In the third quarter, the Federal Reserve raised interest rates for the first time since 2023. Inflation re-accelerated. The 10-year Treasury crossed 5%. Oil ran toward $100 a barrel.

The S&P 500 went up anyway.

That sounds like a contradiction. It isn't. Almost every line below traces back to a single variable, and it isn't growth, earnings, or geopolitics. It's what money costs.

Last quarter we wrote that an accommodative Fed was on hold, but the course hadn't reversed. It reversed. On September 16, the FOMC voted 12-0 to raise the federal funds rate a quarter point to 3.75-4.00%. Chair Warsh described it as removing accommodation rather than tightening, which is a meaningful distinction and also a warning. Sixteen of eighteen participants now see at least one more hike before year-end.

When the price of money changes direction, everything priced off it has to re-rate. Bonds had one of their worst quarters on record. Gold found a floor but couldn't break out. Financials dropped more than 6% in September. And in October, the market started asking a harder question about artificial intelligence spending, not whether it works, but what it costs to finance at 5%.

We've said before: avoid the hype, invest, don't speculate. That process doesn't change when rates rise. What changes is the math on everything you own.

Market Performance

Year-to-date returns through September 30, 2026.

U.S. Equities (S&P 500) YTD: +12.0%

International Equities (MSCI ACWI ex-U.S.) YTD: +13.9%

Fixed Income (Bloomberg U.S. Aggregate Bond Index) YTD: -2.9%

Alternative Market Neutral (Lipper Global Alternative Long/Short Equity Index) YTD: +7.6%

Gold: roughly flat

Bitcoin: -12%

A few numbers behind the numbers: the 10-year Treasury ended the quarter at 5.27%, up roughly 110 basis points on the year. The U.S. Dollar Index gained 3.7%. Commodities rose 29%. Emerging market equities are up 24%, and the energy sector is up 40%.

That last set of figures is the quarter in miniature. The things that did best were the things that benefit from expensive money and expensive oil.

Key Highlights and Commentary

The Fed: The Ground Moved

Last quarter we told you an accommodative Fed was on hold for the moment, but the course hadn't reversed. The ground moved underneath that view over six weeks in August and September, and it moved for reasons that had nothing to do with the Fed's patience and everything to do with energy prices feeding into inflation prints.

Chair Warsh spent August signaling a harder line on inflation, and the September meeting delivered a unanimous quarter-point hike. He declined to call it tightening. He also declined to submit his own rate projection, which leaves the dot plot as a read on the committee rather than on the chair. Sixteen of eighteen participants see at least one more increase this year, and four see two.

Then September CPI came in at 3.7%, the hottest reading in four months, pushed along by energy. Markets now put the odds of a December hike at nearly certain.

Here's the part that kept September from being worse. The bond market had already done the tightening. The 2-year Treasury was yielding about 4.73% going into the meeting while the fed funds rate sat at 3.75%, meaning short rates had moved well ahead of the Fed. By the time the hike arrived, it was priced. Another one or two may be as well.

History offers some comfort on the shape of this. Going back to 1988, the S&P 500 has averaged a 2.5% decline in the 45 days following a first rate hike, positive only 29% of the time. Twelve months out, the average is a gain of 11.7%, positive 86% of the time. Stocks stumble early, then recover, as long as the tightening stays measured and growth holds up. Both of those conditions are doing real work in that sentence.

The lesson here isn't that forecasts are hard. It's that a portfolio built to need a particular Fed outcome is a portfolio with a single point of failure. Ours wasn't, and that's the point.

Fixed Income: The Worst Quarter in Years, and the Best Entry Point

This was the real story of the quarter, and it got far less attention than it deserved.

The 10-year Treasury started the year at roughly 4.2% and ended the quarter at 5.27%. The Aggregate index posted one of its worst quarterly returns on record. This wasn't just a U.S. phenomenon: the European Central Bank hiked, the Bank of Japan followed, and Japan's 10-year crossed 3% for the first time since 1996. Long-term debt sold off globally.

Here's the part that matters more. The Aggregate's yield to maturity is now above 5%, with 30-day SEC yields around 4.75%. Bonds have been painful to own and are now genuinely attractive to buy. Those two things are the same fact viewed from different ends.

We'd rather own an asset that already took its medicine than one that hasn't. We're keeping our position: define the allocation to fixed income, then diversify like crazy inside it. The systematic strategy we added earlier this year earned its seat this quarter.

U.S. Equities: Breadth Is a Position, Not a Prediction

For most of this year, the story was broadening. The other 493 outran the Magnificent Seven, equal-weight beat cap-weight, and the picks-and-shovels names we've talked about carried more of the load.

Underneath the leadership churn, the fundamentals did something unusual. Forward earnings estimates for the S&P 500 rose roughly 30% this year while the index itself gained about 11%. Earnings outran prices by a wide margin, which means the market got cheaper on a forward multiple basis even as it climbed. Valuations compressed about 13%. That's the opposite of what a bubble looks like.

Then September flipped it. The Nasdaq rose 1.9% while the Dow fell 4.3%. The S&P finished within about 1% of its record high while roughly three-quarters of its component stocks declined for the month. The index held up on a handful of names.

Then October flipped it back. On the worst day of the technology selloff, more than two-thirds of S&P 500 constituents actually finished higher, with money rotating toward staples and low-volatility names rather than leaving stocks altogether.

Three reversals in four months. That's the argument against trying to time this. It's also why our U.S. sleeve runs through flexible active strategies rather than a market-cap-weighted index. We don't need to call which side of the market leads each month. We need exposure to both.

Artificial Intelligence: A Different Question Now

We've written that the AI theme is about the companies cashing the checks, not just the ones writing them. That framing is holding. But the question the market is asking has changed.

In early October, a report that OpenAI's annualized revenue was running roughly $20 billion below previously signaled levels set off a sharp selloff in chip and infrastructure names. The hyperscalers are on pace to spend somewhere near $650 to $700 billion this year on data centers and compute. Some of them are now funding that spending through equity issuance rather than operating cash flow, which is a real change from how these companies have historically grown.

Nobody is seriously arguing anymore that AI doesn't work. The argument is about what it costs to finance when money isn't free. At a 5% discount rate, a dollar of capital expenditure eight years out is worth meaningfully less than it was at 2%.

Two things argue the other way, and we take them seriously. Hyperscaler free cash flow appears to be bottoming right about now and is projected to recover from here, meaning the worst of the cash drain may be behind rather than ahead. And the economics of the compute itself keep improving: token usage has grown enormously since early 2025, and each gigawatt of frontier compute now generates several times the annualized revenue that prior-generation capacity did. Smarter models do harder work, harder work drives usage, and usage makes every gigawatt more valuable.

The capital spending is still staggering. Roughly $743 billion across the four largest hyperscalers this year, with estimates above a trillion for next year. Our read is that the AI economy can absorb a moderate tightening cycle. We would feel differently about an aggressive one.

We remain convicted on AI and Defense as long-term structural positions. Both are durable. But we'd be doing you a disservice if we pretended the financing math hadn't changed, because it has.

International Equities: Still Ahead, Just Barely

International equities are up 13.9% for the year and have now outpaced the U.S. for the better part of two years. That's the headline.

The quarter itself was flat, down 0.7%, with September off 2.5%. A dollar that gained 3.7% on the year continues to lean against the asset class. Inside the sleeve, emerging markets are doing the heavy lifting at 24%, helped by the increasingly technology-weighted makeup of those indexes.

The structural case hasn't moved: cheaper valuations, broader industry participation, and far less concentration risk than the U.S. market carries. We're watching the dollar, not the headlines.

Gold: A Floor, Not a Breakout

Gold made a round trip this year. It hit an all-time high above $5,500 in January, fell to roughly $4,000 by midyear, rallied about 10% in August, and gave most of that back in three sessions in early September. It sits near $4,150 today, roughly flat on the year and below its 200-day average.

The short-term story is entirely about real yields. Gold pays nothing. When the competition starts paying 5%, that matters. The long-term case we've laid out before, central bank buying, deficits, geopolitical hedging, currency diversification, is unchanged.

Our disciplined profit-taking earlier in the year and our underweight both look better today than they did in January. We're comfortable here.

Digital Assets: Off the Slide

Last quarter we described Bitcoin as having gone straight down the alpine slides. It stopped. From a low near $58,000 in July, it rallied through August and trades near $82,000 today.

What's interesting is who bought. U.S. spot Bitcoin ETFs took in roughly $3.5 billion in August while long-term holders were net sellers into the strength. The rally was funds absorbing supply from everyone else.

We raised a question last quarter about whether Bitcoin had become a leveraged proxy on the hyperscalers rather than an independent return stream. October offered a partial answer: Bitcoin held up reasonably well through the technology selloff. That argues for some independence. We're not ready to call it settled, and in the meantime we continue to manage the position the way we always have, selling into strength at the ceiling and buying at the floor.

About Those Midterms

You're going to get a lot of election commentary over the next three weeks. Here's what the data actually says, including the part most of it will leave out.

The pattern is weak before, strong after. Across 31 midterm elections from 1900 through 2025, the S&P 500 averaged a gain of just 2.9% in the 12 months before a midterm, against 8.9% for all years. In the 12 months after, it averaged 12.4%. Narrowed to elections since 1980, that post-election figure rises to 17.5%.

A separate analysis going back to 1874 found the market higher in the 12 months after 32 of 38 midterms, an 84% hit rate, averaging 14%. Fidelity's work puts the post-1938 figure at 95% positive. Every one of those studies points the same direction.

The pre-election weakness is just as consistent. Every midterm cycle since 1962 has seen a decline between mid-August and Election Day, averaging about 8%, and ten of the last sixteen cycles bottomed in October. We are, as of this writing, sitting squarely in that window.

Narrowing the window tightens the picture. Measuring just September through March around midterms since 1974, the S&P 500 has averaged a gain of about 11.7%, and the shape is consistent: drift lower through October, bottom somewhere near Election Day, then climb steadily through the winter.

What about who wins? Less than you'd think. Divided government under a Republican president has followed fourteen midterms, more than any other outcome, and markets were positive 93% of the time in the year after. J.P. Morgan found the S&P averaged 16.3% annually under the Obama administration and 16.0% under the first Trump administration. Nearly identical, across opposite political environments.

Now the part that gets left out.

The two midterm years in the modern record that produced negative returns were 2018 and 2022. Both selloffs were driven by Federal Reserve rate hikes, not by elections. We are currently in a Fed hiking cycle. The one condition that has historically broken this pattern is the condition we're in right now.

So take the seasonal data for what it's worth, which is something. Clarity has value, and after November 3 there will be less uncertainty about tax policy, spending, and regulation than there is today. But the mechanism is the removal of uncertainty, not the victory of a party. We don't position portfolios for election outcomes, and we'd encourage you not to either.

Position for the cost of money. Let the election resolve itself.

Key Takeaways: Portfolio Positioning

Four changes this quarter. The through-line: we're keeping our convictions and shrinking the size of the bets behind them.

Holding a modest equity overweight, with tempered magnitude. We trimmed from a 3% overweight to 1% in the spring and we're staying there. The reasoning is straightforward. The earnings backdrop gives equities room to absorb one or more rate hikes. Forward earnings estimates for the S&P 500 have risen roughly 30% this year while prices rose about 11%, which means valuations actually got cheaper as the market advanced. What matters from here is the pace and total magnitude of tightening, not any single meeting, because large moves unsettle valuations well before they affect the real economy. That tension supports staying invested while being more disciplined about position sizes. We're harvesting gains where positions have run and reining in concentration.

Resetting regional positioning to neutral. We've carried a U.S. overweight for some time and we're ending it. Earnings overseas have been strong enough to narrow the case for a large U.S. tilt, and emerging markets in particular are now generating returns through earnings growth rather than multiple expansion. The persistent dispersion between international markets gives us more flexibility to find specific opportunities than one broad regional bet does.

Reshaping the thematic tilts. Three moves here. We're exiting standalone momentum, which whipsawed badly this summer and is exactly the kind of factor bet that needs a tighter leash in sloppy markets. We're trimming core AI and global defense positions, not abandoning them, because both have run and both had grown larger than the role we intended for them. And we're adding to U.S. large caps that benefit from AI adoption rather than AI construction.

That last one is worth dwelling on. We've framed this theme as the check cutters versus the check cashers. There's now a third group: companies that neither build AI nor sell the picks and shovels, but simply use it to run better. That's likely where the next leg of this theme lives, and owning it requires no view whatsoever on whether hyperscaler capital spending is sustainable.

Staying modestly short duration and broadening where the bond sleeve earns its return. With the Fed hiking and the long end already repriced, we don't want a large duration bet in either direction. Instead we're widening the set of return drivers: swapping index for active exposure in mortgages, and adding global sovereigns, convertibles, and liquid alternatives. The aim is to source returns from drivers that aren't the ones already in the equity sleeve, which matters most in bond-heavy portfolios where duration has been doing all the defensive work.

Looking Ahead

Last quarter we said we were expecting cooling inflation, solid earnings, accommodative Fed policy, and stabilizing trade. Here's the scorecard.

Cooling inflation: no. September CPI hit 3.7%. Solid earnings: yes, and they carried the quarter. Accommodative Fed: reversed outright. Stabilizing trade: holding.

One of four intact, one reversed. That's a worse scorecard than we'd like and a better market than that scorecard would suggest. Which is roughly the point of diversification.

What we're watching into year-end: the December Fed meeting, where a hike is now close to fully priced. Third-quarter earnings, underway now, and whether guidance holds up against higher financing costs. Oil, which is doing as much to inflation prints as anything the Fed controls. The AI capital expenditure cycle, and whether the hyperscalers start pacing themselves. And November 3, after which there will be less to guess about.

We closed last quarter by saying noise is expected, and that when it's quiet your radar should go up. It got loud instead. Loud is manageable. Loud is priced.

The thing to watch isn't the noise. It's the cost of money, and what it does to everything that money buys.

As always, if you'd like to discuss how these views impact your portfolio and financial plan specifically, please connect with us here.