As risk professionals, we excel at modeling the probable and the plausible, but can struggle when confronted with scenarios that go beyond our historical framework. The past few years have delivered a sobering reminder of this blind spot.
The Global Financial Crisis and the 2020 pandemic shutdown were not just improbable events: they exceeded the imaginations of most risk managers. Even a foreseeable event, such as the rapid rise in interest rates in 2022, was not anticipated by many institutions, despite extensive monetary policy history. Every crisis teaches us the same uncomfortable truth: the greatest losses often emerge from scenarios we deem too unlikely to prepare for.
Traditional scenario analysis works well for routine business-as-usual planning and stress testing, but may be inadequate for true tail events. Our challenge goes beyond imagining extreme risks. The real value we provide is developing actionable frameworks and models for events that defy typical probability distributions. In short, how do we design scenario planning for the unthinkable?
Start with What Could Break
Most scenario design begins with historical analysis. That is, we extend past patterns incrementally into plausible futures. This approach may work for general stress testing, but misses the mark for tail risks. By definition, genuine tail events have no meaningful historical precedent.
Consider the economic fallout of COVID-19. While previous pandemics exist in the historical record, none occurred in an era of globalized just-in-time supply chains, integrated information technology, and central bank balance sheets that were already stretched by post-2008 interventions. The scenario wasn’t just a pandemic; it was a pandemic plus financial system complexity plus unprecedented policy responses plus digital infrastructure dependencies.
Effective tail-risk scenario design must start with structural analysis rather than statistical extrapolation. Instead of asking, “What happened before?” we should ask, "What could break the fundamental assumptions underlying our current operations?"
Every public or private institution operates on foundational beliefs. We assume markets will remain liquid, counterparties will honor their obligations, information technology systems will function, or regulations will change gradually. Questioning these assumptions can lead to meaningful tail risk scenarios.
Inverting Assumptions
One method for designing tail risk scenarios is to systematically invert our key business assumptions. What if funding markets froze while operational systems failed? What if traditional economic correlations suddenly reversed as geopolitical tensions severed critical business relationships?
Utilizing this process requires discipline. We must resist the natural tendency to move our scenarios back towards something “believable” by asserting that conditions have changed or presuming institutions will be nimble. If a scenario feels comfortable or easily managed, it’s probably not capturing true tail risk. The goal isn't plausibility, it’s identification of core vulnerabilities.
Bottom line: If you can easily explain why a scenario won’t happen, you’re not thinking hard enough about how it could.
Cascading Effects
Tail risks rarely emerge from single-point failures. They develop through cascading effects across financial markets, operational systems, regulatory environments, geopolitical structures, and technological infrastructure.
Cristian deRitis
Consider a typical operational risk scenario: A cyberattack renders our internet service provider inoperable. In isolation, this threat may be manageable through business continuity planning. But cascade modeling may expose deeper vulnerabilities.
Suppose the outage affects not just internal systems, but also those of key counterparties, market data providers, and regulatory reporting platforms. Our settlement systems may become strained as employees switch to manual processes. Our traders may lose access to risk management tools precisely when market volatility spikes due to the widespread operational disruptions.
The ripple effect may extend beyond technology systems. In the case of a bank, media reports may amplify customer concerns, triggering a collapse in stock prices and a run on deposits. Funding pressures unrelated to the institution’s actual financial position may accelerate. Regulators, facing their own system challenges, may delay critical approvals as they demand enhanced reporting due to the heightened risks.
What began as a limited operational event could quickly transform into a reputational crisis, then a liquidity crisis, and then a regulatory crisis. While usually hidden, this interconnectedness may become obvious during times of stress.
Accounting for Human Behavior
The human element is another aspect to consider. Risk models may implicitly assume rational responses to events. Tail risk scenarios must account for amplification effects that can transform manageable problems into existential threats.
During the March 2020 market turmoil, we witnessed how panic selling, credit hoarding, and social media reporting rapidly escalated moderate problems into systemic concerns. Behavioral factors operate differently during tail events. Normal market stress brings measured responses from experienced participants. Genuine tail events can push participants into uncharted territory where established protocols fail, and emotional decision-making dominates.
Tail-risk scenario design should incorporate behavioral assumptions explicitly. How will customers react when facing unprecedented situations? How will regulators respond when normal supervisory tools prove inadequate? How will internal teams perform when standard operating procedures no longer apply?
The COVID-19 experience highlighted these behavioral dynamics. Many institutions had robust capital positions but struggled operationally, given technological limitations, risk models that expected economic indicators to remain well-behaved, and process dependencies that seemed minor during normal operations but became critical during a crisis.
Test Operational Resilience as Well as Capital
Financial losses, while painful, can be absorbed through capital buffers. However, operational paralysis during a crisis can transform survivable events into fatal ones. A firm can rebuild capital, but only if it continues to operate.
As a result, scenario design must stress-test not just balance sheet capacity, but operational capacity under extreme conditions. Can critical functions continue when key systems or personnel are unavailable? Are backup communication systems available? Are processes resilient enough to support emergency decisions that differ from standard procedures?
Backup systems may not be as independent as we think. For example, a “redundant” communication system may rely on the same underlying supplier or infrastructure. Without digging deeper, a small oversight like this could have potentially large consequences.
Focus on Practical Frameworks
Effective tail-risk scenario design requires structured processes that overcome our cognitive biases. “Red team” exercises, where designated groups actively seek to break systems, can surface vulnerabilities that conventional planning may miss.
Including external perspectives in scenario workshops can add value. Former regulators, crisis veterans from other institutions, economists, technology experts, and even science fiction writers who specialize in imagining system failures may identify vulnerabilities that insiders may overlook.
Scenarios should be constructed to maintain a proper balance between detail and adaptability. Excessively prescriptive scenarios risk becoming obsolete, while scenarios that lack sufficient specificity may fail to offer meaningful direction. Recognizing that actual outcomes will not align perfectly with prepared scenarios, emphasis should be placed on establishing adaptable frameworks and models rather than rigid protocols for specific situations.
The Preparedness Paradox
Organizations that prepare for tail risks often avoid them entirely. Their enhanced resilience can transform potential tail events into manageable stress events. This presents a measurement challenge because “success” is a non-event rather than a clearly identifiable achievement.
This paradox means tail-risk scenario planning requires institutional commitment that goes beyond typical performance periods. The benefits of scenario planning only materialize during infrequent crises, which may not arise for many years. Effective implementation requires leadership that views risk management as a form of institutional insurance, rather than solely focusing on short-term profitability.
The Path Forward
In our interconnected world, increased system complexity is leading to exponential risk. The capacity to imagine and prepare for “inconceivable” scenarios has become a core competitive advantage. The institutions that survive the next tail event will be those that spend time designing scenarios they hope will never occur.
The goal isn’t to predict the future – that’s impossible. Rather, it's to build organizational resilience that can adapt when the future delivers a scenario we didn’t anticipate.
Parting Thoughts
Tail-risk scenario design requires us to get out of our comfort zones. We must acknowledge that our most sophisticated models and deepest expertise have blind spots. The scenarios that will matter the most are the ones we’re least prepared to imagine.
The key is not perfection but preparation. Organizations that systematically challenge their assumptions, stress-test their models and operational resilience, and cultivate the capacity for rapid adaptation will find themselves better positioned when the unthinkable becomes reality.
Remember: every tail event in history was once considered “unthinkable” by subject matter experts. The question isn’t whether the next tail event will occur – it’s whether we’ll be ready when it does.
Cristian deRitis is Managing Director and Deputy Chief Economist at Moody's Analytics. As the head of econometric model research and development, he specializes in analyzing current and future economic conditions, scenario design, consumer credit markets and housing. In addition to his published research, Cristian is a co-host of the popular Inside Economics Podcast. He can be reached at cristian.deritis@moodys.com.