COVID-19 Lessons Learned: Nonbinary Risk and the Power of Organization

The rhetoric around the coronavirus pandemic has been intense, creating tension, confusion, and even violence over the right way to handle this crisis. Some have argued that locking down the economy is the only solution, while others contend that doing so will destroy livelihoods.

But if the pandemic experience teaches us anything, it is the deep flaw in a type of binary thinking that not only sows division and acrimony but leads us to make wrong decisions. Risk management is not an “either-or” concept. If I never hear the phrases “risk on” and “risk off” again, I'll be happy.

Taking extreme positions is antithetical to the whole concept of risk management. Whether we are optimizing an investment decision, making a credit decision or managing a pandemic, the answer rarely - if ever - is a single, "all-in" solution. Even poker players use a mixture of strategies.

Decision theory may work better in the extremes, as the math is often much easier there. But reality lives in the spectrum.

Risk management - and decision-making in general - boils down to constrained optimization under uncertainty, accounting for the costs of both action and inaction.

High-Tech vs. OrgTech

In many respects, the COVID-19 crisis is no different from any other risk management problem. There are both technological and organizational strategies for managing risk, whether it's a financial risk or a public health risk.

On the technological side, we may try to invent our way out of the coronavirus crisis by researching new drugs and vaccines. A “miracle cure” would undoubtedly be of great benefit - but the probability of success is far from guaranteed. We may get lucky and discover a pharmacological solution in short order, but past experience suggests this is more likely to be quarters or years away, rather than weeks.

In terms of an organizational solution, distancing, quarantining, testing and contact tracing are some of the oldest methods for dealing with plagues and pandemics, dating back to the Middle Ages. Countries such as New Zealand, Taiwan, South Korea and China have proven that pandemics are manageable, even without a cure or a vaccine. Of course, this effort isn't costless, and recent outbreaks have demonstrated the need for constant vigilance to avoid and address future outbreaks quickly.

The best approach to managing all risks then is to leverage technological as well as organizational solutions. Testing and contact tracing can help keep caseloads down in the short term, while medical research can hopefully deliver a long-term solution. The organizational solution may not be as “sexy” as the high-tech one, but it can save many lives at a fraction of the cost.

Risk management falls apart when we fail to coordinate our efforts, or we become so infatuated with one solution that we lose sight of the bigger picture and our objectivity.

Drawing the analogy to financial risk management, machine learning and artificial intelligence are gaining traction today in financial institutions. But more predictive models alone are insufficient for managing risk. Organizational structure is equally important for leveraging the power of these tools.

People, data, models and systems are all essential components of an effective risk management strategy. Better algorithms are always desirable, but we shouldn't ignore or discount the high returns from investing in organization and good governance. A strong, flexible organization using a simple model will be far more effective than a weak organization with a superior model.

Managing Your Overall Risk Profile: And is greater than Or

Individual investors are notorious for jumping in and out of the market at precisely the wrong time. They may become greedy when stock prices are rising and sell everything when prices are falling. Their behavior is so predictable that professional investors use them as a gauge to manage their own contrarian decisions. Rather than jumping in or out of markets completely, the professionals adjust their positions around the edges with hedges to minimize their risk exposures while still generating positive returns.

Binary solutions such as “sell everything” make for catchy slogans, but they oversimplify. Whether it's credit, insurance or any other risk, the question rarely is “should we be in or out of the market?“ Rather, we should ask: “How do we manage our overall risk profile using the variety of technological and organizational tools at our disposal?”

Given the low-yield environment, institutions today may seek to increase their returns by lending to riskier customers. But this doesn't imply going “all-in” on, say, subprime borrowers or companies.

Increased exposure needs to be complemented with mitigation strategies, such as stronger underwriting, servicing, pricing or re-insurance. At a minimum, firms should invest in data and tools - as well as people and processes - to better measure, monitor, forecast and hedge the risk over time and under a variety of economic scenarios.

Embrace the Risk Spectrum

All of our risk management problems and solutions exist on a spectrum. My own field of economics has had to deal with its own flaws in binary thinking in recent years. A host of “universal truths” have been invalidated, such as the idea that low unemployment necessarily causes inflation, or that budget deficits necessarily lead to higher interest rates, or that tax cuts pay for themselves. As discomforting as complexity may be, the fact is that truth always lies in the middle of the extremes.

There is a reason why risk professionals are called risk managers, as opposed to “risk avoiders” or “risk takers.” Their job is to guide others away from binary, extremist thinking. “Risk on” and “risk off” may make for snappy headlines, but it's terrible advice. A better approach is “risk always, but always in moderation.”

Cristian deRitis is the Deputy Chief Economist at Moody's Analytics. As the head of model research and development, he specializes in the analysis of current and future economic conditions, consumer credit markets and housing. Before joining Moody's Analytics, he worked for Fannie Mae. In addition to his published research, Cristian is named on two U.S. patents for credit modeling techniques. He can be reached at cristian.deritis@moodys.com.