← Analysis page  ·  CNBC Halftime Report hub  ·  Research hub

Actionable insights — Stocks Rise as Oil and Yields Ease

The repeatable analysis behind the committee's calls: not what they bought, but how they got there — written so the process can be rerun later on different names.
2026-SEP-21 · CNBC Halftime Report (audio edition, Monday after the FOMC hike) · Dominic Chu + committee (Terranova, Link, Lebenthal, Talkington) · ▶ Listen · full analysis · transcript
How to read this page: on a day when oil and yields fell together, the committee's useful work was reading positioning and sentiment against fundamentals, re-buying names whose story is changing, measuring mega caps against their own multiple history, and trading a stuck stock instead of waiting on it. Each insight is a procedure you can rerun; the boxed Here: line shows how it played out on 2026-SEP-21. Cues are (m:ss) into the episode.

1. After a scheduled event, check whether the crowded bet failed to pay 2:02

The repeatable method
  1. Before a known catalyst (a Fed meeting), note the consensus speculative position (e.g. short Treasuries for a yield spike).
  2. After the event, check whether the market gave that trade what it wanted. If not, treat it as exhaustion of the crowded side.
  3. Confirm with price: a gap higher and groups that usually diverge (semis and software) rising together.
  4. Pair it with sentiment and positioning: defensive positioning plus bearish surveys means buyers are under-invested.
Here: Terranova: speculators went into the Fed short Treasuries expecting a 5¼% 10-year — "you didn't get that" — and the S&P gapped higher Friday into Monday with semis and software up together: "2 words risk on."
Watch for

2. When analysts pile on, check the fundamentals they are cutting 10:50

The repeatable method
  1. Track the survey sentiment (AAII bears) and the sell-side direction (targets and multiples being cut).
  2. Check the hard data against it: GDP nowcast, jobless claims, retail sales, earnings revisions, credit spreads, breadth of sector earnings growth.
  3. If the data still rises while opinion falls, buy the volatility; name in advance what would change your mind (the economy rolling over, a Fed that hikes more than expected).
Here: Link: AAII bears at a year high, targets being cut, yet GDPNow 5.1%, claims historically low, estimates rising, multiples down from 22× to 19×, 8 of 11 sectors with double-digit earnings growth — the pile-on is "absolutely the wrong decision." Her exit condition: a Fed that goes "three times" this year.
Watch for

3. Plan the seasonal drawdown in advance and buy it 7:59

The repeatable method
  1. Know the historical pattern for this point in the calendar (here: the drawdown between the August high and the midterm-year low).
  2. Size the expected drop from history (7% typical, 25% extreme) and decide in advance what you will buy.
  3. Hold both halves of the market (cap-weight and equal-weight) so you can add to whichever leads.
Here: Talkington: a drawdown every midterm cycle for 6½ decades, followed by the strongest stretch (November–February) — she will "take those draw downs to add to positions on both RSP and your Q holdings."
Watch for

4. Re-buy a trimmed name when the market starts pricing a new business 18:32

The repeatable method
  1. Trim into a parabolic move; note the price.
  2. Re-buy when two things line up: price momentum (a strong one-month move) and a segment the market has not priced (data center, auto, IoT, selling spare compute).
  3. Accept paying above the low — the signal is the market recognizing the new story, not the cheapest price.
Here: Lebenthal trimmed half of QCOM at ~$230 in May, admits he could have re-bought at $160, and doubles it now on +19% in a month and a data-center business — the same pattern he sees in META (selling excess compute). He then picks MU as another "unloved" chip coming back.
Watch for

5. Measure mega caps against their own multiple history 20:28

The repeatable method
  1. Compare the current forward P/E with the stock's 10-year (or long-term) average.
  2. Check that the business is still growing (cloud growth rates) so the discount is not a value trap.
  3. Add when the multiple is at or below half its history and growth holds.
Here: Link: AMZN at 20× vs. a 44× 10-year average with AWS +37%; NVDA at 17× vs. 35× — she "just recently bought NVIDIA when it was really almost at the lows."
Watch for

6. Sell calls on a range-bound holding until it breaks out 16:53

The repeatable method
  1. Identify the range and the ceiling the stock keeps failing at.
  2. While it stays under the ceiling, sell covered calls above the range to earn income on a stock you keep.
  3. Stop selling calls once it closes through the ceiling — that is when the upside you would give away is largest.
Here: Talkington on NVDA: range-bound 210–230 under a ~232 ceiling; "until it can break out, I would just sell calls" — a pattern that "has been very, very solid for the last six months." Her UBER final trade uses the same range logic (buy 70, target 78).
Watch for

7. Judge a rally by what leads it 40:37

The repeatable method
  1. On an up day, check whether cyclicals and equal-weight join in, or only mega caps with defensive qualities lead.
  2. Check what else explains it (oil and yields falling).
  3. Treat a narrow, macro-driven rally as tactical: respect the low it made, but reserve judgment on a new trend.
Here: Santoli: the leadership doesn't "represent genuine risk" — it is the Mag 7 plus semis coming back "first in, first out," lifted by oil and yields; last week's post-Fed low is "meaningful on a tactical basis." Terranova adds that summit optimism may already be priced.
Watch for

Methods distilled from the public CNBC Halftime Report audio episode (Spotify transcript) for personal study. Not investment advice. © CNBC for source material.