US 5-year moved up; the short end was volatile, Canadian bonds behaved almost randomly and the curve is steep.

Interest Rate Weekly · 2026-08-24 to 2026-08-30

One-sentence summary: The main theme last week was US inflation persistence and the Fed’s hawkish tone lifting short-end tightening expectations, but the weekly curve did not move uniformly higher — the clearest signal is that US mid-section rates have remained in an upward channel over the past two months, while Canada shows a coexistence of high-level pullback and a steeper curve.

1. Last week review

The US thread carries the highest information density and most directly affects global rate pricing. The news flow centered on three keywords: inflation has not fallen smoothly, growth prospects are not bad, and the Fed’s new chair delivered an explicitly hawkish policy response function at Jackson Hole.

After August 26, the narrative under the US data lens gradually converged to “growth okay, inflation elevated.” Reuters noted persistent US inflation alongside improved GDP growth prospects; KITCO cited Q2 annualized GDP of 1.5% and PCE month-on-month 0.2%; multiple reports emphasized that consumer spending is slowing but inflationary pressure remains. The implication for rates markets is straightforward: the Fed is not being forced to pivot by a sharp growth deterioration and still needs to prioritize fighting inflation.

What truly drove market repricing was the Fed commentary after Jackson Hole on August 28. Reuters, BBC, PBS, Newsweek, USA Today and others pointed to the same theme: Fed Chair Warsh said that if above-target inflation persists and if the data do not improve, the Fed may need to raise rates, and that there is “more work to do.” The same day also carried reports of a jump in 6-month to 3-year Treasury yields, indicating the shock hit the policy-sensitive front and the short-to-medium end first. On August 29 The Washington Post extended this theme, saying the Fed faces a new strong economic force and that market discussion of “staying tighter for longer” clearly intensified.

But placing the news next to the curve, weekly prices did not form a full-blown bear-flatten or bear-steep move. The US 2-year yield rose only 1bp this week to 4.20%, the 5-year fell 1bp to 4.38%, the 10-year fell 2bp to 4.67%, and the 30-year fell 4bp to 5.19%. In other words, the news shocks were mainly reflected intraday and in localized tenors, while net changes by weekend were modest. If the observation window is extended to 20 and 60 days, the direction becomes clearer: US 5-year, 10-year, 30-year are up 21bp, 21bp, 22bp respectively versus 60 days ago, and the 2-year is up 15bp — the curve overall remains in a mid-term upward move. The 1-year z-scores also corroborate this: US_10Y is 1.78, US_30Y is 1.81, indicating long-end absolute rates remain at relatively high percentiles over the past year.

Supply-side pressure also matters. On August 30 news showed the US government sold 7970 hundred-million-USD of Treasuries this week, and the 10-year yield touched 4.73% while the 30-year touched 5.22%. The magnitude itself implies the Treasury market must digest hawkish policy expectations while continuously absorbing supply. Between August 24 and 25, multiple reports discussed Treasury market stress, the possibility that the TGA might be used to support repo operations, and the idea that “the bond market is reasserting price discovery”; although the accounts differ, the common backdrop is high long-end rates, sensitivity of term premia, and rising market focus on fiscal and supply issues.

Canada’s thread produced fewer headlines than the US, but the policy context of “rate cuts do not necessarily improve affordability” is worth preserving. A summary report on August 23 noted Canadian house prices near highs and the Bank of Canada warning that rate cuts would weaken affordability. On the market side, Canadian yields fell across the curve this week: 2-year down 6bp to 2.96%, 5-year down 5bp to 3.30%, 10-year down 5bp to 3.70%, 30-year down 2bp to 4.13%. However, looking at 20- and 60-day windows, Canada still shows a clear upward pattern, especially stronger at the long end: 10-year 20-day +12bp, 60-day +29bp; 30-year 20-day +14bp, 60-day +34bp. Canada’s 1-year z-score is also high, with the 30-year reaching 2.20, positioned toward the upper end in the sample. In other words, this week’s retreat looks more like a high-level consolidation and is insufficient to rewrite the past two months’ upward backdrop.

Global clues mainly came from the UK and Europe. In the UK, Bloomberg on August 25 reported a rebound in household inflation concerns for August; on August 28 Bank of England Governor Bailey said that, at least for now, second-round inflation effects appear relatively mild and “subdued.” This implies inflation expectations in the UK have not deteriorated further to the point of requiring an immediate stronger response, but the alarm is not fully cleared. In Europe, Schnabel’s remarks leaned more toward institutional and payment-system topics and did not constitute a direct rates-direction signal. Asian news repeatedly mentioned “rising bond market stress”; an Indian think tank flagged the risk of central bank rate hikes, indicating the common global backdrop is “higher rates staying longer,” with US marginal changes being the most important and most able to pull global duration assets.

On the credit side, supporting signals this week pointed to risk-on improvement. US HY OAS narrowed 21bp over the past 20 days, US IG OAS narrowed 1bp. Given the credit spread sample history is only 3.1 years, this can only be descriptive and should not be over-interpreted. At least on the surface, credit markets are not pricing a synchronized recession shock, consistent with the narrative of US growth resilience.

2. Views on future interest rates

First clarify methodological boundaries. Here we can only judge based on given verified signals and the news. Only tenors labeled “verified” can be given directional leanings; tenors labeled “≈random walk” are unpredictable and no directional judgment is made — discussion can only cover static carry and curve position. This boundary is especially important for roughly three-week intermediate holdings, because high carry can improve holding experience, but if yields continue to rise, price losses will quickly overwhelm carry; carry alone cannot be treated as a buy signal.

Within the US short and mid curve, the most informative tenors are the 2-year and the 5-year. The direction signal for US_2Y is [verified] but currently mixed, with a hit rate of 56.1%, implying a statistical edge but not a clear one-sided direction this time — at most one can say the short end is driven by Fed commentary and volatility will likely revolve around the policy path. Combined with the news, September rate-hike expectations are being re-discussed, so the short end is unlikely to strengthen independently of hawkish pricing, but given the model signal is mixed, it is not appropriate to make a firm up-or-down call here.

US_5Y is different. The direction signal for US_5Y is [verified] and dir=up, with a hit rate of 53.9%. Coupled with the 5-year yield’s 60-day rise of 21bp, a 1-year z-score of 1.72, and the post-Jackson Hole reality that 6-month to 3-year yields jumped first, on a three-week view US mid-section rates are still biased upward. In trading terms, this means the 5-year area continues to bear the “policy rates higher for longer” pressure; this tenor combines policy sensitivity and term extension and is often more likely than the 2-year to discount “the path over the next several meetings.”

US_10Y’s signal, while also [verified], is mixed with a hit rate of 52.6% and does not provide enough confidence for a clear directional call. On the news, the long end is driven by inflation, supply, fiscal, and foreign demand; empirically, the 10-year’s net move this week was -2bp. Over a three-week window, the 10-year can be confirmed as having a high absolute yield, positioned at a relatively high percentile over the past year, and the curve does not show clear rapid bull conditions. No strong directional call is made; statically there is relatively high carry: US_10Y carry+roll annualized 5.25%, reliability 0.112, long win rate 56%. That number simply means “when yields do not move, holding a 10-year around this area has some appeal from coupon and roll-down,” and does not imply the long end will fall. If the 10-year resumes probing the August 30 news-mentioned high of 4.73% or higher, a few days of price moves could swallow several weeks of carry.

US_30Y is labeled ≈random walk, so direction is unpredictable and no judgment is made. Although the 30-year has the highest carry, annualized 5.97%, reliability only 0.077, and a long win rate of 53%, the statistical edge is marginal. Considering the 30-year’s sensitivity to supply, fiscal debate, and term premium, high carry alone does not make long bonds a higher-certainty allocation. A more reasonable characterization is: the US ultra-long end is in a combination of high yields, high carry, and high volatility, suitable to be viewed as an asset to “collect coupon and endure volatility,” not as a rates-down trade supported by directional signals.

For Canada, all tenors are labeled ≈random walk, so direction is unpredictable and no judgment is made. Two main points can be made. First, the Canadian curve is steeper than the US: CA 2s10s is +74bp versus the US at +47bp, indicating Canadian markets demand more for far-end nominal growth, inflation compensation, or term premium. Second, Canadian long-end static carry is not poor: CA_10Y annualized 4.5%, CA_30Y annualized 5.126%, and the latter is a verified instrument with a carry win rate of 54%. But it should be emphasized that carry is only a tailwind. Canada’s 30-year has already risen 34bp over the past 60 days, even stronger than the US long end, and high carry is insufficient to offset price risk if yields continue to rise.

On cross-country divergence, the weekly performance this week differed: the US was “hawkish-news-driven with limited net weekend change,” while Canada saw “more intra-week actual retreat.” But on the mid-term dimension, both countries have risen over the past 60 days, with Canada’s long end rising more sharply and its curve becoming steeper. The implication for the next three weeks is that the US front and mid sections are more driven by policy expectations, while Canada looks more pulled by global long-end repricing and domestic curve steepening. Viewed together, the US is more predictable than Canada, especially concentrated in the 5-year area.

On curve shape, both the US and Canada are in positive-slope environments: the US 2s10s is +47bp, Canada is +74bp. Positive slope implies holding mid- to long-end often earns roll-down returns, so carry+roll figures are generally not low. For three-week holdings, this is a tangible buffer, particularly helpful when yields are range-bound. But the market’s core tension now is whether high yields can be pushed even higher. In this environment, positive slope increases holding value but does not eliminate directional risk. The US 5-year has a verified upward tilt and also annualized carry of 4.68% — the proper interpretation of this combination is: even with carry, directional pressure remains and return distributions are not naturally favorable to long-duration bulls.

In one sentence: over the next three weeks, the US short end will be dominated by Fed hawkish remarks with high volatility and event-driven moves; the US 5-year is one of the few tenors that can be clearly judged as “biased higher”; the US 10-year and 30-year warrant no directional call, only recognition of coexistence of high carry and high volatility; the Canadian curve is directionally unpredictable across tenors, and focus should be on the balance among high yields, curve steepness, and static carry.

3. Risk warnings

First, news shocks and weekend closing moves are not always aligned. After the Fed Chair’s August 28 speech, the market clearly repriced but by week’s end US 10-year and 30-year yields had actually retreated slightly. Three-week holdings risk mistaking intraday reactions for trend confirmation.

Second, US supply remains an important disturbance. The August 30 disclosure of 7970 hundred-million-USD Treasury issuance reminds markets that the long end trades not only macro data but also the ability to absorb supply and term premia. Even if Fed expectations stabilize, supply alone can push the long end higher.

Third, carry cannot substitute for directional judgment. Current carry on the US 10-year, 30-year and Canadian long end is high, but if yields continue to rise, price drawdowns will eclipse carry, especially given the longer duration and greater volatility of the 30-year.

Fourth, all Canadian tenors are labeled ≈random walk, so any directional assertion lacks statistical support. US 10-year and 30-year should also not be over-interpreted based on single-week moves.

Fifth, recent credit spread tightening suggests markets are not yet fully pricing a recession, but that credit data history is only 3.1 years and can only serve as auxiliary description, not the core basis for rate judgments.

This report is for research purposes and is not investment advice.