Nobody Rings a Bell at the Top

Every week I check the same four numbers before I read a single headline. High yield spreads. Investment grade spreads. CCC spreads. Financial conditions. That habit has saved me more money than every forecast I have ever ignored, and I have ignored a lot of them.

The scoreboard this week still reads Nirvana. The ICE BofA U.S. High Yield Index option adjusted spread closed Sept. 3 at 265 basis points, up a whisker from 263 the prior week. The ICE BofA AA U.S. Corporate Index spread moved to 61 basis points from 59. The Chicago Fed National Financial Conditions Index sits at minus 0.558, essentially unchanged and still comfortably in loose territory.

My Caution threshold sits near 350 basis points on broad high yield, and the Crisis conversation does not really start until somewhere around 600. At 265, we are nowhere close. Bondholders are a simpler crowd than stock investors. They do not spend months debating artificial intelligence, elections or whatever story Wall Street is selling this week. They ask one question: are we getting our money back? Right now the answer is still yes.

The one number worth real attention is CCC credit. The ICE BofA CCC and Lower U.S. High Yield Index spread widened from 1,031 basis points on Aug. 27 to 1,051 on Sept. 3, part of a steady climb through the week. That is a 20 basis point move in the segment where trouble is always supposed to show up first.

I read this as credit discrimination, not credit contagion. Weak borrowers with too much leverage and thin cash flow are paying up while everyone else refinances on favorable terms. That is a market doing its job, not a dangerous one. A market where every borrower gets cheap money regardless of balance sheet quality would worry me a lot more, and that is not the market we are in.

For community banks, this remains a favorable backdrop, provided you still do the underwriting yourself. Commercial real estate needs to be checked market by market and loan by loan. Construction loans, office exposure and highly leveraged borrowers still deserve scrutiny even in Nirvana. For business development companies and private credit funds, the message is sharper. Underwriting matters more than ever. I want to know how much leverage the portfolio companies carry, where the loans sit in the capital structure, which loans are approaching maturity, and whether nonaccruals are creeping higher. Most of all I want managers who are willing to say no when a tight market tempts everyone else to say yes.

A research report crossed my desk this month built around Apollo economist Torsten Slok's framework for the next two years of interest rates, and it deserves a place in this letter because it changes how I am positioning the income side of the portfolio.

The core argument is not new to readers of this letter. Every major technology in history was inflationary to build and deflationary to use. Railroads, electricity, the telegraph, containerization, computers and the internet all demanded years of heavy spending on labor, materials and financing before the payoff showed up as lower costs across the economy. Artificial intelligence appears to be following the same script, with one difference. It reduces the cost of thinking rather than the cost of lifting, which means it can reach further into the service economy than any prior wave of technology.

Slok lays out two paths. Under the first, AI succeeds: adoption broadens, productivity rises, unit costs fall, inflation cools toward the Federal Reserve's target, and the central bank gets room to ease. Under the second, AI disappoints: technology capital spending falls, richly valued shares reprice, growth weakens, investors run to Treasuries, and the Federal Reserve cuts to offset a deteriorating economy. Both paths can produce lower long term Treasury yields, but they leave very different outcomes for everything else in a portfolio.

The report's base case is productive disinflation, with core inflation drifting toward the 2% to 2.5% range and the 10-year Treasury yield eventually working its way from roughly 4.75% down toward the 3.5% to 4% area over the next two years. A slower adoption path would produce a smaller decline, closer to 4% to 4.5%. The biggest risk to the whole thesis is not AI failing. It is persistent fiscal inflation, where infrastructure spending stays heavy, productivity gains disappoint, deficits keep growing and investors demand a bigger premium for holding long dated government paper.

I want to talk directly to the readers holding business development companies, private credit funds and other alternative income vehicles, because I think this two path debate treats you better than either headline outcome lets on. Most investors hear AI succeeds or AI disappoints and assume one path is good for them and one is bad. For a disciplined alternative lender, both paths pay. They just pay on different schedules.

Take the productive disinflation path first. Falling short rates trim net investment income in the near term, since most BDC portfolios float off SOFR and reset lower before funding costs catch up. That is the visible cost, and it gets all the attention on earnings day. The real payoff sits underneath the coupon. Middle market borrowers refinance more easily, covenant pressure eases, payment in kind interest fades and nonaccruals decline. A calmer loan book protects net asset value, and net asset value is what actually carries a dividend through a full cycle, not one quarter's yield. Banks have spent a decade retreating from middle market lending and handing that business to direct lenders. A friendlier rate backdrop lets that market share keep compounding without the credit losses that usually show up late in a lending boom. Good hitters do not panic when the count runs against them for one pitch. They know the at bat is not over.

The harder path is the one where AI disappoints and growth slows, and it is the scenario a well capitalized private credit investor should actually want, provided the capital is uncommitted and the underwriting discipline stays intact. When banks pull back and public credit markets seize up, direct lenders write the loans nobody else will, at wider spreads, tighter covenants and lower leverage multiples than they could ever get in a calm market. That was true in 2009. It was true in 2020. It will be true the next time fear replaces greed as the dominant emotion in credit. A recession that widens spreads without producing indiscriminate defaults, and my own credit dashboard argues that widening would show up in broad high yield well before it reached the loan book of a careful direct lender, is close to the best entry point disciplined alternative income investors ever get.

The honest bottom line for this corner of the portfolio is better than most investors realize. This is not a coin flip between a good outcome and a bad one. It is a choice between two good outcomes arriving on two different timelines, one built on calm and one built on opportunity. Keep the underwriting standard high either way, keep some capital uncommitted for the second path, and let the AI debate resolve itself without you. A good hitter does not need to guess fastball or breaking ball before the pitch leaves the hand. He watches the ball and adjusts. Alternative income investors (like those using our Alternative Fixed Income portfolio) get to play that same game here, because both pitches are in the strike zone.

I want to be honest with you about how much of this I am willing to bet on. Federal debt sits near $40 trillion, and Treasury reported roughly $1.17 trillion of debt service costs in fiscal 2026, close to 19% of federal spending. Lower policy rates would slow the growth of that bill, first through short term bills and eventually through maturing notes and bonds. Lower rates do not repair a deficit built on persistent spending. The Congressional Budget Office has made that point for years, and no amount of enthusiasm about artificial intelligence changes the arithmetic.

There is also a version of this story that runs backward on the Federal Reserve. The central bank has one legitimate reason to cut rates, and it is not an easier federal budget. It is a credible case that inflation is heading to target and that tight policy is doing needless damage to jobs and credit. If the market ever concludes the Fed is cutting simply to make Washington's bills easier to pay, investors will demand a bigger inflation and term premium on long bonds, and the very yields we are hoping fall would instead rise. That combination has a name, fiscal dominance, and it is the tail risk sitting under this entire thesis. Watch it, but do not build a whole portfolio around fearing it.

Short of outright rate cuts, the Federal Reserve has several levers available if it wants to support a gradual easing path. It can slow or end the runoff of Treasury and agency mortgage securities from its balance sheet, which reduces the duration the private market has to absorb. The Federal Open Market Committee already cut the Treasury runoff cap to $5 billion a month starting in April 2025. It can use forward guidance to move long term yields before a single rate cut is delivered. It can keep repo facilities and the discount window functioning so that a funding squeeze does not turn into a self inflicted crisis. It can reserve outright asset purchases for genuine market dysfunction rather than routine deficit financing.

None of these tools solve the debt math. They can prevent a liquidity accident from making a normal slowdown worse. That distinction matters more than it sounds, because a policy that keeps markets functioning is very different from a policy that pretends deficits do not exist.

A gradual easing cycle paired with the end of quantitative tightening would ripple through fixed income in fairly predictable ways.