Financial War Games?

In my new article Regulation by Hypothetical, I propose that one way to strengthen Dodd-Frank mandated stress testing and living wills (which I call regulation by hypothetical), is to engage in financial war games. You see, the problem with stress tests is that they test a firm’s balance sheet against financial stress at a static point in time to determine whether that balance sheet can withstand certain recessionary forces. But this misses a crucial point: Corporations are people, my friend.

Like Greek tragedies, when crises in financial firms are brought to light, they are often stories of individual hubris. For example, the failure of Lehman brothers was at least as much about Dick Fuld’s mismanagement of the firm as about the failure of the subprime market. Similarly, Bear Stearns suffered at the hands of a disengaged Jim Cayne. Bank of America’s ill-advised purchase of Merrill Lynch has been explained as a Southern outsider’s (Ken Lewis’s) desire to “play with the big boys on Wall Street.” Michael Lewis describes AIG’s irresponsible and market-creating purchases of CDOs from Wall St. as an arrogant and uninformed wager by Joseph Cassano, who has since been labeled “The Man Who Crashed the World.” Going back to previous disasters, the failure of Enron was about “The Smartest Guys in the Room”—Kenneth Lay and Jeffery Skilling’s dishonesty and conceit.

Stress tests and living wills (and the entire risk-management complex) leave this human decision-making out of their models.For example, in the event of a sudden stock market rise or fall, will a fund manager sell, buy, short, or hedge in a particular market? And will they use the opportunity to double down on risk, hoping for a big reward, or accept modest losses in order to prevent what might well be a greater loss to come? Stress tests and living wills are not designed to answer these questions; nor do they pretend to. The Fed states that the models “do not make explicit behavioral assumptions about the possible actions of a BHC’s creditors and counterparties…”

Despite my concerns in the article about regulation by hypothetical generally, how can regulators take account of the risks and problems that involve individual decision-makers if regulators press forward with regulation by hypothetical? Financial war games might be useful. Of course, the more involved and complex the simulation, the higher the cost and the higher the informational value. For example, before the Navy Seals operation that captured Osama Bin-Laden in Pakistan, the Seals practiced the maneuver many times in the United States. The military tried to the best of its ability to replicate the Pakistani compound housing Bin-Laden and cast various players in a realistic simulation to determine ex-ante potential problems they would face during the actual attack. The war gaming proved useful training for the troops. However, the one major hitch in the operation—one helicopter being grounded because of a centripetal air suction—occurred because, when reconstructing the compound in the US, the military surrounded it with a chain-link fence instead of walls, which created different landing conditions for the helicopters. Had they created an exact replica, even that problem could have been avoided.

While financial war games would not require the amount of expensive weaponry and human power that true war gaming requires, creating realistic simulations would be more costly than merely running computer models. Even so, given the dangers that failed financial institutions would pose to the world economy, conducting some high-stakes simulations may well be worth the cost. Indeed, conducting a financial war game is not without precedent. In March of 2009, the Pentagon hosted a two-day financial war game event at the Warfare Analysis Laboratory in Laurel, MD, a facility that is used for conducting military war games. However, the focus of the war games was on national security and not on financial stability. The military was concerned about global financial problems because of their “real world consequences, including failed states.” Bankers and Hedge Fund managers, among others, were invited to the pentagon to role-play a financial disaster. Paul Bracken, a Yale business school professor, who attended the sessions, stated that “The purpose of the game is not really to predict the future, but to discover the issues you need to be thinking about.”

What would financial war gaming look like? Borrowing from the Pentagon’s playbook, regulators would invite experts in the field to assume a role in the financial sector, the regulator would model a financial stressor or firm failure, and each party would react to protect their own interests. The data would be aggregated to get an accurate understanding of the vulnerabilities of the system as a whole and each individual firm. The players could be representatives from the actual firms or industry experts who could vicariously play their roles. Both choices have advantages and flaws. An insider likely knows and reflects firm culture. But the insider might try to game the game. For example, in the event of counter-party failure, a firm might engage in opportunistic behavior that would accelerate a counterparty’s failure or cause damage to other firms. But the insider might not display this sort of behavior in a low-stakes simulation because of public- or regulatory perception concerns. On the other hand, an expert who is not a repeat player in the markets and has no reputational concerns would take actions that actually reflect those made in real crises. Yet the downside of outside experts is that their inability to mimic or portray the culture of a firm that affects how decisions are really made.

So while I make the case in my article that regulation by hypothetical is fundamentally flawed, I think financial war games would provide better information about firm-specific and systemic vulnerabilities and how they might be dealt with in the event of a crisis.