Explainer · Work

A raise is not purchasing power

The offer letter says one number. The receipt says another. Purchasing power is the ratio between them, not the difference.

Published November 13, 20269 min readPayroll earnings for August 2026 (released October 29, 2026); CPI for September 2026 (released October 19, 2026)
A kitchen table in morning light with a pay stub, a calculator and a household budget notebook.
Illustrative photography produced for Boreal Metrics.

A pay increase is a number on a letter. Purchasing power is what that number buys after prices have moved. The distance between the two is not a matter of opinion; it is a division problem — provided both rates describe the same period.

The offer and the receipt

In August 2026, average weekly earnings in Canada were $1,354.55, up 3.7% from a year earlier. In September 2026, the Consumer Price Index rose 3.1% over twelve months. The temptation is obvious: subtract the rates, or divide one by the other, and announce a real wage outcome.

Neither operation is valid with these two figures. The releases cover different reference periods — earnings for August, prices for September — and a real change requires rates measured over the same period. A national earnings average and a national price index also describe different populations and different baskets.

Divide, do not subtract

To see the arithmetic clearly, use a same-period hypothetical. Suppose income grows 5% over a year and prices rise 3% over the same year:

The calculation

real change equals one plus income growth, divided by one plus inflation, minus one. Income growth and inflation are entered as decimals.

g = income growth as a decimal (5% = 0.05). i = inflation as a decimal (3% = 0.03). Rates are divided before subtracting one; this is a ratio, not a difference of percentages.

With g = 0.05 and i = 0.03: (1.05 ÷ 1.03) − 1 = 0.01942 → +1.94%

The ratio says purchasing power rose by about 1.9% in this example — not the 2 percentage points that subtraction suggests. The distinction matters because the two quantities are different units that are constantly compared as if they were the same. The ratio answers a concrete question: how much more of the same basket can this income buy?

When the rates are small, subtraction is a close approximation. As the rates grow, the gap between the methods grows too. A pay increase of 8% against 6% inflation is a 2 percentage-point difference, but the correct real change is about +1.89%. A pay increase of 2% against 3% inflation leaves purchasing power at about −0.97%, slightly better than the −1 point that subtraction suggests. The direction of the answer rarely changes; the size often does.

There is a second reason to use the ratio. Inflation does not tax an income; it reprices a basket. If prices rise 3%, an income must rise by exactly 3% merely to stand still. The ratio captures that; subtracting the rates does not.

What an average wage measures

The earnings figure cited above is not a wage rate. It is average weekly earnings for employees receiving pay and benefits from an employer — gross pay before deductions, including overtime, divided by the number of employees. Statistics Canada notes that changes in the average can reflect wages, hours, the composition of employment and base-year effects. All four are moving at once.

  • Hours. Average weekly hours were 33.4 in August, little changed on the month and up 0.6% year over year. A longer week raises average earnings without any change in the hourly rate.
  • Composition. If lower-paid jobs disappear, the average can rise even if nobody receives a raise. Averages are sensitive to who is in the sample.
  • Coverage. The survey covers payroll employees. The self-employed are not in this average, and neither is anyone outside paid employment.
  • Overtime and bonuses. These are part of gross earnings and can move the average without changing a base wage.

A median would tell a different story from a mean, and an hourly wage rate would tell a different story again. None of these is your pay stub. A useful habit is to ask which of the four forces above is doing the work in any headline about wage growth.

A published check

Statistics Canada published a concrete same-period example of this arithmetic in its October 2026 study of Canada's health workforce: 2019 to 2025, with wages adjusted using the CPI. In nominal terms, the average hourly wage for core-aged employees in health occupations was up 24.6% in 2025 compared with 2019. After adjusting for inflation, the increase was 3.2%.

The aggregate hides divergent paths. Among core-aged registered nurses, real average hourly wages fell by 1.4% over the same period, while core-aged nurse aides, orderlies and patient service associates saw real wages rise by 2.7%. One national average, several occupations, different outcomes — exactly the pattern to expect whenever a ratio is applied to a group.

Work it out

The comparison below is live arithmetic in your browser, not live data. It starts from a hypothetical same-period example — 5% income growth and 3% inflation — precisely because the two dated releases above cannot be combined into a real wage outcome.

Change either field to a pair of rates measured over the same period, if you have them, and read the result as a comparison, not a verdict. The calculator sends and stores nothing.

Purchasing power, worked

Hypothetical example

The calculation

real change equals one plus income growth, divided by one plus inflation, minus one. Income growth and inflation are entered as decimals.

g = income growth as a decimal (5% = 0.05). i = inflation as a decimal (3% = 0.03). Rates are divided before subtracting one; this is a ratio, not a difference of percentages.

Enter as a percentage. 5% becomes 0.05 in the formula.

Prices rising 3% means the formula divides by 1.03.

Example presets

Real change in purchasing power · hypothetical

+1.94%

Income grew faster than prices: purchasing power rose.

Rates as decimals
g = 0.050 · i = 0.030
Subtracting the rates instead
+2.0 percentage points

A difference of percentage points is an approximation. The ratio above is the correct comparison because the basket is priced at the new level.

The fields start from a hypothetical same-period example: 5% income growth and 3% inflation. They are not a pairing of the dated releases. Average weekly earnings for August 2026 (up 3.7% year over year, released October 29, 2026) and the CPI for September 2026 (up 3.1% over twelve months, released October 19, 2026) cover different reference periods, so they are cited separately and are never divided into a real wage outcome here. This is not a forecast or live data.

What the number cannot tell you

A real change in purchasing power is a comparison of two indexes. It is not a budget. It says nothing about taxes, benefits, debt payments, childcare, housing tenure or the specific prices a household actually faces. It also does not account for quality changes in goods, which statistical agencies adjust for in ways that can feel abstract at a kitchen table.

The formula also assumes a single income growing at a single rate, while most households have several streams, irregular hours and one-off costs. Use the result as a way to check a claim — "pay is beating inflation" — not as a way to describe a life.

Used carefully, though, the ratio does something valuable: it forces a precise question. Not "did pay go up?" but "did pay go up by more than the prices this household pays?" Those are different questions. Only one of them has an answer that fits on a pay stub.

Sources

Source

Statistics Canada, Payroll employment, earnings and hours, and job vacancies, August 2026. Released October 29, 2026.

Read the release

Source

Statistics Canada, Consumer Price Index, September 2026. Released October 19, 2026.

Read the release

Source

Statistics Canada, Labour market characteristics of Canada's health workforce, 2025. Released October 19, 2026.

Read the release

Boreal Metrics does not publish forecasts or investment advice. Where a figure is corrected, the correction is noted on the page that carried it.