Rethinking Bank Risk Measurement: The Replicating Portfolio Approach and the Pitfalls of Traditional Metrics

This episode delves into the critical challenge of accurately measuring bank risk, highlighting how traditional methods often lag behind emerging risks and fail to capture the true economic exposures of financial institutions. Jana Bacon, an Assistant Professor of Finance at Stanford Graduate School of Business, introduces a novel "replicating portfolio approach" designed to provide a more precise and forward-looking assessment of bank risk. She argues that regulators and academics frequently find themselves reacting to crises rather than proactively identifying new vulnerabilities, partly due to the reliance on regulatory filings that introduce new variables only after risks have materialized, and accounting values that can mask underlying dangers.

The core of Bacon's proposed method involves reducing the complexity of a bank's diverse portfolio—comprising securities, bonds, Treasury bonds, and loans—into a simpler, equivalent credit market portfolio. This replicating portfolio is constructed to mimic the bank's cash flows and its exposures to term, interest rate, and credit risks. A significant emphasis is placed on the role of leverage, which, despite seemingly low asset risk, can amplify losses dramatically, making it a constant concern for banks. The discussion also touches on how rising interest rates can negatively impact the value of bank portfolios, underscoring the dynamic nature of financial risk.

The podcast critically examines the limitations of accounting-based measures, such as Net Interest Margin (NIM), which, despite remaining flat over time, can misleadingly suggest low interest rate or credit risk. Bacon points out that historical crises, like the interest rate spikes of the 1980s and the credit crisis of 2008-2009, occurred despite stable NIMs, revealing a disconnect between accounting figures and economic reality. She highlights studies demonstrating that banks with high leverage and asset risk before a crisis are consistently the ones that suffer the most, suggesting a predictive power in these economic indicators and a consistency in risky behaviors.

Finally, the episode challenges the notion of banks' "specialness," arguing that their cash flows can indeed be replicated by capital market portfolios, implying that standard financial tools are applicable. While acknowledging banks' unique monopoly power over the payment system, Bacon notes the high operating costs associated with this role. The conversation also addresses the tension between micro-causal and aggregate studies, advocating for the continued value of aggregate data in understanding bank risk, even when it diverges from stock market perceptions. She concludes by pointing out how stock market participants often misprice bank risk, sometimes valuing more levered banks despite their eventual higher losses, underscoring the importance of understanding systemic behavioral inefficiencies in financial markets.

"if you were just to look at market value of equity of banks and the book value of equity of things these are two statistics that are entering and true economic models people are not really agreeing on which one should enter those but they diverge completely during the crisis"

"regulators and academics often are lagging behind whereas the new risk that they are taking spotting the risk"

"it's called kind of the replicating portfolio approach a lot of financial management firms do that that's what we teach also our MBA students can aware how to value companies"

"banks have and in a sense I mean they have a very complicated portfolio but we are trying to do what I'm trying to do is kind of both the reduce the dimensionality of this complicated portfolio and to method into something that is again that we can make sense of an honest answer is exposure"

"anybody that works on creditors most banks have a lot of leverage much more than other financial institutions so even if the asset side doesn't look super risky if you lever it up by a lot even though like the loss is like 1% but you lever 10 times and it's like a 10% loss that could hit you"

"it's concerning that kind of risk management of banks is and not in that they don't use an economic measure but in some sense non-fundamental partial revenue measures in order to figure out what their risk exposure is"

"if banks were special we shouldn't be able to replicate their cash flows with our bank capital market portfolios which kind of mimic the risk and the cash flows of a banking sector"

"a lot of stock market shareholders in general seem to have valued banks more the more levered there were before the crisis although then they expose I mean suffered the highest losses"

"I think this will be an important thinking about systemic ways of behavioral inefficiencies that can come from a source from some and behavioral mechanism will be hugely important I think for economics"