Skip to main content

Fixed Wins Enough of the Time

The debate over fixed or variable rates never ends. But there are now two things that are far less debatable:

1) The Bank of Canada is telling us that prime rate has likely hit bottom, and

2) “Safer” 5-year fixed rates are now as low as “riskier” variable rates.

(Safer, particularly if you choose a fair penalty lender.)

Given the above, if you’re getting a new mortgage you best have a bloody good reason to float your rate in this stage (the trough stage) of the business cycle.

One good reason that many cite is the historical outperformance of variable rates. Those data seem compelling. But what most don’t acknowledge is that said research depends partly on outdated assumptions.

Fixed rates actually win quite a bit…in the right sort of market.

If you run a general back-test to 1970 (as far back as good 5-year fixed data goes), for example, the results are notable.

If you assume, for instance, that:

A)  5-year posted rates sufficiently reflect the direction of fixed rates over time, 

B)  prime rate sufficiently reflects the direction of floating rates over time, and

C)  borrowers in the past received the same discount as people get today 

…then you’d find that 55% of the time variable rates averaged more than a 5-year fixed rate over a 60-month period.¹

Of course, borrowers never used to get the discounts they’re getting today, but it’s overall rate direction (i.e., the amplitude of rate peaks and valleys) that matters most to this sort of analysis.

Even if we adjusted for more historically relevant discounts (like those used by Canada’s father of modern mortgage research, York University professor Moshe Milevsky²) fixed rates would still outperform about one-quarter of the time.

One-quarter may not sound like a lot, until you realize that the one-quarter occurred disproportionately after economic slowdowns when fixed-variable spreads were tight…like today.

That’s why conservative borrowers can’t be faulted for locking in until 2025, assuming again that they choose a fair-penalty lender. There’s always that outside chance the Bank of Canada and/or bond market go negative on rates. In that case, fixing for 5-years might cost you.

But, with rates this low, any such cost won’t be enough to lose sleep over.


¹ Today’s discounts equal 280 bps off the typical 5-year posted rate (for 5-year fixed rates) and 55 bps off prime (for variable rates).

² Milevsky used up to 150 bps off 5-year posted and 75 bps off prime in his study, published in 2008.

Parting Note: The back-of-napkin math above isn’t intended to rival the rigour of Milevsky’s study. For one thing, a simple average of variable rates over 60 months doesn’t return a precise calculation of interest expense. But it’s close enough to the ballpark that one can see the infielders.

compare button


  • Pizderi says:

    Rate isn’t everything right? What happens when you factor in IRDs? I can’t imagine there’s ever been a 5-year period where the *economics* of a fixed beat those of a variable.

  • The Spy says:

    Hi Pizderi

    That’s why we stressed the fair penalty lender (FPL) point.

    There have been plenty of times when a FPL 5yr fixed outperformed. In this sort of rate environment you’ll very like pay a three month interest penalty to break a 5yr fixed ***IF*** you choose the right product.

  • Broker says:


    If you’re worried about penalties take a lender like Merix, Equitable Bank, First National or Manulife. You’ll probably never pay an IRD now because rates can’t fall enough to trigger one.

  • Pizderi says:

    @Broker, @Ratespy. Even with a fair penalty lender the IRD is still likely to kick you in the teeth. Right now those poor schmucks who signed a “great” 2.99% 5-year fixed with FN in late 2019 are paying thru the nose on an IRD (quick check of their comparison rates suggests they’re probably paying something like 4% of their balance as a penalty today).

    To each their own, but to me a fixed rate deal is a complete non-starter.

  • Oiler88 says:

    With banks at prime – .60% at best there is no justifying a variable right now IMHO. There is nothing to gain even if rates stay put for five years.

  • Pizderi says:

    @Ratespy, this is wholly dependent on your rate forecast. You believe, generally, that our yield curve has hit the bottom and will rebound over the next 5-years. I don’t agree with you. That’s allowed, no?

    I also don’t agree that the “cost” of the bet (variable v fixed) has changed – it is the exact same as it always was (ie: based on ones view of future interest rates).

    FYI – I have a “bloody good reason” to go variable: I think at worst (for me as a variable picker) the prome rates stay flat over the next 5-years. With my 20bps discount in hand (fixed v variable) I’ll be ahead in 5 years.

    • The Spy says:

      Pizderi, That characterization would be incorrect. 5-year rate forecasts are not being made here. No one can say that yields have hit bottom. This is merely about helping people calculate the risk-reward based on the available data. If you see the data differently, we wish you the best. Informed counter-points are always appreciated.

      This story was never about existing borrowers with big pre-existing variable discounts. It was intended for new borrowers and for them, the risk-reward has indeed changed.

      In a nutshell, it would be an error to:

      * Suggest that fixed-rate penalty risk is the same at a 0.39% 5-year bond yield yield as it was at 2.00% (it’s not; yields are highly unlikely to drop as much from here, and the effect of that can be seen in hypothetical IRD calculations)
      * Ignore historical performance of fixed rates at this point of the business cycle
      * Ignore that fixed rates have almost no rate premium
      * Ignore that prime rate is unlikely to fall further, per the Bank of Canada
      * Ignore the disproportionate reciprocal rate risk (i.e., that inflation expectations and/or overborrowing could take rates higher by years 4-5)
      * Ignore that it may take as few as one rate hike in year four to put a *new* variable-rate borrower underwater, relative to a 5-year fixed

      The above are precisely why the risk-reward of fixed rates have likely never been more favourable.

  • T says:

    1.65% 5Y fixed looks pretty good.
    Do you think this rate could fall even more?

Leave a Reply

Your email address will not be published.