1.0 DISCOVER — Refined Products (Diesel & Gasoline) Price Elasticity and Recent (2026) Demand Destruction¶
Scope of this discovery pass: fill gaps in the KB's understanding of refined-products price elasticity and quantify the most recent (2026) demand destruction, with diesel as the primary focus and gasoline as the secondary. Prioritize official/primary sources (IEA, EIA, government/statistical agencies, academic research) and clearly separate directly observed volume declines from elasticity-based estimates. Reconcile the headline tension between product scarcity/refinery-capacity loss and crude-supply headlines, and propose three candidate updated "Big Three" monitoring questions that re-center the KB around products.
Relationship to existing KB: the KB already has strong crude-focused material (Q1/Q2/Q3, OMR September, STEO September, IEA physical vs futures disconnect, CSIS, JPM/MS/Goldman). This pass adds the products layer that the refined-products-as-shock-center concept (2026-09-13) flagged but did not quantitatively populate.
1. Directly Observed Volume Declines (2026)¶
These are directly measured volume changes, not model outputs. They reflect real-world consumption data published by official agencies or trade press citing official data.
1.1 United States — Gasoline (weekly observed)¶
| Metric | Value | Period | Baseline | Source |
|---|---|---|---|---|
| Finished motor gasoline product supplied (4-week avg) | 8.779 mb/d | 4 weeks ending 2026-09-18 | 8.848 mb/d (same 4-week 2025) | EIA Weekly Petroleum Status Report |
| YoY change | −69 kb/d (−0.78%) | as above | n/a | calculated from EIA table |
| Single-week product supplied | 8.847 mb/d | week ending 2026-09-18 | 8.798 mb/d prior week | EIA Weekly data |
| US regular gasoline retail price | (forecast $4.07/gal avg 2026; $3.85/gal 2027) | 2026–2027 | pre-war ~$3.50 | EIA STEO Sep 2026 |
Caveats: "product supplied" is EIA's proxy for consumption; it captures implied deliveries, not real-time pump draw. The −0.78% four-week YoY is small enough to be within normal seasonal noise; the year-over-year gap widens in 2026 H2 because the 2025 base included some recovery effects from post-pandemic normalization. Attribution: the small gasoline decline is consistent with sticky-but-rising retail prices, not yet a structural break.
1.2 United States — Distillate/Diesel (inventory + price, no clean weekly volume)¶
EIA does not publish a directly comparable "distillate product supplied" YoY in the same weekly cadence with the same quality of gasoline data. The US-side volume signal for diesel is therefore inferred from inventory draws + export surge + crack spreads, not weekly observed consumption:
| Metric | Value | Period | Source |
|---|---|---|---|
| US distillate inventories | forecast <100 mb | September 2026 | EIA STEO Sep 2026 — Petroleum Products |
| Distillate vs 5-yr (2021–2025) low | below 5-yr low; forecast to remain through end-2026 and most of 2027 | through 2027 | EIA STEO Sep 2026 |
| US net distillate exports | above or near 5-yr high every month since Feb 2026 | Feb–Sep 2026 | EIA STEO Sep 2026 |
| US retail diesel price | forecast $5.07/gal (2026), $4.40/gal (2027); +33¢ upward revision vs prior STEO | 2026–2027 | EIA STEO Sep 2026; Axios 2026-09-09 |
| US diesel crack spread | forecast >$2/gal Aug–Nov 2026 | Aug 2026 – mid-2027 | EIA STEO Sep 2026 |
Attribution: the diesel signal is price + inventory + export behavior, not volume. The implied US consumption change is masked because US refiners are running near capacity to capture record Atlantic Basin margins while exporting the marginal barrel — meaning the US is losing distillate to world markets at the same time its internal price rises.
1.3 Pakistan — Diesel (observed, dramatic)¶
| Metric | Value | Period | Baseline | Source |
|---|---|---|---|---|
| High-speed diesel sales | 422,000 tonnes | August 2026 | n/a (see below) | ProPakistani 2026-09-02 |
| YoY change | −19% | August 2026 | August 2025 | ProPakistani |
| MoM change | −32% | August 2026 | July 2026 | ProPakistani |
| Average diesel retail price | Rs 379/litre | August 2026 | n/a | ProPakistani |
| Price YoY | +36% | August 2026 | August 2025 | ProPakistani |
| OMCs with <20-day diesel cover | 12 of 20 | as of 2026-09-12 | regulatory minimum 20 days | ProPakistani 2026-09-12 |
| Emergency measures | rail transport + depot stocks + OGRA control room activated | 2026-09-24 | n/a | Petroleum Ministry 2026-09-24, PID 2026-09-24 |
| Pricing committee | endorsed emergency diesel-price-intervention principles | 2026-09-03 | n/a | Pakistan Today 2026-09-03 |
Caveats: the −19% YoY decline reflects a combination of price-elasticity-driven demand destruction, transport strikes, monsoon disruption, and pull-forward into July ahead of price hikes (per Pakistan Today, 2026-09-03). The price elasticity component alone cannot account for the full drop. Attribution: this is the strongest single-country "demand destruction" data point in the public record, but it is partly rationing/distribution-driven rather than purely price-driven.
1.4 India — Diesel (counter-evidence: still growing)¶
| Metric | Value | Period | Source |
|---|---|---|---|
| Diesel consumption | 7.02 million tonnes (+6.8% YoY) | August 2026 | Reuters via MarketScreener 2026-09-08 |
| Total petroleum products consumption | −2.8% | August 2026 (MoM) | Reuters |
Significance: India's August diesel demand is still growing YoY even in the context of a global diesel shock. Total product consumption fell because LPG and naphtha weakened sharply, not because diesel collapsed. This is direct evidence that the 2026 diesel shock is not a uniform global demand destruction — emerging Asia diesel demand is resilient at current prices (likely because Indian diesel retail is partially administered/subsidized and Indian growth remains high).
1.5 Africa — Refined products outlook (mixed)¶
| Source | Date | Key claim |
|---|---|---|
| S&P Global Commodity Insights — Africa Refined Products Short-Term Outlook | early 2026 | 1.9% refined products demand growth forecast; modest diesel growth in Algeria, Egypt, South Africa |
| S&P Global — Russian export ban strains African gasoil markets | 2026-07-30 | Russian product export ban strains African gasoil supply; price effect dominates over volume effect |
Caveats: Africa is price-pressured (supply-side) more than volume-destroyed (demand-side). S&P expects 2026 demand growth, not decline.
1.6 Global — IEA Oil Market Report Sep 2026 (the headline numbers)¶
| Metric | Value | Period | Baseline | Source |
|---|---|---|---|---|
| Global oil demand y-o-y change | −2.5 mb/d | 2026 | 2025 | IEA OMR Sep 2026, published 2026-09-11 |
| Demand downgrade vs prior OMR | −940 kb/d deeper | 2026 | August OMR | IEA OMR Sep 2026 |
| Pace of demand decline (q/q) | −5.3 mb/d 2Q26 / −3.4 mb/d 3Q26 / −2.0 mb/d 4Q26 | 2026 quarters | 2025 same quarter | IEA OMR Sep 2026 |
| Global observed inventory draw since war start | −507 mb (avg 2.8 mb/d) | war-to-date through Aug 2026 | pre-war | IEA OMR Sep 2026 |
| August inventory draw alone | −95 mb (3.1 mb/d) | August 2026 | pre-war | IEA OMR Sep 2026 |
| Global oil supply | 100.7 mb/d | 2026 average | 2025 | IEA OMR Sep 2026 |
| Supply y-o-y | −5.7 mb/d | 2026 | 2025 | IEA OMR Sep 2026 |
| Supply downgrade vs prior OMR | −1.3 mb/d | 2026 | August OMR | IEA OMR Sep 2026 |
Attribution: the −2.5 mb/d 2026 demand figure is mostly a model output, not a directly observed aggregate. The IEA explicitly attributes the deepening to "steep losses of petrochemical feedstocks and refined product supplies, along with higher fuel prices, notably for diesel." The August inventory draw of 3.1 mb/d is observed, and supports the demand-destruction read.
2. Refinery Capacity Loss vs Crude Supply Headlines — Reconciliation¶
This is the central reconciliation requested by the task.
2.1 Crude losses narrowed; product losses did not¶
Per IEA OMR Sep 2026:
| Channel | Loss vs February 2026 | Comment |
|---|---|---|
| Crude oil exports from Gulf | ~45% loss (narrowed from prior) | US military escort + STS bypass bringing flow back partially |
| Refined products + LPG from Gulf | ~60% loss (−3.7 mb/d) | Barely improved; refinery damage is structural |
| Gulf diesel/gasoil net exports (Aug) | 390 kb/d (just over ¼ pre-war) | Per IEA, lowest in modern record |
| Gulf + Russia combined diesel/gasoil exports | 1.6 mb/d below February (which was ~45% of global seaborne trade) | Headline product loss |
Mechanism: crude can reroute around Hormuz via STS in the Gulf of Oman and US military escort through the Strait. Diesel cannot reroute in the same way because the refining capacity itself is damaged — Kuwait's Al-Zour (615 kbd), Saudi export refineries, and Russian refining assets are offline or running below nameplate. Diesel demand in importing countries is constrained not by crude unavailability but by absence of the physical refining step.
2.2 EIA STEO framing (independent confirmation)¶
EIA STEO Sep 2026 confirms:
"We forecast inventories of distillate fuel oil—often sold as diesel—in the United States to fall below 100 million barrels in September and remain below the five-year (2021–2025) low through the end of 2026 and most of 2027. Distillate inventories fell below the five-year range in April, coinciding with high U.S. net exports of the fuel following the loss of large amounts of distillate supply from the Middle East, Russia, and China."
"We estimate U.S. average diesel crack spreads to exceed $2 per gallon from August through November, before decreasing steadily through mid-2027. This decrease assumes a return to normal tanker traffic through the Strait of Hormuz in the near term, allowing refineries in Saudia Arabia and Kuwait to increase exports of distillate to the global market."
The EIA explicitly conditions the eventual crack-spread normalization on "a return to normal tanker traffic through the Strait of Hormuz in the near term" — meaning product normalization requires crude normalization first, with an additional refinery-recovery lag.
2.3 Goldman Struyven framing (sell-side confirmation)¶
Per goldman-brent-120-escalation (already in KB): Daan Struyven (Goldman co-head Global Commodities Research) advised investors to "focus on rising natural gas and refined product prices, as supply shocks in those markets are larger than in crude." This is direct sell-side recognition that products are the binding constraint, not crude.
2.4 Reconciliation summary¶
- The headline crude narrative ("Hormuz partial reopening brings crude flows back") is true but misleading for products.
- Crude loss: ~45% of February baseline (narrowed).
- Product loss: ~60% of February baseline (largely unchanged).
- The marginal barrel of crude that is flowing through Hormuz/STS does not produce a marginal barrel of diesel at the same rate as pre-war Gulf refining.
- The 1.6 mb/d diesel/gasoil export loss exceeds the total supply downgrade in the IEA OMR (−1.3 mb/d vs prior report), which is structurally impossible unless offset by inventory draw or other-region refining surge. Both are happening (per IEA, "other regions … pushed to the limit"), but at the cost of record refining margins.
Net: the "crude supply easing" headlines understate the diesel reality. Products are the binding constraint, and they will remain so through 2027 absent either (a) Strait normalization + refinery restart OR (b) demand destruction large enough to clear the gap.
3. Price Elasticity Estimates (Academic)¶
These are model-based estimates, not observed volume changes. They quantify what the demand response should be for a given price increase, derived from historical data and econometric estimation. Use them to back out implied diesel demand destruction at current prices; do not treat them as observed.
3.1 Diesel — direct empirical estimates¶
| Study | Geography / Scope | Method | Short-run | Long-run | Source |
|---|---|---|---|---|---|
| De Borger & Mulalic (2012) | Denmark trucking (1980–2007) | Panel econometrics on firm-level fuel use | −0.13 | −0.22 | CBS Research Portal |
| Wadud (2016) | UK freight rigid trucks | Micro-data fuel consumption | ~−0.15 (rigid trucks); articulated trucks and LGVs not statistically significant | n/a in same form | Applied Energy paper |
| Ramli & Graham (2014) | UK total road diesel demand incl. passenger | Time-series econometrics | −0.11 to −0.16 | −0.21 to −0.30 | Transportation Research D paper |
| Winebrake et al. (2015) | US single-unit trucks (1980–2012) | Activity-based | Not statistically significant on truck activity | n/a | Transportation Research D paper |
Reading the studies together: the consensus short-run diesel demand elasticity is −0.11 to −0.16 for total road diesel demand (passenger + freight), with the trucking-specific subset on the more inelastic end (~−0.13). Winebrake's null result is important: truck activity (km) does not respond much to diesel prices, only truck fuel efficiency does. That means the demand response in 2026 is mostly via efficiency / load-factor / mode-shift (truck → rail) rather than via activity collapse. Long-run elasticities (−0.21 to −0.30) are roughly 2× the short-run, consistent with modal shift and fleet turnover.
3.2 Implied 2026 diesel demand destruction (back-of-envelope)¶
Apply the short-run range to current US retail diesel prices:
- Pre-war US retail diesel baseline: ~$3.50–4.00/gal
- 2026 forecast: $5.07/gal (+25% to +45% vs baseline)
- Apply SR elasticity −0.13 (De Borger trucking, conservative): −3.3% to −5.9% diesel volume decline (price-driven only)
- Apply SR elasticity −0.16 (Ramli total road diesel, upper bound): −4.0% to −7.2% diesel volume decline
Caveats:
- These are price-only estimates. They exclude supply-rationing effects (Pakistan −19%, distribution failures, cash constraints, strikes) which add a separate non-price destruction channel.
- Short-run elasticities underestimate the actual response when prices break expectations upward (i.e., when consumers believe prices will stay high — not yet documented in the studies above).
- The US is unusual in that its diesel market is partially insulated by Atlantic Basin refining surge + record net exports — the consumer price response may be smaller than implied by global diesel prices.
3.3 Gasoline — observed ≈ implied (sanity check)¶
The US gasoline −0.78% YoY (Section 1.1) is consistent with a short-run total petroleum elasticity in the −0.05 to −0.10 range applied to a ~10–15% gasoline price increase, but the relevant elasticities are higher for gasoline than for diesel because of EV substitution. The gasoline demand response is therefore not informative as a diesel proxy.
3.4 Why diesel elasticity matters less than it should¶
The academic literature consistently finds diesel is more inelastic than crude/aggregate oil. This is empirically correct but operationally misleading for the 2026 crisis:
- Short-run inelasticity holds when prices rise gradually.
- 2026 diesel prices have moved in a step (94% in a few months). Step-change price moves trigger larger responses than gradual moves.
- Diesel demand in developing economies (Pakistan, parts of Africa) is being rationed by liquidity and supply, not just by price elasticity.
- Industrial/agricultural diesel demand may be backed out by output reduction rather than captured in fuel-sales data — observed volume numbers understate true destruction.
Implication: the academic elasticities give a floor on diesel demand destruction; the actual 2026 destruction is plausibly higher.
4. Discovered Sources (Tier-Ranked)¶
This is the ranked discovery list. Sources already in KB (compiled/SOURCES.md last updated 2026-09-26) are not duplicated; only net-new sources surfaced by this discovery pass are listed.
Tier 1 — Institutional primary (authoritative, primary data)¶
| Source | Date | URL | Why relevant | Key claims |
|---|---|---|---|---|
| IEA Oil Market Report — September 2026 (full report HTML) | 2026-09-11 | https://www.iea.org/reports/oil-market-report-september-2026 | Direct primary text already in KB at compiled/sources/2026-09-11-iea-omr-september-2026.md; this URL provides the source-of-record text for verification |
Diesel $200/bbl (+94% pre-war); Gulf+Russia diesel/gasoil exports −1.6 mb/d vs Feb; Atlantic Basin refinery margins record; 2026 demand −2.5 mb/d (940 kbd deeper than Aug OMR) |
| EIA Short-Term Energy Outlook — Petroleum Products (Sep 2026) | 2026-09-09 | https://www.eia.gov/outlooks/steo/report/petro_prod.php | US distillate inventory + crack-spread forecast detail; not yet linked as a standalone source in compiled/SOURCES.md |
US distillate <100 mb Sep; below 5-yr low through end-2026 + most of 2027; diesel crack >$2/gal Aug–Nov 2026 |
| EIA Weekly Petroleum Status Report (week ending Sep 18, 2026) | 2026-09-24 (release) | https://www.eia.gov/petroleum/supply/weekly/ ; data: https://www.eia.gov/dnav/pet/pet_cons_wpsup_k_w.htm | Underlying weekly US gasoline product-supplied data; the −0.78% YoY four-week figure is calculable but not yet cited in KB | 4-wk avg gasoline product supplied 8.779 mb/d (−69 kb/d YoY) |
| EIA STEO — Crude Oil and Petroleum Products Price Outlook (Sep 2026) | 2026-09-09 | https://www.eia.gov/outlooks/steo/report/prices.php | US gasoline price forecast detail; pairs with distillate forecast | 2026 US regular gasoline $4.07/gal avg; 2027 $3.85/gal |
Tier 2 — Government/Statistical (national-level demand data)¶
| Source | Date | URL | Why relevant | Key claims |
|---|---|---|---|---|
| ProPakistani — Petrol/diesel sales August 2026 | 2026-09-02 | https://propakistani.pk/2026/09/02/petrol-diesel-sales-decline-sharply-in-august/ | Strongest single-country diesel demand-destruction data point; sources Pakistani OMC data | HSD sales Aug 2026 422 kt (−19% YoY, −32% MoM); diesel Rs 379/L (+36% YoY) |
| ProPakistani — Diesel stocks below 20-day supply | 2026-09-12 | https://propakistani.pk/2026/09/12/diesel-stocks-drop-below-20-day-supply-as-companies-fear-price-cuts/ | Inventory/rationing indicator | 12 of 20 OMCs below 20-day diesel cover |
| Pakistan Today — Petroleum sales Aug 2026 | 2026-09-03 | https://profit.pakistantoday.com.pk/2026/09/03/pakistans-petroleum-sales-fall-3percent-yoy-in-august-as-higher-prices-weigh-on-demand | Attribution nuance: transport strikes + monsoon + pull-forward confound the price-only read | August total petroleum sales −3% YoY |
| Petroleum Division, Govt. of Pakistan — Emergency diesel coordination | 2026-09-24 | https://petroleum.gov.pk/NewsDetail/MzYwM2QwMDQtMzY2YS00NWQ5LWI3NTgtN2I0YWJiYmVmMDMw | Official government source confirming emergency measures | Rail transport, depot stocks, OGRA control room activated |
| Press Information Department (PID), Govt. of Pakistan | 2026-09-24 | https://pid.gov.pk/site/press_detail/34046 | Official government press release on diesel supply measures | n/a (see source) |
| Reuters via MarketScreener — India August fuel demand | 2026-09-08 | https://www.marketscreener.com/news/india-s-august-fuel-demand-falls-month-on-month-to-lowest-level-in-two-years-ce785bd8dc8df420 | Counter-evidence: India diesel +6.8% YoY Aug 2026; total products −2.8% MoM (LPG/naphtha led) | India diesel resilience vs LPG/naphtha weakness |
Tier 2 — Sell-side / analyst commentary on diesel specifically¶
| Source | Date | URL | Why relevant | Key claims |
|---|---|---|---|---|
| Axios — Energy Dept hikes 2027 diesel price outlook | 2026-09-09 | https://www.axios.com/2026/09/09/energy-department-hikes-2027-diesel-price-outlook | Clean secondary cite for the EIA STEO +33¢ diesel revision | 2027 US diesel raised to $4.40/gal (+33¢) |
Tier 2 — Academic / empirical elasticity¶
| Source | Date | URL | Why relevant | Key claims |
|---|---|---|---|---|
| De Borger, B. & Mulalic, I. (2012) — The Determinants of Fuel Use in the Trucking Industry — Volume, Fleet and Fuel Type | 2012 (still cited) | https://research.cbs.dk/en/publications/the-determinants-of-fuel-use-in-the-trucking-industry-volume-flee/ | Direct empirical estimate of diesel demand elasticity for trucking | SR −0.13; LR −0.22 (Danish trucking panel 1980–2007) |
| Wadud, Z. (2016) — Diesel demand in the UK freight sector | Applied Energy | https://eprints.whiterose.ac.uk/93117/1/Diesel%20demand%20accepted%20manuscript.pdf | UK freight-specific diesel elasticity | Rigid trucks ~−0.15; articulated and LGVs not statistically significant |
| Ramli, S. & Graham, D. (2014) — The implied cost of carbon and the demand for road diesel fuel | Transportation Research D | https://www.sciencedirect.com/science/article/pii/S1361920913001351 | UK total road diesel demand elasticity with explicit SR/LR | SR −0.11 to −0.16; LR −0.21 to −0.30 |
| Winebrake, J. et al. (2015) — Fuel efficiency, freight activity, and industrial diesel demand | Transportation Research D | https://www.sciencedirect.com/science/article/pii/S1361920915000711 | US single-unit truck activity response to diesel prices | Activity response not statistically significant (efficiency does the work) |
Tier 3 — Market intel / commentary¶
| Source | Date | URL | Why relevant | Key claims |
|---|---|---|---|---|
| S&P Global — Africa Refined Products Short-Term Outlook | early 2026 | https://www.spglobal.com/content/dam/spglobal/ci/en/documents/general/west-africa/Africa-Refined-Products-Short-Term-Outlook.pdf | Regional demand outlook for refined products in Africa | 1.9% demand growth expected; modest diesel growth |
| S&P Global — Russian export ban strains African gasoil markets | 2026-07-30 | https://www.spglobal.com/energy/en/news-research/latest-news/crude-oil/073026-russian-export-ban-strains-african-gasoil-markets | Supply-side effect on African diesel markets | Russia export ban strains African gasoil supply |
5. Reconciliation: Observed vs Elasticity-Implied¶
| Region | Directly observed volume change (2026) | Implied by short-run elasticity | Reconciliation |
|---|---|---|---|
| US gasoline | −0.78% YoY (4-wk Sep 18, 2026) | SR −0.05 to −0.10 applied to ~10–15% price increase → −0.5% to −1.5% | Consistent; gasoline has higher SR elasticity due to EV substitution and discretionary driving |
| US diesel | Inventory <100 mb Sep 2026; export surge; no clean volume series | SR −0.13 applied to ~25–45% price increase → −3.3% to −5.9% | Inventory behavior consistent with multi-quarter demand/supply imbalance; volume measurement masked by Atlantic Basin refining surge + export pull |
| Pakistan diesel | −19% YoY Aug 2026 | SR −0.13 applied to +36% price increase → −4.7% | Observed is ~4× implied: rationing/distribution/strikes/cash constraints dominate over price elasticity. Pakistan is the canonical case of non-elasticity demand destruction |
| India diesel | +6.8% YoY Aug 2026 | SR −0.13 applied to diesel price increase → mild negative | Sign reversed: subsidized/regulated retail + economic growth + monsoon-agriculture demand overwhelm global diesel price signal |
| Global | IEA −2.5 mb/d 2026 demand (model output, partly observed via inventory) | n/a (total oil, not product-specific) | Globally consistent with both inventory draw and gradual price-driven adjustment, but the decomposition into products is not in the IEA headline |
Reading: the elasticity literature gives a price-driven lower bound on diesel demand destruction. Observed 2026 numbers are higher in Pakistan (where non-price rationing dominates) and effectively absent or reversed in India (where subsidies and growth dominate). The developed-market story (US, EU) lies between these poles — visible in inventories and prices, masked in volumes.
6. Three Candidate Updated "Big Three" Monitoring Questions — Diesel/Products-Centered¶
The current KB Big Three (Q1 Supply Disruption, Q2 Price Impact, Q3 Europe Exposure) was built when the binding constraint was crude. The September 2026 evidence (this discovery pass + the refined-products-as-shock-center concept) supports reframing around products. Below are three candidate replacement questions, diesel/products-centered, ranked by analytical priority.
Candidate 1 (RECOMMENDED) — Distillate Inventory Operational Floor¶
Question: Where are US, European, and Asian distillate inventories relative to operational minimums, and what is the implied multi-quarter diesel demand destruction required to avoid physical rationing?
- Why diesel/products-centered: distillate inventories are the cleanest, most observable signal of how tight the diesel market actually is (EIA STEO Sep 2026 forecasts US distillate below 5-yr low through end-2026 + most of 2027). The question forces direct quantitative monitoring of the product that is the binding constraint.
- Replaces / augments: Q1 Supply Disruption (currently focused on crude).
- Indicator set: EIA Weekly Petroleum Status Report (distillate inventories + net exports), IEA OMR monthly distillate balance, Asian distillate inventories (Singapore/Malaysia).
- Decision trigger: distillate inventories reach operational minimum (EIA ~100 mb threshold is the current marker; EU equivalent to be identified) → physical rationing risk.
Candidate 2 — Refinery Recovery Asymmetry¶
Question: Given that crude can reroute around Hormuz (via STS + US escort) but refining capacity is structurally damaged (Al-Zour, Saudi export refineries, Russian refineries), when does global diesel/gasoil supply normalize — and is the lag months, quarters, or years?
- Why diesel/products-centered: directly addresses the IEA framing that "crude losses narrowed to ~45% but refined product/LPG exports remain ~60% less than February." The question forces tracking of refinery restart timelines rather than just crude flow.
- Replaces / augments: Q2 Price Impact (currently focused on paper vs physical, which is now less informative than products vs crude).
- Indicator set: Kuwait Al-Zour restart timeline, Russian refinery utilization (per S&P/Argus), Saudi export refinery utilization, Atlantic Basin refining margins (EIA STEO forecast $2/gal+ Aug–Nov 2026).
- Decision trigger: diesel crack spread normalizes below $1/gal for >1 month → refinery recovery complete.
Candidate 3 — Diesel Demand Destruction Ceiling¶
Question: Given short-run diesel demand elasticities of −0.11 to −0.16 (Ramli & Graham 2014; De Borger & Mulalic 2012), how much diesel demand destruction is implied by current prices — and at what point does the political economy of rationing (Pakistan-style) become a contagion risk for other emerging markets?
- Why diesel/products-centered: directly answers "how much diesel demand clears at current prices?" using the empirical literature, and surfaces the non-elasticity demand destruction channel (Pakistan −19% YoY, far above any elasticity estimate).
- Replaces / augments: Q3 Europe Exposure (currently focused on European infrastructure, which is the wrong axis for the products-shock phase).
- Indicator set: emerging-market diesel sales (Pakistan OMC data, Indian PPAC monthly data, Brazilian ANP data, South African SAPIA data), retail diesel price changes by region, transport strike frequency, emergency fuel-import tenders (e.g., Pakistan OGRA control room).
- Decision trigger: >2 emerging markets report double-digit YoY diesel volume declines with explicit emergency government response → global demand destruction exceeds the elasticity-implied path and signals broader contagion.
Note on prioritization: Candidate 1 is the highest-leverage because distillate inventory is the cleanest, fastest-updating, and most decision-relevant product signal. Candidate 2 is medium-leverage (refinery restart is slow-moving). Candidate 3 is high-value for emerging-market contagion risk but data-poorer (lower-frequency, less standardized).
7. Gaps and Caveats¶
Gaps not closed by this discovery pass:
- No clean weekly volume data for US or EU diesel. EIA publishes product-supplied for gasoline on a weekly basis but the diesel equivalent is buried in monthly STEO tables and not as current.
- Eurostat / EU diesel consumption for 2026 — web search provider errors prevented direct retrieval. Recommended follow-up: fetch Eurostat
nrg_cb_oilandDS-045352(oil supply by product) datasets for H2 2026. - Recent (2024–2025) academic re-estimates of diesel elasticity under step-change price regimes — the literature I retrieved is dominated by 2012–2016 estimates calibrated to gradual price moves. A 2024–2026 re-estimate under step-change conditions would materially improve the elasticity-based forecast.
- Decomposition of IEA's −2.5 mb/d 2026 demand figure into products (gasoline, diesel, jet, naphtha, LPG) — IEA reports the headline but not a clean product split. The OMR supplementary tables may provide it; worth a follow-up fetch.
- Country-level diesel demand data for emerging markets beyond Pakistan and India (Nigeria, Brazil, Bangladesh, Sri Lanka) — all web searches for these failed during this pass.
Attribution/causation caveats:
- Pakistan −19% YoY is partly rationing/distribution-driven, not purely price-driven. Treat the elasticity-implied decomposition as illustrative, not causal.
- India +6.8% YoY diesel growth reflects partial administered pricing + economic growth — it is not a refutation of the diesel-shock thesis, only of the global-uniformity assumption.
- US gasoline −0.78% YoY is small enough to be inside seasonal noise. Do not over-interpret.
- The IEA −2.5 mb/d global demand figure is a model output, not a directly measured total.
8. Source Confidence and Tier Summary¶
| Tier | Count | Sources |
|---|---|---|
| Tier 1 — Institutional primary | 4 | IEA OMR Sep 2026; EIA STEO Sep 2026 (petroleum products + prices); EIA Weekly Petroleum Status Report Sep 18 2026 |
| Tier 2 — Government/Statistical | 6 | ProPakistani (×2), Pakistan Today, Petroleum Division Pakistan, PID Pakistan, Reuters/MarketScreener India |
| Tier 2 — Sell-side/analyst | 1 | Axios Sep 9 EIA coverage |
| Tier 2 — Academic | 4 | De Borger & Mulalic 2012; Wadud 2016; Ramli & Graham 2014; Winebrake et al. 2015 |
| Tier 3 — Market intel | 2 | S&P Global Africa outlook; S&P Global Russian export ban coverage |
| Total net-new | 17 | (excluding 5 sources already in KB: IEA OMR Sep 2026 file, EIA STEO Sep 2026 file, Goldman Brent 120 file, GasBuddy De Haan, CSIS) |
Quality gate status: minimum 3 sources satisfied; per-source tier, justification, key claims, and URLs all populated. READY for operator review.
9. Suggested Ingest Path (For Operator Approval)¶
If approved, the following would flow into 1.1-ingest:
- Tier 1 batch: EIA STEO Sep 2026 (petroleum products + prices as separate sources); IEA OMR Sep 2026 (already in KB as file, may want full HTML for verification); EIA Weekly Petroleum Status Report Sep 18 2026 (weekly data capture).
- Tier 2 batch: ProPakistani Aug diesel + Sep stocks; PID/Petroleum Division Sep 24 emergency measures; Reuters India Aug diesel; Axios EIA diesel outlook; 4 academic elasticity papers.
- Tier 3 batch: S&P Global Africa outlook + Russian ban coverage.
Not recommended for ingest:
- The academic elasticity papers are pre-2020; recommend treating as reference baseline in a research note rather than as standalone sources in the daily source stream.
Discovered 2026-09-29 by carson subagent for kb-full-ingest / 0-discover. Existing DONE file (2026-06-12) replaced with this run's timestamp.
STATUS: COMPLETE — READY for operator review, NOT ingested