As the world navigates a global pandemic, 2020 has conditioned us all to live with uncertainty in new and challenging ways. In addition to Covid-19, charged racial unrest and a volatile U.S. election cycle have added to a pervasive sense of instability. “Black swan” events, so named because of their rarity, are now the norm in business, financial markets, healthcare, and virtually all facets of society.
Despite the drama of 2020, history suggests that black swan events are not as rare as the name suggests. Consider the dotcom crash in the early 2000s, the 9/11 attacks, the 2008 global financial market crash and the Brexit vote in 2016. According to conventional wisdom, none of these was expected. Yet a more rigorous analysis of factors at the time would suggest their occurrence. Probabilistic, or stochastic, models using Monte Carlo simulations are one such analysis. Monte Carlo simulations examine a large range of possible outcomes – objectively and without bias – with a view to determining how likely each is to occur.
While risk modeling to shape strategic positioning and responses has always been important, this past year has broadened the awareness of the need to evaluate risk. The global pandemic has disrupted industries by introducing economic uncertainty and risks of recession. Geopolitical strife has further added to volatility across the globe. A dynamic risk management approach supported by analytics and modeling enables companies to lead from the front, avoiding pitfalls and seizing opportunities. What’s more, easily accessible desktop technology can equip everyone in an organization to be risk manager.
According to Aon, a leading global professional services firm in the insurance risk industry, client research revealed that 90 percent of global risks are either uninsurable or underinsured. Furthermore, another Aon survey found that risk readiness at the beginning of 2020 had declined to its lowest level in 12 years. Too many organizations have yet to invest in new tools and approaches to identify and assess risks to develop protection and mitigation strategies. Only 24 percent of respondents said they quantify their top 10 risks, only 20 percent use risk modeling at all, and 10 percent do not have a formalized process in place to identify risks.
Technology to support risk management needs to be adaptable and accessible. This past year has clearly demonstrated that models and data evolve in real time, and everyone in an organization has a role in managing risk. Firms can no longer afford to view risk management as siloed, check-the-box function. Black swan events will occur, and may be positive or negative, depending on a firm’s strategic preparedness, perspective and positioning.
For example, in financial markets, forecasting risk and preparing investment portfolios for the unexpected can actually transform risks into opportunities by identifying buy-side or short-sale openings. While in the realm of disaster preparedness, emergency planning and risk forecasting before a natural disaster can save lives.
Investment and financial analysis
Savvy investors and financial analysts know to expect the unexpected. Monte Carlo simulations addressing unexpected market fluctuations provide insights to help manage long-term investment portfolios in ways that conventional static models cannot address.
A primary factor in the success or failure of long-term portfolio outcomes is the allocation among asset categories, according to Joe DiNunno, an asset allocation strategist and consultant to FiduciaryVest, an institutional investment advisory firm. Investment management within an underlying asset class is an important, secondary issue.
By incorporating longer time horizons and realistic probability distributions associated with asset classes, DiNunno said simulations address issues beyond what can be done with static, deterministic financial models. For example, simulated models show probabilities of reaching target returns which would allow a client to meet financial obligations without tapping into portfolio principal. “The models give our clients a strong indication of the impact that each component of asset allocation and diversification delivers in terms of their specific needs for future investment returns and risk management,” DiNunno says.
Emergency management planning
The need to adapt quickly was validated all too often during Covid-19 pandemic. Deterministic models designed to give a single-number answer were used for predicting demand for ICU hospital beds, and vastly overestimated the needs, sometimes by a factor of 10 or more. These overestimates are rooted in uncertainty, or unknown variables. Monte Carlo simulations are designed to address these unknown variables.
Henry Yennie, a program manager at the Louisiana Department of Health and Hospitals, used simulations that more accurately predict and align with what hospitals actually experienced. A veteran of disaster planning, Yennie has used Monte Carlo simulations for decades to help keep hospital generators supplied with fuel and create evacuation plans, serious business considering more than half of the state’s hospitals are within 60 miles of the Gulf of Mexico. Planning for rapid evacuation of up to 40 hospitals and thousands of patients using C130 military airplanes and jet ambulances requires estimating many unknown variables.
The simulations provide a foundation for important, high-value decisions where lives are at stake. “I am not a statistician. By training I am a social worker,” Yennie says. “The models have become key pieces of data our team provides to department and state leadership. In the world of emergency preparedness, we have to avoid the danger of falling into the ’give me a number’ mindset. I need a range of probabilities that I can start with and plan downstream from there. The true scope of the challenge is difficult to look at if we are not using probabilistic methods.”
Yennie and his team used probability analytics following Hurricane Katrina in 2005 when more than 13,000 missing persons were reported. The analysis helped forecast and focus recovery teams on missing people who were likely still alive. According to Yennie, the estimates were over 92 percent successful.
Following Hurricane Gustav in 2008, which left hospitals without electricity, in some cases for weeks, a new statewide system to keep hospital generators supplied was developed. “We were getting calls 24 hours a day, including one hospital that said they needed fuel in the next hour or children will die,” Yennie says. The Department’s probabilistic models equipped the state to manage resources and prioritize fuel needs to keep hospitals operational.
Predictive analytics provides insights
The modern Monte Carlo method of computation was created by a Polish-American scientist and mathematician named Stanislaw Ulam working in the late 1940s as part of the Manhattan Project. He developed the method of applying statistical methods to functions without known answers. Advancements in computing have made Monte Carlo simulation a standard approach to many challenges in the hands of professionals without mathematics degrees.
The simulations involve “rolling the dice” thousands of times (or more) and varying conditions to assess risks and produce insight to explore business implications. Monte Carlo simulation provides a more comprehensive view of what may happen and how likely it is to happen than is ever possible with single-point estimates.
Today, Monte Carlo simulation is widely – and increasingly – used across many industry sectors. For example, the banking and financial sector is using simulation to manage risks and improve processes. The healthcare industry is providing better management of resources and improving clinical performance. Retail companies leverage insights to better understand supply chains, consumer buying and effectively manage merchandise. Financial forecasting, risk management, project management, portfolio optimization, risk-informed pricing, inventory and supply chain optimization are a few more examples of use cases. Practitioners include strategic, financial and risk analysts, engineers, project and program managers, line managers, and many other non-statisticians.
In scientific and technical fields, Monte Carlo simulation is used in many processes ranging from computational physics to aerodynamic design and models for weather forecasting. In the field of engineering, uses range from microelectronics to the design of wireless networks, autonomous robotics and artificial intelligence systems. Environmental firms use simulation for cleanup, preservation, and wildlife protection efforts. Monte Carlo simulations play a role in the creation of photo-realistic images of virtual 3D models with applications in video games, architecture, and cinematic special effects. The U.S. Coast Guard even uses Monte Carlo simulations to calculate the probable locations of victims during search and rescue operations.
Turning risk into opportunity relies on using analytics and data, identifying factors driving risk, and simulating many factors to shape strategy. When (not if) black swans appear, it is not the time to simply hunker down. Monte Carlo simulation can fulfill a key role in increasing the quality of decision-making and helping project teams think clearly, act decisively and feel confident. Predictive analytics help organizations navigate uncertainty, save lives, avoid surprises, make better decisions, and create market advantages that unveil new opportunities.
Randy Heffernan, Chief Executive Officer, Palisade Company