Pre-registration
In Q3 2008 through Q2 2009 the Federal Reserve allowed broad-money M2 growth to slow sharply (annualised QoQ growth fell from ~7% in early 2008 to near zero by late 2008, with M2 outright contracting for several months in late 2008). Had the Fed instead maintained M2 growth at its 1985-2007 mean of ~6% per year through the 2008 financial shock, the trough of US real-GDP would have been shallower by at least 1.5 percentage points and the unemployment-rate peak lower by at least 1.0 percentage point. The hypothesis tests Friedman's (2009 Wall Street Journal) and Anna Schwartz's (2008-09 interviews) public claim that the Great Recession was largely a recurrence of the 1929-1933 monetary policy error: a passive Fed permitting broad money to contract during a banking-system shock.
Falsification criterion — what would disprove this
This hypothesis is considered falsified if:
Not supported if the SVAR-implied counterfactual GDP path under constant-6% M2 growth differs from the actual path by less than 1.5pp (cumulative real-GDP gap) at the 2009Q2 trough, OR if the counterfactual unemployment peak differs by less than 1.0pp from the actual 2009-2010 peak, OR if the SVAR M2-shock impulse response on real GDP is not statistically positive at h=4 quarters and h=8 quarters at 90% confidence bands. A real-side / financial-friction reading wins cleanly if the M2-shock impulse response on GDP is insignificant at h>=4 — i.e. the Fed could not have prevented the recession by holding M2 growth steady because the contraction was driven by housing-collapse and balance-sheet repair, not money supply.
formal test & threshold
test: svar_counterfactual_constant_m2_growth_gdp_unemployment_gap threshold: counterfactual_gdp_gap_2009Q2_trough >= 1.5pp at 90% bands AND counterfactual_unemployment_peak_gap >= 1.0pp at 90% bands AND SVAR M2-shock impulse on real GDP positive at h=4 and h=8 at 90% bands AND result robust to inclusion of housing_starts as additional shock variable
Method
- Template
synthetic_control- Sample
- 1 countries · 1985 – 2024
- Evidence type
- causal
Primary spec: structural-VAR with M2 growth, real GDP growth, unemployment, TED spread, oil prices, housing starts; identification via Cholesky with M2 ordered last among policy variables (so that M2 responds to financial-stress shock first, allowing a measure of exogenous M2 stance shifts within the SVAR). Counterfactual: re-run the SVAR with M2 growth fixed at 6% from 2008Q3 onward, holding other shocks at their realised values. Difference between counterfactual GDP/unemployment paths and actual paths is the estimated cost of the M2 contraction. Robustness: synth-DiD using other advanced-economy quarterly GDP series as donor pool to build a synthetic US.
Data
| Variable | Source | Transform |
|---|---|---|
real_gdp_quarterly outcome | fred:GDPC1tier 1 | log_diff_qoq_annualised |
unemployment_rate_monthly outcome | fred:UNRATEtier 1 | level |
industrial_production_monthly outcome | fred:INDPROtier 1 | log_diff_yoy |
m2_growth_actual treatment | fred:M2SLtier 1 | log_diff_qoq_annualised |
m2_growth_counterfactual_constant_6pct treatment | derived:m2_growth_counterfactual_constant_6pcttier 4 | synthetic_path_constant_6pct_from_2008Q3 |
m2_growth_gap treatment | derived:m2_actual_minus_counterfactualtier 4 | difference |
ted_spread control | fred:TEDRATEtier 1 | level |
oil_price_change control | fred:WTISPLCtier 1 | log_diff_yoy |
housing_starts control | fred:HOUSTtier 1 | log_diff_yoy |
federal_funds_rate control | fred:FEDFUNDStier 1 | level |
● ready · ● pending · ● reconstruct-needed
Detailed result card
Result card — monetarist_fed_2008_great_recession_avoidable_with_constant_m_growth
Verdict: INCONCLUSIVE_DATA_PENDING — no outcome variable loaded
Pre-registration
- Claim: In Q3 2008 through Q2 2009 the Federal Reserve allowed broad-money M2 growth to slow sharply (annualised QoQ growth fell from ~7% in early 2008 to near zero by late 2008, with M2 outright contracting for several months in late 2008). Had the Fed instead maintained M2 growth at its 1985-2007 mean of ~6% per year through the 2008 financial shock, the trough of US real-GDP would have been shallower by at least 1.5 percentage points and the unemployment-rate peak lower by at least 1.0 percentage point. The hypothesis tests Friedman's (2009 Wall Street Journal) and Anna Schwartz's (2008-09 interviews) public claim that the Great Recession was largely a recurrence of the 1929-1933 monetary policy error: a passive Fed permitting broad money to contract during a banking-system shock.
- Falsification rule: Not supported if the SVAR-implied counterfactual GDP path under constant-6% M2 growth differs from the actual path by less than 1.5pp (cumulative real-GDP gap) at the 2009Q2 trough, OR if the counterfactual unemployment peak differs by less than 1.0pp from the actual 2009-2010 peak, OR if the SVAR M2-shock impulse response on real GDP is not statistically positive at h=4 quarters and h=8 quarters at 90% confidence bands. A real-side / financial-friction reading wins cleanly if the M2-shock impulse response on GDP is insignificant at h>=4 — i.e. the Fed could not have prevented the recession by holding M2 growth steady because the contraction was driven by housing-collapse and balance-sheet repair, not money supply.
Synthetic-control estimate
- Error: no outcome variable loaded
Variables resolved
Generated by scripts/run_synth_did.py at 2026-04-30T09:47:44+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.
Notes
Hetzel (2012) "The Great Recession: Market Failure or Policy Failure?" develops the monetarist counterfactual at book length. Sumner's market-monetarist NGDP-targeting critique is closely adjacent. The prior_confidence reflects that the 2008-09 case is far weaker for monetarism than the 1929-1933 case: in 2008 the Fed expanded base money by 100% within 6 months, and M2 growth slowed not because of Fed inaction but because of massive bank-credit contraction. Whether more aggressive M2 targeting (e.g. via Treasury asset purchases that directly increased deposits) would have closed the gap is an open empirical question.