Research campaign · Iteration 1 · unassessed

A Recession Fear Indicator | RecessionAlert — exploratory proxy campaign

What can be learned from the SHA-verified RecessionAlert Google-search fear-gauge narrative using transparent public proxies when its series, thresholds, publication lag, and market rule are incomplete?

Expressions: 6
Logged trials: Not recorded
Independent events: Not assessed
Evidence: unassessed

What the research found

The source describes a Google recession-search fear gauge but provides no machine-readable observations, exact query, native span, or executable market rule.
UNRATE is the longest available public recession-relevant proxy in the tracked cache; SPY/IEF/BIL provide transparent descriptive sleeves, not source fidelity.
All six inferred variants ran successfully; their metrics and weak or negative outcomes remain in the state-space ledger.
A one-available-month lag is the causal baseline, with lag 0 retained only as a research-only timing sensitivity.

Mechanism and falsifiers

Not recorded in this iteration.

Confidence and limitations

No source-linked Google Trends observations, query configuration, native span, publication clock, vintage, or threshold are available.
UNRATE is an aggregate public proxy and cannot establish RecessionAlert source fidelity.
FRED revisions and Yahoo historical adjustments lack point-in-time vintages in the tracked cache.
SPY drawdown is related to the benchmark and is circular as an independent fear measure.
IEF and BIL are descriptive sleeves; source buy/sell, sizing, rebalance, exit, and re-entry rules are undisclosed.
The article-to-archive interval and two-week NBER phrase are not publication lags.

Not recorded in this iteration.

Compare expressions

Download evidence

Exploratory results. Check each period, proxy and cost assumption before comparing. — means not recorded.

Expression / family	CAGR	Sharpe	Max drawdown	Test period	Assessment
recessionalert.1345.unrate_regime.lb3.lag0.t0p0unrate regime	—	—	—	Not recorded	inferred lookahead diagnostic
recessionalert.1345.unrate_regime.lb6.lag1.t0p0unrate regime	—	—	—	Not recorded	inferred causal best guess
recessionalert.1345.unrate_regime.lb12.lag2.t0p5unrate regime	—	—	—	Not recorded	inferred conservative sensitivity
recessionalert.1345.unrate_regime.lb6.lag1.t0p5unrate regime	—	—	—	Not recorded	inferred threshold neighbor
recessionalert.1345.unrate_regime.lb12.lag1.t1p0unrate regime	—	—	—	Not recorded	inferred higher-threshold sensitivity
recessionalert.1345.unrate_regime.lb21.lag1.tm0p5unrate regime	—	—	—	Not recorded	inferred negative long-lookback variant

Open an expression to inspect its rules and request confirmation. The request must be submitted by a trusted repository collaborator.

recessionalert.1345.unrate_regime.lb3.lag0.t0p0

available signal observations

938

claim kind

inferred

classification

inferred lookahead diagnostic

execution

one available monthly bar lag is causal baseline; lag 0 is research-only

exploratory

True

family

unrate_regime

recessionalert.1345.unrate_regime.lb3.lag0.t0p0

input names

UNRATE

lag bars

lookback

metrics spy bil

annualized return: 0.10223803996031511
benchmark total return: 5.775363419995794
end: 2026-06-30
max drawdown: -0.2420584677802402
observations: 228
risk off fraction: 0.29385964912280704
sharpe: 0.8894778625274528
start: 2007-06-30
total return: 5.356711589315653
volatility: 0.11758829494387918

metrics spy ief

annualized return: 0.10231288563106111
benchmark total return: 11.39742578500792
end: 2026-06-30
max drawdown: -0.2542612240937828
observations: 286
risk off fraction: 0.2727272727272727
sharpe: 0.8822748507509688
start: 2002-08-31
total return: 9.192164075887439
volatility: 0.1188110240974314

missingness

warm-up and unavailable states excluded, never converted to risk-on

publication lag months

unknown_native; one available month causal baseline

risk off events

risk off fraction

0.3336886993603412

same close status

research_only_sensitivity

signal span

1948-04-30/2026-06-30

source identity

UNRATE aggregate fallback; not native Google Trends/RecessionAlert

spec id

recessionalert.1345.unrate_regime.lb3.lag0.t0p0

threshold

0.0

unavailable signal observations

Confirm as strategy Confirm as indicator

recessionalert.1345.unrate_regime.lb6.lag1.t0p0

available signal observations

935

claim kind

inferred

classification

inferred causal best guess

execution

one available monthly bar lag is causal baseline; lag 0 is research-only

exploratory

True

family

unrate_regime

recessionalert.1345.unrate_regime.lb6.lag1.t0p0

input names

UNRATE

lag bars

lookback

metrics spy bil

annualized return: 0.07920074261902688
benchmark total return: 5.775363419995794
end: 2026-06-30
max drawdown: -0.23927178828328333
observations: 228
risk off fraction: 0.31140350877192985
sharpe: 0.6896882710580116
start: 2007-06-30
total return: 3.2554204341966066
volatility: 0.12155160802425947

metrics spy ief

annualized return: 0.08726948480189711
benchmark total return: 11.39742578500792
end: 2026-06-30
max drawdown: -0.23927178828328333
observations: 286
risk off fraction: 0.2902097902097902
sharpe: 0.7418373672746367
start: 2002-08-31
total return: 6.345757449032161
volatility: 0.12344110525342702

missingness

warm-up and unavailable states excluded, never converted to risk-on

publication lag months

unknown_native; one available month causal baseline

risk off events

risk off fraction

0.3497326203208556

same close status

not_used

signal span

1948-07-31/2026-06-30

source identity

UNRATE aggregate fallback; not native Google Trends/RecessionAlert

spec id

recessionalert.1345.unrate_regime.lb6.lag1.t0p0

threshold

0.0

unavailable signal observations

Confirm as strategy Confirm as indicator

recessionalert.1345.unrate_regime.lb12.lag2.t0p5

available signal observations

929

claim kind

inferred

classification

inferred conservative sensitivity

execution

one available monthly bar lag is causal baseline; lag 0 is research-only

exploratory

True

family

unrate_regime

recessionalert.1345.unrate_regime.lb12.lag2.t0p5

input names

UNRATE

lag bars

lookback

metrics spy bil

annualized return: 0.0762694432757085
benchmark total return: 5.775363419995794
end: 2026-06-30
max drawdown: -0.23927178828328333
observations: 228
risk off fraction: 0.33771929824561403
sharpe: 0.6907583030545535
start: 2007-06-30
total return: 3.041096252537522
volatility: 0.11653396624608206

metrics spy ief

annualized return: 0.08969910106326662
benchmark total return: 11.39742578500792
end: 2026-06-30
max drawdown: -0.25126795175834027
observations: 286
risk off fraction: 0.32867132867132864
sharpe: 0.802449207817426
start: 2002-08-31
total return: 6.747123114731935
volatility: 0.11576223958507297

missingness

warm-up and unavailable states excluded, never converted to risk-on

publication lag months

unknown_native; one available month causal baseline

risk off events

risk off fraction

0.34768568353067814

same close status

not_used

signal span

1949-01-31/2026-06-30

source identity

UNRATE aggregate fallback; not native Google Trends/RecessionAlert

spec id

recessionalert.1345.unrate_regime.lb12.lag2.t0p5

threshold

0.5

unavailable signal observations

Confirm as strategy Confirm as indicator

recessionalert.1345.unrate_regime.lb6.lag1.t0p5

available signal observations

935

claim kind

inferred

classification

inferred threshold neighbor

execution

one available monthly bar lag is causal baseline; lag 0 is research-only

exploratory

True

family

unrate_regime

recessionalert.1345.unrate_regime.lb6.lag1.t0p5

input names

UNRATE

lag bars

lookback

metrics spy bil

annualized return: 0.07920074261902688
benchmark total return: 5.775363419995794
end: 2026-06-30
max drawdown: -0.23927178828328333
observations: 228
risk off fraction: 0.31140350877192985
sharpe: 0.6896882710580116
start: 2007-06-30
total return: 3.2554204341966066
volatility: 0.12155160802425947

metrics spy ief

annualized return: 0.08726948480189711
benchmark total return: 11.39742578500792
end: 2026-06-30
max drawdown: -0.23927178828328333
observations: 286
risk off fraction: 0.2902097902097902
sharpe: 0.7418373672746367
start: 2002-08-31
total return: 6.345757449032161
volatility: 0.12344110525342702

missingness

warm-up and unavailable states excluded, never converted to risk-on

publication lag months

unknown_native; one available month causal baseline

risk off events

risk off fraction

0.3497326203208556

same close status

not_used

signal span

1948-07-31/2026-06-30

source identity

UNRATE aggregate fallback; not native Google Trends/RecessionAlert

spec id

recessionalert.1345.unrate_regime.lb6.lag1.t0p5

threshold

0.5

unavailable signal observations

Confirm as strategy Confirm as indicator

recessionalert.1345.unrate_regime.lb12.lag1.t1p0

available signal observations

929

claim kind

inferred

classification

inferred higher-threshold sensitivity

execution

one available monthly bar lag is causal baseline; lag 0 is research-only

exploratory

True

family

unrate_regime

recessionalert.1345.unrate_regime.lb12.lag1.t1p0

input names

UNRATE

lag bars

lookback

metrics spy bil

annualized return: 0.07621562357337064
benchmark total return: 5.775363419995794
end: 2026-06-30
max drawdown: -0.2960243788232897
observations: 228
risk off fraction: 0.34210526315789475
sharpe: 0.7064216534755113
start: 2007-06-30
total return: 3.0372584941174665
volatility: 0.11340473832454344

metrics spy ief

annualized return: 0.09016181711337024
benchmark total return: 11.39742578500792
end: 2026-06-30
max drawdown: -0.33226530246303465
observations: 286
risk off fraction: 0.32867132867132864
sharpe: 0.8179606824125406
start: 2002-08-31
total return: 6.8259074733833
volatility: 0.11384310087207021

missingness

warm-up and unavailable states excluded, never converted to risk-on

publication lag months

unknown_native; one available month causal baseline

risk off events

risk off fraction

0.348762109795479

same close status

not_used

signal span

1949-01-31/2026-06-30

source identity

UNRATE aggregate fallback; not native Google Trends/RecessionAlert

spec id

recessionalert.1345.unrate_regime.lb12.lag1.t1p0

threshold

1.0

unavailable signal observations

Confirm as strategy Confirm as indicator

recessionalert.1345.unrate_regime.lb21.lag1.tm0p5

available signal observations

920

claim kind

inferred

classification

inferred negative long-lookback variant

execution

one available monthly bar lag is causal baseline; lag 0 is research-only

exploratory

True

family

unrate_regime

recessionalert.1345.unrate_regime.lb21.lag1.tm0p5

input names

UNRATE

lag bars

lookback

metrics spy bil

annualized return: 0.05463588592459101
benchmark total return: 5.775363419995794
end: 2026-06-30
max drawdown: -0.38809262976694914
observations: 228
risk off fraction: 0.39035087719298245
sharpe: 0.5415486333207293
start: 2007-06-30
total return: 1.7475673909545022
volatility: 0.10954556510390455

metrics spy ief

annualized return: 0.06915131227328475
benchmark total return: 11.39742578500792
end: 2026-06-30
max drawdown: -0.41023092316366905
observations: 286
risk off fraction: 0.3776223776223776
sharpe: 0.6609508380014584
start: 2002-08-31
total return: 3.92153182714026
volatility: 0.11073455419527127

missingness

warm-up and unavailable states excluded, never converted to risk-on

publication lag months

unknown_native; one available month causal baseline

risk off events

risk off fraction

0.38804347826086955

same close status

not_used

signal span

1949-10-31/2026-06-30

source identity

UNRATE aggregate fallback; not native Google Trends/RecessionAlert

spec id

recessionalert.1345.unrate_regime.lb21.lag1.tm0p5

threshold

-0.5

unavailable signal observations

Confirm as strategy Confirm as indicator

Interactive lab

Interactive history unavailable

Loading available evidence…

Research record

Source claims, inferred rules, experiments and the evidence behind the assessment.

Source

article date: 2012-03-06
database: data/recessionalert.sqlite opened read-only
fetched at: 2026-08-22T23:39:13Z
inventory: research/findings/recessionalert_inventory.md
manifest: research/artifacts/recessionalert/manifest.json
native span: undisclosed; one dated article snapshot and no source-linked Fear observations
publication lag: unknown; the article's two-week NBER phrase is not a recurring release lag
redacted url: https://recessionalert.com/a-recession-fear-indicator/
sha256: sha256:2cb8464ee21e8b28af0f8dcbf7bacbd3ddeffc040c53717a198c0b701fa019dd

What the source claims

The article describes global Google search volume for recession as a promising recession indicator and fear gauge, with contextual 2008 NBER and August 2011 observations but no downloadable series.
The article says the 2008 NBER start was within a two-week lag of the search peak; this is event context, not an executable release lag.
Monthly FRED UNRATE rate-of-change is used as a long-history recession-risk proxy with explicitly inferred lookbacks, thresholds, and lags.
SPY risk-on versus IEF/BIL defensive sleeves are descriptive comparison overlays; the source discloses no allocation or target.

Rules actually disclosed

The article describes Google global search volume for recession as a fear gauge and gives only contextual 2008 NBER and August 2011 observations.
No buy/sell/hold side, threshold, query configuration, native observations, recurring publication lag, target, sizing, or exit rule is disclosed.

What had to be inferred

FRED UNRATE monthly rate-of-change was selected as the longest public recession-relevant proxy; SPY/IEF/BIL are descriptive sleeves.
Lookbacks 3/6/12/21, thresholds -0.5/0/0.5/1, and lag 0/1/2 were exploratory choices, not source settings.

Research questions

What can be learned from the SHA-verified RecessionAlert Google-search fear-gauge narrative using transparent public proxies when its series, thresholds, publication lag, and market rule are incomplete?

Data

canonical source: sha256:2cb8464ee21e8b28af0f8dcbf7bacbd3ddeffc040c53717a198c0b701fa019dd
canonical source paths: research/findings/recessionalert_inventory.md
research/artifacts/recessionalert/manifest.json
data/recessionalert.sqlite
database policy: All reads use SQLite mode=ro; no crawl or market cache mutation.
primary common window: 2007-06 through 2026-06 after monthly alignment
proxy inputs: FRED UNRATE snapshot in data/market.sqlite
Yahoo SPY/IEF/BIL snapshots in data/market.sqlite
redistribution: Derived proxy metrics and summaries only; no raw page/media, credentials, cookies, or private account data.

Baseline implementation

available signal observations

938

claim kind

inferred

classification

inferred lookahead diagnostic

execution

one available monthly bar lag is causal baseline; lag 0 is research-only

exploratory

True

family

unrate_regime

recessionalert.1345.unrate_regime.lb3.lag0.t0p0

input names

UNRATE

lag bars

lookback

metrics spy bil

annualized return: 0.10223803996031511
benchmark total return: 5.775363419995794
end: 2026-06-30
max drawdown: -0.2420584677802402
observations: 228
risk off fraction: 0.29385964912280704
sharpe: 0.8894778625274528
start: 2007-06-30
total return: 5.356711589315653
volatility: 0.11758829494387918

metrics spy ief

annualized return: 0.10231288563106111
benchmark total return: 11.39742578500792
end: 2026-06-30
max drawdown: -0.2542612240937828
observations: 286
risk off fraction: 0.2727272727272727
sharpe: 0.8822748507509688
start: 2002-08-31
total return: 9.192164075887439
volatility: 0.1188110240974314

missingness

warm-up and unavailable states excluded, never converted to risk-on

publication lag months

unknown_native; one available month causal baseline

risk off events

risk off fraction

0.3336886993603412

same close status

research_only_sensitivity

signal span

1948-04-30/2026-06-30

source identity

UNRATE aggregate fallback; not native Google Trends/RecessionAlert

spec id

recessionalert.1345.unrate_regime.lb3.lag0.t0p0

threshold

0.0

unavailable signal observations

What to try interactively

default
recessionalert.1345.unrate_regime.lb6.lag1.t0p0
description
Choose an inferred proxy variant; no setting is disclosed by RecessionAlert.
name
variant
values
- recessionalert.1345.unrate_regime.lb3.lag0.t0p0
- recessionalert.1345.unrate_regime.lb6.lag1.t0p0
- recessionalert.1345.unrate_regime.lb12.lag2.t0p5
- recessionalert.1345.unrate_regime.lb6.lag1.t0p5
- recessionalert.1345.unrate_regime.lb12.lag1.t1p0
- recessionalert.1345.unrate_regime.lb21.lag1.tm0p5
default
1
description
Lag 1 is causal baseline; lag 0 is research-only sensitivity.
name
execution_lag
values
- 0
- 1
- 2
default
IEF
description
Descriptive inferred sleeve; source target is undisclosed.
name
defensive_sleeve
values
- IEF
- BIL

Suggested next research

Can an authorized redistributable source-linked Google Trends series and query configuration be obtained?
What exact threshold, smoothing, state, and market mapping did the source use?
What recurring publication timestamps, vintages, revisions, and point-in-time cutoffs apply?
Can the reported 2008 and 2011 narratives be recomputed from source observations?

Trial ledger

iteration
1
status
published
summary
Offline public-proxy campaign; native source-fidelity gaps remain.

Parameter sensitivity

Timing changes the IEF total to 8.4878 (lag 0), 6.3458 (lag 1), and 9.7664 (lag 2), all below SPY; threshold and lookback neighbors are negative or weaker.

State-space exploration

Six inferred variants were executed on the tracked cache and all remain in state-space.json with signal counts, risk-off fractions, events, and IEF/BIL versus SPY totals.
The causal lb6/threshold0/lag1 baseline returned 6.3458 in the IEF sleeve versus 11.3974 for SPY and 3.2554 in BIL versus 5.7754 for SPY.
The lag-0 lb3 diagnostic returned the highest IEF total among tested variants but is lookahead and not deployable.

Timing and cost sensitivity

Native Google Trends publication and vintage timing are unknown. Lag 1 is the conservative causal baseline; lag 0 is shown only as a research-only sensitivity. Transaction costs are not disclosed by the source and are not fabricated.

What did not work

No tested proxy variant beat aligned SPY in either defensive sleeve comparison, and aggregate UNRATE cannot validate the native Google-search identity.

What the signal looks like

The interactive lab plots the inferred UNRATE signal and normalized SPY market path; lag 1 is the causal baseline and lag 0 is research-only.

What worked

The proxy campaign is reproducible offline, has a long UNRATE history, preserves unavailable warm-up values, and exposes a readable interactive signal/equity lab.

Next questions

Can an authorized redistributable source-linked Google Trends series and query configuration be obtained?
What exact threshold, smoothing, state, and market mapping did the source use?
What recurring publication timestamps, vintages, revisions, and point-in-time cutoffs apply?
Can the reported 2008 and 2011 narratives be recomputed from source observations?

Review status

published. Research publication does not imply official admission.

Return to pending research