Founding macro read — 31 May 2026

We are inaugurating this book into an awkward cross-current, not a clean trend. The Fed cut to 3.64% effective in January and has held there four meetings running (Fed funds, FRED, Apr print). The bond market is not ratifying the dovish lean: 10-year Treasury closed 4.45% on 28-May after touching 4.67% mid-month, and 10-year real yield has climbed from 1.90% in mid-April to 2.06% — real rates are rising into a cutting cycle, which is the tape's way of saying it doesn't believe the inflation fight is done. headline CPI ran +0.86% m/m in March and +0.64% in April; the 4-month annualized pace is near 4%. WTI's spike to $112 on 18-May (now back to $98) is not helping the disinflation narrative.

Meanwhile credit is asleep. HY OAS at 272 bps (ICE BofA HY index, 28-May) is well inside its long-run median and VIX at 15.7 says the equity vol complex agrees. UNRATE pinned at 4.3% for five months — labor isn't breaking. So: restrictive real rates, complacent credit, sticky-hot inflation, an unemployment rate that has stopped falling. That is late-cycle, not mid-cycle, and it's late-cycle with the long end doing the Fed's tightening for it.

Implication for how hard to lean out of the gate. Moderately, not heroically, and over a 2–4 week build. The +500 bps mandate gets earned in regimes like this by being selective, not gross-long brave — beta is already priced (HY at 272, VIX at 15, SPX bid). I'd nudge the IPS toward: quality balance sheets over leverage, cash-generative over multiple-expansion stories, energy and short-duration value as an inflation hedge that the market is under-owning, and an underweight to rate-sensitive long-duration growth until 10-year Treasury settles back through 4.30%. Keep dry powder — when credit finally notices real rates, the entry gets better. Hand the factor and sector tilts to Daniel and Margaret; I'm flagging the regime, not the names.

Daniel — IPS round 1

Catherine, here's what I can live with. Folding in Priya's late-cycle / re-accelerating inflation read, the path to +500 bps is active share ~35–45%, 20–28 names, max active sleeve ~250 bps per name over SPY weight. That's a tracking error budget of ~3.5–4.5% and an implied IR of 1.2–1.4 to clear target. Achievable, not heroic.

Factor tilts (regime-aligned): quality + cash-flow-today overweight; energy as inflation hedge (long-side, no leverage); underweight long-duration growth until 10-year Treasury < 4.30%; modest low-vol bias while HY OAS is complacent at 272 bps — that's the spread that snaps first if Priya's call is right.

Limits I'll sign:

Turnover target 60–90% annualized — we are not a HFT desk and we are not buy-and-hold.

Catherine, push back where you need to. Two places I expect friction: (1) you'll want the single-name max tighter than 400 bps — I'll trade you down to 350 if you give me the 1.75x sector ratio. (2) drawdown breaker — I want -800 bps excess, not absolute -8%; this is a relative-return shop.

Daniel — here is what I'll wear, in plain numbers.

Loss path. A 20–28 name book with ~250 bps active per name, beta 1.0–1.1, in Priya's regime: a normal bad week is SPY down 4% and our active sleeve down another 1–1.5%, call it $50–55k, 5–5.5% of NAV. A 2008-style week — SPY -8%, dispersion blows out, our highest-conviction names down 15–20% — is $90–110k, 9–11% of NAV. That is what we are signing up for.

Limits I'll enforce:

• single_name_weight_band {min: 0.0, max: 0.08} — 8% hard cap per name. At 8%, a -20% gap costs 160 bps of NAV in a session. I will not wear 10%.
• top_n_concentration {n: 5, max_weight: 0.32} — top-5 ≤ 32%.
• cash_range {min: 0.0, max: 0.15} — cash/SGOV ≤ 15%; no stealth market-timing.
• max_position_count {28}, min_position_count {18} — matches the active-share frame.
• max_gross_leverage {1.0} — long-only, no margin.
• drawdown_circuit_breaker {peak_to_trough_excess_bps: 400, cut_gross_to: 0.85} — 400 bps behind SPY peak-to-trough cuts gross to 85% and forces a re-vote.
• stop_loss_per_position_bps {300} — any name down 300 bps of NAV from cost forces a review, not an auto-sell.

Sector bands I won't pre-commit — rather flag a 40% tech book in the room than have the engine forced-sell in a drawdown. Earnings-gap freeze I'll add once we see one.

VOTE: approve

Daniel — IPS round 2, closing the deal

Iris, I'll meet you here. Folding your loss-path math into my round-1 numbers:

Sizing. Single-name band 0.005–0.035 (50–350 bps). Trading the 400 cap down to 350 — your tail math on a 2008 week with a 3.5% top name is the right pain threshold. Floor stays at 50 bps; sub-50 slots are signaling, not investing.

Concentration. Top-5 ≤ 22%, top-10 ≤ 38%. Position count 18–30. Active share lands 35–45%, which is what Priya's regime call needs to express itself.

Sector band. 0.5x–1.6x SPY weight. Wide enough for the energy overweight + long-duration-growth underweight to matter; tight enough we don't become a sector fund.

Cash. 1–8%. SGOV is the parking lot.

Beta. 0.85–1.15, 60-day lookback.

Stops. Per-name review at −2000 bps absolute. Earnings freeze ±2000 bps within 2 trading days of a print.

Circuit breaker. −600 bps excess vs SPY peak-to-trough cuts gross to 0.85. Measured in excess — we are paid on relative return; an absolute-drawdown breaker would have us de-risk into a -10% tape when we're +200 bps ahead. Wrong incentive.

Leverage. 1.0x. Long-only by mandate.

Turnover 60–90% annualized. Path to +500 bps: ~35% active share × ~1.4 IR on a 4% TE budget. Achievable, not heroic. That's the deal.

Daniel — I'll meet you at 350 bps single-name and tighten the rest.

Loss path on your frame. 25 names, max 3.5% each, top-5 ~16%, beta 1.05, gross 1.0. Normal bad week (SPY -4%, active sleeve -1 to -1.5%): $50–55k, 5–5.5% of NAV. 2008-style week (SPY -9%, active -2%): $100–110k, 10–11% of NAV. A top name gapping -20% on earnings inside a 3.5% slot costs ~70 bps of NAV in a session — wearable, just.

Limits I'll wear, engine-enforceable:

• single_name_weight_band 0.005–0.035 (50–350 bps per name).
• top_n_concentration n=5, max_weight 0.18 (18%). Stops the book becoming five bets.
• cash_range 0.00–0.15. Priya's late-cycle regime says optionality matters; 15% gives it without forcing a cash drag.
• min_position_count 20, max_position_count 28.
• max_gross_leverage 1.0. Long-only, full stop — no margin in a re-accelerating inflation tape.
• portfolio_beta_band 0.85–1.15, lookback 60d. Keeps us honest to the benchmark we're paid to beat.
• drawdown_circuit_breaker 400 bps peak-to-trough excess vs SPY → cut gross to 0.85. Bleeds risk before it compounds.
• stop_loss_per_position_bps 300 — triggers review, not auto-sell. Iris reads the tape before the engine dumps.

Deferred on purpose: sector_relative_weight_band and earnings_gap_freeze_bps. I'd rather flag a 40% tech book in committee than have the engine forced-sell into a drawdown. We can ratify either later if behavior demands it.

VOTE: approve