David Sacks, a San Francisco venture capitalist who advocated government intervention on behalf of uninsured Silicon Valley Bank (SVB) depositors, has argued that the depositors were not equipped to perform due diligence on the bank. Instead, he suggested that they might be guided by, among other things, the “A” rating from Moody’s Investors Service that SVB carried days before its collapse.

This line of thinking raises a couple of concerns. First an “A” rating is not that high. Moody’s has 21 rating symbols ranging from “Aaa” denoting the safest securities to “C” for the most distressed.

To be more precise, prior to March 8^th, Moody’s maintained an issuer rating of A3 on SVB, which is the seventh highest rating on its scale. On that day, the credit rating agency downgraded SVB one notch to Baa1. Both of these ratings are considered “investment grade” rather than “speculative grade” (or, more colloquially, “junk”), but they do not denote ultrasafe securities.

Before the financial crisis, Moody’s published an “idealized” (or theoretical) expected loss rate table (at page 10 of this document). This table showed how much a fixed income inspector might reasonably expect to lose when holding bonds with various ratings over various time periods. For bonds rated A3, the table shows a one‐​year expected loss of 0.02137% of principal. Over ten years, the expected loss was estimated to be 0.99000%. So, if an investor bought $1 million of SVB bonds, he or she could expect to lose $214 over one year and $9900 over ten years.

By contrast, the expected losses on a “Aaa” security were below 0.0003% and 0.0055% over one year and ten years respectively, implying truly de minimis losses even on very large holdings.

An expected loss represents a mean across various outcomes: depending on circumstances, the realized loss could me zero, lower, or potentially much higher.

So it is not correct to associate an “A” rating with a state of “risklessness”, and, in fact, it is not good financial practice to place working cash needed to make payroll at such a high level of risk.

The second concern is that, by March 8^th, the Moody’s rating was stale. Circumstances at the bank should have triggered a downgrade weeks or even months earlier. The A3 rating had been in place for more than a decade prior to the March 2023 downgrade, despite the bank’s various ups and downs.

And, in the months prior to March 2023, the trend was clearly down. On January 18, Raging Capital Ventures (RCV) tweeted that SVB’s bond portfolio losses had exceeded the value of its Tangible Common Equity meaning that “the bank would be functionally underwater if it were liquidated”.

RCV, the private investment office of William Martin, was shorting SVB stock and thus had a motive for tweeting this negative finding. But Martin used publicly available SVB disclosures that were based on SVB’s position as of September 30, 2022, and were released on November 7, 2022.

Had Moody’s conducted a full analysis of SVB’s September 30, 2022 Form 10‑Q shortly after its release, its analysts may have concluded that ratings downgrades were appropriate.

Moody’s rival S&P, whose ratings could also have been consulted by SVB depositors, did marginally better. Going into the crisis, it rated SVB at BBB, which is the ninth highest rating symbol it applies. This is still an investment grade rating, but still two notches above junk.

Rating agency performance in the SVB saga is reminiscent of their behavior ahead of Enron’s collapse in 2001. They maintained investment grade ratings on the firm after evidence emerged that it was in serious jeopardy and then downgraded just before the company’s failure.

Although Congress has enacted two rounds of credit rating agency reform since 2001, the regulatory changes do not appear to have been sufficient to ensure accurate ratings from SVB or for several distressed Commercial Mortgage‐​Backed Securities as I have discussed elsewhere.

Thus far, rating agencies have received only limited criticism for their performance in this case, as most of the controversy has swirled around the bank’s management and regulators. If attention eventually turns to inadequate ratings, I recommend that policymakers avoid layering on more complex regulations onto the credit rating business.

Instead they should abolish the SEC’s licensing regime for the so‐​called Nationally Recognized Statistical Rating Organizations (NRSROs) as I previously recommended on Cato’s Alt‑M blog. That way, the more incisive credit analysis of a private investor on Twitter would compete on a level playing field with the often‐​outdated assessments from the incumbents. And, perhaps a more competitive credit assessment market can provide high‐​balance depositors with better insight into the financial institutions with which they are banking.