Paulo Macro — Oil & The China Syndrome: Checkmate in Chinese Revisited

Shrub's Razor & "more questions than answers"

• Opens admitting he has "more questions than answers" on oil today, keeping an open mind by dreaming up the most absurd outcomes as the most likely — his Shrub's Razor: "the funniest, most absurd outcome is also the most likely one."

The Kovner signal he ignored

• The April "lack of control" feeling — equity risk ripping while oil failed to hold momentum — was, in hindsight, a bullish commodity consensus the market had stopped confirming: a classic Kovner signal ("a consensus the market is not confirming… a lot of people who are going to be wrong") to get out of the way.
• He didn't listen and "took a nice dose" (a drawdown), but now thinks he understands the two visible factors he got wrong.

Wrong factor #1 — China stepped >4mmbpd out of the market since April

• Beyond the visible SPR/commercial draws, China exited the global market beginning in April to the tune of >4mmbpd, significantly loosening the near-term physical market. China is "at the heart of my questions" and he returns to it later.
• Aside on the title: The China Syndrome (1979 film) — a nuclear-meltdown thriller (Fonda / Douglas / Lemmon) about a plant cover-up — a nod to a prior crazy oil-market era (the 1970s).

Wrong factor #2 — speculators collapsed from near-record long to near-record low

• Speculators took the physical cue and swung from one of their largest net-long positions to one of their lowest exposures in weeks. Combined Brent+WTI managed-money net long (notional US$) is "rarely been this washed out" — near-term lows exceeded on the downside only by 2015 shale, the 2018 surplus, April-2020 Covid, mid-2024, and 4Q25.
• All this "in the teeth of ~5-7mmbpd global inventory draws" — which he calls the "Upside Down": fundamentals and positioning pointing opposite ways.

Positioning at historic extremes (Kemp)

• Per John Kemp (Reuters): WTI & Brent net positions are at the 11th percentile back to 2011; the Brent long/short ratio at the 2nd percentile. WTI non-commercial net long as % of open interest sits below the late-2015 shale blowups and the 2023-24 troughs — only 2025 was lower (back to 2012).

The Pile On & the Soros Misconception

• He started a thread of oil bears "dunking" on experienced barrel-counters — a phenomenon he calls The Pile On (one-sided bearish jubilation as a contrarian tell); his pal Kevin Muir cited it in a bullish energy note (Paulo urges readers to subscribe to Muir).
• Breaking his own trading rules, he has upped oil exposure again over the two-week decline. The bearish jubilation reflects a Soros Misconception — a narrative solving for price. His process/outcome matrix: Good→Good = Deserved success, Good→Bad = Tough beat, Bad→Good = Lucky break, Bad→Bad = Just desserts.
• The bears' foundation is Bad Process → Good Outcome: "there's plenty of oil around" suspends the physics of volume/time; pricing present inventory at $100 but "the future superglut at $70" is inconsistent.

The bear-trap thesis needs physical to tighten

• A bear trap "is being set with current positioning," but for it to spring the physical market must tighten: a loosening market swung positioning down, so shorts/underexposed specs will struggle if it tightens "out of nowhere."
• Even without China the world is drawing >35mm bbl/week since the crisis started (Kpler: -49mmbbl last week, "this is crazy"); if China reemerges as a buyer the market tightens "almost immediately."

The China black box

• Per HFI Research (Kpler data), China is now cycling >-5mmbpd of crude imports vs pre-war and beginning to draw visible onshore inventories, even as global onshore stocks have "never been this seasonally low" (draws mostly outside China so far).
• Did China release hidden underground SPR? He's skeptical (SOE contacts say no; Energy Aspects hears maybe) — "China is a black box." Balances square via SOE run cuts + modest demand destruction (PBOC drained liquidity Mar-May) + product-export restrictions since March.

Chinese refining primer — SOEs vs teapots

• China's refinery capacity ~19mmbpd (2024), capped at 20mmbpd by 2025; actual throughput a record ~14.8mmbpd in 2025 (up ~600kbd, led by the private mega Yulong refinery in Shandong) — the nameplate-vs-runs gap reflects chronic overcapacity.
• SOEs — Sinopec (largest refiner), PetroChina/CNPC, CNOOC — ~75% of capacity (EIA), >800k employees, ~US$1trn revenue. Teapots — ~one-quarter of capacity (~4.5-5mmbpd), Shandong-based, historically 35-40% utilization; they handle discounted/sanctioned crude (Iran, Russia, Venezuela) while SOEs stay insulated.

Teapots as a "product SPR"

• SOE utilization fell from low-80s to upper-60s since the war → a -15% cut on ~15mmbpd implies >2mmbpd of lost product output. Netting exports (~500kbpd) and demand destruction (~500kbpd), China has drawn ≥~1mmbpd of domestic product stocks — >100mmbbls over three months.
• Since products degrade (jet fuel especially), excess/swing refining capacity acts as a form of SPR for products: crude stores indefinitely, but products have a shelf life, so a large crude SPR + spare refining insulates a country from embargos/wars.

US running flat out; cracks imply $125+ crude

• China is the only global refining center with real excess capacity (Mideast product exports 4-5mmbpd offline); the US is the only other, and it's running flat out — timespreads forcing domestic gasoline higher to price out exports into peak driving season.
• Gasoline/diesel Jul-Aug timespreads imply crude "normally" $100-120; the 3-2-1 crack one month out vs WTI is a $56 crack — implying >$125/bbl. The world is in deficit of both crude and products, but products are running out faster (Mideast + Chinese export losses + shelf life). Refineries running so hard they "literally catch fire."

Why won't China buy? (the central mystery)

• Chinese refinery margins are flipping positive; the world needs China to ramp runs and export — yet chatter suggests they'll slow-walk the export-ban loosening (a "creep, not a real lift," per Reuters).
• China is lifting floating storage around Asia and doing most of the Persian Gulf egress buying (halving oil stranded behind Hormuz), but not ramping Saudi/Brazil/West Africa — flipping WAF cargos at -$4-8/bbl discounts to Dated Brent. "Does China want barrels, or not?"
• Electrification isn't the answer: transport fuels are only ~20% of Chinese oil demand; only ~12% of the fleet is EV and demand still grows ~500kbpd/yr.

Geopolitical speculation — embargo trial run, Iran, midterms

• China's non-buying drains global product stocks and keeps prices elevated. Speculation: a Taiwan handshake owed to Trump? A trial run of navigating an embargo (à la Japan in WW2) — proving China is "as powerful on the demand side as OPEC is on the supply side." But "they still need oil… can't draw forever" ("China does not have allies — they have suppliers and customers").
• Iran's North Star: the regime survived and needs to "make this hurt" — higher gasoline hurting Trump in the midterms (126 days out) sends a message.

PBOC liquidity, gold, and the "never happened before" list

• Sellside now calls for a 3mmbpd+ 2027 superglut, but the math says crude-vs-inventories has "never been this far" from regressed price — a Covid-era discount to inventory fair value while drawing, not building.
• PBOC is the "21st-century Bundesbank" (per Louis-Vincent Gave): rising energy is inflationary, so it slammed the brakes in March — but has been injecting again since ~May 21st (a "positive hum"). Gold is now "inextricably linked to Chinese liquidity" (Michael Howell), China being the world's biggest gold buyer.
• The "never happened before" tape: 3-2-1 crack the widest vs Brent since Bloomberg's 1988 inception; gasoline-vs-WTI the widest divergence in energy-market history; timespreads in contango while onshore inventories draw; cracks near all-time wides as crude collapses into contango. "More questions than answers."