Silicon Valley Bank (SVB) failed because it invested too much in long-term bonds that lost value when interest rates went up. That’s what our accounting rules encourage banks to do. As a consequence, U.S. banks, including some of America’s leading banks, are estimated to have over $600 billion of unrecognized losses on the “underwater” securities on their books.
The accounting rules that encourage risk-taking permit banks to show values for bonds on their balance sheets that are not the real values. Instead, they are the prices that the banks paid for the bonds (called “historical cost”), even if the bonds have decreased in value, as they always do when interest rates go up.
Carrying securities at historical cost encourages banks to take risks. Management bonuses usually are based on reported earnings or, in the case of SVB, return on equity (ROE), which is earnings divided by equity capital. Reported earnings include the interest paid on securities that the bank owns. Because the interest rate yield curve usually is upward-sloping, longer-term securities usually pay more interest than shorter-term securities. Therefore, in the short run, management gets bigger bonuses by buying longer-term securities. By buying long-term securities that paid an average of 1.5% instead of safe one-year Treasury Bills, the bank more than doubled SVB’s 2022 income and its ROE. With this system, the extra benefits to CEO Becker and CFO Beck were in the millions.
However, the longer maturities exposed SVB to losses when interest rates went up–except that the accounting rules allowed the bank not to count the losses in their reported income. Nor was the bank required under accounting rules to report the unrealized losses on their securities as the Fed continued to drive interest rates up in its campaign to slay inflation. By Mar. 31, 2022, SVB already had about $7 billion in market value loss. If it sold any of its underwater securities to shorten the average maturity of its holdings and thereby to reduce its downside if rates continued to go up, it would have had to recognize that $7 billion loss. And if it recognized that loss, it would have lost almost half of its equity capital of $16 billion and would have been in danger of failing. Instead, it chose to roll the dice.
By the end of the year, as rates continued to rise, SVB’s market value loss had more than doubled. And the jig truly was up. In banking jargon, balance sheet flexibility is critical, but historical cost accounting for securities inhibits balance sheet flexibility.
What is the excuse for reporting bonds at values that are not real? Bankers advance a few reasons. First, they say, the interest rate risks of assets (like bonds) have to be managed in relation to the interest rate risks of the other side of the balance sheet–liabilities, which are primarily deposits. Interest rates going up may cause assets to decline in value, they say, but interest rates going up may, on the other hand, cause the effective price of deposits to go down. If managed correctly, bankers say, the two should offset each other.
Second, many banks assert that if they are going to hold the bonds to maturity and they are going to be paid at maturity (as government or government-guaranteed bonds will be) then it is misleading–and causes unnecessary gyrations in reported earnings–to show the value going up and down with interest rates.
Using these reasons–and political influence–banks have successfully opposed “fair value accounting” for securities for a long time. Howver, the events of recent weeks have shown that these reasons are flawed.
Deposit rates rarely stay below market rates for very long. And non-interest-bearing accounts, which do have value, and long-term deposits, which also can have value, now account for a small percentage of most banks’ deposits. Even the best liability management simply does not make up for embedded losses on assets.
Moreover, holding long-term bonds to maturity when they earn less than market rates is not without cost. Forgone earnings are irreversibly lost.
Historical cost accounting for banks must change. Large interest rate risks not only threaten individual banks’ solvency, but also, through the possibility of triggering wider runs, the viability of large parts of the banking system.
Moreover, historical cost accounting for securities holdings is inconsistent with transparency, which is central to effective bank supervision. The lack of transparency leads to surprises, surprises lead to lack of confidence, and a banking system loses value when the public lacks confidence in its soundness.
We could wait for the official accounting rules body, the Financial Standards Accounting Board, to change to rules. That should happen–but the process will take time.
There is a better way. The SVB case shows that, in order to maintain the public’s confidence, the federal banking authorities should take charge of the way banks account for bonds and other securities and not leave it to the accountants. The Securities and Exchange Commission (SEC) already has mark-to-market rules for securities brokerages and mutual funds. Under its Securities Exchange Act Rule 15c3-1, the SEC maintains detailed mark-to-market rules for measuring brokerage firm capital. And open-end mutual funds reprice at fair value daily in accordance with the SEC’s Investment Company Act Rule 2a-4.
Federal bank regulators should start by phasing out the crazy distinctions among different categories of securities, which in economic terms, are not different. More specifically, under current accounting rules, banks can assign securities they own to any of three categories: Trading, Available-for-Sale, or Hold-to-Maturity. Trading securities are marked to market and any changes in their value are included in earnings or losses. Available-for-Sale securities are not marked to market–but their changes in value are reported in a strange account called “other comprehensive income”. And for some regulatory capital purposes, those changes are not counted in computing capital. Hold-to-Maturity securities are carried at historical cost and their value changes are not counted in earnings. Their market values are reported–but usually only in a note to the financial statements where they are difficult to find. In a large bank with complex financial disclosures, you have to be looking for it and know where to find it. (Try looking for it in the JPMorgan annual report. We’ll give you a hint: Try Note 10 at p. 218.).
This system was a compromise–and has little logic to it. It is now time for bank regulators to require that all securities be marked to market on a regular basis and reflected in income and, therefore, in the capital account, without resorting to clumsy special exceptions. With the suggested treatment, a bank’s balance sheet will more closely reflect financial reality and banks will be less likely to take excessive interest rate risk.
The digital world does not wait for historical accounting to catch up with reality. If the facts (or non-factual rumors) become known to one person, they quickly become known to many. Secrecy is the friend of rumor; transparency is its enemy.
The debate over the allowed levels of credit risk and interest rate risk in relation to the capital that banks are required to maintain will go on as long as there are banks and bank supervisors. But without good measurement and public reporting systems, whatever the limits are, they will be relatively ineffective in deterring unnecessarily risky conduct.
Robert Litan is a non-resident senior fellow at The Brookings Institution and a shareholder at Berger Montague law firm. Martin Lowy is the author of InStaBiLiTy: Booms, Busts, and the Fragility of Banks, and What To Do about It. Lawrence J. White is the Robert Kavesh Professor of Economics at Stern School of Business, NYU.
The opinions expressed in Fortune.com commentary pieces are solely the views of their authors and do not necessarily reflect the opinions and beliefs of Fortune.
