Financial and national security policy makers both rely on maintaining credible portfolios of assets to counterbalance the threats or liabilities that risk cascading into catastrophic failures.
In March 2023, Silicon Valley Bank collapsed and was seized by its state regulator following a run on its deposits. The Federal Reserve Board and lawmakers were aware of the hazards which precipitate this kind of bank failure. Policies enacted following the 2008 financial crisis prevented banks from over-allocating reserves in riskier asset classes like mortgage-backed securities for that reason. However, risk reduction measures including the 2010 Dodd-Frank act failed to prevent depositor panic and what became the second largest commercial bank failure in U.S. history.
In February 2022, Russian military forces poured into Ukraine in what became the largest escalation of conflict in Europe since the Second World War. NATO national security policymakers understood the threat Russian forces posed to Ukraine and the risk presented by Russian forces massing on Ukraine’s borders. International arms control agreements like the Treaty on Conventional Armed Forces in Europe (CFE) specifically intended to reduce the risk of conflict in Europe. However, deterrence failed despite deliberate risk reduction measures and significant effort by democratic heads of state to convince Moscow of the devastating costs to be imposed by trade and financial sanctions.
Reflecting on these two events reveals common themes which offer a new way to think about integrated deterrence and deterrence failure risk assessment. Financial and national security policy makers both rely on maintaining credible portfolios of assets to counterbalance the threats or liabilities that risk cascading into catastrophic failures. For a large enough bank, failing to deter a bank run can disrupt the monetary system or create a prolonged global financial crisis; for a military power, failing to deter an attack can lead to defeat of allies and partners or unforeseen escalation to a nuclear conflict.
Lessons from big bank failures suggest that for a given level of strategic threat and the elements of national power aligned to deter them, risk can change forms but never be eliminated. Financial risk management concepts like liquidity risk, counterparty risk, and systemic risk are useful to understand and prevent catastrophic deterrence failure. The joint risk analysis methodology framework is insufficient to assess the risk of strategic deterrence failure in a way that is useful for national security policy makers. Strategic deterrence risk assessors should instead consider how total risk is distributed by specific mechanisms of deterrence failure rather than simply using levels like “low,” “moderate, “significant,” or “high.”
The Global Security Balance Sheet
A balance sheet is a measure of the assets and liabilities for a financial entity. For a retail bank, liabilities include debt and cash deposits from account holders which can be withdrawn on demand. Bank assets include cash and securities that the bank can sell for cash or leverage as collateral for short-term loans. Liquidity is the speed at which non-cash assets can be converted to cash. So long as bank depositors are unlikely to withdraw deposits faster than the bank can convert assets to cash in the short term, a bank’s liquidity risk is low. If lenders view a distressed bank’s probability of default as high, they can purchase credit default swaps which function similar to an insurance policy. Backstopping the liquidity of a retail bank is the Federal Reserve in its role as lender of last resort. The central bank can quickly lend cash to distressed banks through a program called the discount window.
While not an asset, the Federal Depository Insurance Corporation (FDIC) insures depositors for limited dollar amounts. FDIC deposit insurance is structured to assure the most vulnerable depositors and to deter panicked withdrawals from ever starting in the first place. The 2022 Nobel Prize winning Diamond-Dybvig model shows that even banks with strong balance sheets are vulnerable to bank runs. Once deposit withdrawals begin forcing a distressed bank to liquidate assets below book value, collapse becomes a self-fulfilling prophecy.
If you were to construct a national security balance sheet for a state actor, the liability side of the ledger would include the full range of military threats in every warfighting domain, including strategic attack. Assets on this balance sheet include all capabilities, military or otherwise, that the state can deploy or leverage to deter these threats regardless of readiness. Some of those assets will be more liquid, like a standing rapid response force, while other assets will be less liquid, like reserve components and auxiliary forces. Backstopping the liquidity of these assets are formal alliances and security partnerships from which smaller military powers “borrow” security to deter threats from larger neighbors, functioning as credit swaps or lenders of last resort do in the financial system. Formal military alliances like NATO are also the lender of last resort in emergencies—not unlike the Federal Reserve system or any other central bank.
Nuclear extended deterrence agreements also function much like deposit insurance does in the Diamond-Dybvig model. The more credible the commitment to defend allies under a nuclear umbrella, the less likely it will be necessary to employ nuclear weapons in defense of allies vulnerable to nuclear coercion. Deposit insurance is by design not a substitute for liquid assets at poorly managed banks, but rather an assurance to the most vulnerable depositors likely to suffer the consequences of a bank failure to lower the liquidity risk of bank runs. Similarly, nuclear extended deterrence agreements are not intended to offset insufficient allied conventional warfighting capacity or underwrite reckless provocations. Nuclear extended deterrence assures adversaries that escalation above a certain threshold will prompt nuclear retaliation so that conventional allied warfighting capabilities more effectively deter conflict at lower levels of escalation.
Aggregate Risk Can Neither be Created nor Destroyed
“So it gets back to this idea I have, which I think others have talked about: you’ve got the conservation of matter. Risk is neither created nor destroyed. It can transform and be chopped up and reallocated, but it all adds up to the same thing.”
Michael Hsu, Office of the Comptroller of the Currency Acting Director, November 2023
Following the 2008 financial crisis, U.S. regulators sought to constrain risks taken on by banks like Lehman Brothers, which invested heavily in assets like mortgage-backed securities before its collapse. The 2010 Dodd-Frank Act increased the reserve requirements of large banks to hold more of the most liquid assets like cash and U.S. government treasuries. Even banks not large enough to trigger all Dodd-Frank reserve requirements, like Silicon Valley Bank, bought up treasuries for years as the Fed’s quantitative easing flattened yields to historic lows.
The combination of rising inflation and Fed rate hikes in the post-pandemic era reduced the market value of these treasuries, which banks previously saw as risk-free assets. When panicked withdrawals forced banks to liquidate treasuries at below book value, the market discovered that many U.S. banks had simply exchanged one category of risk for another and were no less at total risk of failure than before reserve requirements changed. Dodd-Frank possibly presented a moral hazard to Silicon Valley Bank; compliance with regulatory safeguards lowered barriers to taking other kinds of risk.
The 2022 National Security Strategy explicitly called for the U.S. military to reduce the role of its New START Treaty-capped 1,550 deployed nuclear weapons. Some national security policy makers even call for the cutting U.S. nuclear warfighting capacity to reduce the risk of unintended nuclear employment or miscalculation. Meanwhile, Moscow has deployed over 2,000 non-strategic nuclear weapons in excess of its New START Treaty limit and Beijing’s rapidly expanding nuclear stockpile is on pace to reach 1,000 by the end of the decade. In February 2024, United States Strategic Command (USSTRATCOM) Commander, General Anthony Cotton, testified to the Senate Armed Services Committee that the concurrent expansion of adversary nuclear capabilities increased the risk of opportunistic nuclear aggression.
How can nuclear risk be simultaneously increasing due to too many U.S. nuclear weapons but also increasing because there are not enough? To strategic deterrence planners, the risk of strategic deterrence failure is primarily a liquidity risk concern. Deterrence will fail if USSTRATCOM cannot generate proportional response options in time to deter escalation. To nuclear arms control advocates the primary risk is operational, i.e., unintended nuclear employment through accidents or miscalculation. Both liquidity and operational risks are real and can coexist. Strategic deterrence planners and arms control policy makers consistently talk past each other due to the imprecise vocabulary of strategic deterrence risk assessment and preoccupation with characterizing relative levels of risk.
Strategic Deterrence Counterparty Risk
To paraphrase British economist John Maynard Keynes, if a debtor owes a bank $1,00 and cannot pay, that debtor has a problem. If a debtor owes a bank $1 billion and cannot pay, then the lender has a problem. This is the basic idea of counterparty risk in finance: Sometimes risk emerges from a bank’s relationship with its own clients.
For Silicon Valley Bank the counterparty risk was the concentration of high-net-worth clients among tech sector startups. A large share of Silicon Valley Bank’s deposit accounts exceeded $250,000 in value–over the limit for depository insurance protection—which degraded the deterrent value of FDIC insurance. High value account holders were among the first to withdraw their cash, precipitating the bank run and eventual collapse.
Strategic deterrence counterparty risk depends on the level of strategic threat faced by smaller allies. For example, the United States has committed to the defense of Iceland through both NATO and bilateral agreement despite Reykjavik’s minimal funding of its own national defense. However, because Iceland faces low strategic threat, the potential need to offset underinvestment presented by this relationship is probably low, as Iceland’s potential security “debt” to the United States is small. Contrast Iceland with the Republic of the Philippines: Spending less than one percent of GDP on its military in 2024, Manilla now faces the growing threat of military escalation from nuclear-armed China in the Philippines’ own territorial waters.
When a nuclear power like the United States extends security guarantees to many smaller regional powers threatened by nuclear-armed adversaries, a similar counterparty risk can emerge. Each individual ally faces a unique direct military threat and brings into the security relationship its own defense assets. However, if too many allies simultaneously underinvest in defense while strategic threats increase, then the counterparty becomes the mechanism of strategic deterrence failure. Silicon Valley Bank focused its banking business on wealthy tech industry clients resulting in the adverse selection of depositors most likely to panic. The United States as a credible ally and security partner unintentionally attracts the so-called national security “free riders” which pose counterparty risk in the aggregate.
Many other U.S. allies and partners have underinvested in their own defense for years and now similarly face growing strategic threat. To wit, in 2022, only seven of NATO’s thirty members maintained the alliance’s required two percent of GDP defense spending guideline. This could mean that multiple allies count on borrowing the same U.S. and NATO assets to provide liquidity and offset a growing range of strategic threat liabilities.
NATO leaders, like central banks in their lender of last resort role, are familiar with the moral hazard by which the promise to underwrite risk results in poor risk management by the insured party. For big banks, this risk taking could manifest as investments of the type that preceded the 2008 global financial crisis. For U.S. allies and partner nations, this moral hazard counterparty risk can manifest as the atrophied defense industrial base and military capacity which now challenges continued NATO support to Ukraine and deterring Russian strategic attack.
Strategic Deterrence Systemic Risk: Too Big to Fail
The phrase “too big to fail” entered common use by U.S. policy makers following the 2008 financial crisis to describe banks and other private entities so important to the financial system that their bankruptcy threatened to damage the entire financial system. Days after the failure of Silicon Valley Bank in 2023, U.S. Secretary of the Treasury Janet Yellen, in consultation with the Federal Reserve Board and the FDIC, declared Signature Bank a systemic risk and placed it into receivership. A bank large enough to become systemically important can expect some level of assistance from the federal government to be bailed out to prevent collapse from spreading.
Some major U.S. security partners lack explicit bilateral security guarantees or extended assurance agreements to backstop catastrophic deterrence failure. The U.S. policy of strategic ambiguity is structured in a way to create dilemmas of uncertainty for threat actors while retaining the optionality to respond to escalation against security partners on a case-by-case basis. This ambiguity can be sustained until the smaller security partner becomes so systemically important to the critical national interests of the United States that non-support loses credibility as a policy option. The partner becomes too big to fail.
Taiwan, for example, became so critical to global semiconductor supply chains that maintaining strategic ambiguity around U.S. security guarantees is now less viable. In 2022, the White House seemed to edge toward more strategic clarity on the United States’ commitment to defend Taiwan while aware of the moral hazard Taiwan’s “too big to fail” expectation of military bailout brings. Taipei’s defense spending has increased from around two percent in 2022 and is expected to reach 2.45 percent of GDP in 2025. So, while Taiwan’s security balance sheet is improving, a substantial amount of systemic risk is underwritten by U.S. security guarantees.
When a bank is designated a systemic risk, as was the case with Silicon Valley Bank in 2023, the FDIC can preemptively seize that bank and place it in a receivership status that waives the administrative process which is the equivalent of bankruptcy. The FDIC can then find a new buyer and negotiate terms of expanded deposit insurance coverage to sustain the troubled bank as a going concern. The owners of seized big banks do not benefit from this outcome, but the broader financial industry does. Other banks benefit from avoiding the contagion effects of a systemically important bank collapse.
While allies and partners are sovereign entities, there is a collective interest in the mutual defense of a security partner presenting systemic risk. During the Cold War, several direct U.S. military interventions aimed to prevent the contagion effects predicted by Eisenhower’s domino theory. Later, policy makers, including former Secretary of Defense Robert McNamara, criticized domino theory as driving intervention in non-systemic risk-posing conflicts like Vietnam. Lessons from recent banking crises suggest that systemic risk can be managed without Domino Theory-style intervention to prevent every single bank failure. Effective systemic risk management depends on the early identification of only the troubled systemically important institutions while avoiding the moral hazard that accompanies the expectation of bailout.
Applying Bank Failure Lessons Learned to Strategic Deterrence
The most striking difference between how financial risk assessors and strategic deterrence risk assessors discuss risk is how financial risk managers have a larger vocabulary of terminology to precisely characterize risk in financial disclosures. The joint risk assessment methodology categorizes risk to force, risk to mission, and strategic risk. Meanwhile, nuclear arms control advocates continue to focus risk assessment on the prospect of accidental nuclear employment and improbable worst-case scenarios with theatrical devices like the doomsday clock. None of this language is useful for decision makers looking for actionable policy levers to manage the very real risk of strategic deterrence failure.
If integrated deterrence is to remain a viable policy, national security policy makers must acknowledge that combinations of conventional military and non-military options like sanctions to offset nuclear capabilities change the mix of mechanisms by which strategic deterrence is likely to fail. Over-reliance on slow-working and potentially avoidable measures like sanctions presents a possible liquidity risk. Under-provisioning extended deterrence resources to allies in response to growing strategic threat presents possible counterparty risks. The emergence of “too big to fail” security partners facing growing strategic threat—and possibly counting on a U.S. military “bailout”—presents systemic risk.
For any given level of strategic threat, the risk of strategic deterrence failure can neither be created nor destroyed. Strategic deterrence risk assessors should spend less time attempting to measure whether aggregate risks are high, low, or otherwise. Instead, strategic deterrence risk assessors should approach risk assessments like the Federal Reserve Board of Governors examines commercial banks using the Uniform Financial Institution Rating System, not to criticize any specific ally as a risk to strategic deterrence failure, but to facilitate allocating the right mix of integrated deterrence assets needed to improve strategic stability. If policy makers identify a systemic risk, they need a playbook to mitigate possible deterrence failure contagion effects while minimizing the moral hazard an expected bailout can bring.
Steve Speece is an active-duty U.S. Army officer currently assigned to United States Strategic Command (USSTRATCOM) J2.
The views expressed in this article are those of the author and do not necessarily reflect those of the U.S. Army War College, the U.S. Army, U.S. Strategic Command, the Defense Intelligence Agency or the Department of Defense.
Photo Credit: Generated by Gemini and Fotor
