In most quarters, the government bonds of credit worthy countries are considered safe and defensive assets. But as with any asset class, except perhaps for cash, there is a degree of risk, and the term defensive is relative and possibly misleading. The global bond market is approximately 2.5 times larger than equity or share markets. And if I suggest over 95% (possibly 100%) of all bonds bought between mid-2020 and now with a maturity of two years and over are currently trading below the purchase price, people might think commitment to a mental institution is the next logical step (Exhibit 1). The graphic is an example using the 10-year maturity. I suggest all maturities over two years mirror the 10’s situation.

Exhibit 1: 10-year government bond yields (%)

Source: Bloomberg The failure of Silicon Valley Bank is a timely wake up call. The causes are several, though with some more important than others. The depositor base was not diversified, with a meaningful bias to the technology sector and related venture (adventure) capital startups, as the bank’s name and environs infers. Most deposits were well above the US$250,000 level insured by the Federal Deposit Insurance Corporation (FDIC). According to FDIC, SVB had US$209bn in assets end December 2022. As the US Federal Reserve (the Fed) tightened monetary policy to combat rising inflationary pressures, the tech sector, whose valuations are rate sensitive, started to struggle. The deposits were fleeting in nature with either no (at call) or short duration (months). They were certainly not sticky. Given the cash appetite of startups and no cash flow in sight, these deposits funded operations. When SVB launched an equity raising, alarm bells rang and triggered a deposit run. The rest is history. The collapse of the bank was over in a blink. Instead of lending the deposits to credit worthy customers on short durations, say one to five years, SVB bought long-dated US Treasuries and mortgage-back securities committing a cardinal sin. The liabilities (deposits) were at call, the assets were long-dated maturities and purchased around the peak of quantitative easing and yield control asset purchasing by the Fed. With bond yields surging due to the outbreak of not transitory but ingrained inflation bond prices slumped. Unrealised losses in the bond portfolio increased meaningfully. The equity raising failed, and bonds were sold to meet withdrawals by depositors. The balance sheet imploded, and the FDIC was appointed receiver. By buying US Treasuries, SVB effectively lent money to Uncle Sam. But it lent for 10–30 years, not 10–30 days, introducing significant duration risk—safe but not defensive. At maturity, the investor receives the face value of the bond while banking the income over the ownership period. But if the bond is purchased on the open market when yields are low, as was the case with SVB, the price paid is well above the face value and at maturity a capital loss can be realised.

Exhibit 2: US 10-year government bonds; price relative to par value (US$)

* The 2023 bond is converging to par value as it approaches maturity, while bonds that are not maturing for several years have fallen in price and are trading below par value. Source: Refinitiv

Auditors found wanting

The ratings agencies went missing in the lead up to the GFC and now it seems auditors have done the same. The Wall Street Journal reports the big four’s KPMG gave Silicon Valley Bank a clean bill of health just 14 days prior to it spiraling into bankruptcy. While KPMG flagged the potential losses in the loan portfolio as “a so-called critical audit matter” the opinion made no mention of unrealised bond losses nor the likely inability to retain the bonds in the circumstance of deposit flight. In a classic comment, University of Michigan business professor Erik Gordon said, “The auditors failed to mention the fire in the basement or the box of dynamite on the first floor, but they did point out the peeling paint on the flower box.” “How could they miss the interest-rate risk?”

Central banks are probably technically insolvent

Interestingly, in the Reserve Bank’s (RBA) recently released April Financial Stability Review, Box A in the section on The Global Financial Environment provides an in-depth commentary on Recent International Bank Failures—Causes, Regulatory Responses and Implications. The three US banks in question are Silvergate Bank, Silicon Valley Bank and Signature Bank. “A run on these banks’ deposit bases, which were concentrated and largely uninsured, was in part due to concerns that large unrealised losses on banks’ asset holdings would impair their capital positions (particularly for SVB). These failures were in part enabled by poor risk-management practices and a less stringent regulatory regime.” In the case of SVB, the US Treasuries and mortgage-backed securities were primarily classified as held to maturity, which under US and international accounting standards, means changes in value are not required to be recognised. Conditions changed quickly, and the held-to-maturity classification evaporated as securities were sold and heavy losses were realised. This wiped-out capital and insolvency quickly followed. The RBA, along with all central banks, is probably technically insolvent. It too has meaningful asset concentration in government bonds. The RBA outlaid $361bn to support the yield target comprising $36bn on Yield Curve Control, $44bn on market function purchases and $281bn on the bond purchase program. Purchases included maturities from July 2022 to April 2033. While the bonds are classified as held-to-maturity assets and the central bank cannot have a run on deposits, the current mark-to-market losses would swallow the RBA’s capital. It is unlikely central bank portfolios are hedged. The Fed’s current capital of US$42bn would also be wiped out if its US$5.3 trillion US Treasuries portfolio were marked to market. That does not include mortgage-backed securities of US$2.6 trillion.

Banking system issues settled for now

While the upheaval in the US banking system has settled and the concerns of a possible contagion ease, there are still issues that require close monitoring. The big banks are OK, but the deposit flight from small to mid-tier banks was meaningful and materially weakens their ability to offer credit facilities. Unlike Australia’s concentrated banking system, the US regional and community banks are a vital piece of the banking jigsaw. They provide credit to small/medium size businesses who are collectively large employers. There will be a knock-on effect, at a time when economic activity is slowing as tighter monetary policy restricts consumer demand. The small to medium size banks provide about 70% of the banking finance to the commercial real estate sector and with valuations already under scrutiny, the quality of loan portfolios will be questioned. For these banks, it will take time for the outgoing deposit tide to ebb. In addition, regulation across all banks is likely to increase, which will lead to greater caution across the entire banking system. An increased focus on the potential dangers of the hold-to-maturity assets on bank balance sheets could also mean a need for additional stand-by capital. Given the increased uncertainty, who knows when these assets may need to be liquidated. Credit contraction could persist for several years, with economic growth likely to be below the long-term trend.

Data not likely to see the Fed pause just yet

The much-awaited US March CPI did not create any waves. It did however reveal the stubbornness of core inflation and given the forecast of slowing GDP for the remainder of 2023 brings the possibility of stagflation to the fore. In detail, the seasonally adjusted headline CPI rose 0.1% in March, below expectations of 0.2% and 5.0% year-on-year (y/y) was down from 6% y/y in February helped by the first decline in food at home prices since November 2020, a sharp fall in rents inflation from 0.8% m/m to 0.5% and weaker energy prices. Further falls are expected in rents. That is the good news. But energy will rebound in April following the rise in oil prices after OPEC+’s announced production cuts from 1 May. Core CPI remains an issue, coming in at a 0.4% increase m/m and 5.6% y/y, up from 5.5% y/y in February snapping five consecutive months of easing.

Exhibit 3: United States core inflation rate (%)

Source: tradingeconomics.com, U.S. Bureau of Labor Statistics Federal Reserve chair Jerome Powell’s favoured core reading, which excludes shelter remains elevated. While there is evidence of deflation in the goods space over the past year, as both supply disruptions and demand eased, goods prices are showing some resistance to further falls. The core goods, ex food and energy, is up 0.2% m/m, to the highest read since August 2022. After a relatively robust March jobs report, with the additions of 236,000 non-farm jobs, the unemployment rate easing from 3.6% in February to 3.5% and the March CPI reading, the data dependent Fed won’t be convinced the battle against inflation is won. Markets are pricing a 70% chance of a 25-basis point rate increase on 3 May. Fed speak after the CPI data had a hawkish tinge. San Francisco Fed’s Mary Daly said, “while the full impact of this policy tightening is still making its way through the system, the strength of the economy and the elevated readings on inflation suggest that there is more work to do.” Richmond Fed’s Thomas Barkin added “there’s still more to do I think to get core inflation back down to where we’d like it to be.”

Back home

Last weeks’ interest rate pause has been quickly reflected in a sharp improvement in consumer sentiment. The Westpac Melbourne Institute Consumer Sentiment Index jumped 9.4% from 78.4 in March to 85.8 in April. This is encouraging given the near yearlong battering consumers have endured since the tightening cycle began in May 2022. However, confidence remains weak with the April index 10.4% below that of April 2022. The weekly ANZ-Roy Morgan Australian Consumer Confidence Rating also improved after the RBA’s pause and was the most positive result following an RBA meeting since May 2022. Not surprisingly, the gain was led by those paying off a mortgage, but overall confidence is still very weak, with a reading anchored below 80 for the sixth consecutive week. The March Labour Force report revealed 53,000 jobs were added in the month seasonally adjusted, well over twice expectations around 20,000. Unemployment remained steady at 3.5%. Full-time jobs dominated up 72,200, while part-time fell 19,200. The participation rate remained steady at 66.7%. Monthly hours worked eased 0.2%. Clearly the jobs market remains tight, but little different from the RBA’s earlier forecasts in February’s Statement of Monetary Policy. The March quarter CPI to be released on 26 April will be the key to what the RBA does at the next Monetary Policy meeting on 2 May. It will have pricing data for all categories rather than the partial measures in the Monthly CPI Indicators. The headline reading is likely to fall sharply from the December quarter’s 7.8%, perhaps below 7% y/y. However, the more important trimmed mean is likely to remain sticky around December’s 6.9%. Elsewhere, the International Monetary Fund showed how far behind the curve it is by cutting Australia’s 2023 GDP growth forecast from 1.9% to 1.6% and 2024’s to 1.7%. The RBA’s forecast for both 2023 and 2024 has been at 1.5% for two months and most economists are near that mark. Global growth is forecast to fall from 3.4% in 2022 to 2.8% in 2023, stabilising at 3% for the following five years. Share markets continue to ignore the subdued economic and corporate earnings outlook, welcoming the end of the tightening cycle. The possibility of emerging stagflation and the prospect of an extended period of tighter credit conditions are apparently not an issue.