Open data methodology
Every source behind the "World (open data)" composite in Rich or Broke — what's actually integrated, what's real but blocked or wrongly licensed, and what's a genuine dead end. Built because the main World/US tabs run on the Jordà–Schularick–Taylor database, which is licensed non-commercial only.
Why this page exists
The Jordà–Schularick–Taylor Macrohistory Database — the backbone of the regular World and US tabs — is licensed CC BY-NC-SA 4.0, non-commercial only. Its own terms are explicit: "commercial data providers are strictly forbidden to integrate all or parts of the dataset into their services." The World (open data) tab is a from-scratch attempt at something similarly broad, built entirely from sources that are either explicitly openly licensed, government-published, or simply real public market prices — not a proprietary compiled dataset.
That meant sourcing nine countries' and funds' worth of academic papers, central bank archives, open-access journals, and real market prices, one at a time, over several research sessions — finding real data doesn't always mean finding usable data. This page is the honest accounting of that whole process: what made it in, what's real but stuck behind a paywall or a down server, what has the wrong license, and what simply doesn't exist for free anywhere we could find.
Coverage, and where the gaps really are
Every source found across this whole project, plotted by actual year coverage. Broken bars are real gaps in the underlying source — not rendering artifacts. Hover any bar for the detail.
Every source, in detail
Grouped by status. Integrated sources feed the composite directly; everything else is documented so the next pass through this project starts from exactly where the last one stopped.
How the composite's country mix actually evolved
You asked for something like Dimson-Marsh-Staunton's classic "evolution of equity markets" chart — worth being upfront about why this isn't quite that. That chart plots true market-capitalization share by country, sourced from the Credit Suisse Global Investment Returns Yearbook — the same commercial DMS/Credit Suisse data lineage this whole project has been avoiding, for the same non-commercial-license reasons as JST itself. We don't have free market-cap data for these nine countries. What we actually have is equal weighting, which is a genuinely different and much simpler thing: every active country counts the same, regardless of how large its economy or stock market really was. So this chart shows something honest but different — not "how big was each country's market," but "how much of our blend did each country make up, given our own equal-weight methodology and whichever countries had data that year." The bands widening and narrowing reflect countries joining or leaving the available data, not economies growing or shrinking. South Africa's three gaps (1940-44, 1960s, 1980s) show up as literal drops to zero — it's simply excluded from the blend, and every other country's slice grows a little to fill the space. Only through 2002: from 2003 on, the composite switches to real fund data (VTI/EFA/EEM, then VT/ACWI) rather than a country blend, so there's no "composition" to plot the same way.
Equal weight only — not market capitalization. A genuine, if blunt, illustration of the survivor-bias-in-sourcing point made elsewhere on this page: small, easy-to-source economies like Norway, Brazil, and South Africa get the exact same weight here as the US, wildly overrepresenting their true historical share of world equity markets.
Why emerging markets get a rising weight, not a fixed one
This is the actual assumption used in two places in the calculator: the World+EM tab (1988–2025) and the 2003–2008 three-fund reconstruction's 10% EM slice. It's a straight-line approximation, not sourced from an actual historical weight series — anchored to EM's well-documented rise from a tiny sliver of world market cap in the late 1980s to roughly a tenth of it today.