Pre-registration
US CPI inflation was highly persistent (first-order autocorrelation at quarterly frequency above 0.85, sum of AR coefficients above 0.95) during 1960-1979, and substantially less persistent (first-order autocorrelation below 0.6, sum of AR coefficients below 0.8) during 1985-2019. The decline reflects expectations anchoring under the post-1979 Volcker regime: when the central bank is credibly committed to a stable inflation target, inflation shocks no longer feed forward into wage-price-setting expectations, and the AR-1 decay of inflation toward target accelerates. The hypothesis is the Lucas (1973, 1976) prediction that reduced-form inflation dynamics shift when monetary regime shifts.
Falsification criterion — what would disprove this
This hypothesis is considered falsified if:
Not supported if (a) the sum-of-AR-coefficients in 1960-1979 is below 0.85, or (b) the sum in 1985-2019 is above 0.85, or (c) the difference between the two sums is below 0.15, or (d) the Bai-Perron break test fails to detect a break in 1979-1984. A Keynesian / cost-push reading wins cleanly if the persistence drop is fully explained by the absence of 1970s-style oil shocks in the post-1985 sample (i.e. the persistence difference disappears once oil shocks are controlled for at quarterly frequency).
formal test & threshold
test: ar_persistence_pre_post_volcker threshold: sum_of_AR_coeffs(1960-1979) >= 0.85 AND sum_of_AR_coeffs(1985-2019) <= 0.80 AND difference >= 0.15 with bootstrap p<0.05 AND Bai-Perron break detected in [1979Q4, 1984Q4] AND oil-shock-controlled persistence difference still >= 0.10
Method
- Template
panel_fe- Fixed effects
- Clustering
newey_west_4_lags- Sample
- 1 countries · 1960 – 2024
- Evidence type
- associational
Primary spec: AR(p) regression of quarterly inflation on its own lags within each sub-sample (1960Q1-1979Q3 and 1985Q1-2019Q4), with sum-of-AR-coefficients (rho) reported per sub-sample. Bai-Perron structural-break test on inflation persistence is a robustness check; pre-registered prediction is one break in 1979-1984. Local projections of inflation response to a one-SD oil shock are reported pre and post Volcker; impulse response should decay faster post-1985.
Data
| Variable | Source | Transform |
|---|---|---|
cpi_inflation_quarterly outcome | fred:CPIAUCSLtier 1 | log_diff_qoq_annualised |
core_pce_inflation_quarterly outcome | fred:PCEPILFEtier 1 | log_diff_qoq_annualised |
ar1_persistence_rolling_40q outcome | fred:CPIAUCSLtier 1 | rolling_ar1_coefficient_40q |
post_volcker_regime_indicator treatment | derived:volcker_regime_break_1985Q1tier 4 | indicator |
ten_year_inflation_expectations treatment | fred:T10YIEtier 1 | level |
oil_price_change control | fred:WTISPLCtier 1 | log_diff_qoq |
output_gap control | fred:GDPC1tier 1 | cbo_gap_filter |
federal_funds_rate control | fred:FEDFUNDStier 1 | level |
● ready · ● pending · ● reconstruct-needed
Detailed result card
Result card — lucas_expectations_anchoring_post_volcker_us_inflation_persistence
Verdict: REFUTED — coef=+0.5543 (sign opposite claim -), p=3.46e-09
Pre-registration
- Claim: US CPI inflation was highly persistent (first-order autocorrelation at quarterly frequency above 0.85, sum of AR coefficients above 0.95) during 1960-1979, and substantially less persistent (first-order autocorrelation below 0.6, sum of AR coefficients below 0.8) during 1985-2019. The decline reflects expectations anchoring under the post-1979 Volcker regime: when the central bank is credibly committed to a stable inflation target, inflation shocks no longer feed forward into wage-price-setting expectations, and the AR-1 decay of inflation toward target accelerates. The hypothesis is the Lucas (1973, 1976) prediction that reduced-form inflation dynamics shift when monetary regime shifts.
- Falsification rule: Not supported if (a) the sum-of-AR-coefficients in 1960-1979 is below 0.85, or (b) the sum in 1985-2019 is above 0.85, or (c) the difference between the two sums is below 0.15, or (d) the Bai-Perron break test fails to detect a break in 1979-1984. A Keynesian / cost-push reading wins cleanly if the persistence drop is fully explained by the absence of 1970s-style oil shocks in the post-1985 sample (i.e. the persistence difference disappears once oil shocks are controlled for at quarterly frequency).
- Falsification test: ar_persistence_pre_post_volcker
Estimate
- Method: statsmodels OLS time-series fallback
- Coefficient (treatment): +0.5543
- Std error: 0.09383
- p-value: 3.46e-09
- Observations: 65, countries: 1
- Within R²: 0.983
- Fixed effects: entity=False, time=False
- Clustering: HAC(maxlags=4)
Variables resolved
fred:CPIAUCSL→ cpi_inflation_quarterly (outcome, publisher=fred, n=80)fred:PCEPILFE→ core_pce_inflation_quarterly (outcome, publisher=fred, n=68)fred:CPIAUCSL→ ar1_persistence_rolling_40q (outcome, publisher=fred, n=80)derived:volcker_regime_break_1985Q1→ post_volcker_regime_indicator (treatment, publisher=constructed, n=65)fred:T10YIE→ ten_year_inflation_expectations (treatment, publisher=fred, n=24)fred:WTISPLC→ oil_price_change (controls, publisher=fred, n=81)fred:GDPC1→ output_gap (controls, publisher=fred, n=80)fred:FEDFUNDS→ federal_funds_rate (controls, publisher=fred, n=73)
Generated by scripts/run_panel_fe.py at 2026-06-29T17:54:21+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
Stock-Watson (2007) "Why has US inflation become harder to forecast?" is the canonical reduced-form documentation of the persistence decline. Cogley-Sargent (2005) re-derive the result in a Bayesian time-varying- parameter framework. The Lucas-1976 critique is cited as the structural rationale for why pre-1979 reduced-form inflation dynamics should not be expected to survive the Volcker regime change.