Unhedged (or Unhinged) Wrap

The weird rationality of ignoring systemic risk

A Davy Crockett battlefield nuclear weapon

---

The final business I worked in during my Wall Street career was structured derivative products—the stuff Warren Buffett once called “weapons of financial mass destruction” and that, used on toxic subprime mortgages, wound up causing the financial crisis. Most of our portfolio was based on corporate debt (which was my area) rather those mortgages, but the financial technology was the same: constructing a pile of assets and issuing a series of debt “tranches” backed by those assets with varying levels of subordination, and consequently different levels of risk and corresponding credit ratings. Our parent bank sold the debt, and our job was to manage the assets in the pool, as well as various other risks associated with the structures.

Most of these risks could be hedged, and we prided ourselves on running a relatively tight ship in that regard. In particular, we tried to avoid taking “correlation risk,” the factor that was systemically mispriced by the entire market, and which in a theoretical sense got the “blame” for the crisis. Correlation risk is the risk that the values of assets will move in tandem; in the context of defaults, it’s the risk that defaults will happen simultaneously rather than being spread out over time. Structured products like mortgage derivatives and the corporate deals we mostly did were highly sensitive this factor since, if defaults clustered together, then the apparent risk-reduction of building a diversified portfolio would prove illusory in a crisis. (This is the factor that, in the movie Margin Call, the Zachary Quinto character realizes is wrong; once he puts more realistic numbers in, he discovers that the firm would be bankrupt if the securities it holds were properly priced, which, once that information is elevated up the ranks to the head of the company, triggers the decision to liquidate everything before the market gets wise, even though that means destroying their customer base.)

In any event, we tried to avoid holding correlation risk on our books by matching our assets and liabilities as closely and completely as possible. Among the things this meant we did was hedging our so-called “super-senior” risk—stuff that was supposedly rated better than AAA—with financial entities that specialized in this sort of risk that were themselves rated AAA. We were paying money to insure against a risk that, according to our models, was virtually nonexistent, but we did it anyway because it meant we were as completely hedged as we could be. All of which sounds—and felt to us—like a relatively prudent approach to a risky asset class.

Here’s the thing, though. Those entities we traded with did lots of these kinds of deals. That meant that they were effectively enormous warehouses of correlation risk. If that risk was systematically mispriced, then an event which revealed that misplacing—like the crisis that ultimately happened—would also reveal that their portfolios were much riskier than they realized. They would probably be downgraded; they might even go bankrupt.

At some point, we realized this problem, and discussed what we might do about it. The thing was, there was nothing we could do. The same risk we were worried about applied to any other financial institution we might conceivably hedge with. The institutions we were already hedged with ran, ostensibly, extremely low-risk operations. If they went under, everyone would go under. As one of my colleagues put it at the time: “If that happens, we’re all dead anyway.” So it wasn’t worth worrying about.

That sounds like an appalling attitude to take, but it’s actually entirely rational. If there’s a big risk you can’t do anything about, then the rational risk is to ignore it. After all, fretting about the risk won’t change anything. And a risk is not a certainty. If it’s 1900, and there’s a 10% chance that a comet is going to hit the earth in the next decade, the right thing to do is not think about it too much. You have no technology capable of diverting the comet, so there’s no way to improve your situation if it is going to hit, and there’s a 90% chance it won’t hit anyway.

I want to be clear as an aside here that the financial crisis was in an important way not like a comet at all. There were very real things that the ratings agencies and government regulators could have done to prevent or at least mitigate the crisis. But an individual couldn’t do those things, nor could an individual firm. Indeed, they couldn’t even profit from the coming crisis without implicitly relying on the very systems they were betting against ultimately getting bailed out. Late in the movie The Big Short, the Steve Carell character realizes that the trades he did to short these toxic mortgage structures were ultimately hedged with his own parent bank. He was betting against his own firm, in other words, and that firm would only be able to make good on their obligations to him if they got bailed out. That’s why the risk is called systemic: it is pervasive throughout the financial system.

---

Why am I bringing all this up now? Because I think it’s relevant to our fears about nuclear war that have significantly increased due to the conflict in Ukraine, in two important and related ways.

The first of these is in terms of trying to read signals to evaluate the odds of such an end-of-the-world event. Precisely because of the dynamic I describe above, price signals from markets are going to be next to useless for this purpose. If the United States and Russia start lobbing nuclear weapons at each other, the financial markets as we understand them will cease to exist. Indeed, the United States and Russia will likely cease to exist; it’s possible human life itself will cease to exist. Because there’s no way to “hedge” against the end of the world, there are no trades to execute to express anyone’s view that the odds have gone up. And since those trades are what would normally cause prices to move to reflect those changed odds, a rise in the likelihood of the end of the world should have zero impact on financial markets.

This is what the folks at BCA Research put in a recent strategy report that came in for considerable mockery over the weekend. But they’re entirely correct. It’s just important to understand what it really means: that reassuring signs like the stock market continuing to have normal up and down days should not reassure you at all about the risk of nuclear holocaust. They should only reassure you about more mundane things, like the risk that sanctions will have a negative impact on the American economy that the Fed won’t be able to offset.

Robin Wigglesworth @RobinWigg

Amazing note from BCA: "The risk of Armageddon has risen dramatically. Stay bullish on stocks over a 12-month horizon."

I’m going out on a rather considerable limb in saying this, but I think to a large extent the same thing is true of other kinds of signals: diplomatic, military, etc. That sounds crazy, but the thing about a general nuclear exchange is that it is just about the most extreme example I can imagine of a non-rational action. In the best-case scenario, a nuclear war between the United States and Russia would mean that what is left of the world would be inherited by Chinese, Indians, Africans, South Americans. How would that serve either country’s interests, or the interests of either country’s regime? In the worst-case scenario, meanwhile, there would be no rest of the world to inherit. The only reason either Russia or America own strategic nuclear weapons is to prevent the other from using theirs. That’s why it’s called mutually-assured destruction.

Since initiating a nuclear exchange would be wildly irrational under basically any circumstances, though, it’s not clear how one would assess the odds that one was coming. How do you calculate the odds that someone will do something that makes no sense? Nor is it clear what signals would be reliable indicators that a nuclear attack was being planned. No course of action rationally leads to a nuclear exchange, so it’s hard to say what precedent actions would be signs that one was coming.

The corollary to the foregoing is crazier still. If it’s irrational to use strategic nuclear weapons in the context of mutually-assured destruction, then it is equally irrational to expect their use. Threatening to use strategic nuclear weapons in the context of mutually-assured destruction is fundamentally not credible, and threats that aren’t credible can be discounted. Which means one should logically proceed with your military planning as if neither you nor the enemy possessed such weapons.

You can already see this logic already playing out in people’s statements about the conflict. Why aren’t we using every tool at our disposal to stop Russia’s war on Ukraine? Yes, we’re only obliged to defend Estonia and Poland, and not Ukraine, but shouldn’t we defend Ukraine? We defended Kuwait after all; they weren’t a treaty ally either. It can’t be because Russia has nuclear weapons, because we have nuclear weapons too. If we’re so scared of getting into a conflict with Russia because it might lead to the end of the world, then isn’t our commitment to defend Estonia and Poland less than fully credible? If that’s the real reason we’re not willing to shoot down Russian planes that are bombing Ukrainian civilians, then aren’t we undermining NATO by not getting involved, and sacrificing Ukraine for nothing?

If we followed that logic, we’d wind up concluding that, because we shouldn’t worry about the bluster of nuclear threats that aren’t credible, we could consider actions like imposing a no-fly zone over Ukraine. Yes, Russia might well retaliate by attacking Poland or Estonia—but Russia would be substantially overmatched in a conventional conflict with NATO, so they would have to think twice about risking that next step up the ladder of conventional escalation. And if they did take that step, then we could respond in kind. What are they going to do—nuke us and get nuked in return?

We’d keep adding risk into the system, with no feedback that could reliably tell us that we’d triggered precisely the catastrophic event that we said all along wasn’t worth seriously contemplating, because it was irrational. And there’ll be nobody to bail us out.

---

Thankfully, I don’t think the Biden Administration is quite so clueless about where that road ends. That’s why they have been abundantly clear about what NATO will and will not do. They threatened severe economic sanctions, and have delivered on that threat. They said we would not send troops to defend Ukraine, and we haven’t. They know that the risk of accidental launch in times of high tension is very real, and is reason enough to try to avoid such situations and, if they can’t be avoided, to ensure that there is regular communication between both sides.

They also know that there are a number of scenarios for the deliberate use of nuclear weapons that do not involve a general strategic exchange. Multiple countries, for example, have at one time or another been willing to use tactical nuclear weapons defensively, to destroy a superior invading enemy force. America deployed weapons to do exactly that during the Cold War if Soviet tanks rolled into West Germany, and nuclear first use is also part of Pakistan’s planning for repelling an Indian invasion. The most-likely way nuclear weapons could get introduced into the Ukraine war would be if Russia used tactical nuclear weapons against NATO assets to prevent a conventional attack by the alliance. Precisely because it would be irrational to respond to a limited nuclear use of that sort with a strategic exchange, one could convince oneself that the threat to use tactical nuclear weapons to assure regime survival was rational, and therefore credible. But it would be crazy to assume that things would necessarily stop there once the threshold was crossed.

Last point. I’ve noticed a marked eagerness to characterize Putin as fundamentally irrational, and to wield that as a justification for more aggressive action. So I think it’s vital to keep the irrationality of a nuclear exchange in mind when assessing the behavior of other parties, particularly behavior that looks like it is tempting disaster. Was Russia’s invasion of Ukraine irrational, and therefore a sign that we should fear more irrational behavior, including possibly ending the world? I don’t think so: it was extremely risky behavior, and likely a bad miscalculation. But that doesn’t mean it was irrational. Going for broke is a very different matter from deliberate suicide.

We should spend a lot less time worrying about whether Putin is crazy, and more time making sure the Russians in general don’t think we’re crazy, and don’t think we think they are either.

My Week in Punditry

Three pieces this week, two at The Week and one On Here.

• Most recently at The Week, I had a piece on how the war between Russia and Ukraine might well go on much longer than people expect, without an off-ramp or an end-game. I do worry that our desire for some kind of decisive outcome is going to lead us to do something incredibly stupid, so I wish more responsible voices were warning of the possibility that this thing could simply drag on indefinitely. By the same token, I wish we were giving more thought to what isolating Russia for the long term will mean for the shape of the rest of the world. This piece in The Economist discusses that question from an economic perspective; I’d love to see a similar piece tackling the implications for our strategy to contain China.

• Previously at The Week, I had a piece about the Academy Awards, and how even back in the heyday of the blockbuster Oscar tended to favor smaller, human-scale stories.

• Finally, the one post here from earlier this week was about anti-heroes, a response to this piece from Freddie deBoer.