Skip to main content

How does the industrial production index measure the health of manufacturing?

While the retail sales report tells you what consumers are buying, the industrial production index tells you what factories are making. Released monthly by the Federal Reserve, the IP index measures total output from manufacturing plants, mines, and electric utilities across the United States. It's the closest thing economists have to a real-time measure of productive capacity and industrial strength.

The IP index matters because manufacturing is the backbone of any large economy. Factory output drives employment (directly for factory workers, indirectly for truck drivers and warehouse staff), it generates corporate profit (which fuels business investment and stock buybacks), and it creates demand for raw materials (which keeps mining and agriculture busy). When factories are humming, the economy is usually expanding. When factories go quiet, recession is often close behind.

Quick definition: The industrial production index measures the monthly change in total output from U.S. factories, mines, and utilities, adjusted for inflation and seasonal patterns, and serves as a gauge of manufacturing sector health and economic momentum.

Key takeaways​

  • The IP index covers three broad sectors: manufacturing (about 75 percent of the index), utilities (about 14 percent), and mining (about 11 percent).
  • Monthly swings in IP are common—anything from -1.5 to +1.5 percent is normal; sustained moves beyond that signal real economic shifts.
  • Capacity utilization, released alongside the IP index, shows what percentage of factory capacity is in use; 80 percent is historically normal; above 85 percent suggests bottlenecks and pricing pressure.
  • IP lags retail sales by one to three months; strong retail sales eventually show up as higher factory orders and IP growth.
  • Manufacturing IP declined throughout the 2010s in the U.S. as factories automated and offshored; only during energy booms or import surges did it spike.

What the industrial production index actually measures​

The Federal Reserve collects production data from thousands of manufacturing establishments and combines them into a single monthly index. The Fed doesn't measure units produced but rather the inflation-adjusted dollar value of output. This approach allows apples-to-apples comparison: a factory making widgets and a factory making automobiles can be summed into a single number.

The IP index is constructed as a weighted average of production by industry. Heavy industries—steel mills, petrochemical plants, automotive assembly—carry more weight than light manufacturing. This weighting reflects the dollar value of each industry to the overall economy. Automotive manufacturing, for example, represents about 12 percent of the total IP index because cars are expensive and export-heavy.

The Fed also publishes a companion measure: capacity utilization. This is the ratio of actual output to potential output—the maximum the factory could produce if running at full steam. If a car plant normally runs 90 percent of the time (allowing for maintenance and retooling), and actual output is 85 percent of potential, capacity utilization is 94.4 percent. Capacity utilization over 80–82 percent is considered "tight" and suggests pricing power; above 85 percent is very tight and often precedes wage inflation; below 75 percent suggests underutilized capacity and possible price deflation.

The three components: Manufacturing, utilities, and mining​

Manufacturing makes up the bulk of the IP index (roughly 75 percent). It includes durable goods (cars, appliances, machinery) and nondurable goods (food, textiles, chemicals, paper). Durable manufacturing is cyclical—it swings sharply with business confidence and credit conditions. Nondurable manufacturing is less cyclical but moves with consumer spending and inflation.

Utilities account for about 14 percent of the index. Utilities production is driven by temperature swings (hot summers boost air-conditioning demand; cold winters boost heating) and overall industrial activity. Utilities are the most stable component; long-term growth is slow because of energy efficiency. Some months utilities surge due to a cold snap; other months they fall if the weather is mild.

Mining accounts for about 11 percent. This includes coal mines, metal mines, and oil and gas extraction. Mining output swings with commodity prices; when oil prices fall, oil production often falls too (unprofitable wells shut down). When metal prices surge, miners ramped extraction. Mining is the most volatile component of the IP index.

For forecasters, the key is to mentally separate these three. If headline IP falls, was it manufacturing weakness, or just a warm month reducing utility demand? If IP surges, was it a mining boom from oil prices, or genuine factory gains? Breaking the components apart gives a clearer picture.

The lag between orders and production​

There's a crucial timing issue with the IP index: it lags real-time demand. When retail sales surge, manufacturers don't immediately crank up production. They first look at incoming orders. If orders rise 2 percent one month, factories might wait to confirm the trend is real before hiring shifts or running overtime. Then production ramps, which takes another month or two. This lag means that strong retail sales in January might not show up as higher IP until March or April.

Conversely, when demand falls, factories cut production quickly because idle capacity is expensive. But they often cut too much, leading to short-lived inventory shortages. The 2021–2022 period illustrates this: retail sales surged post-stimulus, but manufacturing couldn't keep up, creating supply-chain bottlenecks. IP growth lagged demand for six to nine months. By mid-2022, when retail sales started falling, factories had ramped so hard that they found themselves with excess inventory—and had to cut production sharply, leading to IP declines in mid-2022.

Understanding this lag is crucial for interpreting the IP index. A single month of strong IP growth doesn't mean strong future demand; it often means factories are catching up to past demand that has since weakened.

How recessions show up in the industrial production index​

The IP index is one of the most reliable recession predictors. Here's why: recessions involve a sudden drop in business confidence. Businesses immediately cut orders for parts and materials. Within a month or two, factories see order backlogs shrinking and start cutting production and workforce. The IP index falls sharply—often 3–6 percent over three to six months during recessions.

The Federal Reserve's official recession definition uses multiple data points, but the IP index is closely watched. In the 2001 recession, IP fell 4 percent peak-to-trough and was negative for nine consecutive months. In the 2008 financial crisis, IP fell 9.7 percent peak-to-trough—the worst performance since the 1980–1981 double-dip recession. In the 2020 COVID recession, IP fell 5.8 percent in two months (March–April), the fastest drop on record.

Outside of recessions, the IP index is positive most months, with monthly swings of ±0.5 percent being normal. Three consecutive months of negative IP is a yellow flag; six consecutive months is a recession signal.

Capacity utilization and inflation dynamics​

Capacity utilization is where IP connects directly to inflation and Fed policy. When capacity utilization is high (above 85 percent), factories are running at near-maximum. They can't ramp production without hiring more workers (tight labor market), buying new equipment (capital investment), or raising prices. Higher wages and prices feed inflation. The Fed watches capacity utilization closely; when it exceeds 85 percent, the Fed often starts raising interest rates to cool demand before inflation spirals.

When capacity utilization is low (below 75 percent), factories have plenty of slack. They can ramp production without hiring immediately, without raising wages, and without raising prices. Low capacity utilization is deflationary—businesses compete on price to fill empty plants. This is why recessions, which drive capacity utilization down sharply, lead to price deflation (or at least very low inflation).

The relationship is not mechanical—other factors like commodity prices and global competition matter—but it's strong enough that the Fed builds capacity utilization forecasts into its policy decisions. In 2021, when capacity utilization jumped to 77 percent (high, but not extreme), the Fed correctly expected wage and price pressures to build, and started raising rates in 2022.

Why manufacturing has declined in the U.S. but industrial production still matters​

Since the 1980s, U.S. manufacturing has declined as a share of GDP—from 28 percent in 1970 to 11 percent today. Factories have automated (fewer workers, same output), offshored production to lower-wage countries, and shifted to higher-value-added goods. U.S. IP growth has been low: roughly 0.5–1.0 percent per year on average since 2010.

This long-term trend doesn't make the IP index less important for forecasting. Even though manufacturing is smaller, it still drives industrial employment, capacity investment, and commodity demand. When IP falls sharply, it signals business retrenchment—exactly the kind of confidence loss that precedes recession. When IP rises, it signals business expansion.

Additionally, the COVID pandemic and subsequent supply-chain disruptions have redrawn the geography of manufacturing. U.S. factories (and especially semiconductor fabs and battery plants) have received billions in subsidies to re-shore production. Since 2021, manufacturing IP has been surprisingly resilient, growing faster than the long-term trend. Watching the IP index over the next five years will tell us whether the shift back to U.S. production is real or temporary.

Reading the monthly industrial production report​

The Federal Reserve releases the IP report around the 15th of each month, covering the prior month. The report includes:

  • Total IP (month-over-month percent change). Usually ±0.5 percent; anything beyond ±1.0 percent warrants attention.
  • IP by sector (manufacturing, utilities, mining separately). Helps you see where strength or weakness is concentrated.
  • Capacity utilization rate (percent). 80 percent is neutral; above 85 percent is tight; below 75 percent is slack.
  • Revisions to the prior month. The Fed revises IP figures for the last two months as new data arrive. These revisions are often ±0.3 percent, so don't anchor too hard on the first release.

For a quick read: if month-to-month IP is above +1.0 percent and capacity utilization is rising, manufacturing is accelerating and inflation risk is building. If IP is below -1.0 percent for three consecutive months and capacity utilization is falling, recession risk is rising. If IP is flat and capacity utilization is 78–80 percent, the manufacturing sector is in steady state.

Real-world examples of IP turning points​

In mid-2007, the U.S. economy was running hot. Manufacturing IP had risen steadily for five years, growing at an average +2.5 percent per year. Capacity utilization peaked at 82.7 percent in June 2007. Then, starting in July, orders from financial institutions and home builders began to fall as the subprime mortgage crisis spread. By October, manufacturing IP was negative, and by November 2008, IP had fallen 9.7 percent from its peak. This was the steepest drop since the 1981 recession. The recession was officially underway.

In 2020, COVID caused an immediate shock: IP fell 5.8 percent in March–April, the fastest two-month drop ever recorded. Factories shut due to lockdowns and workers didn't show up. Capacity utilization fell to 68.2 percent (the lowest since 2009). But by May, as factories reopened and stimulus checks landed, orders surged. IP rebounded +5.4 percent in May and stayed positive for the next 24 months. By late 2021, capacity utilization had climbed back to 77 percent, signaling that factories were catching up to demand.

Then, from mid-2022 onward, IP growth slowed. The Fed raised rates aggressively, retail sales weakened, and factories found themselves with excess inventory. By September 2022, IP was negative. Many forecasters predicted a 2023 recession. However, the recession call was wrong; IP stabilized in Q4 2022 and resumed modest growth in early 2023. The episode illustrates that even strong IP signals can be overridden by other dynamics (in this case, persistent labor tightness and government spending kept the economy afloat despite industrial slowdown).

Common mistakes when reading industrial production​

Confusing nominal and real production. IP is inflation-adjusted, so a 1.5 percent IP increase means real output rose, not just prices. This is different from, say, revenue figures, which are often nominal. Always remember: IP is about physical production, not nominal dollars.

Over-interpreting utility swings. If IP rose 1.2 percent month-over-month but utilities surged 3.5 percent due to a cold snap and manufacturing fell -0.3 percent, the headline IP is misleading. Break the components apart before drawing conclusions.

Forgetting the lag to demand. A single month of strong IP doesn't mean demand is strong—it might mean factories are clearing prior backlogs. Similarly, a weak IP month might reflect factories deliberately cutting production in anticipation of weaker future demand. Look at three-month trends, not single months.

Ignoring capacity utilization. High IP growth paired with falling capacity utilization suggests the growth came from a low base (recession ending), not from tightness (pricing power). High IP growth paired with rising capacity utilization is much more bullish because it suggests sustained demand.

Assuming IP growth means GDP growth. IP is one component of GDP, but it's only about 15 percent of the total. Services, government, and investment in other forms (housing, software) matter more. Strong IP paired with weak services or construction tells a different story.

FAQ​

How does the Fed actually measure IP if it doesn't track every factory?​

The Fed uses a combination of surveys (Census Bureau's Manufacturing Survey, trade association reports), administrative data (electric utility billing records, coal shipments), and statistical modeling. For manufacturing, the Census Bureau's Annual Survey of Manufactures gives actual output by facility; the IP index interpolates between these benchmarks using monthly survey data and indicator variables.

What's the difference between IP and production-worker hours?​

IP is the output from all workers and machines combined. Production-worker hours measure only the time spent by factory workers (excluding supervisors and admin). IP growth can exceed production-worker hours growth if factories automate (machines do more, workers stay same or fall). This divergence is common and explains why manufacturing employment has fallen while IP has grown modestly over the past 40 years.

Can the IP index go negative for extended periods?​

Yes, but it's rare outside of recessions. In recessions, IP often falls for six to nine consecutive months (2008, 2001) or briefly (2020). Outside of recessions, IP is usually positive month-to-month, with occasional down months. A sustained period (six-plus months) of negative or near-flat IP growth is a strong recession signal.

How does energy production affect the IP index, especially when oil prices swing?​

Oil and gas extraction is part of the mining component, which is about 11 percent of the total IP index. When oil prices rise sharply, production often falls (because high-cost wells shut down), and IP can fall even though consumer products using energy become more expensive. The relationship is complex and often counterintuitive; always check mining separately from manufacturing when oil prices swing sharply.

How did supply-chain disruptions from COVID affect IP interpretation?​

From 2020 to 2022, supply-chain bottlenecks meant that factories couldn't produce at full potential even though demand was strong. IP growth looked weak because factories were short of inputs, not because demand was weak. This is why the Fed shifted focus to desired production (reflected in backlogs and lead times) rather than actual production alone.

Summary​

The industrial production index is a monthly measure of total output from U.S. factories, mines, and utilities, released by the Federal Reserve. It captures real, inflation-adjusted production and is one of the fastest and most reliable reads of manufacturing sector health and economic momentum. The index includes three components: manufacturing (75 percent), utilities (14 percent), and mining (11 percent). Capacity utilization, released alongside, shows what percentage of potential output is being used; high capacity utilization (above 85 percent) signals tight production and inflation risk, while low capacity utilization (below 75 percent) signals slack and deflationary risk. Manufacturing IP growth has been modest (0.5–1.0 percent annually) over the past decade due to automation and offshoring, but the index remains a crucial recession predictor: three to six months of negative IP is a strong warning sign. The relationship between rising capacity utilization and Fed rate hikes is tight, making capacity utilization a key input to monetary policy.

Next​

→ Housing starts explained