FRED — Industrial Production: Farm Machinery and Equipment

It is the only monthly manufacturing-output series dedicated to US farm machinery in Datadory's catalog of 1,744 datasets across 159 viable industries. FRED — Industrial Production: Farm Machinery and Equipment is FRED's distribution of series IPG33311S from the Federal Reserve Board's G.17 programme, officially titled Industrial Production: Manufacturing: Durable Goods: Agricultural Implement (NAICS = 33311). The title's older 'agricultural implement' wording and the 'farm machinery and equipment' label describe the same thing: establishments located in the United States whose primary output is farm implements.

What separates it from headline industrial production is scope. Where the market-group indexes aggregate entire durable-goods sectors, this one isolates a single four-digit NAICS manufacturing industry, so a move in the series is a move in farm-machinery factory output — not a move in aerospace or autos bleeding through a composite. The index measures real output, seasonally adjusted, on a 2017=100 base, compiled by the Board from physical product data and labour/input estimates. Observations run from January 1972 (76.3386) to the latest month (July 2026: 109.6809) — 655 months of a single, revision-tracked series. That record scores 8/10 against a catalog mean of 7.81. Get a sample of this dataset cut to the date range your model needs.

The full verified window ships as 655 rows — roughly 13 KB once serialized, small enough to hold entirely in memory. Every row carries exactly two populated values:

# Two fields, one row per month — seasonally adjusted index, 2017=100
observation_date=2026-04-01  IPG33311S=110.3253
observation_date=2026-05-01  IPG33311S=108.5661
observation_date=2026-06-01  IPG33311S=106.6745
observation_date=2026-07-01  IPG33311S=109.6809

# The anchor point that opens the panel
observation_date=1972-01-01  IPG33311S=76.3386

# Series identity carried alongside the values
series_id=IPG33311S  naics=33311  adjustment=seasonally_adjusted  base=2017=100  frequency=monthly

Read the recent block as a quarter of movement: April 2026 at 110.3253 eases to 108.5661 in May and 106.6745 in June before rebounding to 109.6809 in July — a three-month dip-and-recover inside a band the industry has occupied since clearing its pre-2017 range. Set those against the 1972 opening value of 76.3386 and the base-year arithmetic does the talking: fifty-four years of real growth in US implement output amount to roughly a third above the 2017 average, with the whole story living in the swings between. Values carry to four decimal places because the Board's compilation keeps them there; treat the trailing digits as compilation precision, not forecasting signal.

Two fields, definitions verified against the live series during the August 2026 research pass — which is the entire appeal for pipeline work. observation_date stamps each row with the first day of its observation month in ISO format, so a July figure arrives keyed 2026-07-01 and sorts correctly under any string or date parser without reformatting. IPG33311S holds the seasonally adjusted industrial production index value itself, expressed on the 2017=100 base the Board rebased its entire G.17 system onto.

There is no third field to negotiate: no preliminary-versus-final flag, no confidence interval, no unit column — the unit is the index definition and the seasonal adjustment is a property of the whole series rather than of individual rows. What sits outside the two columns above — revision history per observation, the unadjusted companion series, and the capacity-utilisation and new-orders siblings for the same NAICS code — folds under additional fields on request and gets confirmed when your sample is cut rather than promised here.

Geography - the United States as one national aggregate: real output of establishments located in the United States, with US territories excluded. There is no state table, no regional split and no company detail — a John Deere plant and a shortline manufacturer contribute to the same index cell. If the question needs sub-national manufacturing resolution, this is not the instrument.

Temporal - monthly observations from January 1972 to the current month: 655 observations as of August 2026, each appended by the G.17 release roughly two to three weeks after month end. Earlier months can be revised as the Board re-estimates, so a snapshot date belongs in any reproducible workflow.

Granularity - one national monthly index for one four-digit NAICS industry (33311). Seasonal adjustment is built in, which strips harvest-driven factory-calendar noise and leaves the cycle you actually want to model; the trade-off is that you cannot un-adjust it from this record alone.

API, files, or your warehouse. Daily, weekly, or hourly.

Name the window when you request the sample — the full 1972-to-current panel, or a decade sliced out for a specific backtest. The sample ships first either way, and the ongoing feed lands on whatever cadence your models need, refreshing as the Board revises the underlying release. The field dictionary above travels with every delivery, so validation is a diff against the sample rather than a project.

A single-series macro index earns its keep in five specific jobs:

• Demand-side feature engineering — join monthly IPG33311S against dealer inventories, auction prices or parts demand; the data scientists use cases page walks the workflow.
• Cycle timing — six decades of monthly frequency make it a core read for timing ag-equipment exposure through farm-income cycles; see investors quants use cases.
• Report anchoring — put a Federal Reserve number behind US farm-implement production claims in client decks and industry studies (market researchers use cases).
• Publication sourcing — a citable official statistic with the citation string supplied (journalists academics use cases).
• Scheduled ingestion — one stable identifier and a fixed schema suit cron-based pipelines (developers builders use cases).

Two personas should look elsewhere: sales teams get no account-level signal from a national index, and e-commerce operations have no assortment or pricing read here. Start from the agricultural farm machinery data hub, then the ranked shortlist for the vertical.

The slice splits into stocks and flows. This record is a flow: monthly US manufacturing output, measured as it happens. Everything else in the family is stock — machines accumulated in fleets over decades. FAOSTAT – Agricultural Machinery counts tractor, harvester and milking-machine stocks across roughly 258 countries but runs annually and stops at 2009; World Bank WDI – Agricultural machinery, tractors mirrors tractor counts across 292 areas from 1961; Eurostat – Agriculture & Farm Structure database counts EU machinery by NUTS 2 region in survey years only. CEMA – Business Barometer measures sentiment, not output. For stock-versus-flow questions the head-to-head sits on the vs FAOSTAT comparison page.

Within the Federal Reserve family, IPG33311S is the narrowest useful cut: pair it with capacity utilisation and new orders for the same NAICS code to cover utilization pressure and forward orders, and the US farm-machinery cycle reads end to end. Marketplace records — TractorHouse, MachineryPete, TractorData — answer what individual machines sell for, which this index deliberately does not.

Three things, stated upfront.

First, respect the grain. This is one national monthly index — the right tool for modelling the US farm-machinery cycle, the wrong tool for anything needing geography below the national border, company attribution or machine-level prices. Say what you need in the request; if a fleet-stock or transaction-price question is the real target, the sample gets routed to the FAOSTAT or marketplace records instead.

Second, revisions are part of the deal. The Board re-estimates the G.17 as fuller source data arrive, so historical values move — capture a snapshot date if reproducibility matters, and expect the most recent few months to be the softest.

Third, mind the title. The official G.17 title uses 'Agricultural Implement', the older wording for the same NAICS 33311 scope this page calls farm machinery and equipment. Both names map to the same series; confirm the exact industry scope against the Board's documentation before citing verbatim. All three points confirm at sampling, which is also where the field list locks down for your pipeline.

Provenance note — compiled by the Board of Governors of the Federal Reserve System under the G.17 industrial production and capacity utilisation programme, from physical product data and Fed labour/input estimates for establishments located in the United States, and distributed through FRED, the Federal Reserve Bank of St. Louis' economic-data platform. One methodology across the whole 1972-present window, with the 2017=100 rebasing applied uniformly.

Methodology note — the index measures real output, seasonally adjusted, meaning harvest-timed factory-calendar patterns are removed by construction. Revisions follow the Board's re-estimation cycle; reconcile year-over-year moves against revised history rather than first prints, or noise becomes trend.

Completeness note — Datadory scores this record 8/10 against a catalog mean of 7.81 across 1,744 datasets. The two-field dictionary above is complete for the verified window; per-observation revision history, the unadjusted companion and the utilisation/orders siblings fold under additional fields on request.

Where to go next — pair the flow measured here with the stocks counted elsewhere in the slice, and with the capacity utilization glossary entry and seasonal adjustment glossary entry for the concepts doing the work underneath the numbers.