Open data: UK macroeconomic series from 1086 to present. Bank of England Millennium of Macroeconomic Data. (Month)

Rows: 41,950
Columns: 8

Hundreds of annual time series covering nominal and real GDP, CPI, wages, interest rates, public debt, money supply, and equity markets for the UK. Some series go back to 1086. Most are continuous from at least 1700.

AI analysis

Dataset Analysis: Bank of England Millennium of Macroeconomic Data

1. Dataset Overview and Research Value

This dataset represents one of the most historically comprehensive macroeconomic records publicly available, spanning nearly a millennium of UK economic history from 1086 to 2017. Published by the Bank of England as part of their "Millennium of Macroeconomic Data" initiative, it covers critical economic indicators including nominal and real GDP, CPI, wages, interest rates, public debt, money supply, and equity markets. The monthly granularity of the join-key range (1694-11-01 to 2017-02-01) makes it exceptionally powerful for long-run correlation studies, policy analysis, and business cycle research. Its breadth — spanning wars, industrial revolutions, financial crises, and monetary regime changes — gives analysts a rare opportunity to test economic relationships across radically different structural environments. No source URL was provided, but the Bank of England is the publisher, lending high institutional credibility to the data.

2. Data Quality Observations

Data quality is notably strong for a dataset of this historical scope. There are zero null cells across all 41,950 rows and 8 columns, which is remarkable given that many series necessarily begin centuries apart. This likely means missing historical values have been handled upstream — either imputed, excluded from the long-format structure, or represented as coded placeholders — so analysts should verify whether any value entries use sentinel values (e.g., 0 or -1) to represent true missingness before proceeding. Duplicate row counts are pending final recomputation (Phase D), so deduplication validation should be confirmed before any aggregation. No type mismatches are flagged across the columns, and the clean integer range on month (1–12, mean 6.499, median 6.5) confirms no malformed date components. Overall, the dataset is production-ready for exploratory analysis with only the sentinel-value caveat warranting further inspection.

3. Key Column Distributions and Notable Patterns

The value column is the analytical centerpiece and exhibits extreme distributional characteristics that demand attention. Its range spans -0.26 to 1,914,210, with a mean of 10,406.9 dramatically exceeding the median of just 5.0 — a gap signaling that a small number of very large values (likely nominal GDP or money supply figures in millions of pounds) are heavily skewing the distribution. The skewness of 13.7 confirms severe right-skew, and 8,515 outlier rows (~20% of the dataset) flag values far beyond typical ranges. The interquartile range (Q1=3, Q3=27.05) suggests the majority of observations are small, unit-normalized figures such as index values or percentage rates, while the extreme upper tail likely reflects monetary aggregates in absolute currency terms. This heterogeneity is expected given that 26 distinct variables across 5 sections and 13 unit types are stacked in long format — analysts must filter by variableid or unit before any statistical analysis to avoid comparing apples to pounds sterling. The month column's near-perfect symmetry (skew = -0.000287) confirms even temporal coverage across calendar months.

4. Recommended Join Key Columns

The period column (monthly Date, 3,868 distinct values, zero nulls) is the primary recommended join key for time-series alignment with external datasets, offering monthly precision from November 1694 onward. The year column (324 distinct annual values) provides a coarser but more universally compatible join key for datasets that are only available at annual frequency — particularly useful given that many of the deeper historical series (pre-1700) will only have annual observations. For multi-variable within-dataset pivoting (e.g., constructing a wide-format panel for correlation matrices), variableid paired with period forms a reliable composite key given its 26 distinct, null-free values. When joining to international datasets, analysts should note that the dataset's UK-centric scope means a country code constant ("GBR") would need to be added as a synthetic join key.

5. Recommended Pairing Datasets for Correlation Discovery

Several dataset categories would pair powerfully with this data. UK political and fiscal history datasets — covering general elections, government spending, and Budget announcements — would allow testing of fiscal multiplier effects and political business cycle theories across centuries. Global commodity price series (e.g., oil, wheat, coal) would help isolate supply-side inflation drivers in the CPI series, particularly around the 1970s oil shocks and pre-industrial grain crises. Bank of England policy rate decisions and Monetary Policy Committee minutes (post-1997) would enable modern monetary transmission analysis. For equity market correlations, FTSE historical index data or the Barclays Equity Gilt Study would complement the dataset's equity series with dividend and total return decompositions. Finally, pairing with cross-country macroeconomic panels such as the Maddison Project Database or IMF World Economic Outlook historical data would unlock comparative analysis of UK economic performance relative to peers across key historical episodes like the Great Depression, Bretton Woods, and post-2008 austerity.

Columns

period (date)
year (date)
month (integer)
variable_id (string)
variable (string)
section (string)
unit (string)
value (decimal)