The S&P/TSX composite had 11 sessions between September 1 and October 14 in which it moved 0.9% or more in either direction, against 4 in the 31 sessions from July 17 to August 31. Added up without regard to direction, the daily moves total 21.2 percentage points. The index closed October 14 at 35,584.46, which is 1.9% below its August 31 close of 36,270.50.
Several of the largest moves followed single data releases. The research on how people read small samples explains why each release felt decisive when it arrived.
Three Reversals in Two Weeks
XTB put the probability of an October rate increase by the U.S. Federal Reserve at almost 70% in the last week of September, and near 20% by October 6. U.S. payrolls for September then added only about 29,000 jobs, according to investinglive, and the odds sat at roughly one in four. On October 14, a U.S. consumer price report showed prices up 3.7% from a year earlier, with core prices up 0.3% on the month against 0.2% expected. That lifted the odds to 42% on the FXStreet reading, 46% on Bloomberg and about 50% on CME FedWatch.
The underlying pressure did not change by anything like that much. The Federal Reserve raised its target range to 3.75% to 4.00% on September 16, and Brent futures closed between US$104.72 and US$106.65 in every session from October 9 to October 14, according to Yahoo Finance.
Amos Tversky and Daniel Kahneman described the mechanism in a 1971 paper, Belief in the Law of Small Numbers. People expect a small sample to resemble the population it came from, so they give one reading the weight that only a long run of readings has earned. A monthly payroll estimate and a monthly price index are each a single draw from a noisy process.
Strings of News Produce the Overreaction
Nicholas Barberis, Andrei Shleifer and Robert Vishny modelled the next step in the Journal of Financial Economics in 1998. In their framework, investors underreact to any one announcement and overreact to a run of announcements pointing the same way, because a run looks like a trend. Werner De Bondt and Richard Thaler reported in the Journal of Finance in 1985 that portfolios of prior losers outperformed prior winners over the following three years, a pattern they linked to overreaction to dramatic news.
The tape in October fits the shape. On October 2 the composite rose 0.99% as the soft payrolls report cut Fed hike odds, according to Trading Economics. On October 7 it fell 1.70% as oil stayed elevated and Canadian bond yields rose. On October 9, the day Statistics Canada reported about 68,000 lost jobs, it rose 1.48%, because weak employment cut the odds of a Bank of Canada increase from roughly 40% to about 25%, according to LSEG data cited by the Financial Post. On October 14 the CPI report took back 1.11%.
The composite spent 30 sessions between 35,042 and 36,633 and finished 686 points below its August 31 close, with three marked releases each moving it by about 1% or more.
The dashed green line marks the August 31 close of 36,270.50. Daily moves summed without regard to direction total 21.2 percentage points against a net change of negative 1.9%.
What the Next Two Releases Will and Will Not Settle
Two more small samples arrive before the decisions. Statistics Canada reports September consumer prices on October 19, and Continuum Economics expects headline inflation to hold at 3.0% for a third month, with the Bank of Canada core measures edging slightly lower. The Federal Reserve and the Bank of Canada then both announce decisions on October 28, with the Canadian overnight rate at 2.25%, according to the Bank of Canada.
Professional forecasters do not agree on the outcome. TD Economics expects the Bank of Canada to hold in October, CIBC expects holds at both remaining 2026 meetings, and Capital Economics expects an increase soon to 2.75%, the midpoint of the neutral range. A spread like that is itself a statement about sample size: economists with full models and full data histories still disagree after one more print.
The behavioural point is narrow. Each of these releases is informative and none of them is conclusive. The sessions on October 9 and October 14 moved the composite 1.48% higher and 1.11% lower on releases that bore on the same policy question, with both central bank decisions still 13 days away.