IESET.
Hypotheses·monetary·qe_asset_inflation_vs_cpi_divergence_post_2008

Post-2008 large-scale asset purchase programmes by the Federal Reserve, ECB, Bank of England, and Bank of Japan produced a measurable divergence between asset-price inflation (equities and residential real estate) and headline consumer-price inflation until roughly 2021.

Specifically, cumulative real asset-price appreciation in these economies exceeded cumulative real consumer-price change by a large margin over 2009-2020, with the gap only closing once the 2021-2023 CPI shock materialised. The divergence is consistent with portfolio-rebalancing and safe-asset channels operating without the textbook quantity-theoretic pass-through to consumer prices, and is compatible with the scope caveat carried in the Austrian D.1.5 spec and with New Keynesian models that assign limited direct CPI pass-through to reserve expansion at the zero lower bound.

REFUTEDengine/runs/qe_asset_inflation_vs_cpi_divergence_post_2008

refuted — Only 2 of 8 countries had even a 0.10 log-point asset-vs-CPI gap by 2020 (mean GAP_2020 = -0.02). The post-2008 divergence story does not survive a panel test.

confidence cueThis test cuts against the claim as written or misses its pre-declared threshold.

policy briefNeeds review

In ordinary language

In plain terms, this asks whether central bank balance sheet to income is actually linked to better or worse cumulative real equity return from 2008 to 2023.

plain answer

The data did not support the prediction. Only 2 of 8 countries had even a 0.10 log-point asset-vs-CPI gap by 2020 (mean GAP_2020 = -0.02).

why it matters

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

how the test works

It compares 8 country or place units from 2008 to 2023, using a local projections design, with fixed effects for country and year.

what was measured
What changed
  • Central bank balance sheet to income
What we checked
  • Cumulative real equity return
  • Cumulative real house price
  • Cumulative core cpi
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

No evidence packet has been generated yet.

Results

engine/runs/qe_asset_inflation_vs_cpi_divergence_post_2008
Loading chart…

Who has skin in the game — schools predicting on this

5 schools 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.

Austrian Business Cycle Theory (ABCT)wins if SUPPORTEDschool refutedChicago School Monetarismwins if MIXEDpartial/no-callEmpirical Pragmatist (the framework's house position)wins if SUPPORTEDschool refutedMarxian Political Economywins if SUPPORTEDschool refutedPost-Keynesianwins if SUPPORTEDschool refuted

Pre-registration

pre-registered
first-spec commit 4c8ce8e · 2026-07-18T22:11:21Z

Post-2008 large-scale asset purchase programmes by the Federal Reserve, ECB, Bank of England, and Bank of Japan produced a measurable divergence between asset-price inflation (equities and residential real estate) and headline consumer-price inflation until roughly 2021. Specifically, cumulative real asset-price appreciation in these economies exceeded cumulative real consumer-price change by a large margin over 2009-2020, with the gap only closing once the 2021-2023 CPI shock materialised. The divergence is consistent with portfolio-rebalancing and safe-asset channels operating without the textbook quantity-theoretic pass-through to consumer prices, and is compatible with the scope caveat carried in the Austrian D.1.5 spec and with New Keynesian models that assign limited direct CPI pass-through to reserve expansion at the zero lower bound.

Falsification criterion — what would disprove this

set before the run · honoured after

This hypothesis is considered falsified if:

PRIMARY A (dispositive): for each country i in {USA, GBR, JPN, DEU, FRA, ITA, ESP, NLD}, define GAP_i(h) = log(real_asset_index_i(h) / real_asset_index_i(2008)) - log(cpi_index_i(h) / cpi_index_i(2008)) where the asset index is the equal-weighted REAL ASSET BASKET of three classes — (1) equity total return, (2) residential property price, (3) long-dated sovereign bond price index. A country counts toward the panel ONLY if at least 2 of 3 asset-class series are on disk; single-asset substitution does NOT satisfy the basket requirement and the country is dropped with a method-validity flag. The v1 spec (single-asset using Shiller for USA + BIS residential property for the other seven) was an unfair test because European house prices were weak 2008-2020 for idiosyncratic reasons (Spain crash, Italy flat, Germany sub-trend until 2015) while DAX/CAC/FTSE/Nikkei roughly doubled-to-tripled over the same window — i.e. the QE-portfolio-rebalancing channel operates precisely through the equity / long-bond legs that v1 omitted. v2 corrects this. The hypothesis is SUPPORTED iff at least 5 of 8 countries have GAP_i(2020) >= 0.30 log-points (~+35 pp asset-vs-CPI excess) on a qualifying basket. PRIMARY B (sanity check on regime-conditionality): at least 4 of the eligible countries (those with GAP_i(2020) > 0) must narrow the gap by at least 25 % between 2020 and 2023, i.e. GAP_i(2023) <= 0.75 * GAP_i(2020). The 25 % bar is weaker than the original 50 % because the post-2023 asset rally re-widened the gap in the US; promotion methodology_note records the change. REFUTED iff fewer than 3 of 8 countries cleared even a 0.10 log-point gap by 2020 (i.e. the divergence story does not survive a panel test). INFORMATIVE: per-country dispersion of GAP_2020; whether the USA central-bank balance sheet (FRED:WALCL / FRED:GDP) crossed 15 % in 2009-2020 (regime-confirmation indicator). METHOD_VALID: at least 5 of 8 countries must have BIS WS_LONG_CPI plus a qualifying asset basket (>=2 of {equity total return, residential property, sovereign bond price}) covering 2008-2020. If the gate fails — i.e. only residential property is available cross-country — emit `inconclusive — data gap on cross-country equity total returns and sovereign bond price indices`.

formal test & threshold
test:      cumulative_log_gap_panel_h2020_with_h2023_reconvergence
threshold: PRIMARY A: count(GAP_i(2020) >= 0.30) >= 5 of 8. PRIMARY B: count(GAP_i(2023) <= 0.75 * GAP_i(2020) | GAP_i(2020) > 0) >= 4. REFUTED if count(GAP_i(2020) >= 0.10) < 3. METHOD_VALID if at least 5 of 8 countries have both an asset and a CPI series covering 2008-2020.

Method

Template
local_projections
Fixed effects
country, year
Clustering
country
Sample
8 countries · 2008 – 2023
Evidence type
associational

Jordà local projections of cumulative real asset-price response and cumulative core-CPI response to a unit change in central-bank balance sheet / GDP, horizon 1-8 years. Primary comparison is the gap between asset-price impulse response and CPI impulse response at each horizon. Robustness: narrative identification using pure QE-announcement windows (Fawley-Neely 2013 coding) reported as a v2 companion specification.

Data

VariableSourceTransform
cumulative_real_equity_return
outcome
shiller:ie_data (USA)tier 3
bis:WS_SPPtier 2
real_cumulative_index_base_2008
cumulative_real_house_price
outcome
bis:WS_SPPtier 2
real_cumulative_index_base_2008
cumulative_core_cpi
outcome
oecd:OECD.SDD.TPStier 2
fred:CPILFESLtier 1
real_cumulative_index_base_2008
central_bank_balance_sheet_to_gdp
treatment
fred:WALCLtier 1
fred:GDPtier 1
boj:monetary_basetier 1
fred:JPNASSETStier 1
ratio_to_nominal_gdp
short_rate
control
fred:DFF (USA)tier 1
ecb:FM (EUR)tier 1
boe:IUDSOIA (GBR)tier 1
boj:policy_rate (JPN)tier 1
level
output_gap
control
oecd:OutputGaptier 2
level

● ready  · ● pending  · ● reconstruct-needed

Detailed result card

Post-2008 QE asset inflation vs CPI divergence

Verdict: refuted — Only 2 of 8 countries had even a 0.10 log-point asset-vs-CPI gap by 2020 (mean GAP_2020 = -0.02). The post-2008 divergence story does not survive a panel test.

Summary

  • Sample: 8 advanced QE economies (USA, GBR, JPN, DEU, FRA, ITA, ESP, NLD).
  • Cumulative-log gap between asset prices (equities for USA, BIS residential property for others) and CPI, 2008 base.
  • Mean GAP_2020 = -0.02 log-points (~-2% asset-vs-CPI excess).
  • 1 of 8 countries cleared the 0.3 log-point threshold (PRIMARY A; need 5).
  • 1 of 3 eligible countries narrowed the gap by >= 25% by 2023 (PRIMARY B; need 4).
  • USA WALCL/GDP > 15% in post-2008 window: True (2009 ratio = 14.39%).

Method

  1. CPI cumulative log-index from BIS WS_LONG_CPI annual %change.
  2. Asset cumulative log-index from Shiller real total return (USA primary) or BIS WS_SPP residential property (other countries).
  3. GAP_h = asset_log(h) - cpi_log(h). Compute at h=2020 (PRIMARY A) and h=2023 (PRIMARY B - sanity check that the 2021-23 CPI shock narrowed the gap).
  4. Verdict requires PRIMARY A AND PRIMARY B. REFUTED if fewer than 3 countries cleared the 0.10 log-point floor.

Caveats

  • The spec called for cross-country central-bank balance-sheet treatment series (ECB, BoE, BoJ); only FRED:WALCL is on disk for the USA. The hypothesis is therefore tested as an associational post-2008 divergence, not as an LP impulse response. The QE-active indicator confirms regime classification for the USA only.
  • Property-price index used as the asset proxy for non-US countries. This is conservative for the hypothesis: equity returns over 2008-2020 typically exceeded property returns, so the asset side of the gap is understated for the euro-area panel.
  • 2020 used as the primary horizon to match the spec's '2009-2020 window only closing once the 2021-2023 CPI shock materialised'.

Data

  • bis:WS_LONG_CPI (cross-country annual CPI %change)
  • bis:WS_SPP (residential property index — euro panel)
  • shiller:ie_data (US S&P composite real total return)
  • shiller:home_price_index (US real home price)
  • fred:WALCL + fred:GDP (US Fed balance-sheet ratio)
  • world_bank_wdi:NY.GDP.MKTP.KD (real GDP panel control)

Strongest opposing argument

Every hypothesis ships with its charitable opposing argument. The framework earns credibility by handling objections at their strongest, not weakest.

Notes

Per D.1.5 the hypothesis does NOT claim that QE caused the 2021-2023 CPI shock; causality of that shock is out of scope. The claim is narrowly that the 2009-2020 divergence between asset and consumer inflation is documentable and consistent with specific transmission channels, not with strong quantity-theoretic pass-through at the ZLB.

Authored framework. Read the transparency note.