How close to nuclear war did we get over Cuba?

Cross posted to LessWrong and The Good blog

The Cuban missile crisis was probably the closest we came to nuclear war. There were two types of close call in Cuba. There were moments of tension between the US and Soviet militaries, all of which tell us something interesting about how different patterns of civil-military relations influence the outcomes of moments of crisis. But, there was also the risk of nuclear not by accident but by intention. Throughout the crisis, the US government kept an invasion of Cuba or airstrikes on the missile sites as options on the table, and the former was pushed very aggressively by the joint chiefs of staff. The joint chiefs of staff were the most senior officers in the US army and provided its strategic direction. The influence of these most senior members of the military bureaucracy has been pervasive throughout the history of nuclear weapons, and their role in nuclear risk will be something I’ll return to in a later post.

Cuba: some background

I gave some background on the strategic situation in Cuba in this post, but for this story it’s laying out the timeline of events in a little more detail. From the end of the Spanish-American war in 1898 until Castro’s revolution toppled the dictator Hugo Baptista, Cuba was the playground of the American elite and a quasi-colonial possession. American companies owned the sugar plantations, and the mines and the casinos and Baptista’s thugs beat up anyone who tried to change things. You can imagine then that it came as quite a shock when the bearded, fiery, Jesuit educated Castro succeeded in launching a nationalist revolution and overthrowing the Baptista government. Initially Castro made overtures to the Americans but these were rebuffed as Ho Chi Minh’s attempts at alliance—another nationalist leader who turned to communism—had been 4 years earlier in 1956. So, Castro turned to the Soviets and so, with Castro now firmly in the Soviet camp and expropriating all of the American owned businesses, it was time for Castro to go.

Another theme in the history of nuclear weapons has been the importance of intelligence and intelligence failures and they will get their own post later. By 1961 the CIA had successfully orchestrated changes of government in Iran and Guatemala and expected to be able to do the same with Cuba and that the Cuban refugees they’d armed would be welcomed with open arms. This did not happen. The force that landed in the Bay of Pigs was roundly defeated by the Cuban army and Cubans celebrated their victory over the Yanqui imperialists. Two things came about as a result of this. Firstly, President Kennedy became all the more desperate to depose Castro and avenge his humiliation. Secondly, Castro was able to convince his Soviet backers to place nuclear missiles (although, there were other reasons Krushchev wanted missiles in Cuba.)

This brings us to the start of the crisis. U2 planes, American spy planes that flew much higher than conventional aircraft, photographed missile sites on the 14th of October 1962. By 16th CIA analysts realised what the U2 had photographed and estimated that missiles could be ready to hit the US within 18 hours. Over the whole course of the crisis the Americans acted on the assumption that there were nuclear warheads on Cuba that could be mated with the missiles, while never actually finding the sites where the warheads were kept. They were correct as it turned out, with the first missiles being fully assembled and able to launch within 8 hours by the 25th of October.

Now Kennedy had a choice to make. There were essentially 4 choices on the table: invade Cuba, launch an airstrike against the missile sites, blockade Cuba, or do nothing. He chose to blockade, leading to the first of 4 brushes with nuclear war.

The Invasion

The Joint chiefs of staff were pushing aggressively for an invasion, led by Air force chief of staff Curtis LeMay. LeMay in particular believed that the US vast advantage in nuclear capabilities meant that they could invade and the Soviets would be forced to back down. The Joint chiefs believed that the missiles in Cuba represented an important strategic threat because they the missiles would have been able to hit airforce bases where the planes that carried the majority of US nuclear arsenal were housed. I talk about the credibility of that belief here, but for the purposes of this story, all you need to know is that Kennedy was convinced by Robert McNamera, the secretary of defence, to hold off on the invasion. This was supported by internal defence department documents showing that simulations run on their state of the art computers showed that the missiles would have a very small effect in a nuclear war. Despite this, Kennedy accepted his General’s arguments that if the missiles weren’t gone by the end of the Monday the 29th of October, they’d invade Cuba on the 30th. This decision in my mind took us closer to nuclear war than any other. U2s had by the 27th found Luna anti-tank missiles on Cuba, which were nuclear tipped. It wasn’t known to the Americans that the missiles were nuclear tipped, but it was known that they had the capacity to be. However, what the Americans didn’t know was that there were also nuclear tipped cruise missiles on Cuba aimed at the Guantanamo bay naval base with orders to fire in response to a US invasion.

The Submarine

I’ll start with the most famous of the incidents, the night when Arkhipov maybe—or maybe didn’t—save the world. Under the conditions of the quarantine—what the Americans called the blockade, as blockades are technically acts of war—no Soviet sea traffic was allowed into Cuba. The Soviets had sent a fleet of submarines to Cuba, which Vassily Arkhipov commanded, although he was second in command of the one he was on, the B-59 . Prior to the Cuban missile crisis Robert McNamera had instituted a new method for signalling to submarines to surface—dropping depth charges. However, the Soviet Union hadn’t accepted this and so hadn’t passed the information onto their officers. Therefore, when a flotilla of American destroyers located the Soviet sub and started dropping depth charges, the exhausted, exhausted, exhausted crew of the B-59 didn’t know what hell was going on.

I think this was a much less dangerous event than I think it is commonly believed, or at least that I believed until I started to read more deeply about the missile crisis. The first thing to note is that the B-59 carried a tactical nuclear weapon not a strategic one. Tactical nuclear weapons were battlefield weapons so if the weapon had been launched it would have destroyed the American floatilla but it would not have hit the American mainland. It’s not at all clear that this would have escalated to full scale nuclear war—there were no American war plans for instance that escalated from the use of tactical nuclear weapons by the Soviets to firing nuclear missiles. There have also been a surprisingly large number of conflicts between nuclear powers where one side killed the other’s soldiers including a small-scale border war between China and the Soviet Union in 1969, and numerous incidents between Indian and Pakistani soldiers. Finally, it’s just not clear why the use of a tactical weapon would escalate to use to a thermonuclear weapon. It’s not clear that it wouldn’t—maybe the taboo against the use of nuclear weapons is the only thing holding back their use and this would have broken it, or maybe this would have just escalated into a full scale war between the US and USSR. But maybe it wouldn’t and we know that both Kennedy and Khrushchev desperately wanted to avoid war.

There are conflicting reports about what actually happened in the submarine after the depth charges started to drop. There are some reports that it really only was Arkhipov opposing the bellicose captain, but there are other reports that the captain was opposed by all of the officers in the submarine. But all the reports agree that the captain wasn’t proposing firing the missile and certainly hadn’t turned his key in the firing system—he wanted to arm the missile. Granted, this means that the next step would be firing the missile, but I think the history of nuclear close calls suggests that at every step on the path to actually firing a nuclear missile any individual is unlikely to proceed to the next step. Whatever really happened, the Captain didn’t arm the missile and surfaced. The nuclear danger though, didn’t end there. Navy pilots dropped photographic canisters over the surfaced sub which the crew interpreted as a practice bombing run. Assuming they were about to be blown out of the water they started to return below deck to fire their nuclear tipped torpedo before a Navy messenger was able to signal to Soviets that they weren’t under attack.

The U2s

Two American spy planes moved the world closer to nuclear war during October of 1962. The first was flying over Cuba when it was shot down by Soviet anti-aircraft fire, killing the pilot. The Soviet military was not authorised to shoot down US planes over Cuba and the specific officer who ordered the planes to be shot down was acting independently of his commanding officer. However, there was no American reprisal and this was a move supported by the Joint chiefs, the arch hawks of the crisis. The reason was that a proportional response wouldn’t have destroyed the nuclear missiles in Cuba, the key target, while still dramatically increasing the risk of a Soviet response. Therefore the decision was taken to wrap the reprisal into the invasion that was scheduled for 3 days later.

The second U2 was engaging in a routine mission over the north pole. Because of how light the planes needed the U2s had very limited navigational equipment on board, meaning the pilots often had to use the stars to navigate. Unfortunately on that day the Northern lights were active on the plane’s flight path, meaning the pilot lost his bearings and ended up 300 miles into the Soviet union. Soviet fighter jets were scrambled with orders to shoot him down. American fighters were sent in response. But the real kicker, from a nuclear risk perspective, is that all of the fighter jets were armed exclusively with nuclear tipped missiles meaning a dogfight would have been fought with tactical nukes. The context for this moment of danger was that Kennedy had ordered the strategic air command to put on defcon 2 - the state of alert before a war footing. For our purposes this meant two things: firstly that US planes would carry tactical nuclear weapons and secondly that there be US nuclear bombers constantly in the air. We now have documents showing Kennedy did order this move—what he didn’t order was for Thomas Powers, the commander of SAC, to effectively tell the Soviets that his forces had been moved to the state of readiness before nuclear war. Nor did Kennedy or McNamara know that routine U2 missions near the Soviet union were continuing during the crisis. When the Soviets picked up the U2 on their radars they assumed that it was scoping out Soviet territory in preparation for it to be bombed and so the orders of the scambled fighters was to the shoot the plane down first and ask questions later.

There are two final cases of nuclear close calls which I’m including for completeness sake. The first came when a missile regiment received an order to fire its missiles on the Soviet union. The commanding officer refused the order, believing it not to be genuine as the US military was on defcon 3 and SAC on defcon 2 and they should only expect to receive orders to fire once at defcon 1. The officer threatened to have anyone who tried to follow the order shot, a threat that may have been necessary. The reason I’m not going into this in more detail is that there’s no historical consensus around whether or not it happened or not. The second case was a classic failure in an early warning system at the end of the crisis after Kennedy and Krushechev had agreed to a deal. Again, the evidence here is scant, and there are no reports as to the degree of seriousness with which the false alarm was treated. However, it mustn’t have been taken that seriously as there are no reports that any of civilians it would have gone through before reaching the President were notified.

Why

I think the Cuban missile crisis really shows well a few of the key themes of nuclear risk:

The existence of tactical nuclear weapons
The breakdown of civilian control over the military
Intelligence failures

I think the first really notable thing is how all of these moments of danger could have been avoided if none of the actors had tactical nuclear weapons. It’s possible that the invasion of Cuba would have triggered a conflict between Soviet and American soldiers and this would have escalated, but it seems like it was vastly more likely to escalate to a nuclear war if the Soviets used a nuclear weapon to destroy an American military base. I think there are a couple of reasons why tactical nuclear weapons are so dangerous. The first is that their use is delegated to the individual officers and sometimes individual soldiers. There are just a lot of individual officers and so it’s much more likely that one of these soldiers will get into a high stakes situation and that one of individuals in high stakes situations will choose to use their nuclear weapon. The second reason why I think they’re especially dangerous is that the threshold for their use is much lower than with strategic weapons. If the Americans had invaded Cuba is seems extremely likely that the Soviets would have used their tactical weopons whereas they’d never respond to an invasion of Cuba by bombing the continential US. It would have become just another proxy battle in the cold war.

The second theme I think is the military apparatus getting out of control and acting as a fundamentally bellicose force. The classic theory of civil-military relations comes from Samuel Huntington’s book the Soldier and the state in which Huntington lays out a vision of what civil military relations should look like. This ideal is the military as a tool of the state that does the states the exact bidding while the civilians leave the details of how to implement their vision operationally to the professionals in the officer corps. The classic case of this relationship breaking down was General Moltke the Younger going ahead with the Schilefeen plan in the first world war against the wishes of the Kaiser, invading neutral Belgium and Luxembourg and so bringing Britain into the war. The Soviet officers shooting down the American U2 and General Powers informing the Soviets of the move to defcon 2 are classic examples of the military acting independently and aggressively. This is the fundamental problem of bureaucracy—to have an effective state there must be delegation but delegation leaves open the possibility of independent actions against the wishes of the principle. However, the submarine and the U2 over Russia point to a deeper problem with Huntington’s model of civil-military relations in the nuclear age. When tactical nuclear weapons are used it’s not clear where the line between operational and strategic decision falls. Finally, in a return to a failing that falls much more neatly in Huntington’s paradigm is the decision by McNamara to institute a new way of signalling to enemy submarines. This is a classic case of civilians getting involved in operational matters that professional soldiers are much better equipped to deal with. Any submarine officer would have known a submarine captain would sooner die than be forced to surface.

Finally, intelligence failures form the backdrop of the whole crisis. The first and probably most egregious was not seeing that Castro was a nationalist not a communist and really would have allied himself with the US had they accepted his initial advances. The second was not understanding that Castro had broad support from the Cuban people in his capacity as a nationalist liberator and a pretty transparently US backed invasion would not endear the US to the Cubans. The Bay of pigs debacle was the inciting incident for bringing the missiles to Cuba. The next in this list of failures was failing to see that missiles were being installed in Cuba despite it taking over a month and the missile sites having extremely poor cover from the natural environment. This one, to be fair, seems much harder to solve and much less obvious at the time—photo interpretation is extremely hard! Finally, we have the failure to realise that there were nuclear tipped missiles both on Soviet submarines and on Cuba. Unlike the missiles being moved to Cuba, the use of tactical missiles wasn’t a radically new tactic and seems like a classic case of where probabilistic thinking could have led to a better informed US response. I think the Cuban missile crisis is the best example I know of where an organisation that could have feasibly used Tetlock style forecasting could have reduced existential risk.