Manufacturing Costs

Steel rose 18.66% while fasteners fell 5.69%. A common purchasing escalator missed the split.

January to July 2026 steel rose 18.66%, fasteners fell 5.69% and machinery rose 3.53%. A disclosed 50/30/20 machinery/fastener/steel basket rises 3.79% under full-pass-through assumptions.

Editorial evidence cutoff: September 9, 2026. Published September 29, 2026. Observation periods are stated throughout; older figures are retrospective evidence.

From January to July 2026, the steel mill products price index rose 18.66%. The index for bolts, nuts, screws, rivets and washers fell 5.69%. Machinery and equipment rose 3.53%. All three changes use the same beginning and ending months.

For September purchasing reviews, the divergence challenges a convenient assumption: that a movement in steel supplies the right adjustment for every metal-containing component or capital purchase. The record does not support that shortcut. It also does not reveal the supplier margins or cost transmission needed to replace it with a universal rule.

Three baskets took three different paths

The calculation compares each BLS index with its own January value. Steel rose substantially, machinery increased more modestly and fasteners declined. Rebase January to 100 and their July readings become 118.66, 103.53 and 94.31 respectively.

That presentation avoids comparing raw index levels as if they were prices in the same units. A higher raw index number does not establish that fasteners cost more than machinery or that one product category is intrinsically more expensive. The index base is a reference for change within the series.

The matched dates are just as important. A supplier's annual comparison and a buyer's recent monthly comparison can produce different conclusions. This investigation fixes January and July for all three categories so that the headline contrast is not created by selecting a different horizon for each one.

The broad machinery number cannot price a specific machine

Machinery and equipment is a broad product basket. Its 3.53% increase does not establish the change in a specific replacement machine's delivered price. Engineering scope, controls, installation, tooling and other terms can differ from one quotation to another.

The machinery index can still serve as context. If a supplier requests a larger increase, it invites a documented explanation of the relevant product and costs. It does not prove the excess is padding. An aggregate benchmark is not a ceiling on every individual transaction within a changing market.

For a capital decision in September, the comparison should therefore sit beside actual comparable quotations. Holding specifications and included services constant is essential. Otherwise, a change in what is being purchased can masquerade as inflation, regardless of how accurately the national index has been calculated.

Fasteners do not provide a clean test of steel pass-through

Steel can be relevant to the cost of many fasteners, but the two PPIs do not cover the same product basket or stage of production. Fastener selling prices also reflect processing, specifications and the market conditions of the products represented in that series.

The observed divergence does not identify a delayed steel increase waiting to arrive. Demonstrating a stable lag would require a separate analysis, appropriate controls and evidence that the relationship holds beyond the period used to choose it. A plausible production story is not a measured forecast.

Nor can the difference be labeled an assembler margin. The indexes do not supply an identical input quantity, matching output price and full conversion cost for the same firms. The useful result is a warning about benchmark selection, not proof that one group captured a particular amount of profit.

A disclosed budget shows the consequence

Consider a hypothetical $100,000 January purchasing budget consisting of separate purchases: 50% machinery, 30% fasteners and 20% steel mill products. Applying each observed index change fully to its own fixed base amount produces $103,788.60 in July, an increase of about 3.79%.

Applying the steel change to every dollar would instead imply an 18.66% increase. The difference is not a measured saving available to the buyer. It is the effect of choosing a benchmark that does not match most of the illustrative basket.

The example treats the buckets as separate purchases. It does not add steel contained inside a machine to the machine's full price again. That distinction prevents double-counting an input and the finished item that already includes it. Quantities, contract timing and pass-through are assumed fixed solely to make the arithmetic visible.

The weights are a business fact, not a convenient default

Another buyer with a much larger steel share would produce a different weighted result. A buyer purchasing mostly fasteners could see the benchmark calculation move downward. Publishing one blended adjustment without its purchase weights would hide the assumption that largely determines the answer.

The correct weights also depend on the decision. A spending forecast may need expected quantities and new projects. A price-variance review may hold quantities constant to isolate price changes. Those are different exercises and should not be combined under an undefined inflation allowance.

A September review should therefore state the purchase categories, base period, cost shares and treatment of outstanding commitments. The national indexes then become transparent inputs to the calculation. They cannot supply the company's purchasing structure merely because a weighted average is easy to compute.

Check the claimed exposure before arguing about the percentage

The steel history, fastener history and machinery history allow the three paths to be inspected separately. The relevant question is which one fits the actual order and what share of its price remains exposed.

A fixed material commitment may differ from a purchase yet to be placed. A distributor's quote can contain costs outside a domestic producer-price measure. Those differences deserve explanation from the transaction itself rather than assumptions about how a national index must flow through it.

The January-to-July record makes a defensible point without predicting the next move: steel, fasteners and machinery did not share one price trajectory. September procurement becomes more accurate when the adjustment follows the product and the period, instead of allowing the most dramatic material headline to stand in for the entire budget.

Sources and calculation

The analysis uses BLS machinery WPU11, fasteners WPU1081 and steel mill products WPU1017, January to July 2026. The version available September 9 is retained. July PPI was released August 13; August PPI was not available by the cutoff. The basket is a disclosed fixed-weight illustration, not an observed buyer cost index.

Sources and evidence

Evidence period: January to July 2026; seven matched observations. The frozen evidence record lists the source files and verified hashes available September 9, 2026. Source revision: 5e4fb7726c3d40060c0151c086c903baae856cab. Later live-data updates do not alter the historical evidence in this article.

fred.stlouisfed.org/series/WPU11

fred.stlouisfed.org/series/WPU1081

fred.stlouisfed.org/series/WPU1017

Published 2026-09-29.