Per the Aug 2026 Pakistan OMC data (2026-09-02-propakistani-pakistan-august-2026-petroleum-sales, 2026-09-03-pakistan-today-pakistan-petroleum-sales-august), furnace oil (RFO / residual fuel oil) sales rose more than 4× YoY (and +26% MoM) to approximately 98–107 kt in August 2026 — the opposite direction from the −19% YoY HSD collapse in the same data set.

Definition

Pakistan RFO furnace-oil power substitution is the mechanism by which Pakistan's power-sector fuel dispatch shifted from RLNG (regasified LNG, the marginal fuel in normal operations) to residual fuel oil (RFO, the alternative fuel) during the August 2026 month, in response to delayed RLNG cargo deliveries caused by the Hormuz disruption. The result was a more-than-4× YoY increase in furnace oil sales to Pakistani power plants, even as transport-sector diesel demand collapsed.

Headline Numbers

Source Aug 2026 RFO YoY change MoM change
2026-09-02-propakistani-pakistan-august-2026-petroleum-sales (Topline) not separately quantified, but rose more than 4× YoY and +26% MoM +4×+ YoY +26% MoM
2026-09-03-pakistan-today-pakistan-petroleum-sales-august (AKD) ~98,000 tonnes +5× YoY (i.e., roughly 500% of Aug 2025) consistent with Topline

The two sources triangulate to ~98–107 kt of RFO sold in Aug 2026, of which the majority went to power-sector dispatch per AKD's attribution. The two attributions differ slightly (Topline says "4× YoY", AKD says "5× YoY") — this is consistent with rounding differences across two secondary broker reports (Topline Securities vs AKD Securities), both reporting on the same OMC primary disclosures.

Why the Mechanism Operates

Pakistan's power sector has fuel-flexibility by design. Many Pakistani power plants (especially the IPP fleet built in the 1990s–2010s) can burn either natural gas or RFO, with fuel choice driven by relative cost and availability of delivered fuel. In normal operations, RLNG (regasified LNG imported via Qatar / global LNG spot) is the marginal fuel because:

  1. RLNG has lower variable cost than RFO (per BTU).
  2. RLNG burns cleaner, avoiding Pakistan's SOx/NOx emissions constraints on coastal plants.
  3. RLNG is contracted on long-term agreements with Pakistan's two LNG terminals (Engro Elengy Terminal and Pakistan GasPort) plus spot cargoes.

When RLNG cargoes are delayed (as happened in mid-2026 because the Hormuz disruption lengthened LNG voyage times and raised shipping costs), Pakistani power plants face a dispatch problem: they must either (a) run on reduced load, or (b) switch to RFO, which is more expensive but available from Pakistan's domestic refineries (Pakistan Refinery Limited, Attock Refinery, Cnergyico/Byco) and from imports via PSO.

The Aug 2026 surge in RFO sales reflects option (b) at scale.

Why It Matters

The RFO substitution is analytically important for three reasons:

1. It Cushions the Power Sector

Pakistan avoided load-shedding escalation in August 2026 by switching to RFO. The fiscal cost is high (RFO is more expensive than RLNG per BTU; the subsidy bill rises), but the physical availability of electricity was maintained. Other EMs without RFO capacity (e.g., Egypt, Bangladesh, Vietnam) faced worse load-shedding in the same period because they lack the fuel-flexibility option.

2. It Is the Inverse of the Diesel Signal

Within the same Pakistan data set:
- Transport diesel (HSD): −19% YoY / −32% MoM → demand destruction
- Power-sector RFO: +4× YoY / +26% MoM → demand substitution

The total petroleum sales number (−3% YoY) is therefore misleading as a measure of distress. The HSD decline shows transport-sector stress; the RFO surge shows power-sector fuel-mix substitution. Aggregating them into "total petroleum sales" hides both signals. Distillate-aggregate framing is insufficient for EM analysis. Product-level disaggregation is required.

3. It Demonstrates a New Mechanism Not in Pre-2026 Models

Standard pre-2020 oil-shock models assume:
- A global crude price shock → retail fuel prices rise → demand falls uniformly across products.
- Power-sector demand for liquid fuels (RFO, diesel for gensets) is price-elastic and small relative to transport-sector demand.

The Pakistan RFO surge breaks both assumptions:
- Power-sector demand for liquid fuels rose despite the price shock, because the substitute fuel (RLNG) was unavailable rather than expensive.
- The substitution is physical availability-driven, not price-driven.

Reconciliation With the Pakistan Cluster

Source What it says about RFO
2026-09-02-propakistani-pakistan-august-2026-petroleum-sales (Topline) RFO rose more than 4× YoY; +26% MoM; attribution: RLNG shortage
2026-09-03-pakistan-today-pakistan-petroleum-sales-august (AKD) RFO 98 kt in Aug 2026, more than 5× YoY; explicit "RLNG cargo shortage redirected power-sector dispatch to RFO"
2026-09-12-propakistani-pakistan-omc-diesel-stocks (OCAC) (No direct RFO data — focuses on diesel inventory)
2026-09-03-pakistan-petroleum-division-pricing-committee-september-3-2026 (No RFO data — focuses on diesel pricing framework)
2026-09-24-pakistan-pid-fuel-relief-scheme-emergency-coordination-september-24-2026 (No RFO data — focuses on emergency coordination)
pakistan-demand-destruction-emergency-2026-09-18 (Dawn package) (No RFO data — focuses on austerity package)

The RFO signal is concentrated in the two OMC monthly sales reports (sources 4 and 6). It is not mentioned in the policy/government response sources (sources 7, 8), which focus on diesel pricing and supply coordination. This is consistent with the RFO surge being a dispatch-level operational adjustment rather than a policy-driven outcome.

Implications

For the KB's Demand Destruction Framework

The RFO surge does NOT count as demand destruction in the demand-destruction-dual-risk framework. It is a fuel-substitution signal: power-sector demand for liquid fuels rose, but the underlying electricity demand did not change. The RFO growth offsets some of the HSD collapse in product-level terms but does not reduce aggregate petroleum consumption as much as the headline HSD −19% YoY number might suggest. The product-mix shift matters for refining margins and refinery utilization (more RFO demand → higher bottom-of-barrel run rates) but not for total barrels consumed.

For Cross-Country Comparisons

Pakistan's RFO flex is not replicable in most EMs. Countries with limited power-sector fuel-flexibility (e.g., gas-only generation in Bangladesh; nuclear-heavy France; renewables-heavy Brazil) cannot substitute RFO for unavailable gas. The Pakistan RFO surge therefore exaggerates Pakistan's apparent resilience in total-petroleum terms relative to EMs without the same flex. Cross-country comparisons must control for power-sector fuel mix.

For 2027 Forecasting

If the Hormuz disruption persists and RLNG cargoes remain delayed, Pakistan's RFO demand could stay elevated through Q4 2026 and into 2027. This would:
- Tighten the global residual fuel oil market (Pakistan is now a marginal buyer)
- Raise Pakistan's RFO import bill (subsidy pressure)
- Maintain Pakistani power-sector output (avoid load-shedding) at the cost of higher fiscal subsidy

Why This Concept Exists

This concept isolates the RFO power-substitution mechanism as a first-class analytical object because:

  1. The Pakistan data set cannot be properly interpreted without understanding the inverse RFO signal. The total-petroleum −3% YoY headline obscures both the HSD destruction and the RFO substitution.
  2. The mechanism is novel in 2026 oil-shock literature. Pre-2020 oil-shock models assume uniform cross-product demand response; the 2026 Pakistan RFO surge demonstrates product-specific substitution.
  3. The mechanism is cross-country relevant. Any EM with RFO-capable power plants facing LNG/RLNG shortage could exhibit the same pattern. The KB should be alert to similar signals in Bangladesh, Egypt, Vietnam, and Pakistan's neighbors.

Caveats

  • Single-country, single-month evidence. Pakistan Aug 2026 is the only data point. The mechanism cannot be confirmed across other EMs without their RFO data.
  • Topline vs AKD attribution gap. Topline says 4× YoY; AKD says 5× YoY. The discrepancy is small (rounding + different reporting cutoffs) but should be flagged. Downstream users should cite both.
  • Power-sector attribution is not direct. Both Topline and AKD attribute the RFO surge to power-sector RLNG substitution by inference, not by direct power-plant dispatch data. Pakistan's NPCC (National Power Control Company) does not publish disaggregated monthly fuel-mix data publicly.
  • Refinery-side capacity. Pakistan's domestic refineries cannot supply 98–107 kt of RFO from inventory alone. The implied import volume (likely 50–80% of the surge) must come from PSO imports, which adds to the FX-reserve drawdown that triggered the Sep 17-18 austerity package. The RFO surge is therefore partially causal of the broader Pakistan crisis, not just a parallel signal.

Referenced from: 2026-09-02-propakistani-pakistan-august-2026-petroleum-sales