Article by Louis-Vincent Gave, founder & CEO of Gavekal
So far, 2023 has been an unusual year for investors in the US fixed income market. In the last four months, they have seen: Volatility at the front end of the US treasury curve of a magnitude unseen since the 2008 mortgage crisis, with implied volatility going through the roof.
An unprecedented widening of the spread between four-week and 13-week T-bill yields, largely on fears that with the US government likely to hit its debt ceiling in June or July, the risk of a technical default, while small, is not zero.
Poor growth numbers, with the OECD leading indicators for the US below par for the last 11 months, bank loans collapsing, and M2 contracting. Yet there has been no meaningful widening of corporate spreads.
The failure of Silicon Valley Bank and Signature Bank, and a rout in the shares of US regional banks, yet no rally in long-dated US treasuries.
In short, there has been massive turbulence at the short end of the treasury curve, but everywhere else—the long end of the curve, corporate bonds, mortgage bonds — there has been a surprising calm. How long will this last?
Surging US treasury supply
The second half of 2023 is set to see a surge in the supply of new US treasuries. Firstly, this is because the US Treasury’s own “checking account” has almost been run dry. The US government is preparing for a bruising battle on Capitol Hill over its debt ceiling with very little spare cash in the bank.
Secondly, it is because despite near record low unemployment, the US budget deficit is projected to remain between 6% and 7% of GDP for the next few years—assuming there is no recession. And it could be worse; in the first half of the current fiscal year, the US deficit came in at a much bigger than expected US$1.1trn.
This means US$2trn, and probably more, in additional debt issuance per year for the next five years or so. And again this assumes no recession. If the US were to experience a recession, then the annual increase in debt for that year would likely be closer to US$3trn. So, the supply of debt is unlikely to run dry. And with the US government’s outstanding debt increasing by at least US$2trn a year, it’s safe to say there will be enough debt to go around for anyone who wants to buy some. So, who will buy this debt?
Foreign appetites waning
For the last month, the financial buzzword has been dedollarization. Admittedly, dedollarization can mean very different things to different people. But it’s safe to say that foreign central banks have not exactly been falling over each other to accumulate more US treasuries lately. Despite current account deficits of roughly US$200bn a quarter, little of the money the US consumer has pumped abroad in the last decade seems to be coming back into US treasuries held at the Fed for foreign central banks.
This may be one of the reasons that US treasuries have delivered nominal returns of zero (and deeply negative real returns) over the last eight years. Regardless how one feels about the whole dedollarization zeitgeist, the fact remains that the countries running large current account surpluses today — China, Saudi Arabia, Russia — are unlikely to increase their demand for US treasuries any time soon.
This leaves two big potential buyers for the coming supply surge in US treasuries. The first is the US banking sector. The second is the Fed.
Will US commercial banks load up on US treasuries?
If commercial banks are to step up and buy the additional treasuries, what money will they use? The trend in deposits over recent months has been for money to flow out of regional banks and into very large banks (for depositors worried about the safety of their deposits) or into money market funds (for depositors seeking higher yields). Arguably, this is a “pass the parcel” zero-sum game. As money leaves regional banks, that money ends up elsewhere. Nonetheless, this makes for an environment in which monetary aggregates are now collapsing.
At the same time, bank credit has grown by 9% over the last 12 months. This credit expansion may be a temporary phenomenon, with worried corporates tapping credit lines while they are still available. Or perhaps they have simply been using bank credit to fund the rising cost of inventories (inflation impacts everyone) and unexpected increases in payroll costs.
With this in mind, it seems unlikely that the demand from US commercial banks will surge over the coming months, especially at the long end of the curve. Why would commercial banks rush to buy long-dated treasuries yielding from 3.5% for the 10-year to 3.75% for the 30-year when their own cost of funding is now likely to be higher than that? Without a steepening of the yield curve, banks are unlikely to rush to buy long-dated US treasuries.
Does this leave the Fed as the only meaningful buyer?
Fed policymakers spent much of the last year telegraphing to the market, and to th US government, that the days of Fed balance sheet expansion were over. Instead, the Fed would look to “normalize” its balance sheet and withdraw the excess liquidity it injected during the Covid years (and during the years of quantitative easing before that. But this turned out to be easier said than done. As the implosion of Silicon Valley Bank and Signature Bank highlighted, large parts of the US economy have become dependent on easy Fed liquidity. Removing this excess liquidity suddenly might threaten financial stability.
This was the lesson of fall 2022 for the Bank of England. And it increasingly seems as if it is the lesson of spring 2023 for the Fed. For all the talk of quantitative tightening, as soon as SVB ran into trouble, the Fed immediately grew its balance sheet by US$400bn. To put this number in perspective, it is more than half the size of the initial US$700bn troubled asset relief program authorized by Congress in October 2008 in response to the Lehman bankruptcy and AIG debacle.
It is hard not to conclude from this that US quantitative tightening is something akin to a financial Keyser Söze: “a spook story that criminals tell their kids at night,” but no one has ever seen.
This brings us to the real problem with the US treasury market. If we add the US$751bn of US defense spending in the fiscal year that ended on September 30, 2022 to the federal government’s US$4.1trn of mandatory spending on Medicare, Medicaid, social security and the like, then the US government’s “essential” spending already outweighs the entirety of its tax revenues (and eats up 99% of its total revenues).
Meanwhile, because so much of the US government’s debt has been issued at the short end, its interest costs are now starting to rise parabolically, as debt issued in recent years at 1% is replaced by debt yielding 4% or even 5%.
Incidentally, this concentration of US government debt at the short end may help to explain the deep inversion in the US yield curve. Regulation forces insurers and pension funds to buy long-dated bonds even if supply is tight, pushing up their price. Meanwhile, most of the issuance is at the short end, meaning little pricing power in the shorter maturities.
Looking at the record 168bp spread between the three-month and 10-year US treasury yields alongside the rapidly surging interest rates costs on US government debt, it stands to reason that the US Treasury might want to start replacing shorter-maturity debt with longer-dated bonds.
Beyond the short term debt-management implications of higher interest rates and the inverted yield curve, there is a bigger question. Can the US government continue to increase its debt by US$2trn or more every year for the next five years, while the Federal Reserve simultaneously shrinks its balance sheet, without these two opposing forces creating massive tensions in the US government debt market?
It seems unlikely. This leaves us with three possible scenarios:
1) The US government brings its debt expansion under control as the Fed shrinks its balance sheet.
This is a very long odds scenario. A reduction in US government spending would imply a dramatic cut in the defense budget (unlikely given the cold-war backdrop), massive reform of social spending (unlikely given the demography of US voters), or a collapse in interest rates (unlikely given inflation). But if by a miracle this were to occur, perhaps because of the development of new energy sources or because of productivity gains driven by artificial intelligence, US treasuries and the US dollar would both deliver solid returns.
2) The US government continues to spend money like a sailor on shore leave, while the Fed resumes its balance sheet contraction.
In this scenario, yields would rise meaningfully, and the US dollar would likely appreciate as the US government started to crowd out most other global capital expenditure. This would be very bearish for global growth.
3) The US government continues to spend money, with this money provided by the Fed, which eventually ends up embracing yield curve control.
In this scenario, bond yields may not rise much, but the US dollar will undergo a structural decline. Inflation in the US will remain high, and will be exacerbated by US dollars sitting outside of the US flowing back to the domestic market. In this scenario, US asset prices might hold up in US dollar terms, supported by the US dollar inflows. In foreign currency terms, their performance will be less solid.