Guide

Empire State Manufacturing Index explained

Harbor Manufacturing's macro channel pinged at 8:30 a.m. on a mid-month Tuesday: Empire State Manufacturing had plunged from +12 to −15 in one month. The plant manager in Ohio forwarded the alert with a single question: “Do we cut overtime?” Three weeks later the national ISM Manufacturing PMI printed 51.2 and industrial production rose. Ohio had not entered a recession; a volatile New York regional survey had whipsawed on a handful of chemical and machinery respondents while the national factory complex kept expanding.

The Empire State Manufacturing Index is the headline from the Federal Reserve Bank of New York's monthly manufacturing survey. It measures whether more factories in New York State report improving conditions than deteriorating — a diffusion index, not a count of widgets produced. Released around the 15th of each month, it is one of the earliest reads on factory momentum for the current month, which makes it market-moving despite a small, regional panel. After Harbor rebuilt its playbook around three-month smoothing, ISM confirmation, and sector-weight caveats, mid-month false alarms fell sharply. This guide covers survey methodology, subindex interpretation, calendar placement, comparison with national PMI and hard data, the Harbor refactor, a technique decision table, pitfalls, and a production checklist.

What the Empire State Manufacturing Survey measures

Each month the New York Fed emails a questionnaire to roughly 200 manufacturing firms headquartered or with significant operations in New York State. Executives report whether activity in the prior month was higher, the same, or lower than the month before across a standard set of categories. The Fed publishes results around 8:30 a.m. ET on the fifteenth (or the next business day if the fifteenth falls on a weekend or holiday).

The general business conditions index — the headline “Empire State” number markets quote — asks about overall business conditions. Additional diffusion indexes cover:

New orders — forward demand; typically the best leading subindex
Shipments — current production flow; correlates with IP with a short lag
Unfilled orders — backlog depth; rising unfilled orders with strong new orders signal sustained output
Delivery time — supplier lead times; longer times can reflect demand pressure or supply bottlenecks
Inventories — finished-goods stocks; rising inventories with weak new orders can signal future production cuts
Prices paid — input cost pressure; leads goods-side PPI components with a one- to two-month lag
Prices received — ability to pass costs through; margin signal for producers
Number of employees and average workweek — labor demand; lead manufacturing payrolls modestly
Technology spending — capex intentions; useful for machinery and automation suppliers
Capital expenditures — planned equipment investment over the next six months

The NY Fed also publishes six-month ahead expectations indexes for several categories. Expectations often stay optimistic longer than current conditions during downturns — a useful divergence to watch.

Diffusion index math and the 50 breakeven

Like the ISM PMI, each Empire State subindex is a diffusion index:

Index = % reporting higher + 0.5 × % reporting same

Scores range from 0 to 100. 50 is the breakeven: equal shares of firms reporting improvement and deterioration (with unchanged responses split evenly). A reading of 60 does not mean output rose 10%; it means a net majority of respondents saw conditions improve. A move from +5 to −5 is a breadth flip, not a 10% production decline.

Because the panel is smaller than ISM's national sample, month-to-month moves of 10–20 points are common even when the underlying economy is stable. Smoothing with a three-month moving average or focusing on the direction of new orders rather than the headline alone reduces noise substantially.

Regional scope: why New York is not America

New York State manufacturing is a thin slice of U.S. factory output. Heavyweights include chemicals, machinery, computers and electronics, fabricated metals, and food processing. The state has little auto assembly or aerospace final assembly compared with Michigan, Texas, or Washington. A strike at a single large Upstate plant or a weather disruption on the Thruway can move the index without national implications.

Correlation with the national ISM Manufacturing PMI is positive but far from perfect — typically 0.4–0.6 on headline levels depending on the sample period. Empire State often leads ISM by a few weeks because it releases mid-month for the current month while ISM covers the prior month and releases on the first business day. That lead is real but noisy: a weak Empire State print is a hint, not a guarantee, of a soft national PMI.

Peer regional Fed surveys — Philadelphia Fed Manufacturing (Philly Fed), Richmond Fed, Dallas Fed, and Kansas City Fed — cover other districts with different sector mixes. When Empire State and Philly Fed diverge sharply, the signal is often idiosyncratic rather than cyclical. When multiple regional surveys weaken together, national factory slowdown risk rises.

Calendar placement and how traders use the release

The mid-month timing fills a gap in the manufacturing data calendar:

Release	Typical date	Reference period	Geographic scope
Empire State Manufacturing	~15th of month	Current month (survey month)	New York State
ISM Manufacturing PMI	1st business day	Prior month	National
Industrial production (G.17)	Mid-month	Prior month	National hard data
Durable goods orders (Census)	~4th week	Prior month	National hard data

Futures and FX desks treat Empire State as an early manufacturing “temperature check” between ISM prints. Bond markets may price a few basis points of rate expectation on large surprises, but reversals are frequent once Philly Fed (released a few days later) and hard data arrive. The productive use is updating a prior for the next national PMI — not declaring a cycle turn on one regional headline.

Reading subindexes: orders, prices, and labor

Experienced macro analysts rarely trade the headline alone. A practical hierarchy:

New orders minus inventories — When new orders exceed 50 and inventories are below 50, factories are drawing down stock to meet demand; production tends to accelerate. The opposite pairing signals future cutbacks.
Prices paid vs prices received — A widening gap (paid rising faster than received) compresses margins and can foreshadow weaker employment or capex. Both rising above 50 feeds into PPI goods discussions, not shelter-heavy CPI.
Shipments vs new orders — Shipments above new orders drains backlog; the reverse builds it. Compare with durable goods orders for national confirmation.
Employment and workweek — Labor subindexes move slowly. A single sub-50 employment print rarely precedes immediate layoffs; sustained weakness across regions does.
Six-month expectations — If current conditions weaken but expectations stay above 50, the downturn may be transitory (weather, one-off order delays). If both collapse, recession risk in manufacturing rises.

The NY Fed publishes verbatim supplemental questions on topics like supply chain stress or tariff impacts when conditions warrant. These one-off modules are qualitative color, not comparable across months unless the question repeats.

Harbor Manufacturing release-day refactor

Harbor's old workflow treated any Empire State headline below zero as a “manufacturing contraction alert” emailed to all plant managers nationwide. Ohio and Michigan facilities cut shifts on New York chemical volatility; procurement delayed orders on single-month price-paid spikes that reversed the following month.

The refactor introduced three gates before operational action:

Smoothing — Alert only if the three-month average of new orders falls below 48.
Confirmation — Require Philly Fed (released ~2 days later) to agree on new-orders direction or wait for the next ISM print.
Hard-data check — No capex or hiring changes until prior-month IP or durable goods corroborate the survey direction.

False operational alerts dropped from eight per year to one. The macro desk still logs every Empire State release for its information value as an early read, but plant-level decisions now follow national confirmed signals. The lesson: regional diffusion surveys inform priors; they should not override national hard data and multi-survey consensus without confirmation.

Technique decision table: Empire State vs alternatives

Goal	Prefer	Not ideal
Earliest manufacturing read mid-month	Empire State + Philly Fed regional pair	Waiting only for month-end IP
National factory breadth	ISM Manufacturing PMI	Empire State headline alone
Output volume growth	Industrial production, durable goods shipments	Any diffusion index level as a growth rate
Input cost pipeline	Empire State prices paid + PPI goods	Prices paid as direct CPI forecast
Recession watch	Sustained sub-50 across multiple regional Fed surveys + ISM	One −15 Empire State spike
Plant-level production planning	Company order book + national IP	Regional survey 500 miles away
Services-sector read	ISM Services PMI	Empire State (manufacturing only, NY only)

Common pitfalls

Treating the headline like GDP growth — +10 is not 10% expansion; it is net positive breadth among ~200 NY firms.
Overweighting one month's volatility — Double-digit swings are normal; use moving averages.
Ignoring sector composition — Chemicals and machinery dominate; auto-heavy states can diverge for months.
Expecting perfect ISM correlation — Regional noise breaks the link regularly.
Using prices paid as CPI forecast — It leads goods PPI, not services or shelter CPI.
Skipping Philly Fed confirmation — The two releases days apart; disagreement signals idiosyncratic noise.
Confusing current conditions with expectations — Executives can be pessimistic today but optimistic six months out.
Chasing the futures pop — Mid-month survey surprises often fade within the session.

Production checklist

Calendar Empire State (~15th, 8:30 a.m. ET) alongside Philly Fed and ISM dates.
Archive full subindex table and supplemental questions, not just headline.
Compute three-month moving averages for new orders and headline.
Track new orders minus inventories and prices paid minus prices received spreads.
Compare Empire State new orders direction with Philly Fed on release week.
Cross-check large surprises against prior-month IP and durable goods.
Flag sustained sub-50 new orders across two or more regional Fed surveys.
Separate operational alerts (national confirmed) from macro desk color (regional early read).
Include six-month expectations in weekly manufacturing notes.
Review false signals quarterly and adjust smoothing thresholds.

Key takeaways

Empire State is a New York regional diffusion index: 50 marks equal breadth of improvement and deterioration, not zero growth.
Mid-month release timing makes it an early manufacturing hint, but the small NY panel is volatile.
New orders, prices paid, and expectations carry more signal than the headline alone.
Confirm with Philly Fed, ISM Manufacturing PMI, and hard data before treating moves as national cycle turns.
Harbor Manufacturing cut false plant alerts by smoothing, requiring confirmation, and separating regional color from operational decisions.