TSX Large-Cap Turn-of-the-Month Window: Long an Equal-Weight 16-Name Canadian Basket From the Open of the Month's Last Session to the Open of the 4th New-Month Session (CAEQ 1-DAY, calendar-only, long/flat)

Original Research Lead hypothesis, built on Cadsby & Ratner (1992, Journal of Banking & Finance, 'Turn-of-month and pre-holiday effects on stock returns: some international evidence'). They found a significant turn-of-the-month (TOM) effect in Canada, the UK, Australia, Switzerland and West Germany, and none in Japan, Hong Kong, Italy or France. Lakonishok & Smidt (1988) and McConnell & Xu (2008) give the canonical [-1,+3] window and the payroll and pension flow explanation. This is the first test on the platform's native Toronto venue. CAEQ currently has 0 experiments, and no non-US/CN/HK/JP equity venue has ever been used. Related work on the platform: the SPY TOM window was promoted (401 trades, PF 1.34, 0.25% per trade), and so was the emerging-market country-ETF basket (1,379 trades, 0.44%). The developed-market country-ETF basket (EWG/EWJ/EWU/EFA) ended no_edge at 0.087% per trade. Because of that, this is framed as a fresh test of a market the literature names (Canada), on local shares in local hours. It is not a transfer of the US result (L228). Universe: a fixed panel of 16 domestic TSX large caps, all continuously listed and TSX-60 members over the whole catalog span (2018-01-02 to present): RY, TD, BMO, BNS, CM (banks); MFC, SLF (insurers); ENB, TRP, FTS (pipelines/utilities); CNQ, SU (energy); CNR, CP (rails); BN (Brookfield); ATD (Couche-Tard). CDRs, US cross-listings and preferred-share lines in the universe file are excluded on purpose. Survivorship caveat: the platform has no point-in-time TSX-60 membership file. The panel is chosen from names that exist today, which biases the basket's level return upward. The claim tested is therefore the TOM window's return against the same basket's unconditional 4-session drift. The analyst should compare the per-window mean to the measured unconditional 4-session open-to-open mean of this basket, 0.191%. Pre-registered measurement (frozen rule, catalog daily opens, equal weight, gross): 105 distinct month windows from 2018-01 to 2026-09. Mean per window +0.339%, std 1.75%, t = 1.99, 67% of windows positive. Max drawdown of the compounded window series after 0.07% cost is -8.7%; the worst window is -5.69%. By year: 2018 -0.67%, 2019 -0.12%, 2020 +1.26%, 2021 +1.05%, 2022 +0.40%, 2023 +0.13%, 2024 +0.26%, 2025 +0.15%, 2026 YTD +0.67%. Neighbouring windows ([-1,+2] 0.23%, [-1,+4] 0.46%, [-3,+3] 0.43%) are smooth, with no cliff. Independent evidence = 105 month periods, not about 1,680 per-leg trades (L238). The measurement documents firing rate and economics. It was not used to accept the rule. Execution: CAEQ is a session venue, so a decision at bar N's close fills at the next session's open (next_open_fill, supported). To own the basket from the OPEN of the month's last trading session, the entry decision is made at the close of the second-to-last session. The exit decision is made at the close of the 3rd trading session of the new month and fills at the 4th session's open. Month boundaries and trading days come from the platform session calendar (src.strategies.helpers.sessions / FactoryStrategy session helpers), never hard-coded dates; Canadian holidays come from the calendar. Costs: Wealthsimple $0 commission, 3 bp half-spread per fill, so about 0.06% round trip plus impact. Net expectancy is about 0.34% - 0.07% = 0.27% per window per leg, about 4x the round-trip cost. Dividends are not in PnL (split-adjusted prices). Ex-dividend drops inside the window bias the test against the strategy, slightly and conservatively.