Pre-registration
The 2007-2009 financial crisis originated in endogenous leverage build-up (Minsky hedge→speculative→Ponzi financing transition) within financial sector balance sheets, not in exogenous shocks. Tested by examining whether sectoral leverage and maturity-mismatch indicators rose systematically over the 2000-2007 pre-crisis window and crossed Minsky-style thresholds prior to the 2008 rupture, independent of any named deregulation episode.
Falsification criterion — what would disprove this
This hypothesis is considered falsified if:
The hypothesis is SUPPORTED if at least 3 of 5 pre-registered Minsky- indicator series (broker-dealer leverage ratio, shadow-bank maturity mismatch, MBS issuance growth, household-debt-to-GDI ratio, residential mortgage debt-to-income) show systematic monotonic rise over 2000-2007 and cross their pre-specified Minsky thresholds before Q1 2008. REFUTED if fewer than 2 indicators meet the threshold or if the rise is non-monotonic.
formal test & threshold
test: Pre/post-crisis trajectory of 5 Minsky indicators (broker-dealer leverage, shadow-bank maturity mismatch, MBS issuance, household DTI, mortgage debt/income) over 2000-2007; >=3/5 show monotonic rise crossing thresholds before Q1-2008 supports.
Method
- Template
descriptive- Sample
- 9 countries · 2000 – 2009
- Evidence type
- associational
Multi-indicator time-series scorecard 2000-2007 across USA + UK + IRL + ISL + ESP. For each pre-registered Minsky indicator (broker-dealer leverage, household debt-to-GDI, mortgage-growth, MBS issuance, house- price-to-income), test for systematic monotonic rise and threshold- crossing pre-Q1-2008. Hypothesis supported if >=3 of 5 indicators meet monotonicity + threshold criteria. Caveat: thresholds are stipulated not estimated — flagged TODO swarm-6e for sensitivity.
Data
| Variable | Source | Transform |
|---|---|---|
broker_dealer_leverage_ratio outcome | world_bank_wdi:FS.AST.PRVT.GD.ZStier 2 | level |
household_debt_to_gdi_ratio outcome | world_bank_wdi:FS.AST.PRVT.GD.ZStier 2 | level |
mortgage_debt_share_gdp outcome | fred:MORTGAGEtier 1 | level |
real_house_price_index outcome | oecd:OECD.SDD.PINtier 2 | log |
minsky_threshold_crossing_indicator treatment | constructed:indicator = 1 for country-quarter when at least 3 of 5 pre-registered Minsky indicators cross their thresholds (broker-dtier 5 | indicator |
federal_funds_rate control | fred:FEDFUNDStier 1 | level |
log_gdp_per_capita control | world_bank_wdi:NY.GDP.PCAP.KDtier 2 | log |
● ready · ● pending · ● reconstruct-needed
Detailed result card
GFC endogenous Minsky leverage build-up — pre-crisis scorecard
Verdict: SUPPORTED — 3 of 4 Minsky indicators passed (>=5 of 9 sample countries showing BOTH significant upward 2000-2007 trend AND magnitude threshold). I1: 7/8 PASS | I2: 4/8 fail | I3: 6/7 PASS | I4: 8/9 PASS
Summary
- Sample: 9-country panel (USA, GBR, IRL, ISL, ESP, NLD, DEU, FRA, ITA), 2000Q1-2007Q4 quarterly.
- 3 of 4 Minsky indicators passed the cross-country test (need >=3 for SUPPORTED, <2 = REFUTED).
- METHOD_VALID gate: 8/9 countries have >=3 indicators on disk (floor: 6). PASS.
Per-indicator results
- I1 BIS private non-fin credit-to-GDP ratio (level) — 7/8 countries passed (need >=5). PASS — passing: ESP, FRA, GBR, IRL, ITA, NLD, USA.
- I2 BIS credit-to-GDP gap (deviation from HP-trend) — 4/8 countries passed (need >=5). FAIL — passing: ESP, IRL, ITA, USA.
- I3 BIS household debt-service ratio — 6/7 countries passed (need >=5). PASS — passing: ESP, FRA, GBR, ITA, NLD, USA.
- I4 BIS real residential property price index — 8/9 countries passed (need >=5). PASS — passing: ESP, FRA, GBR, IRL, ISL, ITA, NLD, USA.
Method
For each (country, indicator) cell over 2000Q1-2007Q4: fit a plain OLS time trend (in quarter-index units; log-transformed for the real-house-price indicator I4) and test whether the slope is positive at p<0.10. Separately compute the magnitude change from 2000Q1 to 2007Q4 and compare against an indicator-specific BIS early-warning-literature threshold:
- I1 credit-to-GDP level: rise >= 20.0pp.
- I2 credit-to-GDP gap: peak >= 10.0pp during 2005-2007 (Drehmann-Borio-Tsatsaronis 2011 'warning' band).
- I3 household debt-service ratio: rise >= 1.0pp.
- I4 real residential property price: cumulative log-rise >= 0.18 (~20%).
Country-indicator cell passes iff BOTH conditions hold. Indicator passes iff >=5 of 9 countries pass. Hypothesis SUPPORTED iff >=3 of 4 indicators pass; REFUTED iff <2 pass; PARTIAL if exactly 2 pass.
P-values are computed from a normal-tail approximation to the OLS-slope t-statistic (sample n~32 quarters; the threshold p<0.10 is far from any boundary case in this run, so the t-vs-normal gap is immaterial).
Data
- bis:WS_CREDIT_GAP — credit-to-GDP ratio (CG_DTYPE=A) and gap (CG_DTYPE=C).
- bis:WS_DSR — household debt-service ratio (DSR_BORROWERS=H).
- bis:WS_SPP — real residential property prices (VALUE=R).
Caveats
- Iceland (ISL) is not in BIS WS_CREDIT_GAP or WS_DSR; it carries only the real-house-price indicator. It is treated as missing (not as a failed test) per the METHOD_VALID rule.
- Ireland (IRL) is not in WS_DSR.
- Indicators are stipulated proxies for the original spec's broker- dealer leverage and MBS-issuance series, neither of which is on disk. If those become available in a future fetcher pass, the spec should be re-promoted to v2.
- This is an associational pre-crisis trajectory test, not a causal identification of crisis origin. Even a clean SUPPORTED verdict is consistent with deregulation-driven (rather than purely endogenous) leverage build-up.
Notes
Origin is auto-generated Phase-4E stub seeded from post-Keynesian framing of GFC as endogenous-Minsky leverage build-up rather than deregulation-episode-to-crisis. Human review required to pin indicator series and threshold-crossing methodology.