A century: XEQT's world mix vs US stocks, in Canadian dollars (yearly, 1926 - Nov 4 2026)

Top: growth of $1 (log scale). Bottom: % below its own previous year-end high. Shaded: crash periods.

source: Jorda-Schularick-Taylor Macrohistory Database R6 (eq_tr total returns, xrusd, rgdpmad x pop), 1926-2020; World Bank CM.MKT.LCAP.CD; OECD share prices for Japan via FRED; Yahoo Finance SPY, ACWI adjusted closes x CAD=X for 2021-2026 (2026 = to Nov 4) • transform: CANADA IS LEFT OUT: JST has no Canadian stock returns and I found no free yearly Canadian series back to the 1920s, so the blue line is XEQT's 75% world part, not the whole fund. Weights reset yearly. US share = US share of world stock-market value at the prior year-end (World Bank) from 1976; before that there is no market-value data, so it is the US share of real GDP among the 16 JST countries (about 40%); the 1976 jump to 61% is that switch. The other 15 developed countries (JST minus US, Canada, Ireland) are weighted by real GDP throughout, so Japan's 1989 bubble shows only through the falling US share. No emerging markets before 2021 (JST has none); from 2021 the world part is ACWI. Converted to CAD with JST's exchange rates; the CAD was fixed or pegged in some years. Fixes to the raw JST data: Germany is left out of 1948-49 (its Reichsmark-to-Mark reform puts the returns and the exchange rate on different bases, a fake gain); Japan's JST returns follow yearly AVERAGE prices, so 1960-2020 Japan uses the OECD share-price index (FRED SPASTT01JPM661N) December to December plus JST's dividend return. In this data the world mix pulled ahead from 1950 to 1989 (post-war Europe and Japan, then Japan's boom) and has lost ground since. Year-end data hide much of a crash's depth: falls inside a year that recover by December do not show at all. • drawn 2026-11-05 04:10 UTC