---
ticker: "USADS"
title: "U.S. ADS Business Conditions Index"
unit: ""
frequency: "Daily"
source: "Stock and Watson (1989); Stock and Watson (1991)"
release: "Updated Daily"
category: "Real Activity & Business Cycle"
country: "US"
language: "en"
canonical: "https://kred.dev/en/series/USADS"
license: "https://creativecommons.org/licenses/by-nc-nd/4.0/"
latest_value: 0.15
latest_date: "2026-10-17"
first_date: "1960-03-01"
observations_total: 24337
observations_shown: 120
---

# U.S. ADS Business Conditions Index

## Overview

A daily read on whether US business conditions sit above or below their own average, distilled from a panel of activity indicators.

## Key Figures

|  | Value | Date |
|---|---|---|
| Latest | 0.15 | 2026-10-17 |
| Change from previous | 0.00 | 2026-10-16 |
| Change over one year | +0.94 | 2025-10-17 |
| Highest on record | 9.64 | 2020-06-09 |
| Lowest on record | -26.64 | 2020-04-07 |
| Period covered | 1960-03-01 – 2026-10-17 |  |
| Observations | 24337 |  |

## Recent observations

| Date | Value | Change |
|---|---|---|
| 2026-06-20 | -0.04 | 0.00 |
| 2026-06-21 | -0.04 | 0.00 |
| 2026-06-22 | -0.04 | 0.00 |
| 2026-06-23 | -0.04 | 0.00 |
| 2026-06-24 | -0.04 | 0.00 |
| 2026-06-25 | -0.03 | 0.00 |
| 2026-06-26 | -0.03 | 0.00 |
| 2026-06-27 | -0.03 | 0.00 |
| 2026-06-28 | -0.02 | 0.00 |
| 2026-06-29 | -0.02 | 0.00 |
| 2026-06-30 | -0.02 | 0.00 |
| 2026-07-01 | -0.02 | 0.00 |
| 2026-07-02 | -0.01 | 0.00 |
| 2026-07-03 | -0.01 | 0.00 |
| 2026-07-04 | -0.01 | 0.00 |
| 2026-07-05 | -0.01 | 0.00 |
| 2026-07-06 | -0.00 | 0.00 |
| 2026-07-07 | -0.00 | 0.00 |
| 2026-07-08 | -0.00 | 0.00 |
| 2026-07-09 | -0.00 | 0.00 |
| 2026-07-10 | -0.00 | 0.00 |
| 2026-07-11 | -0.00 | 0.00 |
| 2026-07-12 | -0.00 | 0.00 |
| 2026-07-13 | -0.00 | 0.00 |
| 2026-07-14 | -0.01 | 0.00 |
| 2026-07-15 | -0.01 | 0.00 |
| 2026-07-16 | -0.02 | 0.00 |
| 2026-07-17 | -0.02 | -0.01 |
| 2026-07-18 | -0.03 | -0.01 |
| 2026-07-19 | -0.04 | -0.01 |
| 2026-07-20 | -0.04 | -0.01 |
| 2026-07-21 | -0.05 | -0.01 |
| 2026-07-22 | -0.06 | -0.01 |
| 2026-07-23 | -0.07 | -0.01 |
| 2026-07-24 | -0.07 | -0.01 |
| 2026-07-25 | -0.08 | -0.01 |
| 2026-07-26 | -0.09 | -0.01 |
| 2026-07-27 | -0.09 | -0.01 |
| 2026-07-28 | -0.10 | -0.01 |
| 2026-07-29 | -0.11 | -0.01 |
| 2026-07-30 | -0.11 | -0.01 |
| 2026-07-31 | -0.12 | -0.01 |
| 2026-08-01 | -0.13 | -0.01 |
| 2026-08-02 | -0.13 | -0.01 |
| 2026-08-03 | -0.14 | -0.01 |
| 2026-08-04 | -0.14 | 0.00 |
| 2026-08-05 | -0.14 | 0.00 |
| 2026-08-06 | -0.15 | 0.00 |
| 2026-08-07 | -0.15 | 0.00 |
| 2026-08-08 | -0.15 | 0.00 |
| 2026-08-09 | -0.14 | 0.00 |
| 2026-08-10 | -0.14 | 0.00 |
| 2026-08-11 | -0.14 | 0.00 |
| 2026-08-12 | -0.14 | 0.00 |
| 2026-08-13 | -0.14 | 0.00 |
| 2026-08-14 | -0.13 | 0.00 |
| 2026-08-15 | -0.13 | 0.00 |
| 2026-08-16 | -0.13 | 0.00 |
| 2026-08-17 | -0.12 | 0.00 |
| 2026-08-18 | -0.12 | 0.00 |
| 2026-08-19 | -0.11 | +0.01 |
| 2026-08-20 | -0.11 | +0.01 |
| 2026-08-21 | -0.10 | +0.01 |
| 2026-08-22 | -0.09 | +0.01 |
| 2026-08-23 | -0.09 | +0.01 |
| 2026-08-24 | -0.08 | +0.01 |
| 2026-08-25 | -0.07 | +0.01 |
| 2026-08-26 | -0.06 | +0.01 |
| 2026-08-27 | -0.05 | +0.01 |
| 2026-08-28 | -0.04 | +0.01 |
| 2026-08-29 | -0.03 | +0.01 |
| 2026-08-30 | -0.02 | +0.01 |
| 2026-08-31 | -0.01 | +0.01 |
| 2026-09-01 | 0.01 | +0.01 |
| 2026-09-02 | 0.02 | +0.01 |
| 2026-09-03 | 0.03 | +0.01 |
| 2026-09-04 | 0.04 | +0.01 |
| 2026-09-05 | 0.05 | +0.01 |
| 2026-09-06 | 0.07 | +0.01 |
| 2026-09-07 | 0.08 | +0.01 |
| 2026-09-08 | 0.09 | +0.01 |
| 2026-09-09 | 0.09 | +0.01 |
| 2026-09-10 | 0.10 | +0.01 |
| 2026-09-11 | 0.11 | +0.01 |
| 2026-09-12 | 0.12 | +0.01 |
| 2026-09-13 | 0.12 | +0.01 |
| 2026-09-14 | 0.13 | +0.01 |
| 2026-09-15 | 0.13 | 0.00 |
| 2026-09-16 | 0.14 | 0.00 |
| 2026-09-17 | 0.14 | 0.00 |
| 2026-09-18 | 0.14 | 0.00 |
| 2026-09-19 | 0.15 | 0.00 |
| 2026-09-20 | 0.15 | 0.00 |
| 2026-09-21 | 0.15 | 0.00 |
| 2026-09-22 | 0.16 | 0.00 |
| 2026-09-23 | 0.16 | 0.00 |
| 2026-09-24 | 0.16 | 0.00 |
| 2026-09-25 | 0.16 | 0.00 |
| 2026-09-26 | 0.16 | 0.00 |
| 2026-09-27 | 0.16 | 0.00 |
| 2026-09-28 | 0.16 | 0.00 |
| 2026-09-29 | 0.17 | 0.00 |
| 2026-09-30 | 0.16 | 0.00 |
| 2026-10-01 | 0.16 | 0.00 |
| 2026-10-02 | 0.16 | 0.00 |
| 2026-10-03 | 0.16 | 0.00 |
| 2026-10-04 | 0.16 | 0.00 |
| 2026-10-05 | 0.16 | 0.00 |
| 2026-10-06 | 0.16 | 0.00 |
| 2026-10-07 | 0.16 | 0.00 |
| 2026-10-08 | 0.16 | 0.00 |
| 2026-10-09 | 0.16 | 0.00 |
| 2026-10-10 | 0.16 | 0.00 |
| 2026-10-11 | 0.16 | 0.00 |
| 2026-10-12 | 0.16 | 0.00 |
| 2026-10-13 | 0.15 | 0.00 |
| 2026-10-14 | 0.15 | 0.00 |
| 2026-10-15 | 0.15 | 0.00 |
| 2026-10-16 | 0.15 | 0.00 |
| 2026-10-17 | 0.15 | 0.00 |

## Definition

USADS is the published ADS business conditions index recorded without transformation, a single latent business-conditions factor that a mixed-frequency dynamic factor model extracts from a panel of real-activity indicators observed at weekly, monthly, and quarterly cadence. KRED serves the index verbatim and applies no transformation of its own, so the object on this page is the factor as its authors compute it and not a KRED estimate. The construction is the Aruoba-Diebold-Scotti mixed-frequency dynamic factor model, the factor is centered on zero by construction rather than by an empirical finding, and positive values stand for above-average business conditions.

Reading one number as the state of activity belongs to the reference-cycle tradition, where many comoving series turn together around aggregate cycles (Burns and Mitchell 1946) and each series is screened and classified by its timing against those turns (Mitchell and Burns 1938). The composite ancestor of such an index combined activity series after detrending, seasonal adjustment, and standardization (Persons 1923), the diffusion and composite apparatus that grew out of it was codified for routine use (Moore 1961), and monthly composites were operationalized as real-time signals of turning points (Shiskin 1961).

The statistical form of the index is the single-index dynamic factor model, which recasts the composite as the latent common factor of several activity aggregates (Stock and Watson 1989) and states it as an explicit probability model (Stock and Watson 1991). Its inputs are themselves measured constructions rather than direct readings, since employment counts rest on the operational labor-force framework (Hussmanns, Mehran, and Verma 1990) whose stocks are the net outcome of entry, exit, and duration (Clark and Summers 1979), production rests on the physical-output index tradition (Fabricant 1940), and the quarterly output measure rests on the national-accounting framework (Kuznets 1941; Stone 1947).

A reading of zero marks average business conditions rather than zero growth, values above zero mark better-than-average conditions, and values below zero mark worse-than-average conditions, with zero drawn on the chart as the reference line. The daily cadence makes the index a current-conditions gauge whose turning points can be dated by the same reproducible algorithm applied to any activity series (Bry and Boschan 1971), and the measurement and timing properties of such gauges are catalogued in the indicator literature (Zarnowitz 1992), though an index built as measurement alone acquires its meaning only against an explicit model of the economy (Koopmans 1947).

## Methodology

KRED applies no transformation to USADS and stores the published index exactly as recorded, performing none of rescaling, deflating, smoothing, or annualizing, so every step below belongs to the compilation at source rather than to this platform. Extracting a latent factor from a mixed-frequency panel, recovering the state at each date with a two-sided smoother, and centering the factor on zero are all properties of that compilation, which is why a passthrough of an unchanged record and a heavily processed construction hold together without contradiction for the same daily figure.

The construction is the Aruoba-Diebold-Scotti mixed-frequency dynamic factor model. One latent factor loads on every indicator in the panel, each indicator carries its own idiosyncratic dynamics, and the state is recovered over the whole sample by a two-sided Kalman smoother, which is the form the single-index dynamic factor model takes once it is written as a probability model of coincident indicators (Stock and Watson 1989; Stock and Watson 1991). The differing cadences of the panel are handled inside the state space rather than by aggregating the faster series down to the slowest, so a weekly or monthly reading and a quarterly flow each enter at their own timing.

The inputs are measured rather than observed directly. Employment aggregates follow the operational definitions of activity status (Hussmanns, Mehran, and Verma 1990) and summarize flows of entry, exit, and duration (Clark and Summers 1979), production is a quantity index built from physical output (Fabricant 1940) under fixed-base or chained aggregation rules (Laspeyres 1871; Frisch 1936), and the quarterly output measure follows the double-entry national accounts (Meade and Stone 1941; Stone 1947) whose estimates carry a precision of their own (Stone, Champernowne, and Meade 1942).

The scale and the sign are conventions of the construction rather than results. Standardization leaves the factor without a unit and fixes its mean at zero, so comparing two dates compares each against the same average rather than against a fixed reference level, and the detrended, standardized comparison basis this presumes comes from the composite tradition (Persons 1923; Moore 1961). Because the smoother is re-run over the whole history at each release, the series served here is the latest vintage and revises with it, and the measurement and timing properties of such gauges are catalogued in the indicator literature (Zarnowitz 1992), while the caveat that a measurement procedure presupposes a model for its interpretation remains in force (Koopmans 1947).

## Applications in Economics

USADS states how far current business conditions sit above or below their own average, which is why it is read as a coincident gauge rather than as a forecast. The reference-cycle tradition it belongs to defines the cycle through the joint turning of many activity series (Burns and Mitchell 1946), classifies each series by whether it leads, coincides with, or lags those turns (Mitchell and Burns 1938), and evaluates the resulting gauges as a body of measurement practice (Zarnowitz 1992).

Because the factor is the component common to indicators of employment, production, and output rather than anything specific to one of them, it moves when the shared signal moves and stays quiet when a single indicator is idiosyncratic, which is the property the single-index representation was built to isolate (Stock and Watson 1989; Stock and Watson 1991). The labor-market side of such a panel is a stock produced by flows of entry, exit, and duration (Clark and Summers 1979) under fixed operational definitions of activity status (Hussmanns, Mehran, and Verma 1990), the production side is a physical-output index (Fabricant 1940), and the output side is the national-accounts aggregate (Kuznets 1941).

The daily cadence is what the mixed-frequency form buys, so the index can be read between the release dates of its slower inputs, and its turning points can be dated by a reproducible algorithm (Bry and Boschan 1971) in the register that monthly composite indicators established for real-time signaling (Shiskin 1961; Moore 1961; Persons 1923).

Two limits belong with the reading. The index is centered on its own average, so it carries neither a unit nor a level, and the zero line is that average rather than a neutral rate of growth. The whole history is also re-estimated at each release, so the value shown for a past date is the latest vintage of that value, and an index of this kind takes its economic meaning from an explicit model of the economy rather than from the measurement itself (Koopmans 1947).

## Applications in Financial Markets

On desks the index functions as one current-conditions level for the United States, and its use is bounded by what it measures. A reading above zero places business conditions above their own average and a reading below zero places them under it, and because the factor is coincident by construction it confirms rather than anticipates a turn dated by the reference-cycle apparatus (Burns and Mitchell 1946; Mitchell and Burns 1938; Moore 1961).

Release-day interpretation follows the current-conditions tradition of composite indicators (Shiskin 1961) and the reproducible turning-point mechanics that keep successive readings comparable (Bry and Boschan 1971), while the catalogued timing properties of coincident and leading indicators set expectations for what such an index can and cannot precede (Zarnowitz 1992). The single-index representation is also what lets one number stand in for a panel of activity series inside a risk model (Stock and Watson 1989; Stock and Watson 1991).

Three operational cautions apply. The state at each date is estimated with information from both sides of that date, so the path is smoother than its inputs and a one-day move carries less independent information than its variation suggests, and the whole history is re-estimated at each release, so a value read today for a past date can differ from the value read for that same date earlier. The index carries no unit and no level, so it cannot be compared with an output level or a growth rate directly, and the comparison basis is the detrended, standardized composite convention rather than an economic norm (Persons 1923). Its inputs are revised at source under their own definitions, in the labor-force framework (Hussmanns, Mehran, and Verma 1990) and its flow accounting (Clark and Summers 1979), in the physical-output index tradition (Fabricant 1940), and in the national accounts whose estimates carry a stated precision (Stone, Champernowne, and Meade 1942).

The gauge is descriptive and carries no probability, so a reading below zero is a statement about conditions now rather than a forecast of a recession, and the interpretation of any such measurement rests on an explicit model (Koopmans 1947).

## Statistical Tests

USADS is the single latent factor of the Aruoba-Diebold-Scotti mixed-frequency dynamic factor model, a business-conditions state recovered by a two-sided Kalman smoother and served here on a calendar-daily grid. Over 24,246 daily observations from 1960-03-01 to 2026-07-18, its measured serial correlation is dominated by the smoother's gain rather than by the data-generating process, so the object framed here is a persistence summary of the smoothed path and not an integration order of the data, and the centering of the factor on zero is likewise a property of the construction rather than an empirical finding.

The integration-order battery is therefore deliberately not run. The augmented unit-root regression of Dickey and Fuller (1979) with the lag augmentation of Said and Dickey (1984), the semiparametric Phillips and Perron (1988) test, the KPSS stationarity test (Kwiatkowski et al. 1992), the efficient GLS-detrended test of Elliott, Rothenberg, and Stock (1996), and the modified M-tests of Ng and Perron (2001) are all excluded, because a symmetric two-sided filter manufactures the persistence an integration test reads and the random-walk-versus-constant character of the latent state is not point-identified by the likelihood (Stock and Watson 1998; Orphanides and van Norden 2002). The level mean-break search of Bai and Perron (1998) is likewise not run, since its asymptotics require a stationary object and the filter-persistent path spuriously segments (Perron 1989). No order of integration is assigned to USADS, by ruling rather than by an inconclusive test.

The matrix-mandated replacement is a descriptive persistence summary labeled as a property of the smoothed series. The lag-one autoregressive coefficient of the level is 0.9993 and the implied half-life is about 992.1 days, which describes how slowly the smoothed path decays and carries neither an integration-order claim nor any claim about United States business conditions themselves. That figure is read with the small-sample downward bias of Andrews (1993) in mind, and a level portmanteau such as that of Ljung and Box (1978) would read the same filter gain rather than evidence about the process.

Real-time behavior is stated as a channel rather than as a magnitude. The smoother is re-run over the whole history at each release, so the state at a fixed past date is rewritten as the sample grows, and the loader here upserts the full history on every run, which leaves the served path as the latest vintage and revising with it. No dated vintage panel is stored, so no revision figure accompanies this field and none is quoted, a limitation familiar from real-time work on latent-state measures (Orphanides and van Norden 2002; Edge and Meisenzahl 2011). The model-implied content stays honest as an assumption rather than a discovery, namely a single mean-zero latent factor with autoregressive dynamics estimated from a mixed-frequency panel of activity indicators (Stock and Watson 1998).

## Frequently Asked Questions

### What do the US business cycle indicators cover?

United States real output, industrial production, the published unemployment-gap rule reading and the published business conditions index, all carried as released.

### Is the US ADS business conditions index a KRED estimate?

No. It is a passthrough of the index its original authors publish. KRED separately estimates business conditions for Korea, but the two are distinct series produced by different parties for different economies.

### Does KRED process the US industrial production index further?

None. Several of these are seasonally adjusted at source and KRED applies no further treatment to any of them.