Why might traditional defensives offer limited value at current prices? Lou Miller, global head of Custom Basket Trading in Goldman Sachs FICC and Equities, discusses on The Markets:

Carson’s Ryan Detrick (@RyanDetrick) warns improving market breadth could be a bull trap if defensives are leading. “Be careful what you wish for,” as strategists point to a similar rotation during the dot-com unwind.

BofA’s Hartnett on Warsh at Jackson Hole: “Bull success” requires Warsh to reinforce his inflation-fighting credentials while supporting Bessent’s push to contain long-term yields. A “policy failure” could send yields back above pre-buyback levels, weigh on the dollar & favor defensives over cyclicals.

Rough week so far. Thanks to macro headwinds, we're currently seeing some momentum unwind—semis down, defensives up. I believe that institutions are simply hedging their high beta exposure right now. I mentioned earlier that there's plenty of institutional headroom for semis/tech exposure to increase. After today, I'm even more sure of this, with many funds positioning a little more defensively with higher yields as the backdrop. Personally, I'm dreaming of a scenario where $NVDA earnings will be so powerful that AI-related names resume their upwards trajectory, resulting in increased momentum as funds become more risk-on after the July de-leveraging. I feel like Jensen's remarks next week will be ultra critical to settle some nerves and to get the AI trade back on track.

A few thoughts from last night’s member video. It’s an ugly market, and the pain is concentrated in the stuff a lot of us actually trade. AI infrastructure, software, AI applications, plenty of biotech. Look at how few names in those groups are above the 20, 50 or 200-day averages. It looks horrible. Zoom out and there’s still strength elsewhere. Equal weight is doing well, RSP is outperforming the cap-weighted S&P, and defensives, financials and metals have held up. A lot of what we’re seeing is rotation out of the names that were working. A couple things stand out. Oil came off hard and equities didn’t respond. That inverse relationship has been a big part of the market lately, and this time we didn’t get the relief. Earnings aren’t helping either. Apple and Palantir both beat and both traded down after hours. The frustrating part is seeing enough to get interested, then talking yourself into the rest. I did that Monday. Most names were above VWAP going into the open, volume was lighter than Friday, and it looked like we could get a bounce. We pushed through Friday’s highs, failed and rotated lower. I lowered my bar and got long anyway. Then the amended Anthropic S-1 hit and I got stopped out across the board. I can’t predict that headline, but I can control whether I take a trade that meets my rules. This one didn’t. That part is on me. What I’m looking for now is a more decisive flush at the index level. The S&P is still at the bottom of multi-month value and hasn’t really gone anywhere. VIX has been in the 20s, but we haven’t seen a real spike at the index level, capitulation volume or wide-range blowout candles. Correlations are starting to pick up, so I’d rather focus on an index trade here than keep digging through individual names. If we get that flush, with a vol spike, higher and real volume at the lows, I’m interested in a bounce that can last a few sessions. I’d want tech participating. My bigger view is still that we make new lows after that. Midterms might make things clearer. Until then, I’m keeping risk down and waiting for the trade I actually want. Monday was a reminder of what happens when I start making exceptions.

The S&P 500’s unfavorable seasonal stretch ended at the October 14 close, opening the historically favorable year-end window. Since 1985 the index rose through this window in 33 of 41 years, an 80% win rate. The lone caution is small caps, now in a correction. Five reports this week. Weight of Evidence: Constructive, selective. U.S. Stocks • Defensives • Year-End Tilt • Transports • Small Caps Read the full Weekly Research Signals: