What is a Black Swan event, and can technology help mitigate it? “Black Swan” is a metaphor that usually describes a large-scale, large-impact, rare event (LSLIRE) that is extremely difficult to predict. Such events are unexpected, and when they occur, the complete lack of anticipation of such a significant happening tends to cause chaos, confusion and disruption. These may persist long after the event, adding to the initial damage caused. But exactly what constitutes a Black Swan?
Some say the September 11, 2001, incidents in the US are an example; others cite the Asian Financial Crisis of the late 1990s and the Global Financial Crisis of 2008. More recently, the COVID-19 pandemic was definitely regarded as one, drawing many comparisons with the Spanish Flu epidemic of the early 20th century. The Dotcom Shakeout and Brexit have been cited by many as well. However, as unprecedented as these events were at the time of their occurrence, there were indicators that they were going to happen.
The perpetrators of the September 11th terrorist attacks, for example, were already ‘persons of interest’ known to the authorities. In many financial systems, there were signs that all was not well before the 2008 financial crisis hit. In Wuhan, where the COVID-19 virus is thought to have originated, there were indications of a new viral strain with more serious consequences. Although Black Swans are unpredictable, even with technology, just knowing that such an event can occur helps to focus attitudes and mindsets on how to prepare and respond appropriately.
But one major challenge is perhaps the lack of data. Such events usually do not repeat themselves, making accurate modelling difficult. One advantage, however, is that being relatively high-profile, there is probably documentation of other similar events. Appropriate due diligence and the application of technology may unearth more information on events after they have happened. This big-picture information is crucial, as it provides an idea of the conditions when the event occurred. Many things are forecastable, and technology today can be used quite effectively for this.
New technology such as AI and generative AI, for example, can enhance the accuracy of assessments and analysis when records and inputs such as satellite imagery are combined. Data is often fragmented, unstructured, unstandardised or difficult to collect. Even within organisations, data may be siloed, with separate departments making their respective analyses and perhaps delivering conflicting information. This complicates decision-making. AI, on the other hand, is able to consider more factors in a shorter time span, providing stronger support during critical, time-sensitive periods.
Black Swans are difficult to predict, but there is a need to distinguish between a Black Swan and an LSLIRE so that they can be controlled and reduced. Risk professionals should try as far as possible to leverage mainstream analysts and model failure dynamics to reduce such events, and help in LSLIRE forecasting. They could ask qualitative questions; as Risk managers, they need to always step back and look at the bigger picture.
Qualitative thinking needs to be merged with quantitative thinking, and organisations should learn to deal with the question of arithmetic or quantitative and qualitative interface change issues just prior to a LISLIRE. What does this mean? They need to look for changes in what they believe the norm to be, understand and contextualize those changes, and then try to look at what to expect. There is a definite need for organisations to think outside the box.
They should also encourage the development of a culture where everyone not only learns to observe what is happening internally and externally but is also able to contextualise it. This is similar to the iterative process of ERM in many ways. However, most models and scenarios apply a retrospective perspective; they look at past incidents in order to forecast a future event. This does not work with Black Swans or LSLIREs. Research has also shown that as the LSLIRE or Black Swan event gets closer, the data models being relied on are more likely to become unreliable.
Experts believe that if the norms which are expected cannot explain what is happening, that is an indicator that an LSLIRE is close. Dealing with the unpredictable requires robust intelligence and data for critical support, hence the need for the application of the appropriate technology, to derive in-depth knowledge and understanding of the firm and its requirements at such crucial times. Equally robust internal systems that enable quick analysis, due diligence and data with integrity are must-haves. Big data sets can be managed with both human intelligence and AI.
Ultimately, a Black Swan cannot be predicted, nor can anyone determine how the event will play out, but knowing that such an event can occur helps to focus attitudes and mindsets on how to prepare and respond properly to disasters. In some cases, people may have been aware of the approaching incidents while many others may have just ignored the warning signs. Whether an event can be classified as a Black Swan or not, depends on the information that is available at the material time; the Black Swan concept is actually an encouragement not to attempt to predict the improbable.
Rather, it drives organisations to examine the measures already in place and determine how to limit possible losses should such an event occur. The positive effect of Black Swan events tends to be overlooked, perhaps because the negative outcomes have longer, more extensive repercussions. Some management experts have even singled out the Dot-Com Era as a positive Black Swan event, as it ushered in global digitisation, gave rise to the Internet, and jump-started connectivity between devices on an unprecedented scale.
