The Fed Runs a Fool’s Errand on Bank Capital Rules

There are many ways a bank can lose money. Top of the list are credit and financial market losses but beneath, there’s a whole taxonomy of issues that industry insiders brand “operational risk.” These result from bad internal processes, people and systems or from external events.

Banks aren’t alone here, but because they look after the money, they are particularly exposed. The largest such loss so far this year happened at Saigon Commercial Joint Stock Bank in Vietnam, where a majority shareholder embezzled $12.5 billion, aided by board members and executives. Elsewhere, Canada’s Toronto-Dominion Bank paid $1.2 billion to resolve claims it enabled a Ponzi scheme perpetrated by Allen Stanford over 10 years ago, and Canadian Imperial Bank of Commerce paid $848 million following a contract dispute over loan repayments.

So prolific are these mishaps that Swiss research firm ORX keeps a database of them. Since it started compiling records 20 years ago, it has registered over a million such events across 105 banks, with total losses amounting to €592 billion ($640 billion). To date, 2023 has been relatively mild, although settlements often ramp up towards year end. In December last year, Wells Fargo & Co. and Danske Bank A/S each incurred penalties in excess of $1 billion following instances of consumer abuse and money laundering, respectively.

To mitigate against the risk, regulators insist banks maintain a capital cushion – one they are currently recalibrating. The idea emerged following a series of trading scandals in the 1990s, the biggest of which led to the downfall of Barings Bank. In 1998, the Basel Committee on Banking Supervision introduced the concept of operational risk while working on a new framework (Basel II) and proposed developing an explicit capital charge to cover it.

The problem is that it’s extremely difficult to measure, much less anticipate. Unlike credit and market risk, it’s hard to distill governance and control practices into a quantitative framework. Regulators outlined a crude model tied to gross income but allowed banks to fashion their own models as an alternative. JPMorgan Chase & Co. deployed a fancy tool called a Loss Distribution Approach, which simulates the frequency and severity of future operational risk loss projections based on historical data to estimate an aggregate one-year loss at a 99.9% confidence level. At the end of September, the model directed the group to allocate just over a quarter of its regulatory capital to operational risk, in addition to the 70% being allocated to credit risk and 5% to market risk.

Yet policymakers harbor reservations over internal models and new rules propose replacing them. “These models can present substantial uncertainty and volatility,” said Federal Reserve Vice Chair for Supervision Michael Barr in October. “In the agencies' proposal, the operational risk capital requirements would be standardized rather than modeled.”