IESET.
Hypotheses·regulatory·banking_crisis_laeven_valencia_predictors_panel

In the Laeven-Valencia systemic-banking-crisis panel covering 1980-2020, four ex-ante observables jointly predict crisis onset: (i) high private-credit-to-GDP, (ii) negative current-account balance, (iii) real-effective-exchange-rate appreciation versus a 3-year trailing average, and (iv) low foreign-exchange reserve coverage of short-term external liabilities.

A logit on these four predictors plus country and year fixed effects achieves out-of-sample crisis-prediction AUC of at least 0.75 on a rolling-window cross-validation across the full advanced-and-emerging panel.

SUPPORTEDengine/runs/banking_crisis_laeven_valencia_predictors_panel

SUPPORTED — coef=+0.0006819 (sign matches claim +), p=0.000139

confidence cueThis is a clear pass for the claim as written. It still applies only to this sample, period, and method.

policy briefNeeds review

In ordinary language

In plain terms, this asks whether private credit to income is actually linked to better or worse systemic banking crisis onset from 1980 to 2020.

plain answer

The data clearly moved in the predicted direction. coef=+0.0006819 (sign matches claim +), p=0.000139

why it matters

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

how the test works

It compares 34 country or place units from 1980 to 2020, using a panel fe design, with fixed effects for country and year.

what was measured
What changed
  • Private credit to income
  • Current account balance pct income
What we checked
  • Systemic banking crisis onset
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

0 input datasets, 0 unresolved missing series, provenance status: no input vintages recorded.

Results

engine/runs/banking_crisis_laeven_valencia_predictors_panel
1007550250198020002020USAGBRFRADEUITAESPNLD
illustrative sketch · run pending
No coefficients yet. When the model fires, this chart will show systemic_banking_crisis_onset across 34 sampled countries over 1980–2020.
The shapes above are stylised — none of the lines are real data.
Placeholder for banking_crisis_laeven_valencia_predictors_panel. Published chart will be generated from engine/runs/banking_crisis_laeven_valencia_predictors_panel/chart_data.json.

Who has skin in the game — schools predicting on this

4 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.

Democratic Socialism (Bernsteinian revisionism + modern)wins if SUPPORTEDschool supportedMarxian Political Economywins if SUPPORTEDschool supportedModern Monetary Theorywins if SUPPORTEDschool supportedPost-Keynesianwins if SUPPORTEDschool supported

Pre-registration

registration ordering unverified
first-spec commit 4c8ce8e · 2026-07-18T22:11:21Z
run generated · 2026-06-29T17:54:27Z
Run timestamp predates this path's first git-add commit (rebase, rename, or pre-git local run). Spec hash is still the path's first-add commit — not repository HEAD — but ordering is not a clean pre-registration proof.

In the Laeven-Valencia systemic-banking-crisis panel covering 1980-2020, four ex-ante observables jointly predict crisis onset: (i) high private-credit-to-GDP, (ii) negative current-account balance, (iii) real-effective-exchange-rate appreciation versus a 3-year trailing average, and (iv) low foreign-exchange reserve coverage of short-term external liabilities. A logit on these four predictors plus country and year fixed effects achieves out-of-sample crisis-prediction AUC of at least 0.75 on a rolling-window cross-validation across the full advanced-and-emerging panel.

Falsification criterion — what would disprove this

set before the run · honoured after

This hypothesis is considered falsified if:

SUPPORTED if out-of-sample AUC across rolling-origin CV >= 0.75 AND each of the four predictors carries a marginal coefficient with sign matching the literature (credit +, CAB -, REER appreciation +, reserves -) at p < 0.10. PARTIAL if AUC in [0.65, 0.75). REFUTED if AUC < 0.65 OR if at least two predictors flip sign.

formal test & threshold
test:      panel_logit_rolling_origin_cv
threshold: out_of_sample_AUC >= 0.75 AND four_predictor_signs_match

Method

Template
panel_fe
Fixed effects
country, year
Clustering
country
Sample
34 countries · 1980 – 2020
Evidence type
associational

Pooled-logit predictor model with country and year FE. Out-of-sample AUC by 5-year rolling-origin cross-validation. Robustness: random-forest with same predictors; pre-2000 vs post-2000 split.

Data

VariableSourceTransform
systemic_banking_crisis_onset
outcome
owid:systemic-banking-crisestier 2
binary_first_year_of_episode
private_credit_to_gdp
treatment
world_bank_wdi:FS.AST.PRVT.GD.ZStier 2
level_pct_of_gdp
current_account_balance_pct_gdp
treatment
world_bank_wdi:BN.CAB.XOKA.GD.ZStier 2
level_pct_of_gdp
reer_3y_change
treatment
bis:WS_EERtier 2
change_vs_3y_trailing_mean
reserves_to_short_term_external_debt
treatment
world_bank_wdi:FI.RES.TOTL.MOtier 2
imf:NGDPDPCtier 2
ratio_in_months_of_imports
real_gdp_growth
control
world_bank_wdi:NY.GDP.MKTP.KD.ZGtier 2
yoy_pct
cpi_inflation
control
world_bank_wdi:FP.CPI.TOTL.ZGtier 2
yoy_pct
wgi_regulatory_quality
control
wgi:GOV_WGI_RQ.ESTtier 4
level

● ready  · ● pending  · ● reconstruct-needed

Detailed result card

Result card — banking_crisis_laeven_valencia_predictors_panel

Verdict: SUPPORTED — coef=+0.0006819 (sign matches claim +), p=0.000139

Pre-registration

  • Claim: In the Laeven-Valencia systemic-banking-crisis panel covering 1980-2020, four ex-ante observables jointly predict crisis onset: (i) high private-credit-to-GDP, (ii) negative current-account balance, (iii) real-effective-exchange-rate appreciation versus a 3-year trailing average, and (iv) low foreign-exchange reserve coverage of short-term external liabilities. A logit on these four predictors plus country and year fixed effects achieves out-of-sample crisis-prediction AUC of at least 0.75 on a rolling-window cross-validation across the full advanced-and-emerging panel.
  • Falsification rule: SUPPORTED if out-of-sample AUC across rolling-origin CV >= 0.75 AND each of the four predictors carries a marginal coefficient with sign matching the literature (credit +, CAB -, REER appreciation +, reserves -) at p < 0.10. PARTIAL if AUC in [0.65, 0.75). REFUTED if AUC < 0.65 OR if at least two predictors flip sign.
  • Falsification test: panel_logit_rolling_origin_cv

Estimate

  • Method: linearmodels.PanelOLS
  • Coefficient (treatment): +0.0006819
  • Std error: 0.0001758
  • p-value: 0.000139
  • Observations: 256, countries: 14
  • Within R²: 0.000163
  • Fixed effects: entity=True, time=True
  • Clustering: country

Variables resolved

  • owid:systemic-banking-crises → systemic_banking_crisis_onset (outcome, publisher=owid, n=2766)
  • world_bank_wdi:FS.AST.PRVT.GD.ZS → private_credit_to_gdp (treatment, publisher=world_bank_wdi, n=9562)
  • world_bank_wdi:BN.CAB.XOKA.GD.ZS → current_account_balance_pct_gdp (treatment, publisher=world_bank_wdi, n=7621)
  • bis:WS_EER → reer_3y_change (treatment, publisher=bis, n=2112)
  • world_bank_wdi:FI.RES.TOTL.MO; imf:NGDPDPC → reserves_to_short_term_external_debt (treatment, publisher=world_bank_wdi, n=8998)
  • world_bank_wdi:NY.GDP.MKTP.KD.ZG → real_gdp_growth (controls, publisher=world_bank_wdi, n=13897)
  • world_bank_wdi:FP.CPI.TOTL.ZG → cpi_inflation (controls, publisher=world_bank_wdi, n=7550)
  • wgi:GOV_WGI_RQ.EST → wgi_regulatory_quality (controls, publisher=wgi, n=5169)

Generated by scripts/run_panel_fe.py at 2026-06-29T17:54:27+00:00

Strongest opposing argument

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

Authored framework. Read the transparency note.