You're conflating two different types of portfolio leverage. What you're describing is a pledged securities line of credit. These take the portfolio assets as collateral, with the custodian bank providing a "non-purpose loan" from its own balance sheet that can be used for most purposes, except buying additional securities. Most cited investment leverage would be using the aggregate margin debit balances, which is different. This is the custodian lending cash to the investor in exchange for allowing the custodian to lend out the securities to those wishing to borrow or short them. This leverage is intended to be used to buy additional securities on margin, but it can also be withdrawn and used for other purposes. To give you historical context, there is roughly $1.5 trillion in margin debt today, near an all-time high.  As a percentage of GDP, it’s above what we saw in both 2000 and 2021.  So, regardless of whether it was used to buy additional securities or other things (which is more difficult to quantify), the more leverage in the system, the more speculation being taken and the more embedded risk. When you hear of broad market "dislocation" as we saw in the GFC and during the pandemic, it's almost always massive deleveraging as securities get indiscriminately sold to pay off excessive margin balances.

Since the financial crisis of 2008, any use of the words “off balance sheet” conjures up the feeling of something sinister, hidden, nefarious, or significantly concerning. Indeed, the SPVs and other vehicles where borrowing had been placed in the mortgage market of 2003-2007 proved to be a debacle that intensified actual leverage and exposure when everything hit the fan in 2008. That said, there is some need for clarification here.

First of all, the use of joint ventures, special purpose vehicles, or other structures does not mean that the borrowings are “off balance sheet.” In some cases they are, and in others they are not, but what we are mostly talking about is accounting treatment that is more complicated than headline summaries allow. What we do know to be true is that the contractual obligations many of these firms have taken on are simply outrageous, and the reported leverage does not always capture the full thrust of exposure. That is true enough. Future payments of lease obligations are far above the balance sheet debt that we see. But there is a perfectly good reason that long-term future-year lease payments are not captured in the present debt obligations on the balance sheet, and contingent exposures should be treated differently.

The main concern I have is for proper identification of who holds the risk. If a creditworthy borrower has an obligation but is assuming contractual payments from a weaker counter-party, there exists domino-like fragility in the model that is not visible in the headlines and summary ratios.

My summary is this: The vocabulary can be used to make something sound worse than it is and imply “financial engineering” when it is really standard accounting rules. And yet, each deal requires an understanding of the contracts involved, the counterparties and the economic modeling. An overbuild of AI infrastructure and poor returns on invested capital will not end well for anyone involved, regardless of accounting treatment.

When I mentioned CCC bond spreads blowing out over 300 basis points, that was about bond prices and yields, which is different from direct financial stress in the underlying companies. Now, many times those two things can become tethered to one another, and it's somewhat "chicken or the egg" as to the order. High-yield (aka junk) bond spreads at the lowest credit tier of CCC started in September near 10% and are now over 13%. Add the risk-free Treasury rate to the spread, and it takes yields of nearly 18% to clear that market. Will having to refinance debt borrowed in 2021 at roughly half that rate cause real financial stress for the underlying businesses? For some, you bet. Since there remains a large maturity wall with roughly 30% of the $150 billion in CCC debt maturing over the next two years, I imagine we'll find out soon enough. As for those cash-rich balance sheets you mentioned coming in to buy these businesses now (which would include the debt)? Unlikely before that maturity wall hits and actual distress occurs. Once it does, yes. That's the natural order of the business cycle and the definition of creative destruction.

“Public debt” is the amount the federal government owes to “someone else.” That number is where the 100% debt-to-GDP ratio comes from. “Total debt” includes intergovernmental transfers, and that is where the 120% or so number comes from. The reason “public debt” is used is to avoid dishonest weights and measures. Including intergovernmental debt in the numerator of debt-to-GDP is problematic because

a. It is NOT the historical measure used, so it is misleading,
b. It is NOT the measure other countries use, so it creates an apples-to-oranges comparison, and
c. The issue in 120% debt-to-GDP counting intergovernmental debt is NOT the same thing as social security obligations; that number is a LOT more than $7 trillion … It is notes to and from the left pocket and right pocket; there is no maturity date, coupon, or amortization. They could write a note back from one pocket to the other and level it out. It is all shenanigans.

The Social Security obligations do have to be paid, but they are unfunded liabilities that count on a different part of the balance sheet and have to be weighed against revenue, which is not a set asset. The public debt and GDP are KNOWN numbers, and they are brutal. I think the public debt ratio is more telling, more contextual, and bad enough as is without the hyperbolic juice of the other number.