Mutually Assured Destruction: A False Phenomenon

With this blog post, I would like to take some time to coalesce and present some thoughts and conclusions that have possessed me of late, concerning the phenomenon of Mutually Assured Destruction.

For those unfamiliar, Mutually Assured Destruction, or MAD, is the idea that if I have nuclear weapons and delivery systems, and you have nuclear weapons and delivery systems, then we can create stability if we just promise each other that if the other tries to take advantage of us (the exact criteria here differs depending on who uses the theory) we will annihilate them.

In this post, I am not saying that standoffs are not possible, but only refuting the idea that an engineered standoff between real adversaries can be stable in the long term, which is essentially how MAD is supposed to work.

Before we discuss why MAD does not work, I need to lay some groundwork on how nations work. Nations, fundamentally, are enterprises with the goal of increasing their national worth. One of the most basic ways to do this, and one of the most lucrative, is to conquer other nations. In short, every nation wants to rule the world, even if they don't want to admit it, either because they're playing coy (The US) or are a client state of some larger power (everyone else, except maybe China). This model of the motives of nations will become important later.

To understand why MAD does not work, we must first understand the nature of threats. I do not know about other creatures, but I do know that hominids will not concede merely based on the word of a threat. This seems obvious: If you threaten me with nuclear annihilation, but you don't seem to possess any nuclear weapons, I may try my luck anyway. This is solved by simply doing regular demonstrations of power.

Except, it's also not that simple. Hominids evolved to compete primarily with other hominids, which is why we have huge, energy sucking brains but can't seem to do algebra to save our lives. This means that even if the other hominid has a big stick, and waves it around a lot, we may still try our luck if we believe they don't have the will to use it on us. Given a long enough timeline, the chances of this situation coming to pass approaches 1.

In addition, however, technology is not a constant. It moves, and old technologies get obsolesced. In 1945, the Soviets had virtually nothing that could deal with the threat of nuclear-armed B-29s. Today, swatting B-29s would be child's play for any two-bit nation. This means that as new technologies and capabilities are developed, old technologies are obsolesced. And what are the chances that every new game changing capability is developed by every nation on Earth simultaneously? Essentially zero. For hominids, that means that in a MAD scenario the time to strike is as soon as you have a game-changing advantage. Just developed an anti-ICBM, hydrogen-bomb-tipped, satellite-based, hypersonic missile? Attack now! Attack for God's sake, while you have the edge! This means that MAD actually encourages offense, if only in those areas where a significant technological gap exists.

Further, a policy of MAD is one that virtually no entirely rational, well-informed authority will enact. Why? Because it sacrifices large long term potential rewards for a relatively short-term one. Fundamentally, the nations of the world don't want to obliterate one another so that they alone remain atop a blasted hill overlooking a wasteland. They want to conquer one another, absorbing their population, materiel, and production capability (or, in other words, a nation is a predator, not an arsonist). A policy of MAD says that you will sacrifice that potential goal for the purposes of keeping your neighboring countries off your back. (You cannot enact MAD against only one nation, because then your MAD partner is greatly incentivized to equip your non-MAD-partners with nuclear weapons and get them to launch them at you). Not only is this tradeoff not very rewarding, it also, as I demonstrated earlier, doesn't work.

"But," you exclaim, "the US and Russia kept each other out of war for years using MAD! Clearly it works!" To which I'd respond: The US and USSR post-WWII were not true adversaries. They were sister states, and they knew it. In a true MAD policy, you would have to be very ready indeed to annihilate the enemy whenever they made a move. We don't see this with the US and Russia. The US doesn't, as soon as they get Pershing II missiles based in Turkey, strike at the Soviet Union, knowing the USSR can't do anything about it. They don't undergo crash anti-ballistic missile research programs, even though they knew full well that the Soviets very likely couldn't match them in research capability. With Saturn V launch systems, it would have been quite straightforward to lift a missile defense system into geosynchronous orbit. Yet, they didn't. In short, I do not think the US and USSR took MAD seriously. Neither really wanted to destroy the other* (which you must want for MAD to work), which is why they engaged in some petty proxy wars, and always tried to de-escalate** whenever the situation reached high tensions.

*Though it certainly does seem as though their generals did not know this.

**It does seem counter-intuitive that de-escalation would be bad for MAD, but it is. MAD is essentially the art of posturing with nuclear weapons, and seeking to de-escalate tells the other hominids in the room wearing red stars that you don't mean it, you won't destroy them, and they can attack you with impunity. Of course, they won't, as long as they're just as much a bunch of wet noodles as you are.