is the only one to leave one on the table.

What could no entrepreneur have done alone? He does not get into the room: the seats belong to administrations. Nor does he invent the rule that organizes the room. Alcatel got that week the best thing an industrialist can get, namely the instruction of a head of government, carried into the room by two national delegations and defended to the end. Temple got a method of arbitration. The method won.

The club

Back up twenty months. An agreement existed before Funchal that no European text records: France, Germany and Italy had signed in Nice, in 1985, a cooperation on digital cellular, outside the European Community and outside CEPT, the conference that brings together Europe’s telecommunications administrations. Getting in was not free: Temple writes that he learned later that STET, the Italian public telecommunications holding company, had had to pay something to the French administration to be admitted. This agreement of the three let them arrive at meetings with positions already settled. Temple writes that it was becoming harder and harder to fight, meeting after meeting, against views pre-coordinated by three administrations that talked to each other before talking to him.

The pebble beach along the Promenade des Anglais in Nice, under a blue sky.
Nice, where France, Germany and Italy signed their cooperation on digital cellular radio in 1985. Photograph Uhooep, license CC BY-SA 4.0, via Wikimedia Commons.

Mauro Sentinelli, the Italian delegate, wants the British brought in and drops the idea to Temple one evening, taking him aside on the way into the dinner the United Kingdom is hosting for a GSM working party. Temple is willing to join, but only once the group has adopted the five evaluation criteria he has been defending for months, the grid on which the competing solutions will have to be compared. Sentinelli smiles, and the two men agree on how to proceed: the Italian minister will make the suggestion on his visit to London, and Temple will write his own minister’s brief so that the answer is that this becomes possible once the criteria are adopted. The two ministers meet, say what was written for them, and there the matter rests.

The United Kingdom will end up joining this club, and it is not Temple who opens the door. It gets in over three moments, between the autumn of 1985 and the summer of 1986, and none of the three looks like what one imagines of an international negotiation: an office accident, a legal opinion set aside, and a corridor bargain that will not be honored. It will take a fourth moment, and six months more, before the last signatory obliges.

An office accident in Bonn

The first moment is the office accident. A minister on a visit speaks from a brief his officials prepare, which sets what he may say. In the autumn of 1985, a reshuffle puts Leon Brittan at the head of the Department of Trade and Industry, and the new minister leaves at once on a courtesy visit to Bonn. Some months earlier an Italian minister had come to London, and it was Temple who had drafted the brief for his British host. He had written in it that the United Kingdom would not join the agreement of the three until the GSM group had adopted his five criteria. For Bonn, Brittan’s office asks for a brief by the next day, marked urgent. Temple and his colleague Greg Faulkner, the diplomat who follows the file with him, are both abroad and unreachable, in an age when mobile phones stopped working at the border. Two junior officials who do not know the file dig out the Italian brief, replace Italy with Germany, and send it to the embassy.

The entrance of the former Bundesrat in Bonn, with glass doors and gold lettering.
The former Bundesrat in Bonn, then the capital of West Germany. Photograph Dietmar Rabich, license CC BY-SA 4.0, via Wikimedia Commons.

In Bonn, the diplomats find an agenda with nothing positive in it: on market opening, London and Bonn are poles apart. The only subject that might please the German minister is digital cellular, and even there the brief bristles with reservations. Nobody answers in London. In the well-understood interest of Anglo-German relations, the embassy removes the reservations and adds a little enthusiasm. Brittan reads his brief, walks into the office of Christian Schwarz-Schilling, the German posts minister, and declares that the United Kingdom would be delighted to join the agreement, the very one Temple had been refusing to join for months. The Germans, whose delegate had been describing a reluctant United Kingdom for months, are caught off guard. The minister says out loud that a system running from Milan to London would be a fine thing for Europe, and agrees on the spot, in front of his own officials, that the British should come in. Back at his desk, Temple reads the report telegram with disbelief.

The legal opinion set aside

The second moment is a legal opinion set aside. In the months after Brittan’s visit, the United Kingdom joins the agreement of the three not through a new text but through an annexe which the four ministers have to sign. As a matter of routine, the annexe goes to the department’s lawyers. The answer comes a few days later and does not target the annexe: it targets the original agreement. Three administrations in a monopoly position coordinating on a technique is the kind of arrangement Europe’s competition authorities watch for, and signing the annexe would amount to subscribing to it. The lawyer who handled the file recommends not signing. Temple calls her. His argument: a newcomer cannot demand that the club change its rules before admitting him. Can she at least put a figure on the risk of legal action? She declines to give a figure and declines to cover any risk-taking at all. Following her advice would mean staying outside. A note goes up to the minister responsible, sets out the case for and against, and recommends overriding it and signing. The annexe comes back with the British signature.

The Berlaymont building, headquarters of the European Commission in Brussels.
The Berlaymont, headquarters of the European Commission in Brussels. Photograph Andersen Pecorone, license CC BY 2.0, via Wikimedia Commons.

Odense, a chairmanship against a signature

The third moment is a coffee break. In July 1986, Temple and Faulkner sit at Odense, in Denmark, at the meeting that oversees CEPT’s telecommunications work. Two matters bring them there. The first is a new agreement on terminal approval, the procedure by which one specimen of a device is tested so that the whole series is authorized across Europe, instead of one test per country. Temple spent a year writing that text and stands for the chairmanship of the committee that will apply it; France puts up a rival candidate, but the room leans towards him. The second is the French signature on the cellular annexe, which still has not arrived. During a break, Faulkner offers him a trade: withdraw your candidacy in favor of the Frenchman, in exchange for France’s signature.

Odense city hall, in red brick, with Danish flags.
Odense city hall, in Denmark. Photograph Kåre Thor Olsen, license CC BY-SA 3.0, via Wikimedia Commons.

Temple lets his coffee go cold. Then he withdraws his candidacy in favor of the French candidate, under the disbelieving eyes of the other delegations, and settles for the vice-chairmanship, which is to say, he writes, all the work and none of the power. Before leaving Odense, the head of the French delegation assures the British that the agreement has gone up to the director general and has been signed.

A month later, no signed copy has arrived in London. Temple calls Dupuis, who admits, embarrassed: the director general of French telecommunications, on whom his delegation depends, has refused to sign, because he wants to extract one more concession from British Telecom. The word given at Odense bound only the man who gave it. The chairmanship is surrendered, the card is played, and only the bluff is left. The head of the French delegation had said the agreement was signed; Temple and Faulkner take him at his word. Faulkner sends the three capitals a warm telegram rejoicing that everyone has signed, and then Temple turns up at the next meeting, in Rome, as a full member.

January 1987, the sentence written in advance

The fourth moment is a sentence written in advance. In January 1987, Geoffrey Pattie, the British minister for information technology, pays a courtesy call in Paris on his French opposite number at the posts ministry, Gérard Longuet. Temple has drafted the brief for the meeting and has placed one line in it: tell the French minister how pleased the United Kingdom is to be party to the cooperation agreement on digital cellular. Pattie says it. The conversation stops. Longuet turns to one of his most senior officials, M. Grenier, and asks why he has not been told. The next day Grenier walks into the director general’s office with the copies of the agreement and tells him the game is up. The documents are signed. A week later, in Paris, Temple’s new chief is handed them by Grenier, who does not blush, and passes them on to London. One month was left before Funchal.

Gérard Longuet in profile, smiling.
Gérard Longuet, French minister of posts and telecommunications in 1987, here in 2012. Photograph Jebulon, public domain (CC0), via Wikimedia Commons.
A line drawing in three numbered columns. 1, an office accident, autumn 1985: junior officials reuse a brief written for an Italian minister and swap Italy for Germany; the embassy removes the reservations and adds a little enthusiasm; the minister declares that London would be delighted to join. 2, a legal opinion set aside: the lawyers recommend not signing, are asked to put a figure on the risk and refuse, a note goes up to the minister and recommends overriding them, and the annexe comes back signed. 3, a corridor bargain, July 1986 at Odense: a trade over a coffee break, the chairmanship candidacy against the French signature; Temple withdraws, a month later Paris has not signed, and they send a warm telegram and turn up in Rome anyway. At the bottom: the United Kingdom joins that club without anyone having decided it, and none of the three cost a penny.
Three doors, and none of them looks like what one imagines of an international negotiation. That is what access to the room where the rule is written runs through.

Five weeks after Madeira

Three frames show the same agreement at three moments. 1985: France, Germany and Italy, the agreement of the three which no European text records. January 1987: France finally signs the annexe, the United Kingdom joins, they are four. March 1987: the United Kingdom uses their own agreement against France and Germany.
The same text at three moments. In 1985 it exists only between three administrations, in January 1987 it binds four, and in March 1987 the last to join uses it against the first two.

Since Funchal, Paris and Bonn have been negotiating as a pair, and Dupuis is soon pushed out: on 18 March he drafts with Silberhorn a joint technical report that favors narrowband, his chief rewrites it before sending it to the minister, and the political level then keeps him out of the file. On 25 March 1987, the German minister sends his senior officials, led by Waldemar Haist, to wring a common line out of the French, led by Marcel Roulet. The Germans arrive determined: they are about to write to the other CEPT administrations that Germany will adopt narrowband, leaving France to associate itself with that or go its own way. It all holds until lunch. There, the French propose new comparative trials between the two solutions, and a German procurement official replies at once that he has budget left. The head of the German team wavers, the hard line falls apart, and the two delegations leave with trials to pay for. The United Kingdom, Temple writes, is reduced to the role of spectator. Shut out of the tête-à-tête, Temple looks for something that could bring the other two back to the table, and he goes to reread the agreement of the three, the one his country joined through the annexe. The obligation is written there in black and white: each administration undertook to work with the other three. Not with one of them.

On Monday 30 March, Temple (United Kingdom) writes it to Silberhorn (Germany) in a note in diplomatic style whose substance fits in one sentence: a Franco-German arrangement breaks the undertaking given between four. Silberhorn receives it without pleasure, concedes that Temple is right, and asks which of the two agreements prevails, the one between four or the one he has with the French. The next day he calls Temple back and arranges the order of the calls so that his own minister hears the British argument before the French one: the French minister is due to call him on Friday afternoon. Geoffrey Pattie, the British minister, obliges on Friday 3 April, shortly after eight in the morning. On the other end of the line, Christian Schwarz-Schilling leaves the wideband camp and states his preference for narrowband, while refusing to make it public, out of regard for his French colleague. Then he makes two commitments: France and Germany stop negotiating as a pair and everything will from now on be handled between the four countries; and if the senior officials of the four cannot find a solution, the four ministers will meet in Bonn before the end of May. It is out of that Friday morning call that the meeting of 19 May comes.

What is left is the instruction that came from the prime minister’s office, and it is industry that lifts it. Léonard Laborie establishes that the reversal came from industry itself: the chairman of Alcatel-Thomson-Radiotéléphone realizes, in circumstances the historian says remain to be established, that the process set in motion by the majority of the GSM group has become inescapable, reports this to the chairmanship of the Alcatel group, meeting in Brussels, which agrees to drop its demands. The plenary had no power to settle anything. It made the majority irreversible, and it was that fact which made the instruction give way.

Temple, for his part, puts a name to those circumstances. He recounts relaying the two ministers’ telephone call in detail, on the margins of a conference in Brussels, to Philippe Glotin, the Alcatel executive in charge of mobile communications, who replies that he had guessed as much, but that the group’s German subsidiary was still promising to turn Bonn around. The next day, Temple writes, Glotin reported back to his own company that support for narrowband was overwhelming, and Alcatel stopped defending wideband. This is a negotiator’s account of his own maneuver, and it does not replace the archive the historian is missing.

Bonn, 19 May 1987

The text the four sign on 19 May is written on paper twenty-one centimetres by thirty, and their four countries undertake in it to support the GSM specification at home. It is addressed to the ministers responsible for telecommunications or their representatives: Geoffrey Pattie signs for the United Kingdom, Christian Schwarz-Schilling for Germany, R. Panella for Italy, and for France Marcel Roulet, the director general of French telecommunications. Temple, who is there, writes that 19 May was more of a ceremony than a meeting: each party makes a short speech, and then they sign. Two further signatures sit at the foot of the text on the British side, those of John Carrington for British Telecom and Gerry Whent for Vodafone: the United Kingdom’s seat in that room is shared with its two operators. Pattie uses his turn to speak to have the ministers instruct their officials to draw up an agreement between operators, ready for signature in September. Look for the law in that object and you will not find it.

That is how Alcatel lost the standard. The network that prevailed did not belong to the industrialist, richer though he was, better connected, able to reach a head of government on a Friday evening. It belonged to four civil servants who had tied their hands to one another, three of them in Nice in 1985, the fourth through an annexe that France signed only a month before Funchal.

The man who wrote a decision the state signed

Back up to the 1960s, change country, keep the same house.

France is funding a national computing industry in the early 1970s. The Compagnie internationale pour l’informatique, the CII, which the state brings into being in 1966 by forcing the merger of the joint computing subsidiary of the Compagnie générale d’électricité and the CSF with the one belonging to the Schneider group, serves as the instrument of the Plan Calcul, and Georges Pompidou and his government push a European alliance, Unidata, which joins it to Siemens in January 1972, then to Philips in July 1973. On paper, France has an Airbus of the computer.

The front panel of the CII IRIS 80 computer.
The front panel of the IRIS 80, a mainframe computer from the Compagnie internationale pour l’informatique, first delivered to customers at the end of 1971. Photograph Damien.b, license CC BY 4.0, via Wikimedia Commons.

Ambroise Roux runs the Compagnie générale d’électricité, and he does not want this alliance. His house is a shareholder in the CII without commanding it. Thomson was not around the table in 1966; it comes in in September 1967 by absorbing the CSF, and since June 1969 the company has been held seventy per cent by Thomson-CSF and the CGE together, through a joint holding called Fininfor, of which Thomson-CSF controls fifty-two per cent and through which it has installed its own man at the head of the CII. The business historian Tristan Gaston-Breton reports Roux’s reasoning in one sentence: the alliance starts in computing, it will end up reaching components, and even the telephone. Roux is fighting a rapprochement between his competitor Thomson, Philips and Siemens, in sectors where he operates himself and from which he expects most of his growth.

Jean-Pierre Brulé, who runs Honeywell-Bull, prefers a merger with the CII to the arrival of a European competitor, and Michel d’Ornano, the industry minister, leans towards the same merger, which promises the CII access to the American market.

Roux opens no public campaign. He exercises his house’s shareholder position from the inside and without a press release, and he also has direct access to those who decide. He had advised Pompidou on economic questions. He chaired the economic committee of the main employers’ federation.

1974, the American camp wins

Valéry Giscard d’Estaing enters the Élysée in May 1974. He had already come down in favor of Honeywell in July 1970, as finance minister, when General Electric pulled out of computing and Bull needed a buyer. That same year, 1974, Maurice Allègre, who had led the computing delegation since 1968, resigns in discouragement, and the delegation disappears with him. The American camp then has its arguments: Jean-Pierre Brulé lets it be known that he will ask for no subsidy to take over the CII, and Siemens buying up the computing arm of a struggling Telefunken supplies the pretext of an imbalance inside Unidata. In May 1975 the government agrees to the absorption of the CII by Honeywell-Bull and announces the French withdrawal from Unidata. The consortium disappears in December. The German partners speak of betrayal.

Black and white portrait of Valéry Giscard d'Estaing.
Valéry Giscard d’Estaing, President of the French Republic, in 1975. Photograph Christian Lambiotte, license CC BY 4.0, via Wikimedia Commons.

Two things.

The French computing industry passes under an American umbrella, and the public bill does not get any lighter. The French Senate went back over these accounts twenty years later. The first convention between the state and the CII, from 1967 to 1971, cost it around one billion francs. The second, from 1972 to 1975, is endowed with more than twice that and in fact comes to three times the amount, while the contribution of the private shareholders falls to two and a half times less than in 1967. The report puts the bill for the two conventions and for the exit from them, for the taxpayer, at around thirteen billion 1992 francs. The operation brings Bull back under French majority ownership in 1976, and Jean-Pierre Brulé, who had wanted it and whom the report quotes, judges that the same result was probably within reach as early as 1966, for ten to a hundred times less money.

Ambroise Roux had said publicly that he would leave the chairmanship of the CGE if it were nationalized. The CGE passes into state hands in 1982, he obliges, and he founds the Association française des entreprises privées, whose founding members hold their first council on 12 October 1982. It refounds the Association des grandes entreprises françaises faisant appel à l’épargne, the Agref, created in 1969 and thinned out by the nationalizations, and it counts thirty-six affiliates at the time. Jacques Chirac, François Pinault, Michel Pébereau and Jean-Louis Beffa are all at his funeral, in April 1999. The press called him the godfather of French capitalism. A man had just turned his personal influence into a permanent institution, and the institution outlived him.

Alcatel comes from that house. In 1968, the CGE takes control of a small Alsatian telecommunications equipment company called Alcatel. Two years later Roux, by then chairman of the CGE, merges it with the CIT, the group’s telephone arm, to refocus it on the telephone. The whole becomes CIT-Alcatel. In 1985 it absorbs Thomson’s telecommunications and takes the name Alcatel. It is this house that walks into the prime minister’s office in February 1987.

A vertical timeline with two tracks. On the left the French house: 1966 the CII is born as the arm of the Plan Calcul, 1968-1970 the CGE buys Alcatel and folds it into the CIT, 1972-1973 Unidata with Siemens then Philips, 1975 the CII passes to Honeywell-Bull and Unidata ends, 1982 the CGE is nationalized and Roux leaves to found the AFEP, 1985 CIT-Alcatel takes the name Alcatel, 1999 his funeral. On the right the European negotiation: 1985 Nice and the agreement of the three outside the EEC and CEPT, autumn 1985 a recycled brief brings London in, July 1986 Odense where a chairmanship is traded for a signature, January 1987 Paris finally signs the British annexe, May 1987 Bonn where four ministers sign. In the middle, in red, February 1987: Alcatel walks into the prime minister's office and the room in Funchal decides nothing.
Two stories run separately for twenty years, one inside a French company, the other in European meeting rooms. They touch on a single day, in February 1987.

Three questions for your next product review

The previous episode proposed one question to add to the agenda, right after the customer question: which public decision opens or closes my market, on what horizon, and who holds the pen? Here are three more, which come before it, and which fit on half a page of a notebook.

Where is the rule of my market written? A technical committee, a trade association, an administrative working group, a qualification scheme. Name the room. Write its name down. If it does not exist yet, your market is young, and you have just learned something useful.

Who sits there for my industry, and since when? Duration counts for more than size. Three administrations bound to each other for twenty months, and a fourth that spent fifteen months forcing their door, beat an industrialist who knew how to walk into the prime minister’s office on a Friday evening.

By what date will it be too late? A specification that will appear in the Official Journal in four years is closing today, in a meeting room where the coffee is lukewarm and where somebody is settling a parameter you will discover too late. The day you read it, all that will be left is to comply with it.

Gather the answers in a four-column grid: the forum, who sits there for your industry, the closing date, and the name of the person who follows the file at your company. One line per rule, and one read-through at the product review twice a year. If the fourth column stays empty, you already know the result.

This grid does not call for a lobbyist. It calls for someone whose job is to reread it twice a year, in front of the product review.

One thing those three questions do not say: having a seat is not binary. This episode’s narrative gives five of them, and one man climbed the first four before selling the fifth. You are outside, and the others coordinate before they talk to you. You are inside without being a full member, and it sometimes takes a warm telegram to settle in. You are a full member. You are vice-chairman, which is to say, as Temple writes himself, all the work and none of the power. You chair the committee that applies the text you wrote. The first two degrees cost nothing beyond knowing where the room is. The last one is negotiated, and Temple traded his away for a signature.

A line drawing. A staircase of five rising steps, from red to green, each numbered. Step 1, outside: the others pre-coordinate before talking to you. Step 2, inside but not quite: the annexe is unsigned, you settle in on a telegram. Step 3, full member: you sit in Rome and the room accepts it. Step 4, vice-chairmanship: all the work and none of the power. Step 5, chairmanship: you apply the text you wrote yourself. A dashed red arrow points at the fifth step: Temple gave this one up for a signature. At the bottom: the first two rungs cost no budget at all, the last three are negotiated.
Having a seat is not binary. The five rungs all come from this episode’s narrative, and one man occupied the first four.

The forums and rungs in this article occupy the second line of a sheet that gathers the series. It fits on one page.

What remains is how one gets into that room at all. Nobody goes alone: it takes suppliers who attest to a capability, customers who attest to a need, institutions that lend their letterhead. That is the subject of the next episode.

What this article rests on

The back room and the Madeira package

The club

Five weeks after Madeira

The man who wrote a decision the state signed

Partager

ARTICLES CONNEXESPLUS DE L'AUTEUR