Digitalisation cuts hours – wage growth eats the gain

Two observations sit almost side by side in the latest round of Norges Bank's Regional Network survey, published on 11 June 2026: fewer companies report a shortage of IT skills, and some say that the use of digital solutions is reducing their need for employees somewhat. In the same report, expectations for annual wage growth have been revised upwards, and overall the companies expect employment to rise slightly in both the second and third quarter (Norges Bank, Regional Network 2/2026).

Those two sentences are the entire productivity debate in miniature. Digitalisation has begun to affect staffing needs. It simply isn't moving fast enough to keep up with the price of labour.

The threshold is 4.2 per cent

In the March round, the companies in the network expected annual wage growth of 4.2 per cent in 2026 and 3.9 per cent in 2027 (Regional Network 1/2026). The June round points upwards from that level.

4.2 per cent is not merely a figure in a macro report. It is the threshold the productivity gain has to clear for the wage cost ratio to stand still.

A calculation makes it concrete. A business with 40 full-time equivalents and NOK 80 million in revenue generates NOK 2 million in revenue per FTE. With NOK 28 million in wage costs, the wage cost ratio sits at 35 per cent. Wage growth of 4.2 per cent adds roughly NOK 1.2 million. If revenue stands still, the ratio moves to 36.5 per cent – which in practice means the entire earnings improvement for the year is gone before it has been earned.

The alternative is to take the gain on the staffing side. If headcount falls from 40 to 38 FTEs, the wage cost ratio lands at around 34.7 per cent even with full wage growth and unchanged revenue. In other words, digitalisation has to free up something in the order of every twenty-fifth FTE – every year – simply to neutralise the wage settlement.

That is the real ambition threshold for a digitalisation project in 2026. Not "efficiency", but four per cent of the payroll, every year, documented.

The aggregate hides the spread

Norges Bank describes the reduction in staffing needs as somewhat, at some companies – while the network as a whole expects employment to rise slightly ahead (Regional Network 2/2026). The report also notes that it remains difficult to recruit project managers and skilled workers such as machine operators, carpenters, electricians, chefs and drivers.

That produces an interesting picture: technology dampens demand where competence is easiest to automate or support, while the bottlenecks sit where the work is physically carried out. A company may therefore find that digitalisation solves its capacity problem in administration and reporting, while the expensive scarcity – and the wage driver – sits in the production line.

The fact that employment overall is expected to rise slightly also means that total payroll in many businesses grows faster than the 4.2 per cent per employee. Wage growth multiplied by more heads is a different quantity from wage growth alone, and that difference is easily lost in a budget discussion conducted in percentages.

The pricing route out is narrower than in 2022

The simplest way to finance wage growth is to pass it on in prices. That room is shrinking. In June, Norges Bank assumed that activity will continue to rise, but that growth in 2026 will be moderate and lower than projected in March, partly because of weaker growth in the first quarter. Capacity utilisation is described as close to a normal level, but edging down (interest rate decision, June 2026). The March round pointed the same way: the share of companies operating at full capacity utilisation had fallen slightly (Regional Network 1/2026).

Falling capacity utilisation and moderate demand growth are not an environment in which price increases pass through unchallenged. That leaves two sources to cover wage growth: volume or productivity. And volume growth is precisely what Norges Bank has just revised down.

Three metrics – and the traps inside them

This is where the Capassa perspective departs from the macro analysis. Norges Bank measures the nation. Margin is lost or won in the individual income statement, and there three quantities determine whether digitalisation has actually moved anything.

Revenue per FTE is the simplest indicator of whether freed-up time has been converted into income. The trap is that the figure rises on its own when prices rise. If prices have gone up three per cent, then three per cent growth in revenue per FTE is zero productivity growth. Without a rough adjustment for your own price increases, you are measuring inflation and calling it efficiency.

The wage cost ratio is the most honest of the three, because it puts wage growth and revenue growth in the same fraction. The trap is that cost is moved rather than removed. Replace employees with hired labour or consultants, and the wage cost ratio falls while total cost is unchanged or higher. The same happens when automation shifts money from the payroll line to software licences and consumption of AI services. A licence cost growing faster than payroll is a productivity gain on paper and a margin leak in practice.

Gross profit per employee separates your own value creation from purchase prices. For trading and project businesses, this is the figure that shows whether the company has become better at creating value, or has simply invoiced more expensive goods onwards.

Why quarterly is something other than annual

The wage settlement hits the accounts over spring and summer. That means the effect of wage growth – and of any digital gain – only becomes visible in the second and third quarter, precisely the quarters in which Norges Bank expects employment to rise slightly.

A business that reads its wage cost ratio in the annual accounts gets the answer in February of the following year. By then eleven months of margin development is history, and the next settlement is already being prepared. A business reading the same figure on a rolling twelve months each quarter sees the direction while it can still be influenced.

The difference between the two is not reporting hygiene. It is how many decision points you have in a year.

The question a board is really asking

Around the board table in 2026, "have we adopted AI" is a low-information question. The answer is almost always yes, in some form.

The question that actually separates businesses is how much of this year's wage growth has been financed by productivity, and how much has been financed by margin. That is a calculation, not an assessment – and it requires knowing your starting point before the settlement, not after.

Norges Bank's figures document that the gain from digital solutions has begun to show up in staffing needs, and that wage expectations are pointing upwards at the same time. Which of the two forces wins is not decided in the macro statistics. It is decided in three metrics, quarter by quarter, in each individual company – and gut feeling has a systematic tendency to rule in favour of the project you launched yourself.