Richmond Fed Manufacturing Index

The Richmond Fed Manufacturing Index is the common name of the composite index in the Fifth District Survey of Manufacturing Activity. The Federal Reserve Bank of Richmond (Richmond Fed) publishes it in its regional business surveys. The manufacturing survey began in June 1986 and took its current monthly form in November 1993.

It covers establishments with a NAICS code in the 300s. That is the District of Columbia, Maryland, North Carolina, South Carolina, Virginia and most of West Virginia. On August 25, 2025 the panel held 522 active businesses: 120 manufacturing and 402 non-manufacturing. Active means taking part at least once in the prior 12 months, in a panel mirroring the district.

A typical month brings about 100 questionnaires; since 2021 replies averaged 76 for manufacturing and 153 for non-manufacturing. Since January 2021 response rates averaged 39 percent for manufacturing and 36 percent for non-manufacturing. Taking part is voluntary and unpaid, via an online form for senior managers.

The index gauges neither American industry nor output volume, only how widespread district factory change is. Participants also give expectations for the next 6 months and sometimes answer special questions. Zero, not 50, is neutral; readings between -5 and +5 are noise at 68 percent confidence.

It comes out monthly at 10:00 AM on the fourth Tuesday, though holidays can shift the date. Collection spans about 3 weeks: two business days after the prior release until the Wednesday before publication. Links go out before 10:00 AM Eastern time on day one, with three reminders.

The parallel non-manufacturing survey appears at the same time; state detail follows Thursday at 10:00 AM. Each release brings a full report, tables, the history file, special questions and quarterly pricing notes.

Seasonal factors update each July; the whole series is revised in the data file only. Error fixes land in the history file, not old reports, as in the July 25, 2017 errata.

The composite tracks 3 parts on fixed ISM weights; other series give context only.

• Shipments, 33%: change in finished-product shipment volumes.
• Volume of new orders, 40%: change in new-order volumes, the largest weight.
• Employment, 27%: change in headcount.

Other rows cover backlogs, capacity use, lead times, business conditions, capital spending, inventories, wages, hours and prices.

The July 28, 2020 release, from 56 firms, lifted the composite from 0 in June to 10, its first positive reading since March. Shipments reached 23 and new orders 9, while employment stayed at -3, up from -6 in June. The math was 0.33 times 23 plus 0.40 times 9 plus 0.27 times -3, or about 10.

A positive headline can thus coexist with falling headcount, as breadth decides and employment is only 27 percent. It then rose to 18 in August 2020, confirming the rebound. Releases call gains improved, losses weakened, and flat spells little changed.

Local conditions scored 10, capacity use 20, backlogs 5, lead times 12 and capital spending -15. Prices paid rose 0.93 percent, prices received 0.45 percent annualized, while employment expectations hit 28.

1. Diffusion index: Each question sorts replies into increase, decrease or no change, then subtracts decline share from growth share.

For example, 78 of 120 contacts reporting growth and 24 reporting decline gives 65 minus 20, or 45, on a -100 to 100 scale.

1. Seasonal adjustment: Each diffusion index loses its seasonal pattern through an Arima-13 process, with factors updated each July.

Non-adjusted series are published too.

1. Composite: The headline is 0.33 times shipments plus 0.40 times new orders plus 0.27 times employment after seasonal adjustment.

So 18 for shipments, 10 for orders and -3 for employment gives about 9.

1. Prices: Average price changes are trimmed at 5 percent before averaging to curb outliers.