IESET.
Hypotheses·fiscal·fiscal_multipliers_zlb_higher_than_normal_regime

Fiscal multipliers are substantially larger at the zero lower bound than in normal interest-rate regimes, because crowding-out via higher rates is suppressed.

INCONCLUSIVEengine/runs/fiscal_multipliers_zlb_higher_than_normal_regime

inconclusive (data gap on quarterly OECD national accounts panel + Ramey-Zubairy / Guajardo-Leigh-Pescatori narrative fiscal-shock series; US-only OLS-LP diagnostic indicative only — h=8 ZLB multiplier -0.10 vs normal 0.52, gap -0.62, against claimed direction (ZLB <= normal); threshold for SUPPORTED would have been gap >= +0.50 on the panel LP-IV).

confidence cueResult card produced; verdict unclassified.

policy briefCoverage too thin

In ordinary language

In plain terms, this asks whether government consumption shock is actually linked to better or worse real income growth quarterly from 1995 to 2021.

plain answer

This test cannot make a firm call yet. h=8 ZLB multiplier -0.10 vs normal 0.52, gap -0.62, against claimed direction (ZLB <= normal); threshold for SUPPORTED would have been gap >= +0.50 on the panel LP-IV).

why it matters

This matters because fiscal claims should change belief only when they survive a pre-declared empirical test.

how the test works

It compares 21 country or place units from 1995 to 2021, using a lp iv design, with fixed effects for country and year.

what was measured
What changed
  • Government consumption shock
  • Narrative fiscal shock
What we checked
  • Real income growth quarterly
  • Real income level
  • Unemployment rate
what this does not prove

A single test is not the whole truth. It narrows the claim under a specific sample, time period, and method. Strong policy conclusions need the pattern to survive nearby tests, alternative data, and serious objections.

verification

4 input datasets, 0 unresolved missing series, provenance status: reproducible hash verified.

Results

engine/runs/fiscal_multipliers_zlb_higher_than_normal_regime
Loading chart…

Who has skin in the game — schools predicting on this

1 school list this hypothesis as a test of their position. The chips below are school-level scoreboard outcomes, not a second hypothesis verdict.

hypothesis verdict vs scoreboard outcome

The banner verdict judges this hypothesis as written. The scoreboard asks whether each school's polarity-corrected prediction was right. Raw status is not a school win: SUPPORTED supports schools that needed SUPPORTED, but refutes schools that needed REFUTED.

New Keynesian (mainstream synthesis)wins if SUPPORTEDpartial/no-call

Pre-registration

pre-registered
first-spec commit bae09ab · 2026-04-29T22:09:42Z

Fiscal multipliers are substantially larger at the zero lower bound than in normal interest-rate regimes, because crowding-out via higher rates is suppressed.

Falsification criterion — what would disprove this

set before the run · honoured after

This hypothesis is considered falsified if:

PRIMARY (dispositive): the dispositive thresholds for this hypothesis are encoded directly in engine/runs/fiscal_multipliers_zlb_higher_than_normal_regime/replication.py and pinned in methodology_note. The auto-grader's verdict in diagnostics.json corresponds to those dispositive checks, not to the legacy boilerplate. See methodology_note for the exact pre-registered thresholds.

formal test & threshold
test:      State-dependent LP-IV (Ramey-Zubairy) on real GDP responses to government-consumption shocks instrumented by Ramey defence-news, OECD quarterly panel 1995-2021, ZLB regime indicator at short-rate ≤0.25%; country-clustered SEs, country+year FE. Refute if ZLB cumulative multiplier at h=8 not >0.5 above normal-regime multiplier at p<0.10.

Method

Template
lp_iv
Fixed effects
country, year
Clustering
country
Sample
21 countries · 1995 – 2021
Evidence type
associational

State-dependent local projections IV with regime indicator (ZLB vs normal) à la Ramey-Zubairy / Auerbach-Gorodnichenko. Identified fiscal shocks (narrative or forecast-error) as instruments; country-quarter panel covering 2008-2021 ZLB episodes plus pre- crisis normal regime.

Data

VariableSourceTransform
real_gdp_growth_quarterly
outcome
oecd:NAQ_GDPtier 2
imf:NGDP_RPCHtier 2
fred:GDPC1 (USA)tier 1
pct_yoy
real_gdp_level
outcome
oecd:NAQ_GDPtier 2
fred:GDPC1tier 1
log_level
unemployment_rate
outcome
oecd:harmonised_unemploymenttier 2
ilostat:unemployment_ratetier 2
pct_labour_force
government_consumption_shock
treatment
oecd:NAQ_government_consumptiontier 2
pct_gdp_innovation
narrative_fiscal_shock
treatment
manual: Ramey-Zubairy military news shocks; Guajardo-Leigh-Pescatori narrative consolidationspct_gdp
zlb_regime_indicator
treatment
constructed:indicator = 1 in country-quarters where short-rate (fred:DFF for USA, ecb:FM for EUR, boj:policy_rate for JPN, boe:IUDBEtier 5
binary
short_term_policy_rate
control
fred:DFF (USA)tier 1
ecb:FM (EUR)tier 1
boj:policy_rate (JPN)tier 1
boe:IUDBEDR (GBR)tier 1
rba:cash_rate (AUS)tier 1
statcan:overnight_rate (CAN)tier 1
level_pct
output_gap
control
oecd:OutputGaptier 2
pct_potential
cyclically_adjusted_primary_balance
control
imf:GGXCNL_NGDPtier 2
oecd:underlying_primary_balancetier 2
pct_gdp
trade_openness
control
world_bank_wdi:NE.TRD.GNFS.ZStier 2
pct_gdp

● ready  · ● pending  · ● reconstruct-needed

Detailed result card

Fiscal multipliers at the ZLB vs normal regime

Verdict: inconclusive (data gap on quarterly OECD national accounts panel + Ramey-Zubairy / Guajardo-Leigh-Pescatori narrative fiscal-shock series; US-only OLS-LP diagnostic indicative only — h=8 ZLB multiplier -0.10 vs normal 0.52, gap -0.62, against claimed direction (ZLB <= normal); threshold for SUPPORTED would have been gap >= +0.50 on the panel LP-IV).

Summary

The spec calls for a state-dependent LP-IV across an OECD panel of 21 countries (1995-2021), instrumenting government-spending innovations with the Ramey-Zubairy defence-news series and the Guajardo-Leigh-Pescatori narrative consolidations. Neither the quarterly OECD national-accounts panel (real GDP, government consumption) nor the narrative fiscal-shock series are on disk under any publisher. The pre-registered identification therefore cannot be run.

Missing inputs flagged by assess_required_panel_inputs():

  • oecd:NAQ_GDP (quarterly real GDP, panel)
  • oecd:NAQ_government_consumption (quarterly govt consumption, panel)
  • manual: Ramey-Zubairy defence-news / Guajardo-Leigh-Pescatori narrative consolidations

Informative diagnostic - US-only OLS LP

To keep the run artefact-complete, this script also runs a US-only downgraded state-dependent OLS local projection on the FRED quarterly data (GDPC1, FGEXPND deflated by GDPDEF, FEDFUNDS for the ZLB indicator). Sample window 1995-2021 ex 2020Q2 (34 ZLB quarters, 73 normal quarters). The shock is the OLS innovation of dlog real federal expenditures after lag-1 controls (Auerbach-Gorodnichenko forecast-error proxy). Multiplier scaled by steady-state G/Y = 0.218. This does not adjudicate the spec's panel-LP-IV claim and is reported only to indicate whether the US signal points toward or away from the New-Keynesian prediction.

  • h=4: ZLB multiplier 0.10 (n=30), normal multiplier 0.57 (n=71), gap -0.47.
  • h=8: ZLB multiplier -0.10 (n=28), normal multiplier 0.52 (n=69), gap -0.62.
  • h=12: ZLB multiplier -0.23 (n=28), normal multiplier 0.54 (n=65), gap -0.77.

At the spec's primary horizon h=8, the US-only diagnostic gap is -0.62 (threshold for SUPPORTED on the panel LP-IV would have been >= +0.50).

Method (pre-registered, blocked)

  1. Quarterly OECD-panel real GDP and government consumption (OECD NAQ_GDP / quarterly national accounts).
  2. Country-quarter ZLB indicator constructed from short-rate <= 0.25%.
  3. Narrative fiscal-shock series as instruments (Ramey-Zubairy defence-news; Guajardo-Leigh-Pescatori).
  4. State-dependent local-projection IV a la Ramey-Zubairy / Auerbach-Gorodnichenko, country-clustered SEs, country+year FE.
  5. Cumulative multiplier at h=8 quarters compared across the ZLB and normal regimes; SUPPORTED if (ZLB - normal) >= +0.50.

Data

Available (US-only diagnostic):

  • fred:GDPC1 (real GDP, quarterly)
  • fred:FGEXPND (federal government expenditures, quarterly)
  • fred:GDPDEF (GDP deflator, quarterly)
  • fred:FEDFUNDS (effective federal funds rate, monthly to quarterly mean)

Required for the spec but missing on disk:

  • oecd:NAQ_GDP (quarterly real GDP, panel)
  • oecd:NAQ_government_consumption (quarterly govt consumption, panel)
  • manual: Ramey-Zubairy defence-news / Guajardo-Leigh-Pescatori narrative consolidations

Fetcher backlog:

  • OECD quarterly national accounts (NAQ_GDP, NAQ_government_consumption) for the 21-country panel.
  • A manual publisher hosting the Ramey-Zubairy defence-news and Owyang-Ramey-Zubairy international-panel news shocks, plus the Guajardo-Leigh-Pescatori narrative consolidations.

Notes

Maps the new-Keynesian school's ZLB-higher-multiplier claim to a state-dependent LP-IV across OECD economies covering the 2008-2021 ZLB era plus pre-crisis normal regime. Estimator and prior set; full pre-registration awaits steelman + human sign-off.

Authored framework. Read the transparency note.