Opinion: How the pandemic is changing views on financial risk

Historian and author of Whom Fortune Favours: The Bank of Montreal and the Rise of North American Finance

The great pandemic of 2020 has administered a series of brutal, real-time tests of how we understand and manage risk. How do you understand or manage something that has never happened before? In finance, the pandemic has demonstrated the value of, and the need for, human judgment in an increasingly algorithm-driven sector.

In fact, the pandemic shock could and should lead to important changes in how we handle risk in finance.

Canadian finance has a very specific risk profile. The evolution of Canadian finance resulted in a corporate and regulatory architecture that was highly effective at dealing with systemic risks, especially against mortgage and investment exposure. Canadian banks favoured stability, conservatism and large reserves. In the 20th century, national institutions like the Bank of Canada and supranational agreements, such as Basel Accords on liquidity and capital requirements, were put into place to co-ordinate risk within and beyond national borders.

Up until the early 20th century, bankers relied on personal and professional networks to assess the three Cs of banking credit – character, capacity and capital of the borrower. In the postwar decades, risk management became more systematic and eventually model-driven.

But the real transformations came with the advent of Big Tech and especially Big Data. Cloud computing, major advancements in processing power and speed and falling costs all drove technology adoption in the last decade. The industry became increasingly dependent on both machine-assisted and in many cases fully automated risk adjudication as financial institutions expanded their reach.

That’s what risk management systems roughly looked like on the eve of the pandemic of 2020.

But the systems are only half the story. The other half is all about the way we think about risk.

In the broader context, scholars of risk point to two major waves that changed the way we think about risk in the North Atlantic world. The Chernobyl nuclear incident in 1986 caused a complete reassessment of the need to manage risk associated with human-made environmental disasters. The Great Financial Crisis (GFC) of 2008 was another watershed as it challenged the deep relationship of trust between individuals and financial systems.

The pandemic of 2020 has been a big enough and deep enough shock to the global financial system that it could well herald a “third wave” of change in the way we think about risk. Thus far, the pandemic has outstripped the GFC in its impact. The events of 2020 have also shown that current risk management tools failed to acknowledge the true range of the possible. In spite of all our prior experience, our expensive technology and our risk models, nobody saw this one coming.

So how can financial institutions possibly prepare for the next crisis?

First, human judgment must reclaim a more prominent role. Without it, we are courting new and bigger dangers. Big Data is doing remarkable things for risk management, but if the data inputs are too narrow or too broad, conclusions are meaningless.

Second, applying a long-run lens to risk is vital. Executives understandably focus on the day, the week, the quarter, but a systematic effort to incorporate context is crucial to understanding and managing risk. It’s a simple idea that is not so simple to implement.

Third, a multifaceted assessment of risk in the post-pandemic world must address a perceived wider problem in corporate decision-making: incorporating more complex and even contradictory views in the decision engine. As Philip Tetlock, a professor at the University of Pennsylvania and co-author of Superforecasting: The Art and Science of Prediction, has observed, decision-making systems need to privilege and protect independent judgments, uncontaminated by conformity pressure and buoyed by free critique. Hence, diversity of thought, as well as diversity of decision-makers, will be an effective way to drive performance compared with machine output only.

Fourth, understand where your decision-makers are coming from. Their personal history matters as much as your institutional history. A recent paper in The National Bureau of Economic Research found “that managers who enter the labour market during recessions exhibit a strong proclivity to reduce their firm’s systematic risk.” The research implies that controlling for those biases is important.

Finally, the pandemic has pointed to the need for a broader and more systematic use of the “reverse stress test.” Risk managers need to understand the key risk parameters and reverse engineer scenarios that deliver aggregate losses in excess of the maximum tolerable loss. Machines are unequivocally better equipped for this task. Human judgment should then be deployed to determine how probable the scenario is and whether the risk of occurrence is acceptable. The other lesson is that the second-best defence against the unknown is the ability to respond quickly. That way, rather than run your business based on the worst case which might only happen once every 10-15 years, you can run your business with only a modest nod to the possibility of the unknown because you can react to it quickly when it happens.

The pandemic of 2020 has appeared in many guises: the destroyer (of lives and livelihoods), or the divider (generating tension, highlighting inequality). When it comes to risk and finance, the pandemic has come in the guise of the Great Revealer by exposing the system’s decision-making overreliance on Big Data. It has also underscored the urgency of reinforcing the organizational behaviours that support decision-making in a complex and transformational era.

In a nutshell, the pandemic points to four key insights: 1) acknowledging strong capital and liquidity as the first and strongest defence; 2) reinforcing financial risk management systems appropriately with sophisticated human judgment; 3) avoiding the hubris of technological infallibility; and 4) ultimately finding a new equilibrium between harnessing the awesome power of algorithms and artificial intelligence, and the irreplaceable, inimitable power of human judgment.