US Macro

Industrial Production Index

The Industrial Production Index measures the real output of US factories, mines, and electric/gas utilities. It's the most direct read on the goods-producing side of the US economy and a critical component of NBER's recession-dating methodology. Note: this dashboard tile is labeled 'ISM Manufacturing PMI' historically, but the actual data source is the Federal Reserve's Industrial Production Index (FRED INDPRO) — the ISM PMI series was discontinued from free FRED publication in 2016.

May 14, 2026INDPROmanufacturing · us-macro · industrial-production · leading-indicator

ISM_PMI

—

The most direct measure of US factory output. When economists talk about "is the economy producing more or less stuff than last month," industrial production is the answer. It's not the headline number most casual observers watch (that's GDP, or unemployment, or CPI), but for serious cyclical analysis — for NBER recession dating, for Fed policy calibration, for understanding goods-producing sector dynamics — industrial production is one of the irreplaceable inputs.

ISM_PMI

—

What it measures

The Industrial Production Index (FRED INDPRO) is published monthly by the Federal Reserve's G.17 statistical release:

Important note on this dashboard tile: it's labeled "ISM Manufacturing PMI" for historical reasons (that was the original intended indicator when the dashboard was scoped), but the actual underlying data is the Federal Reserve's Industrial Production Index. The ISM PMI series was discontinued from free FRED republication in 2016; INDPRO is the closest free public alternative and is published by the Federal Reserve directly.

The release schedule: approximately the 17th of each month at 9:15 AM ET as part of the Federal Reserve's G.17 Industrial Production statistical release. The release also includes capacity utilization (the percentage of total industrial capacity that's currently in use) — a complementary indicator useful for assessing whether the goods-producing sector has slack or is overheating.

Recent readings (mid-2025) have been in the 103-104 range (against the 2017 = 100 baseline), reflecting a manufacturing sector that has roughly recovered to pre-COVID levels but hasn't grown meaningfully above them.

Why it matters

Two angles.

The NBER-recession-dating angle. Industrial Production is one of NBER's six primary recession-dating indicators (alongside real GDP, real GDI, real personal income excluding transfers, nonfarm employment, real consumer spending, and household survey employment). NBER weights all of these in determining when a recession begins and ends. Sustained declines in industrial production (3-6 consecutive months) are one of the cleanest recession-precursor signals; the 2008 recession saw industrial production decline for 12 consecutive months before stabilizing. The 2020 COVID recession saw industrial production decline by approximately 17% peak-to-trough in 2-3 months, an extraordinary speed of contraction.

The Fed-policy-and-monetary-transmission angle. Manufacturing is one of the most interest-rate-sensitive sectors of the US economy. When Fed tightening begins working through the economy, the goods-producing sector typically declines before services. The 2022-2023 Fed hiking cycle produced visible industrial-production weakness in 2023 (industrial production declined by approximately 1.5-2% from peak through 2023) even as services-sector activity remained strong. This goods-vs-services divergence is now a routine feature of US cyclical analysis.

What moves it, and what it moves

Moves industrial production:

Goods consumption demand. When consumer spending shifts toward goods (as in 2020-2021 with COVID-era spending rotation), manufacturers respond by increasing production.
Business investment cycles. Capital expenditures on equipment and structures drive manufacturing of those goods (turbines, semiconductors, construction equipment).
Inventory cycles. When wholesale and retail inventories fall, manufacturers ramp up production to refill them. When inventories overshoot, production falls.
Interest rate sensitivity. Higher rates reduce demand for housing-related goods (appliances, furniture, building materials), autos (durable goods), and capital equipment.
Energy and commodity prices. Mining output specifically (oil, gas, coal, minerals) responds directly to commodity prices.
The US dollar. A stronger DXY hurts US export competitiveness and thus manufacturing demand.

Industrial production moves:

Manufacturing employment (with 1-3 month lag).
Construction-equipment and capital-goods stocks.
Industrial commodity prices (copper, steel, lumber).
Transportation sector (rail freight, trucking) — moves goods that industrial production creates.
Federal Reserve policy considerations (industrial weakness supports cut expectations).
Recession probability models that include industrial production as an input.

A worked example: the 2020 COVID collapse and recovery

US Industrial Production entered 2020 at approximately 109 (vs the 2017 baseline of 100). The COVID crash:

March 2020: -4.4% MoM as supply chains began breaking and lockdowns started
April 2020: -12.5% MoM — the trough; industrial production reached approximately 84 (the lowest level since 1985)
The peak-to-trough decline was approximately -17% in 2-3 months — the steepest contraction in any modern peacetime period

The recovery was V-shaped:

May-September 2020: aggressive snap-back as reopening occurred — industrial production recovered to approximately 102 by year-end 2020, recovering most of the COVID decline
2021-2022: gradual continued recovery to approximately 105 by mid-2022
2022 supply-chain catchup: post-COVID demand shift to goods kept manufacturing running hot through much of 2022; industrial production was operating at high capacity utilization (around 79%, vs the 78% long-run average)

The Fed-tightening cycle reversal:

Late 2022 - 2023: industrial production peaked around 104.5 in mid-2022 and declined modestly through 2023 — a "manufacturing recession" of sorts, with industrial production declining by 1.5-2% from peak. Manufacturing employment also softened. The decline was concentrated in housing-related goods and durable goods (where rate sensitivity is highest).
2024-2025: stabilization in the 103-104 range with modest growth

The post-COVID industrial production trajectory illustrates the typical pattern: rapid collapse in stress, V-shaped recovery, then a longer multi-quarter cycle as Fed policy and broader economic conditions normalize. Industrial production is now back to roughly pre-COVID levels but hasn't grown meaningfully above them — consistent with the broader story of US economy being increasingly services-driven.

The current cycle, and the open question

The structural debates:

Manufacturing reshoring. Industrial policy (CHIPS Act semiconductor subsidies, IRA battery and clean-energy provisions, broader tariff regime) is incentivizing US manufacturing capacity expansion. If durable, this should support industrial production growth in coming years. Early data shows some response — semiconductor manufacturing capacity is expanding (Arizona TSMC, Intel Ohio, Samsung Texas).
AI-driven productivity. AI-driven productivity gains in manufacturing (process optimization, predictive maintenance, automation) could lift the index even with stable labor inputs. The actual effect on physical output is uncertain and early data is mixed.
Goods-vs-services demand balance. Post-COVID, consumer spending is gradually rotating back toward services. This is a structural headwind for industrial production growth.
Capacity utilization implications. Currently around 77-78% (vs long-run average of ~80%), suggesting some slack in the industrial sector. Higher utilization would support increased production; below 75% historically signals weak demand.

What you watch: the monthly Federal Reserve G.17 release; the manufacturing component specifically (separate from mining and utilities); capacity utilization (published in the same release); the ISM Manufacturing PMI (still relevant via direct ISM publication); and durable-goods orders (a leading indicator of future industrial production).

FAQ

What's the difference between Industrial Production and the ISM PMI?

Industrial Production (FRED INDPRO) is the actual output measure — the Federal Reserve's monthly index of real production across manufacturing, mining, and utilities, in physical-output terms (units of goods produced, kilowatt-hours generated, etc., aggregated and indexed). The ISM PMI (Purchasing Managers' Index) is a survey-based measure where the Institute for Supply Management asks ~300 purchasing managers about new orders, production, employment, supplier deliveries, and inventories at their firms. PMI is a 'diffusion index' centered at 50 (above 50 = expansion; below 50 = contraction). They measure related but different things: Industrial Production measures actual production volume; PMI measures sentiment and order momentum. Both are useful; both are correlated. Our dashboard's data source is INDPRO because ISM discontinued free FRED republication of its PMI series in 2016; the historical label 'ISM Manufacturing PMI' has been retained for continuity but the underlying data is Industrial Production.

Why did ISM discontinue free publication of the PMI series?

Commercial reasons. The PMI is a proprietary survey product produced by ISM (a private membership organization for supply-chain professionals); ISM monetized the data more aggressively starting in 2015-2016. The series remains widely cited in financial press and is published by ISM directly (free real-time press release with the headline figure on the first business day after each month), but the FRED republication arrangement ended. The Federal Reserve's Industrial Production Index (INDPRO) is a Fed-published, freely available alternative that captures roughly the same economic phenomenon (US goods-producing-sector activity) with somewhat different methodology. We use INDPRO; researchers needing the ISM PMI specifically can access it via ISM's official channels or paid third-party data providers.

How is Industrial Production actually measured?

By the Federal Reserve's Industrial Production statistical program — a monthly indicator constructed from approximately 300 individual data sources covering manufacturing (~78% of the index), mining (~15%), and utilities (~7%). Each component is measured in physical-output units where possible (tons of steel produced, megawatt-hours generated, vehicles assembled, etc.) and aggregated using value-added weights. The aggregate is indexed to a base year (currently 2017 = 100). The series captures real (inflation-adjusted) physical output, not nominal dollar values — important because nominal industrial production can rise just through inflation even when physical production is flat or declining. The release schedule: mid-month, approximately the 17th, at 9:15 AM ET as part of the Fed's G.17 statistical release.

Why is industrial production used in NBER recession dating?

Because the goods-producing sector — particularly manufacturing — is one of the most cyclically-sensitive parts of the US economy. Manufacturing employment and output respond to interest-rate changes, demand cycles, and inventory swings faster than services-sector activity. Of NBER's six primary recession-dating indicators, two are industrial-production-related: industrial production itself and aggregate employment (which includes manufacturing). When industrial production declines for several consecutive months, it's one of the most persistent recession signals. The 2008-2009 industrial production decline of approximately -17% peak-to-trough was the largest in the post-WWII era and was a primary input to NBER's recession dating.

What's the manufacturing-services divide and why does it matter?

The US economy is structurally services-heavy (~78% of GDP is services; ~22% is goods-producing). But industrial production captures only the goods side. This means industrial production can be weak (signaling manufacturing recession) while the broader economy continues growing (services-led growth). The 2015-2016 period is the canonical example: industrial production declined for 12+ consecutive months (a 'manufacturing recession' driven by oil-price collapse, dollar strength, and weak global demand), but the US economy as a whole continued growing. NBER did not date a recession. So industrial production weakness alone isn't a sufficient recession signal anymore — services-sector indicators (PCE services, services employment, ISM Services PMI) need to corroborate. Industrial Production remains one of the cleanest reads on the manufacturing cycle specifically.