Upside favored for short- and medium-term U.S. Treasuries, while 30-year Treasuries and Canadian bonds follow a random walk.

Interest Rate Weekly · 2026-09-21 to 2026-09-27

One-line summary: The main theme over the past week was clear — the US growth-and-inflation trade regained the upper hand, global long-end yields followed higher, and the US curve continued to steepen at elevated levels; looking ahead three weeks, US short- and mid-ends retain a statistical edge for further upside, but the long end is unpredictable, so carry should be treated as a static buffer, not a directional signal.

I. Weekly review

The US thread was the strongest driver. News flow concentrated on the same fact: growth has not been quickly suppressed by higher rates, and inflation worries have been re-elevated. On 9月23日, Reuters noted that rising inflation pressures increased the chance of another rate hike before the election; on the same day MarketWatch said US business activity accelerated at the fastest pace in more than four years, and reports from MarketWatch, CNBC, Wolf Street and others described bonds suffering continued selling after a "hot PMI," with the 10-year Treasury yield spiking to around 5.10% at one point. By 9月25日, WSJ, FT, and Reuters continued to reinforce the same narrative: the US economy is showing resilience to high rates and high bond yields, and the market is repricing the next policy path toward "longer and higher, possibly another hike." On 9月26日, Yahoo Finance again emphasized consumption has not clearly cooled and the economy remains relatively strong; at the same time, discussion around diesel prices, geopolitical conflicts, and imported inflation picked up, so inflation tail risks have not exited.

Curve data lined up perfectly with this round of news. US 2Y rose 20bp over the week, +68bp over 20 days, +73bp over 60 days, now 4.87%, 1-year z value 2.93; 5Y rose 25bp over the week, +66bp over 20 days, +84bp over 60 days, to 5.03%, z value 3.08; 10Y rose 24bp over the week, +52bp over 20 days, +74bp over 60 days, to 5.18%, z value 3.17; 30Y rose 18bp over the week, +29bp over 20 days, +56bp over 60 days, to 5.47%, z value 2.73. Several key signals deserve separate attention.

First, the entire US curve is at elevated levels, but the mid-section is strongest. Significant 20-day and 60-day gains in the 5Y and 10Y indicate that last month’s core trade was not only a widening of term premium but also an upward revision of the policy-rate terminal and the time spent at elevated levels. Second, the curve is positively sloped and continuing to steepen. US 2s10s has reached +31bp, which in the current environment is more consistent with a repricing toward "higher growth + sticky inflation + longer and higher policy" than with the classic pre-recession signal. Third, although 30Y yields have risen materially, they are relatively weaker versus the 2Y–10Y segment, indicating the ultra-long end is simultaneously influenced by growth expectations, fiscal supply, and risk appetite — momentum alone is insufficient to explain its move.

Canada’s thread was clearly weaker than the US but moved in the same direction. In the 9月24日 news, the most informative items were BoC-related statements: most Canadian households have a cash-flow shortfall in coping with inflation, implying Canadian domestic demand is more sensitive to financial conditions. Compared with the US story of "growth remains firm despite high rates," Canada is more like "inflation pressure persists, but household balance sheets show clearer stress." These macro features mean Canadian bond yields are being pulled up by the global rate environment, but domestic fundamentals do not present as strong a narrative for additional hikes as in the US.

On the data, Canada’s 2Y to 30Y rose over the past 5 days by 12bp, 15bp, 14bp, 10bp respectively; over 20 days they rose 46bp, 43bp, 31bp, 18bp respectively; over 60 days they rose 66bp, 69bp, 60bp, 50bp respectively. Latest levels are 2Y 3.39%, 5Y 3.69%, 10Y 3.97%, 30Y 4.27%, CA 2s10s +58bp. Several features stand out: first, Canada is also in a bear-steepening regime and its curve is steeper than the US; second, absolute yield levels are clearly below the US, indicating cross-border spreads still favor dollar assets; third, the Canadian long end’s last ~20-day rise is smaller than the US’s, showing markets acknowledge a global upward shift in rate centers but have not simply copied the US policy-path narrative for Canada.

The global thread’s keyword is "long-end selling pressure spreading." On 9月24日 Reuters directly pointed out global bond sell-offs continued and US 30-year yields hit highs not seen since 2004; the same day Asian equities and bonds fell together, oil rose, and the dollar climbed to a two-month high. In the UK and Europe, news focused in two directions: first, UK activity is slowing, but inflation and wage pressures still make it difficult for the BoE to fully ease its rhetoric — on 9月25日 there were reports the Bank of England governor warned rates could rise; second, bond markets are repricing fiscal and growth constraints, with commentary in the UK and France suggesting "the bond market is forcing fiscal tightening." On the Continent, Reuters noted ECB Executive Board member Schnabel will leave to take a post at the IMF; this item itself has limited direct rate implications, but against a backdrop of increased global long-end volatility, uncertainty in European policy communication also amplifies term-premium swings.

Credit markets provided cautious auxiliary signals. US HY OAS widened 17bp over the past 20 days to 2.80%, while IG OAS remained unchanged at 0.79%. This indicates risk assets have not entered a full disorderly phase; credit spreads are only modestly reflecting tighter financing conditions. The core shock currently comes from rising risk-free rates rather than a sudden worsening of credit risk. For rate trading, this combination usually implies that "the discount rate rising" remains the dominant theme.

II. Views on future rates

Set method boundaries first. Directional judgement for the next three weeks can only be based on momentum segments already validated by the data. Currently only US 2Y, 5Y, and 10Y direction signals are labeled as [Proven]; US 30Y and the full Canadian curve are labeled as ≈ random walk. For those random-walk segments, direction is unpredictable and we do not make directional calls — we only discuss carry and roll-down when yields are unchanged. Carry is a static tailwind; if yields continue to rise, price losses can fully offset carry, which must not be overlooked.

Start with the US short and mid ends. US 2Y, 5Y, and 10Y currently all have [Proven] upward direction signals; the statistical edge is not overwhelming but is sufficient to support a baseline view that "upward pressure remains" over the next three weeks. In terms of strength, 5Y and 10Y have higher z values at 3.08 and 3.17 respectively, indicating current positions are quite extreme relative to the past year but momentum has not naturally faded. Combined with the macro background in the news flow — growth resilience, stronger consumption, hot PMI, markets re-discussing additional hikes, and oil/diesel-related inflation disruptions — the short and mid ends are unlikely to fall back quickly without new weakening data.

For the 2Y the logic is most direct. The current 4.87% level along with 20-day +68bp and 60-day +73bp moves reflect ongoing revision in policy expectations. The validated signal gives dir=up, with a hit rate of 56.2%, above the base-rate of 50.5%. Over the next three weeks, the 2Y is still biased toward continued pressure. Note that although 2Y carry is high — annualized 5.59 — and its win rate is 68% (a reasonably good static return profile), this does not mean "long 2Y is safe"; if yields keep rising, short-duration bonds will still record price losses, albeit generally smaller than those of longer durations.

For the 5Y and 10Y the direction is also upward, but the implications differ. The 5Y is closer to the intersection of policy path and the medium-term neutral rate: at 5.03% with 20-day +66bp and 60-day +84bp, momentum is quite strong. The 10Y embeds policy path, nominal growth, and term premium: currently 5.18% with 20-day +52bp and 60-day +74bp. Both validated signals support continued upside, with hit rates of 53.9% and 52.7% respectively. Therefore, there is data-backed reason to expect US mid-term rates to keep rising over the next three weeks. Within the short-to-mid bucket, the 5Y’s elasticity has been more pronounced in this move; the 10Y is more sensitive to macro data and supply sentiment and may show larger volatility.

The US long end cannot be given a directional call. US 30Y, though at a high 5.47% and accompanied by news noting "near 20-year highs," is explicitly labeled by the model as ≈ random walk; the current dir=up does not provide a usable predictive edge, so we make no directional judgement for the next three weeks. What can be said is the static return profile: 30Y carry annualized 5.905 is the highest across maturities, but reliability (corr) is only 0.077 and long-only hit rate is 53%, which is only marginally effective. In other words, if the 30Y yield trades sideways, coupon and roll-down provide some buffer; if the long end rises another dozen basis points, duration losses are likely to wipe out that carry. Global bond markets are still adjusting their views on fiscal, growth, and inflation, so the ultra-long end can easily undergo large two-way swings — high carry should not be misread as a directional safety cushion.

The full Canadian curve likewise does not provide directional predictability. Whether 2Y, 5Y, 10Y, or 30Y, the signals are all labeled ≈ random walk, so we do not make directional calls for the next three weeks. In static terms, Canadian 10Y and 30Y carry are 4.53 and 4.966 annualized, with long-only hit rates of 56% and 54%, offering slightly more coupon attractiveness versus holding pure cash, but their nature remains "expected returns if yields do not move." Given Canadian households’ greater sensitivity to high rates and news showing increased household cash-flow stress, Canadian bonds may not be able to sustain the same intensity of policy repricing as US Treasuries; however, that relative view is not proven by directional signals, so it can only be expressed as "US and Canada may continue to diverge," not as a definitive yield path.

Cross-country divergence is a key focus to watch over the next three weeks. US news and data jointly point to "stronger growth resilience, more persistent inflation risk, and policy that is harder to ease," while Canadian news emphasizes household stress. On yield levels, US 10Y is 121bp above Canada 10Y, and US 30Y is 120bp above Canada 30Y — such wide spreads are themselves an important variable for global capital allocation. If US data remains hot, the dollar rate center may stay higher for longer than Canada’s; but at the trade level, only the US short and mid ends currently have statistically meaningful directional signals — Canada is mostly following external environment moves.

On curve shape, the clearest feature is bear steepening. US 2s10s at +31bp and Canada 2s10s at +58bp show markets have moved away from the old framework of "front end high, back end low ahead of a recession" toward a new framework of "nominal growth and sticky inflation lifting the whole curve, with the back end reluctant to fall." Over the next three weeks, continued upside in US short and mid ends implies the curve will not immediately re-flatten; but since the long-end direction is unpredictable, there is insufficient evidence to judge whether steepening will keep widening. The prudent statement is: the curve is in a positively sloped region, and if macro data remains strong, the front and mid ends still have room to reprice upward; the long end is more influenced by supply, risk appetite, and overseas synchronization.

From a holding-period returns perspective, over a 3-week medium-short holding window carry should be seen as a buffer, not a directional bet. US 2Y, 5Y, and 10Y all have non-negligible annualized carry, especially the 2Y at 5.59, and carry win rates are relatively high, meaning if yields are unchanged or only mildly move, short-duration assets offer a relatively better holding experience. Although long-ends have higher carry, their prices are too sensitive to multi-week volatility and currently lack validated directional advantage, so static yield alone is insufficient to offset directional uncertainty.

III. Risk warnings

The biggest risk over the next three weeks remains US data continuing to surprise hot. Reuters and WSJ have both noted that employment and inflation prints will test the US rate path; if forthcoming employment, price, and activity data continue last week’s strength, the market’s pricing for another hike and an extended period at high rates could be revised higher and US 2Y to 10Y risk further upside.

The second risk comes from energy and geopolitical conflict. Bloomberg mentioned diesel prices’ disturbance to inflation expectations, and reports also describe Middle East conflict putting pressure on equities and bonds. If oil rises further, inflation compensation and term premium could passively rise, especially amplifying global long-end volatility.

The third risk is fiscal and supply pressure. Multiple media outlets link this round of long-end rises to government debt and fiscal sustainability discussions. Even absent new policy changes, high-yield environments alone can challenge sovereign debt absorption capacity and keep ultra-long ends like the 30Y highly volatile. Since these maturities lack a validated directional edge, any one-sided judgement should be held with caution.

The fourth risk is credit and risk-asset linkage. Currently HY spreads are only modestly wider and IG is basically stable, indicating the move so far remains in the "discount-rate rise" phase. If risk-free rates continue to climb and begin to materially suppress growth, credit spreads could fall back, which would then feed into long-end safe-haven demand and curve shape. Data has not reached that point yet, but this development needs monitoring.

Finally, restating the methodological constraints: US 2Y, 5Y, and 10Y can be given a directional judgement of "upward pressure remains" based on [Proven] signals; US 30Y and the full Canadian curve are directionally unpredictable and we do not make directional calls — only carry and other static factors can be discussed, and carry alone cannot constitute a bullish rationale. This report is for research purposes and is not investment advice.