What can refinery runs and product supplied tell a researcher?

Are higher refinery runs the same as stronger demand?

No. Higher processing activity can build product inventories, serve exports or meet domestic use. Lower runs can reflect maintenance or unavailable capacity rather than weaker end-user demand. Begin by separating the crude-input question from the finished-product disposition question. A single headline about refinery activity cannot identify where the resulting products went or why the processing decision changed.

EIA's glossary defines refinery inputs and capacity utilization, while its refining explainer describes transformation into different products. These are distinct physical concepts. Keep total feedstock separate from crude-only input, and retain the exact utilization denominator. A changing estimate of operable capacity can alter a utilization ratio even without the same proportional change in measured throughput.

Explain volumetric processing gain correctly

Refining can produce a larger measured liquid volume than its input volume because outputs have different densities. That volumetric gain is not creation of mass, energy or investment profit. Record whether a balance uses barrels, tonnes or energy equivalents before interpreting an output-to-input ratio. A unit change can alter the appearance of gain while describing the same physical transformation.

In a synthetic illustration, 100 volume units of feedstock become 103 volume units of products under a simplified boundary. The volume gain is three units, or 3% of input. This is not a plant engineering model and omits detailed fuel and loss accounting. It demonstrates only why adding product barrels cannot be interpreted as a mass-conservation violation or financial margin.

Reconcile an invented product disposition

Use a separate synthetic finished-product balance with all quantities in million barrels for the same week. Production is 60, imports are eight, exports are ten and inventory rises by three. Under a deliberately simplified identity with no other terms, product supplied is 60 plus eight minus ten minus three, or 55 million barrels.

An inventory build reduces this implied disposition relative to supply available after exports. Reversing the stock-change sign would incorrectly raise the estimate to 61. The table checks a fictional balance, not a reported weekly release. An actual agency table may contain transfers, adjustments or differently defined components, so reproduce its published identity rather than forcing every release into this simplified equation.

Keep product supplied within its evidence boundary

EIA's table definitions describe product supplied as a calculation from supply and disposition components. It is useful as a consumption-related proxy, but it is not a direct count of every final purchase. Timing, reporting scope and revisions can affect a weekly estimate. Avoid translating a one-week movement into an exact change in retail demand without additional supporting evidence.

Compare compatible products and periods. A gasoline component is not interchangeable with all petroleum products; a four-week average has a different question from a single-week estimate. Preserve forecast, weekly estimate and revised monthly data as separate vintages where relevant. A smoother series can aid interpretation while concealing short-lived events and cannot retroactively establish the original information available to a researcher.

Separate physical yield from refinery economics

A product yield share says how much of an output basket falls into a category under a declared convention. Profitability requires product realizations, feedstock acquisition, energy, operating costs and other terms. An indicative price spread does not contain that complete ledger. Never multiply a simple yield number by a benchmark gap and present the result as an observed refinery margin.

Product mix also matters for inventories. High total throughput can coexist with tight availability in one product if the output mix and demand mix differ. Investigate the relevant cut, specifications and region rather than assuming all refined barrels substitute for one another. Seasonal fuel requirements and logistics can change the interpretation without implying that a processor has changed its overall operational efficiency.

What belongs in a useful refinery research note?

Ask which inputs are measured, which capacity denominator is used, and where products are disposed. Reconcile stock changes with trade and production before naming a demand explanation. Ask whether a claimed improvement concerns physical throughput, product availability or financial profitability. These are different questions and require different evidence; a note should not move silently from one to another.

Failure modes include treating processing gain as profit, reversing stock-change signs, mixing daily rates with weekly quantities, and reading a national balance as local availability. Retain assumptions, exact source definitions and unresolved adjustments. This guide adds no observed refinery dataset, instrument strategy, fitted forecast or provider capability; its numerical examples are explicitly synthetic checks of interpretation.

Keep in the decision record

• Throughput, product disposition and profitability require different ledgers.
• Volumetric processing gain is not mass creation or profit.
• Use the source table's actual balance identity and stock-change sign.

Source notes

1. EIA Glossary, R - refinery input, output, capacity utilization and gains; public definitions reviewed
2. EIA - Refining crude oil: the refining process; public explainer reviewed
3. EIA - Weekly petroleum supply/disposition table definitions; public full text reviewed

Source dates and scope matter. Historical inventory observations are retained snapshots; forecasts and synthetic curve assumptions are labelled separately. No investment outcome or product coverage is implied.

Continue the analysis