The economy behind the budget

Long annual histories for structure, quarters for the cycle and months for turning points. Every observation remains tied to its source, unit and transformation.

Global data layer—— countries and economies · — definitions
Sample depth

How much history each analysis needs

Twenty annual points describe one generation of public finance, but usually only two or three large shocks. Structural relationships need more regimes.

20× A20

Descriptive base

Useful for trends and accounting comparisons. Weak for stable regression conclusions.

40–60× A40+

Structural analysis

Better for regimes, lags and robustness checks—if definitions remain comparable.

80× Q20y

Cycle analysis

The same twenty years yield roughly 80 quarters or 240 months.

Reporting rule

Use 1995–2025 for annual comparisons, 2000–2025 for the quarterly cycle and the longest consistent series for structural tests. Length alone does not repair methodology breaks or endogeneity.

Cycle and budget

Growth, inflation, unemployment, balance and debt

Growth, inflation and unemployment describe the cycle; balance and debt show the fiscal footprint. Causality cannot be read straight off the chart.

CzechiaCZE
Source series

Several measurements of the same concept

The filter preserves frequency, transformation and seasonal adjustment. Future reports must disclose these distinctions.

—

—
—
Data coverage

How many observations each country has

The table counts actual points and first/last periods for the selected country. Zero means a missing series, not an economic value of zero.

TopicAvailable seriesLongest historyObservations
Data contract

One fact table behind every report

Every point carries country, indicator, full source-series key, frequency, period, unit, seasonal adjustment, transformation, vintage and source URL.

01

Raw snapshot

Dated OECD, BIS, World Bank and IMF files with checksums.

→
02

Observation fact

1.4+ million rows in one long format, without imputation.

→
03

Report marts

Compact extracts for charts, coverage and latest values.

Important boundary

More observations improve description but do not create causal identification. Policy-impact claims require a separate design: natural experiment, instrument, difference-in-differences or another defensible strategy.