Pre-registration
Brazil's Novo Arcabouço Fiscal (Lei Complementar 200 of 2023) holds the federal primary-balance path and the real-spending growth rate within its statutory band (real primary spending growth 0.6-2.5% per year, conditional on primary-balance target attainment) through 2026, and the cumulative 2023-2026 primary- balance outcome does not diverge from target by more than one standard deviation of the historical primary-balance forecast error. The hypothesis is testable because the framework statutorily specifies a band; the test is whether the realised fiscal path falls inside or outside that band. The null is not that Lula 3.0 is fiscally expansionary, but rather that the institutional framework it created is binding on the executive's ex-ante preferences. A positive result updates toward institutional-constraint effectiveness; a falsification updates toward the post-Keynesian and developmentalist reading that PT-era fiscal commitments erode under pressure.
Falsification criterion — what would disprove this
This hypothesis is considered falsified if:
Not supported if EITHER (a) in two or more of the years 2023, 2024, 2025, 2026, real primary-spending growth falls outside the 0.6-2.5% statutory band AS MEASURED IN REALISED SPENDING (not as measured in the budget ex-ante), OR (b) the cumulative 2023-2026 primary-balance outcome deviates from the cumulative pre-Lula IMF-projection baseline by more than two standard deviations of the historical (2015-2019) annual primary-balance forecast error, OR (c) the Novo Arcabouço Fiscal is statutorily replaced or suspended within the 2023-2026 window (a measurable institutional-reversal event).
formal test & threshold
test: fiscal_framework_compliance_and_primary_balance_variance threshold: years_in_band >= 3 of 4 (2023-2026 inclusive) AND |cumulative_primary_balance_deviation| <= 2 * SD(historical_forecast_error) AND fiscal_framework_statute_replaced = FALSE within window
Method
- Template
event_study- Clustering
year- Sample
- 1 countries · 2015 – 2027
- Evidence type
- causal
Single-unit event study around the 2023 start of Lula III and the August 2023 activation of the new fiscal framework. Primary presentation: realised primary-balance path and real-spending growth vs the Lei Complementar 200 statutory band over 2023- 2026. Historical comparison: Rousseff era (2011-2015) same- window path for qualitative context. Formal statistical test uses the forecast-error variance of primary-balance IMF projections (published pre-Lula 2022 for fiscal years 2023- 2026); deviations beyond ±1SD are treated as breaches.
Data
| Variable | Source | Transform |
|---|---|---|
primary_balance_pct_gdp_federal outcome | imf:GGXCNL_NGDPtier 2 brazil_stn:Resultadotier 2 | annual_pct_gdp |
real_primary_spending_growth_rate outcome | brazil_stn:Despesatier 2 imf:PCPIPCHtier 2 | annual_real_growth_pct |
gross_debt_gdp outcome | imf:GGXWDG_NGDPtier 2 world_bank_wdi:GC.DOD.TOTL.GD.ZStier 2 | level |
fiscal_framework_compliance_indicator outcome | constructed:1 if both real spending growth in band AND primary-balance within target +/- 1SD of forecast error; 0 otherwisetier 5 | annual_binary |
lula_third_term_indicator treatment | constructed:indicator = 1 from 2023-01-01 onward; 0 priortier 5 | binary |
marco_fiscal_regime_indicator treatment | constructed:indicator = 1 from 2023-08-30 (LC 200/2023 publication) onward; 0 priortier 5 | binary |
pt_prior_era_indicator treatment | constructed:indicator for Rousseff era 2011-2016 as a historical comparatortier 5 | binary |
commodity_price_index control | world_bank_wdi:TT.PRI.MRCH.XD.WDtier 2 | annual_log_change |
us_policy_rate control | fred:FEDFUNDStier 1 | annual_average |
real_gdp_growth control | imf:NGDP_RPCHtier 2 | annual_pct |
output_gap control | oecd:OutputGaptier 2 | level |
● ready · ● pending · ● reconstruct-needed
Strongest opposing argument
Every hypothesis ships with its charitable opposing argument. The framework earns credibility by handling objections at their strongest, not weakest.
Notes
Data-gated: Brazil STN (Secretaria do Tesouro Nacional) monthly Resultado Primário and Despesa Primária series required; IMF WEO annual general-government fiscal aggregates are the automated fallback. Historical forecast-error variance computed from IMF WEO vintage series (April 2022 vintage for 2023-2026 projections as pre-Lula baseline). Pre-registration is particularly important here because the political tendency will be to re-characterise the statutory band mid-stream if realised spending exceeds it. V1 commits to the band as published in LC 200/2023; any post-2023 statutory amendment widening the band is coded as a framework erosion, not as compliance.