← Analysis page  ·  CNBC Halftime Report hub  ·  Research hub

Actionable insights — Inflation Risks to the Rally

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-18 · CNBC Halftime Report (audio edition, Friday after the FOMC hike) · Leslie Picker + committee (Raskin, Lebenthal, Snipe, Sechan) + Bill Baruch by phone · ▶ Listen · full analysis · transcript
How to read this page: with the 10-year at 5% and the index stalled, the committee's useful work is sizing and sorting rather than calling direction — trimming a great company when its multiple hits an extreme, testing a capex boom against its revenue, splitting a sector by business model, and waiting for estimates (not prices) to turn. Each insight is a procedure you can rerun; the boxed Here: line shows how it played out on 2026-SEP-18. Cues are (m:ss) into the episode.

1. Trim a core holding at a valuation extreme, and name where the money goes 24:55

The repeatable method
  1. Measure the stock's current multiple against its own history (median and standard deviation), not against the market.
  2. At about two standard deviations above the median, cut the weight back to a normal size — don't exit a quality business outright.
  3. Redeploy the proceeds into the names you expect to lead the next leg, and say which ones in advance.
  4. Keep a low-basis core if taxes make selling expensive; trade the extra position around it.
Here: Baruch cuts AAPL from 14% to 8% of a 10-name portfolio at "2 standard deviations above its median valuation" and buys more NVDA and AMZN. Lebenthal trimmed half at ~35× forward but keeps the $15-basis half: "it has made a lot of sense to have a trading position in addition to the core position." Sechan's counter: for an active manager, being underweight the company that "owns the customer" is the bigger risk.
Watch for

2. Test a capex boom against the revenue it has produced so far 11:30

The repeatable method
  1. Put cumulative revenue from the new technology next to one year of capex spent on it.
  2. List what is raising the required return: cost of capital (yields), component prices, power costs, number of competitors.
  3. Ask whether the multiple has fallen enough to reflect a move from a capital-light to a capital-heavy business model.
  4. Separate "will we overbuild?" (almost always yes) from "is it this year's problem?" — the timing decides the trade.
Here: Raskin: ~$200 billion of cumulative AI revenue (per Roger McNamee) against ~$1 trillion of capex this year — "a financially unsustainable arms race," with yields and component costs raising the bar. Sechan agrees "we will overbuild" but "that is not today's problem," noting the mega-caps de-rated from 34× to 24×. Snipe's numbers: capex $900B this year, $1.6T next.
Watch for

3. Split a falling sector by business model before judging it 30:55

The repeatable method
  1. When a whole sector sells off on a macro event, sort its members by what actually drives their earnings (here: capital markets versus deposits/net interest margin).
  2. Ask which bucket gains from the current conditions (deal activity, capital raising, volatile bond markets) and which loses (margin compression).
  3. Own the winners of that split and treat their drawdown as "more of a scare."
Here: Sechan: "there's going to be haves or have nots" — JPM, MS, GS gain from deals and trading desks; deposit-driven banks may see margins squeezed. Snipe keeps GS (best of breed in IB) and APO (its insurance arm sets it apart from other alt managers).
Watch for

4. On a de-rated growth stock, wait for estimates to turn, not the price 36:25

The repeatable method
  1. When a former premium-multiple stock looks cheap, check whether the multiple fell because the price fell or because earnings estimates fell.
  2. If estimates are still falling, a low multiple is not value — stay out regardless of product quality or headlines.
  3. Re-enter only once consensus earnings revisions turn up; pair this with a macro check on the customer (real income).
Here: Lebenthal on ONON after the Mbappé signing: now "very much value territory," but "until those earnings estimates start going up, this is a no touch." Raskin: ~18 months of negative real income. Sechan on the group (NKE, DECK): real disposable income must rise first — "now is not that time."
Watch for

5. Add duration when yields are high enough to hedge again 7:50

The repeatable method
  1. Decide whether the inflation driving yields is transitory (a supply shock) or embedded.
  2. If transitory and the Fed has already tightened, treat high yields as a chance to buy bonds that will now offset equity losses.
  3. Build the hedge while the "we'll lose control of the curve" story is loudest.
Here: Sechan: inflation "is completely based on what is happening in the Middle East," the hike "reloaded the gun," and "we have been buying bonds like maniacs" because yields "can insulate portfolios and now is the time." Epperson's 401(k) version: keep near-term spending in cash and bonds.
Watch for

6. When leadership narrows, broaden with what your plan already offers 41:10

The repeatable method
  1. Check how much of your index fund sits in the top five names (~30% here) and whether target-date funds repeat that exposure.
  2. If the rally is broadening (equal-weight and small caps rising under a flat index), add the plan's equal-weight, value, international or small-cap options.
  3. Rebalance rather than abandon the core index fund.
Here: Lebenthal: the S&P fund was the right call in 2023–24, but since mid-2025 "we've had a little bit of a broadening" — look for equal-weight (RSP-style), value, international or small-cap options.
Watch for

Methods distilled from the public CNBC Halftime Report audio episode of 2026-SEP-18 (transcript in transcript.txt). Not investment advice. © CNBC for source material.