Quick answer: The Bank of Canada sets the overnight rate, which anchors variable and prime. Your 5-year fixed is priced off the 5-year Government of Canada bond yield, and that yield is being pulled up by a global bond selloff that has almost nothing to do with the Bank of Canada. Two rival explanations for the selloff have emerged. Only one of them reaches a Canadian fixed mortgage — and the Bank of Canada’s own May 2026 neutral-rate assessment quietly told us which. If you are renewing a mortgage in the next 12 to 24 months, the practical read is that the floor under 5-year fixed rates has drifted structurally higher, and “wait for fixed to fall” is now a weaker bet than it has been at any point this cycle.
The Bank held. Your rate hold expired higher anyway.
On September 2, 2026, the Bank of Canada held its policy rate at 2.25% — exactly as every economist Reuters surveyed expected. The Governing Council’s own statement contained a sentence worth reading twice:
“Financial conditions have tightened since July. Long-term bond yields have moved up globally, including in Canada.”
— Bank of Canada, September 2, 2026 policy statement
Translation: we did not tighten policy. Something else did. That “something else” is why a homeowner whose rate hold expired last week is looking at a 5-year fixed quote that is higher than the one from six weeks ago, on a day the Bank did nothing. It is the single most confusing part of this rate cycle for clients, and the reason it is confusing is that most rate commentary treats “interest rates” as one thing. They are not.
The Bank of Canada controls exactly one rate — the overnight target — and every retail rate that follows it (prime, and therefore variable-rate mortgages and HELOCs) moves in near-lockstep. Fixed mortgage rates are a completely different animal. A 5-year fixed is priced off the 5-year Government of Canada bond yield, plus a lender spread. That GoC yield gets set every morning by bond investors trading against every other government bond on earth — the US 10-year, German bunds, UK gilts, Japanese JGBs. When those global yields move, ours move with them. When the Bank of Canada holds, that has essentially no direct effect on where the 5-year GoC settles that afternoon.
So here is the more useful question, and it happens to be the one this essay is actually about: up where, and why? Because the answer decides whether “just wait, fixed rates will come back down” is a strategy or a fantasy for a 2026 renewal.
Up where — the shape of the move that matters
Government bonds trade at different maturities — 2-year, 5-year, 10-year, 30-year — and they move independently. Each maturity gets its yield from a different mix of forces. The short end (2-year) tracks what investors think the central bank will do over the next year or two. The long end (30-year) barely reacts to next-meeting policy and is dominated by inflation expectations and the “term premium” — the extra compensation investors demand for locking money up for three decades. The 5-year, which is what matters for a Canadian fixed mortgage, sits somewhere in the middle: it reflects the average expected policy rate over the next five years, plus its own share of term premium.
Since the summer of 2024, the Canadian curve has moved in a distinctive way. The chart below shows the average monthly yield on the 2-year, 5-year, 10-year and long GoC bonds, alongside the US 10-year as a sanity check. Hover any month to see all five values.
Two years ago the curve was inverted — the 2-year was above the 5-year, because markets were pricing aggressive Bank of Canada cuts. Those cuts happened; the 2-year is essentially back where it was, at 2.96%. But every longer tenor moved up. The 5-year is 51 basis points higher. The 10-year is 72 higher. The long bond is 96 higher. The US 10-year — the yield that most closely anchors the global term-premium story — is up 96 basis points, from about 3.72% to 4.68%.
| Bond | Sep 2024 | Aug 2026 | Change | What it prices |
|---|---|---|---|---|
| GoC 2-year | 2.99% | 2.95% | −4 bp | Near-term BoC policy path |
| GoC 5-year | 2.77% | 3.28% | +51 bp | Your 5-year fixed mortgage |
| GoC 10-year | 2.95% | 3.67% | +72 bp | Corporate bonds, government finance |
| GoC long | 3.12% | 4.08% | +96 bp | Pension funds, insurers, headlines |
| US 10-year (context) | 3.72% | 4.68% | +96 bp | Global term-premium anchor |
The 2-year is flat. Everything longer is up, and the further out you go, the more it has moved. That is a textbook steepening of the yield curve. It is also the shape that tells you the move is not about the Bank of Canada: if it were, the 2-year would be moving too. Something further out along the curve is pulling every longer maturity up with it, and the 5-year — the piece that matters to a mortgage — is being dragged along in the wake.
So the question narrows: what is happening at the long end of global bond markets, and how much of it reaches the 5-year?
Two rival explanations, and they don’t imply the same mortgage rate
Serious analysts have offered two different reads of the 2026 selloff. They point at the same yield curve and describe two different reasons it steepened. The difference matters — because one story reaches your mortgage and one mostly doesn’t.
Explanation 1: A higher term premium at the long end
Capital Economics’ Global Economics Focus on August 18 argued that the selloff is mainly a term-premium repricing — investors demanding more compensation to hold long-dated government debt because fiscal, geopolitical and policy uncertainty have all gotten worse. In their words, “higher long-end yields appear to mainly reflect rising real term premia.” They point out that long-run inflation compensation — what the bond market implies about inflation five to ten years out — has barely moved, so it is not a loss of confidence in central-bank inflation targets. It is simply a bigger risk premium for locking your money in a US, UK or French government bond for 30 years while those governments show no appetite for repairing their fiscal positions.
Capital Economics is explicit about what this doesn’t mean for a mortgage:
“The direct impact is reduced by the fact that the sell-off has been focused on the very long end, which has a limited influence on borrowing costs for firms and households.”
— Capital Economics, How to think about the bond market sell-off, August 18, 2026
Read that quote next to the yield-curve chart above and the connection is obvious. Term-premium repricing lives at the 30-year. The 30-year has very little to do with a 5-year Canadian fixed. If this is the whole story, the mortgage-relevant part of the move is a spillover — real, but modest and eventually contained.
Explanation 2: A structurally higher R-star
A different explanation surfaced this week. Reuters reported on September 3 that a meaningful chunk of the bond selloff is investors repricing R-star. R-star (written r* by economists) is the neutral real rate of interest — the inflation-adjusted policy rate at which the central bank is neither stimulating nor restraining the economy. It is the rate the central bank would settle at once every business cycle has washed out and the world sits in some kind of long-run equilibrium.
R-star is not observed. It is estimated with models, and the most-cited estimate is the New York Fed’s Laubach-Williams (LW) series. Below is the LW estimate for the United States going back to 2000. The line is famously noisy — a rolling one-sided filter that updates in real time as new data lands — which is exactly why market participants pay attention to which way it is drifting.
| Quarter | R-star estimate | Context |
|---|---|---|
| Q1 2015 | 0.88% | Coming out of the 2010s “secular stagnation” debate. |
| Q1 2019 | 1.50% | Pre-COVID normalisation. |
| Q2 2023 | 1.62% | Model resumed after being suspended for the pandemic. |
| Q1 2025 | 1.36% | Recent low — the base Reuters compares against. |
| Q1 2026 | 1.73% | Sharp move higher over one year. |
| Q2 2026 | 1.65% | Latest — down slightly, still well above the Q1 2025 base. |
The LW estimate rose from 1.36% to as high as 1.73% inside a year. That is a large move by R-star standards. And Reuters reports that market participants believe the true level is higher still — nobody will name a number, but the direction of the guessing is up.
Two structural drivers get named. The first is the artificial-intelligence capital-expenditure buildout — hyperscalers such as Amazon, Microsoft and Alphabet issuing very large amounts of long-dated corporate debt, which competes directly with Treasuries for the same pool of investor capital. The Economist puts the projected AI data-centre spend at over US$5 trillion between 2025 and 2030, with related long-dated corporate issuance running about US$250 billion this year and a projected US$400 billion in 2027. The second is the US federal debt itself — around US$40 trillion — with no meaningful bipartisan appetite to slow the growth of it. What is unusual is that both pressures are landing at the same time: heavy public and private demand for capital, competing.
“The likely culprits for higher R-star could be the heavy investment around the AI build-out and higher government debt levels, both of which can boost demand for capital and lift real yields.”
— Chip Hughey, managing director of fixed income, Truist Wealth, quoted in Reuters
The mechanics matter for a mortgage. A higher R-star implies the central bank’s policy rate needs to settle at a higher level to remain neutral once the cycle plays out. That expectation gets built into the 2-year and 5-year yields immediately, because those tenors are essentially the market’s forecast of the average policy rate over the next 2 to 5 years. Then it flows through to the 10-year and beyond as well. CreditSights, quoted in the same Reuters piece, put it plainly: “A higher R-star is pushing rates across the curve higher.”
That is the crucial difference between the two stories. Term premium mostly bites the 30-year and has little pull on household borrowing. R-star bites the 2s and 5s directly — the mortgage-relevant part of the curve.
The popular take: “Bond yields are up, but they always come back down — take variable and ride the cuts, or wait a few months and fixed will drop.” This has been the default broker script for three years, and until this summer it was defensible. The problem is it treats the 5-year yield as a byproduct of the Bank of Canada’s cycle. It isn’t. It is a byproduct of a global bond market that has just given us two different reasons — one structural, one persistent — to expect the floor under 5-year fixed to be higher for longer than that script assumes.
The Bank of Canada’s split verdict
Here is the piece almost nobody in the retail mortgage conversation has picked up on. The Bank of Canada publishes its own assessment of the neutral rate every spring. The May 2026 update — Staff Analytical Paper 2026-21, authored by Alves, Beaudoin, Desgagnés, Dong, Schneider, Trostin, Toktamyssov and Twieling — did something quietly interesting. It raised the Bank’s assessed range for the US nominal neutral rate to 2.50%–3.50%, from 2.25%–3.25% a year earlier. And it held the Canadian range at 2.25%–3.25%, unchanged since 2024.
| Country | 2025 assessment | 2026 assessment | Change | Driver cited |
|---|---|---|---|---|
| United States | 2.25%–3.25% | 2.50%–3.50% | +25 bp (revised up) | “A stronger outlook for potential output growth.” |
| Canada | 2.25%–3.25% | 2.25%–3.25% | Unchanged | Lower long-term population growth offsets higher productivity. |
The Bank of Canada looked at exactly the questions the Reuters piece is asking — is the world settling at a higher neutral rate, and if so, by how much — and answered them country by country. For the United States: yes, we now think the neutral rate is 25 basis points higher than we thought a year ago. For Canada: no revision. Same range as 2024.
Which means that whatever is happening to the 5-year GoC is being imported — pulled up by its tether to US Treasuries, not by any change in the Bank’s view of where Canadian rates should ultimately settle. This is not a subtle point. A retail broker who says “the neutral rate is going up so fixed rates are going up” is wrong on the Canadian side of the border. What is actually going up is the American neutral rate, and the American 5-year yield with it, and the American 5-year drags the Canadian 5-year along because they trade against each other every morning.
The practical implication is uncomfortable and worth being honest about: this pressure on Canadian fixed rates is not something the Bank of Canada can meaningfully offset. Its September 2 opening statement put it in the plainest possible language: “Monetary policy cannot offset the effects of tariffs or influence global energy prices.” You can extend the same sentence to global bond markets. If US Treasury yields keep drifting up because R-star is drifting up in the United States, the Bank of Canada can hold, cut, or hike the overnight rate — and the 5-year GoC will still mostly follow the US 5-year.
Don’t conflate real and nominal — this is where every retail article gets it wrong
One last piece of housekeeping before we get to the renewal decision. R-star is the neutral real rate — the inflation-adjusted policy rate. The Bank of Canada’s 2.25%–3.25% range is nominal. Roughly, nominal neutral equals R-star plus expected inflation:
nominal neutral ≈ R-star + expected inflation
With Canadian inflation targeted at 2%, the implied Canadian R-star is somewhere around 0.25%–1.25% — nowhere near the 1.65% the NY Fed’s LW model shows for the United States. So the two numbers are not comparable at face value, and mixing them produces some genuinely bad takes: “R-star is 1.65% and the BoC is at 2.25%, so the Bank is already at neutral” (wrong: those are different rates for different countries in different units).
The correct read is more subtle and more useful. The BoC’s 2.25% policy rate sits at the bottom of its own assessed nominal neutral range — modestly stimulative, in the Bank’s own framework. This is a completely different claim from anything the NY Fed’s R-star estimate says. And it is why Capital Economics, in its September 2 analysis of the BoC decision, describes the Bank’s next move — if it moves at all this cycle — as “stage one: normalising to the 2.75% neutral midpoint” via 50 basis points of hikes, spread across two 25-basis-point steps. Not a hiking cycle. A normalisation to neutral, at most.
What it means for a 2026 or 2027 renewal
Rate mechanics aside, this is where a broker earns rent. Here is my read of it, honestly held and worth challenging.
1. The base case for 5-year fixed has drifted up, not down.
Both of the two rival explanations for the bond selloff point at a higher floor under 5-year fixed rates for the rest of 2026 and into 2027. Capital Economics thinks term premium settles permanently higher because governments “will do the bare minimum” on fiscal repair. Reuters thinks R-star has drifted up and market participants suspect further. The Bank of Canada is telling us both channels are, at the moment, mostly a US story that reaches us through the bond market. The 5-year GoC at 3.30% is up 51 basis points from where it was two years ago, and the best 5-year fixed at 4.24% reflects that.
What that means practically: if your rate hold is live, do not casually let it expire on the theory that you will get a better quote in two weeks. In a world where the 2-year GoC is flat and the 5-year is up 51 basis points, “wait for fixed to fall” is a bet against a structural drift, not a small tactical call.
2. Variable is not a slam-dunk anymore either.
The old “take variable and ride the cuts” pitch assumed the Bank of Canada’s next moves were down. That view is no longer credible. Capital Economics has just pulled its first-hike call forward from June 2027 to potentially December 2026. Consensus is now that every meeting from here is live for a hike, not a cut — and markets (via overnight index swap forwards) are pricing more than 100 basis points of cumulative tightening across two years, versus Capital Economics’s 50 basis points. The realistic scenario band for variable-rate mortgage costs, then, is somewhere between +50 basis points and +100 basis points over the next 18 to 24 months.
The current best 5-year variable rate is around prime minus 0.90%, or roughly 3.55%. The current best 5-year fixed is around 4.24%. Variable is starting the race about 70 basis points ahead. Capital Economics’s Canada base case is +50 basis points of Bank of Canada hikes across the cycle; markets are pricing more than +100. At +50, that head start is mostly gone by the end of your term (variable ends about 20 basis points cheaper than today’s fixed). At +100, the head start is fully consumed and overshot (variable ends about 30 basis points above today’s fixed). The realistic scenario band, in other words, is now “roughly a wash” to “materially worse than locking today.”
3. Get the rate hold. Get the pre-approval. Renew early if it’s within reach.
The most valuable, least-glamorous piece of paperwork in a bond selloff is a live rate hold. Most Canadian lenders will hold a 5-year fixed rate for 90–120 days, some longer. It costs nothing. It is a free option — heads, rates drop and you re-price; tails, you close at your held rate. In an environment where the 5-year GoC has moved 51 basis points in two years and one honest school of thought thinks it has further to go, that option is worth more than it has been at any point this cycle. Same logic applies to early renewal: if you are within about six months of a maturity, it is worth pulling a fresh quote and comparing the cost of blending your current rate with today’s rate versus rolling the dice on where rates are in six months.
Renewing in 2026 or 2027? The right lender for your file depends on your equity, income, and what your current lender will actually offer you at maturity. Book a 20-minute call and I’ll show you the numbers side-by-side.
Honest limits
Every part of this essay could be wrong for reasons that are already known. In no particular order:
R-star is unobservable. It is estimated with statistical models that rely on trailing data, get revised as new data lands, and produce estimates that reasonable economists disagree about. One quote in the Reuters piece calls estimating it “equal parts art and science.” The move from 1.36% to 1.73% could be signal; it could also be a filter noise pattern that the smoothed two-sided estimate (which barely budged) is telling us to discount.
The AI effect could flip. CreditSights’s Zachary Griffiths, in the same Reuters piece, flagged the counter-scenario: higher R-star now because AI is capital-hungry, but potentially a lower nominal policy rate later if AI proves broadly disinflationary. On a 5-year mortgage renewal horizon, that flip probably does not happen fast enough to matter. On a 10- or 20-year mortgage-planning horizon, it could.
Bond markets can also just calm down. Capital Economics’s own base case says the move is smaller than 2022 or autumn 2023 and has not — yet — triggered a broader tightening of financial conditions. The Globe and Mail carried the thinnest, most honest caveat: it is a summer market, volumes are low, and a few sessions do not make a regime. All true.
The trade war is the wild card. The one path back to Bank of Canada cuts runs through the US-Canada trade dispute showing up in hard investment and hiring data. Governor Macklem flagged that channel on September 2, and it is why the June “we would cut further” pre-commitment is still technically live. So far the data has not turned. If it does, the whole framework in this piece shifts and a variable-with-a-fixed-conversion clause becomes worth another look.
Frequently asked
What is R-star (the neutral rate) in plain language?
R-star is the interest rate — adjusted for inflation — at which a central bank’s policy is neither speeding the economy up nor slowing it down. It is the rate the policy rate is expected to settle at once the current business cycle has washed out. It cannot be observed directly; it is estimated with models. The most-cited estimate is the New York Fed’s Laubach-Williams model, which currently puts US R-star at about 1.65% (Q2 2026).
Why is my fixed mortgage rate going up when the Bank of Canada isn’t hiking?
Because the Bank of Canada sets the overnight rate, which anchors variable-rate mortgages and prime. Your 5-year fixed is priced off the 5-year Government of Canada bond yield, which is set every day by global bond markets. Long-term bond yields have moved up globally since July 2026, including in Canada — the Bank of Canada’s own September 2 statement says so explicitly.
Should I lock in a fixed rate now or wait for rates to drop?
The honest answer depends on your file, but the base case for waiting has gotten weaker. Two credible independent explanations — a higher R-star (Reuters) and rising real term premia (Capital Economics) — both point at a higher floor under 5-year fixed rates for the rest of 2026. If you have a live rate hold, do not casually let it expire. If you are within a few months of a renewal, get a fresh quote and compare it to your current lender’s offer.
Will 5-year fixed mortgage rates drop in 2027?
Possibly, but the path to that outcome now requires the US-Canada trade dispute to show up in hard investment and hiring data forcing the Bank of Canada to cut, and the global term-premium repricing to reverse. Neither is the current base case among named forecasters. Capital Economics’s Canada base case is 50 basis points of Bank of Canada hikes across 2027, not cuts.
What is the difference between R-star and the Bank of Canada’s neutral rate?
R-star is a real (inflation-adjusted) rate. The Bank of Canada’s assessed neutral range of 2.25%–3.25% is a nominal rate. Roughly, nominal neutral equals R-star plus expected inflation. With Canadian inflation targeted at 2%, an implied Canadian R-star of about 0.25%–1.25% is consistent with the Bank’s nominal range — far below the ~1.65% the New York Fed estimates for the United States. The Bank of Canada’s May 2026 update raised its US neutral range by 25 basis points but left the Canadian range unchanged.
Does the Bank of Canada control fixed mortgage rates?
No. It sets the overnight rate, which anchors prime and therefore variable-rate mortgages. Fixed mortgage rates track Government of Canada bond yields (5-year GoC for a 5-year fixed, 3-year GoC for a 3-year fixed, and so on), and those yields are set by global bond investors trading against every other government bond on earth.
How does a US bond yield affect a Canadian mortgage?
Government bonds of similar quality and maturity trade against each other globally. If the US 5-year Treasury yield moves up, the Canadian 5-year GoC yield tends to move up with it — otherwise investors would sell one and buy the other until the spread closed. Since 5-year fixed mortgage rates are priced off the 5-year GoC, a persistent move in US Treasury yields eventually shows up in Canadian mortgage pricing. The US 10-year is up about 96 basis points over the past two years; the GoC 10-year is up 72 basis points over the same window.
Is a global bond selloff a mortgage crisis?
Not by the numbers so far. Capital Economics called the August 2026 selloff “smaller than 2022 or autumn 2023” and noted bond-market volatility (measured by the MOVE index) had risen only modestly. What this is is a structural drift higher in the floor under 5-year fixed rates, which is a different and slower story than a crisis. The right response is a paperwork response — rate holds, pre-approvals, early renewal quotes — not a panic response.
This is commentary, not personalised mortgage or investment advice. Rate and yield figures are as of the dates cited (Bank of Canada policy rate: 2.25% as of the September 2, 2026 decision; Government of Canada yields: August 27, 2026, from the Bank of Canada Valet API; best available 5-year mortgage rates: September 3, 2026, from lender sources tracked by Mortgages for Less). Your rate, product and eligibility depend on your file. Josh Tagg is a licensed mortgage broker with Mortgages for Less (Alberta) and Indi Mortgage (Ontario).
