• Search in book...
• Toggle Font Controls

Chapter 10Game Theory – It's Not a Game!

It is relatively easy to produce a cyclical, constraint-based net assessment. Sure, it takes a lot of research, but the subject is simple and discrete. The real challenge is to operationalize competing sets of constraints, such as when two countries or multiple sets of policymakers face off against one another.

The multiplayer and competitive factors in these circumstances make game theory a useful tool of analysis. Game theory allows the forecaster to formalize choices and the conditional probability of events. However, a formalized model where each choice is assigned a precise mathematical value does not guarantee a correct – accurate – forecast. Precision is not the same as accuracy, and the presence of one (mathematical precision) does not guarantee the presence of the other.1 Game theory–based predictions are only as good as their empirical and contextual knowledge foundations. Without a solid base, they can be detrimental, even downright stupid.

In graduate school, I read some really stupid game theory papers. Political scientists – some of whom are essentially failed economists or mathematicians – would pick a topic, seemingly at random, to “formalize.” The outcome was often mathematically elegant, but empirically … well, inaccurate, and therefore useless in application.

From their mistakes, I learned how important it is to know something about the topic at hand – something concrete and empirical – before attempting to formalize it. As such, I hesitate to discuss game theory in this book. I don't actually use it that often. But because I see investors use it all the time when navigating geopolitics, I decided to include this chapter as a guide for the dos and don'ts of game theory. At the very least, it provides the observer with some ways to recognize a well-researched game theory analysis.

Before You Play the Game, Set the Gameboard

The key to game theory is in assessing which actor holds the greater capability to change its rival's expected utility function. To calculate expected utility, take the weighted average of all possible outcomes under certain circumstances.2

Investors often overstate the role of material power in forcing capitulation and changing the rival's expected utility. In the contest between the US and China, investors throughout 2018 and 2019 argued that the US had enough material leverage to force China to accept onerous terms in the trade war. This did not happen. The Phase One deal, concluded in late 2019, was not much different from the May 2018 agreement that Beijing offered Washington. In other words, the extra 18 months did not really budge Beijing from its original position. Why? Because material power was not the only factor on the gameboard.

When assessing relative power, investors should consider three factors that measure an actor's relative capability:

• Material balance of power: In the case of the US–China trade war, the material balance of power favors the economy that is less leveraged to trade – clearly the US (Figure 10.1) – and the country that is the “customer” in the relationship – again, the US (Figure 10.2).
• Risk tolerance: The actor willing to risk more is the one with greater interests at stake. For this factor, the balance of power is more equally distributed. China had reason to stress stability in 2019 given its 70-year anniversary celebration as the People's Republic of China. But the US was heading toward a general election, meaning it had the greater interests at stake should US–China relations go south. It is not clear what the transmission mechanism between a US recession and a trade war would be, but the economy was clearly nearing the end of the cycle even before the COVID-19 outbreak. The yield curve inverted in 2019, signaling approaching recession, and animal spirits were already weak.
• Credibility: The reputations of the players matter. Do the actors involved have follow-through? President Trump gained global street cred early on in his term by being especially tough on North Korea. However, the renegotiated US–Korea Free Trade Agreement only marginally benefited the US while the United States–Mexico–Canada Agreement (USMCA; “NAFTA 2.0”) fell far short of the White House's initial “ask.” Therefore, President Trump's bark appeared stronger than his bite from Beijing's vantage point. For China, establishing credibility and signaling a high risk tolerance were critical to national security. As a result, it maintained its credibility throughout negotiations. Any sign that it had reached its pain threshold on trade could prove fodder for US decision-makers eager to force a Beijing behavioral change in future conflicts, whether on trade or geopolitics.

With the addition of two factors, China wins the capability game two to one. But balance of capabilities alone does not help resolve the trade war puzzle. China and the US are not only confronting each other, but also the competing interests in geopolitical and domestic arenas. They are playing a “two-level game.”3 Two-level game theory posits that domestic politics creates acceptable “win-sets,” which are then transported to the geopolitical theatre. These win-sets, if lost, become constraints to policymaker action. As Figure 10.3 illustrates, politicians cannot conclude foreign agreements that are outside of those domestic win-sets. For a win-win outcome, all four arenas must align.

All of these points combined result in Figure 10.4, which illustrates the competing “tolerance curves” of China and the US. The figure suggests that the Chinese economy – saddled with the constraints of export dependence and an addiction to the American consumer – has a lower pain threshold than the US economy. The difference between the pain thresholds is determined purely by material power.

Although China's pain threshold may be lower, its internal politics, focus on the real economy (as opposed to the equity market), and lack of an election calendar mean it has fewer domestic political constraints than the US.4 As a result, it has sufficient political capital to delay when it reaches the capitulation point.

Game theory is a powerful tool, but investors have to be careful in wielding it. Setting the gameboard is as important as playing on it, as it reveals material factors are not the only constraints in play. Often, analysts have to play three-dimensional chess when determining which side is more constrained on the domestic front, especially when considering domestic politics and interest groups.

Beware Simplistic Games

In early 2015, Greek voters gave SYRIZA and its leader, Alexis Tsipras, the mandate to govern Greece and negotiate with Europe. Tsipras appointed Yanis Varoufakis, an economist and game theory expert, to be his finance minister. At the height of their battle with Brussels, many market commentators flagged their “game of chicken” as likely to produce a negative outcome.

For those who are not fans of James Dean movies, a “game of chicken” refers to a scenario where two young men, each behind the wheel of a hot rod, drive straight at one another to prove their bravery and win the heart of a young woman. The one that swerves – the “chicken” – proves unworthy of wooing the lady thanks to his high levels of risk aversion and self-preservation (a signal to the potential mate that he will fail later in life, obviously). The one that keeps driving straight wins.

The game of chicken is indeed the most dangerous form of game theory. In a world where both actors assume rationality in their opponent, each has an incentive to drive straight and not swerve. This brand of reckless obstinacy means that the bottom-right outcome in Figure 10.5 – the costliest and least optimal for both actors – has a high probability of occurring.

In the case of Greece in 2015, however, Athens was massively constrained by its material reality: a subpar mode of transport. Using constraint-based analysis, Figure 10.6 illustrates the game much more accurately. The adjusted outcomes show that game theory is useless without first correctly assessing each actor's relative power, which includes material balance of power, risk tolerance, and credibility.

In the case of Greece, Tsipras was facing considerable political, economic, and geopolitical constraints to leaving the EU. These constraints lowered his vehicle's horsepower relative to Angela Merkel's.

On the political front, polls throughout 2015 showed that over 60% of the Greek public supported Euro Area membership. On the economic front, an exit from the Euro Area would have produced a massive depreciation of the drachma, accompanied by a massive rise in prices because the Greek economy imported its energy needs, as we discussed in Chapter 5. Yes, there was the off chance that a depreciation would lead to foreign investment and an export boom, but not before the angry median voter forced Tsipras out of power. On the geopolitical front, Greece had no alternative to Europe, no other ally to fall back on. When Tsipras signaled that Greece would consider Russian financial support in exchange for a geopolitical alliance, President Vladimir Putin rebuffed him.

The logic behind my view that Greece was bluffing in 2015 is strong, and steadfast empirical evidence supports it. German Chancellor Angela Merkel and Tsipras may have been heading straight for each other, but they were doing so in different vehicles. While Merkel was driving a G-Class Mercedes-Benz SUV, Tsipras and Varoufakis were pedaling a tricycle.

Game Theory: The Takeaway

Game theory is often mentioned but rarely applied correctly in the investment community. There is too much prima facie theorizing going on. I have read too many game theory–reliant reports on investment that go little beyond material power.

To properly set up a game, investors first have to produce a comprehensive net assessment of the actors whose interaction is being formalized. Accurate information about the actors is key. The game of chicken between Greece and Germany relies on the net assessments in Chapter 5, where I explained the constraints faced by policymakers in both Berlin and Athens.

In Chapter 5, I showed that during the sovereign debt crisis, a Euro Area exit for Greece would have been folly. At the same time, I illustrated that Berlin did not want the Euro Area to collapse. By 2015, the odds that a Greek exit would produce such an outcome were much lower. Thanks to the ECB's actions to ring-fence Mediterranean economies against bond vigilantes, combined with the fact that Greek sovereign debt had been cleansed from Europe's banking sector, Berlin felt it could play hardball with Athens and not risk a dissolution of the entire Euro Area. A constraint-based analysis bolstered by deep contextual knowledge of each actor's relative capabilities proved the game of chicken was not quite so high-stakes as it seemed in 2015. Those who invested based on this dire Euro Area prediction lost the game.

A way to update this game for 2020–2021 conditions would be to consider a similar confrontation between Italy and the EU. Italian policymakers are certainly not riding a tricycle. They are behind the wheel of a finely crafted Italian supercar. Much safer than a tricycle, for sure, but still no match for the G-Class Benz.

While Italy has a much larger economy than Greece, the reality of material constraints is still cogent. As such, investors should expect to see Italian populists consistently bend their Euroscepticism once they come to power, as the Lega party did during its brief time in power between June 2018 and September 2019. The main difference between Greece and Italy is that Germany will have to accommodate Italy more than it did Greece. As such, the COVID-19 recession is likely to see Germany cross a number of red lines that it never did with Greece, including the adoption of some form of debt mutualization.

Without a thorough net assessment and in-depth knowledge of the matter at hand, game theory is dangerous. The mathematics can produce a false sense of confidence in the forecast. An accurate assessment of the players and gameboard is key. Investors should therefore only engage in game theory once they have produced painstaking research to underpin their net assessment.

Notes

1. 1   Matt Parker, Humble Pi: A Comedy of Maths Errors (New York: Penguin, 2019).
2. 2   , assuming that the utility is logarithmic, which is most often the case where material outcomes are concerned.
3. 3   Robert Putnam, “Diplomacy,” 427.
4. 4   In Chapter 4, I noted that the strongest constraints to policymaker action are political, and the two-level game theory supports this view. They play a key role in determining an actor's capability on the global stage.
• No Comment
..................Content has been hidden....................