Chapter 05 of 13

Suspect 5: Protected profits vs. spiraling wages

The orthodox debate about persistent inflation has two camps. Monetarists say it is wage-push: a tight labor market forces nominal wages up, firms pass them through, the price level chases. Heterodox critics — Isabella Weber's "sellers' inflation in emergencies" framing is the canonical one — say it is profits-push: firms with pricing power use a legitimate cost shock as cover for margin expansion. Both stories are about distribution, and both can be tested. In Iceland's case the answer is more interesting than either camp's slogan, because neither side of the bargain behaved the way the orthodox models predict, and the part of the system that captured the resulting transfer is not the part doing the bargaining.

The wages side

Iceland has had an extraordinarily tight labor market over 2020–2025. Unemployment among Icelandic nationals has been below 3% almost continuously. Construction has been running at capacity for years. Tourism boomed beyond any pre-pandemic baseline. In the orthodox wage-push story this is exactly the setting in which workers should be capturing rents from employers and bidding the price level up.

In real terms, that is not what happened.

Six years — covering almost the entire tight-labor, high-inflation cycle — and the weighted aggregate real wage is up about 9%. That works out to roughly 1.5% per year. Finance and insurance workers are real-flat: 2% total over six years. The only group clearly outrunning inflation is hospitality, with the highest foreign-labor share, and even there the bulk of the 19.5% real gain is 2022–2023 catch-up from pandemic-suppressed wages rather than a 2025 wage regime running ahead of the CPI. The shape of this table is wages catching up with prices, with a lag.

That ordering matters. If workers were pricing in next year's inflation ahead of the CPI print, wage settlements would move first and CPI would follow. In Iceland they don't. The kjarasamningar are written against realized inflation and include forsendur clauses — inflation-review triggers that let the contract be reopened if CPI overshoots an agreed threshold. The causal arrow runs from the CPI print into the wage deal, not the other way around.

BHM — the federation of Icelandic university-educated employees — has made the same arithmetic point with a single statistic: labor costs are only about 25% of total firm costs on average in Iceland, 15% in retail, 9% in metal production. A 5% wage raise on a cost base that is 15% labor contributes 0.75 percentage points to price. A 5% markup on revenue contributes the whole 5%. The arithmetic is not in the wage-push story's favor.

The aggregate wage line and the headline housing line run roughly together over the cycle: wages held real value, housing costs ran a couple of points ahead. But the housing-cost shock is heterogeneously distributed. The imputed-rent component (CP042) — which is calculated from house prices and is the channel through which housing inflation propagates into the rest of the CPI basket via verðtrygging — is up roughly 30% in real terms. Sitting tenants on multi-year leases (CP041) are up only ~7% real, because rent indexation under existing contracts is sticky. The marginal experience — anyone signing a new lease — sits much closer to the imputed line than the paid-rent line. The HMS national leiguvísitala on newly registered leases ran at +14.8% nominal between April 2024 and February 2026 alone, against CPI of +8.2% over the same window. Real new-lease rent is climbing at roughly 3–4% per year for movers.

So the honest version is: wages, on average, kept pace. Housing costs, on average, ran a little ahead. But the variance is enormous, and the people getting hit are not the average.

The 2024 Stöðugleikasamningurinn

Viðskiptaráð — the Iceland Chamber of Commerce — had spent 2023 and early 2024 running the orthodox line: inflation was a wage-push problem, there was a svigrúm (room) for wage growth defined by productivity plus the inflation target, and any settlement exceeding that room would produce a spiral. The unions delivered exactly the svigrúm the Chamber had asked for. On 7 March 2024, the ASÍ coalition Breiðfylkingin — SGS, Efling, and Samiðn — signed a four-year Stöðugleika- og velferðarsamningur with SA; VR followed on parallel terms later in March. The wage path was 3.25% in 2024 and 3.5% in each of 2025, 2026 and 2027, with a forsendur clause letting unions walk if headline inflation was still above 4.95% by August 2025. Against the 7.3% and 7.8% labor-cost growth the CBI attributed to the 2022 and 2023 rounds, this was roughly half the velocity. Inflation did not fall.

Union leadership closed ranks around the template. Sólveig Anna Jónsdóttir (Efling), Vilhjálmur Birgisson (SGS) and Hilmar Harðarson (Samiðn) — some of whom had spent the prior cycle as the most militant voices in Icelandic labor — framed restraint as the route to lower inflation and a lower policy rate. No named bargaining-committee member broke ranks at signing. You would have to go back to the national-wage-accord politics of the 1990s to find comparable discipline.

The government layered a roughly 80-billion-króna four-year welfare package (Vaxandi velsæld) onto the deal, aimed mostly at renting families with children — the segment of the workforce most exposed to housing-cost inflation. The IMF's 2024 Article IV treats it as a fiscal loosening; Suspect 6 examines its actual 2024 footprint and why the headline number overstates the near-term impulse. This is best read as the government taking on the distributive function the wage agreement could not, using fiscal capacity to compensate renters for a wage path set deliberately below prevailing inflation.

One item on the union demand list did not survive negotiation: the leigubremsa, a temporary cap on rent increases that would have bound landlord pricing on the housing component of CPI. It was excluded from the government's housing bill of 7 March 2024 — the same day the Breiðfylkingin contract was signed. The wage side of the bargain was disciplined; the rent side was not.

The CBI's read

At the 8 May 2024 MPC press conference, Governor Ásgeir Jónsson called the settlement "gott innlegg í baráttuna við verðbólguna" — a good contribution to the fight against inflation — and said outright "hætta á víxlverkun launa og verðlags hefur þar með liðið hjá í bili": the risk of a wage-price spiral has passed for now. The CBI's own labor-cost forecast for the contract period was around 4% per year, described as "much more in line with the inflation target relative to productivity growth than previous years' agreements." If the central bank's own read is that the 2024 wage round removed the spiral risk, then a 2025–2026 inflation that keeps printing above target cannot be a wage-push phenomenon from that round. The 2025 sectoral data confirms the settlement is biting — year-on-year nominal wage growth across sectors clusters in a narrow 5.4–7.0% band, half the 2022–2023 pace, and only 0.9–2.4% in real terms. Construction workers, at the heart of a supposedly overheated construction boom, are at 5.4% nominal and 0.9% real. Nobody with real-wage growth under 1% is producing a wage-push inflation.

The protected-profits side, done carefully

If wages are not the source, the heterodox alternative is profits. The "sellers' inflation in emergencies" framing argues that firms with pricing power use a legitimate cost shock as cover for margin expansion, especially in sectors where consumers cannot exit. The European Central Bank's 2023 unit-profits work documented this for the eurozone: in 2021–2022 corporate gross operating surplus per unit of real GDP rose materially, before mean-reverting as wages caught up. The question for Iceland is whether the same pattern shows up.

At the aggregate level, partly. The share of Iceland's national income going to wages (Hagstofan's framleiðsluuppgjör / production account) fell from 60.7% in 2019 to 58.3% at the 2022 peak, then fully unwound to 60.1% by 2025. The mirror-image profit share rose +2.4 percentage points and reverted. The point here is narrower than it may first sound, and it is worth being precise about. Icelandic capital did not stop extracting rent — the baseline profit share sits at roughly 39% of national income throughout the window, which is itself the product of a long-run distributional settlement and by any historical standard substantial. What the data rules out is the specific claim that the share grew during the inflation episode and stayed up. It grew briefly in 2021–2022 and then shrank back. So whatever is driving the 2021–2025 price level is not visible in the aggregate numbers as capital permanently enlarging its slice of the pie. The "greedflation across the Icelandic economy" reading, in the specific form that says rent extraction accelerated and stuck, is weak at the aggregate.

The sectoral picture, however, is much more concentrated.

Two things stand out. Finance and insurance is the only major sector whose profit share climbed monotonically through the entire rate-hike cycle — from 37.9% of sector GVA in 2019 to 49.4% in 2024, an 11.4-percentage-point swing that has not reverted. Construction is striking in the opposite direction: profit share compressed from 40.5% to 36.7% as the wage bill rose faster than output. The popular "greedy developers gouging on housing" narrative is not in this data. Builders absorbed the construction-wage shock; their margins did not widen.

Manufacturing is the other big number, but it is dominated by aluminium smelting, and the swing lines up with the 2021–2022 global aluminium price spike and the European power-price episode — a global commodity windfall, not pricing power over Icelandic consumers. Wholesale and retail spiked in 2021 and unwound completely. Real estate (which is mostly imputed rent and thus mechanically GOS-heavy) is essentially flat at ~90% throughout. Hospitality is down. The persistent sector-level rent capture is in finance, and only in finance.

The bank story

The finance-sector profit expansion is concrete and traceable. The three large Icelandic banks — Landsbankinn, Íslandsbanki, Arion — produced combined profits of 32 bn ISK in 2019 and 83 bn in 2023. Their net interest margin widened from ~2.6% in 2019 to 3.0% in 2024 — roughly three times the Nordic peer average. Total dividends and buybacks ran to ~210 bn ISK over 2020–2024 against ~42 bn in 2015–2019: a five-fold increase in shareholder distributions during the same five years the central bank was telling the public it was fighting inflation. Policy-rate pass-through to deposit rates was slow; pass-through to loan rates was fast; the spread between the two was the channel.

Bank ownership during this period was mixed — Landsbankinn ~98% state, Arion fully private, Íslandsbanki state-dominant until its residual 45.2% stake was sold on 16 May 2025 for ~90.58 bn ISK. Roughly half of the combined profits accrued to the Treasury, half to private shareholders. The state privatised its residual position at the cyclical peak.

This is not "profits-push across the Icelandic economy." It is narrower and more mechanical: a rate-tool transmission failure that turned the central bank's main instrument into a rent stream for the financial sector.

Where the squeeze landed

The aggregate real wage held value. The aggregate paid-rent index for sitting tenants ran modestly. Where the squeeze actually fell was at the intersection of three things: renting, moving (or being on a short lease), and being at the bottom of the wage distribution. The bottom four income deciles in Iceland are overwhelmingly renters — homeownership rates of 3%, 7%, 12%, 29% by decile. Foreign nationals (20.4% of the population, the demographic absorbing the labor-intensive hiring boom) rent at much higher rates than Icelanders, and HMS data shows they pay roughly 14% more per square meter for smaller units. Stefán Ólafsson, Efling's in-house economist, summarised the period in DV in February 2026: "Það er ekki láglaunafólk eða skuldsett ungt fólk sem er aðalorsök verðbólgu." Over 2020–2024, by his figures, capital and asset income grew +172% while wage income grew +50.8%.

Verdict

Suspect 5 is not the culprit. Wages are not the source of Iceland's inflation: the aggregate real-wage data, the structural restraint of the 2024 settlement, the CBI's own "spiral has passed" assessment, and the absence of forward-looking wage-setting all point in the same direction. Workers are absorbing inflation, not producing it.

Profits, on average, are not taking a larger slice of the pie either — capital's baseline ~39% share of national income is still there, still substantial, but it did not grow during the inflation. The important carveout is the financial sector, where the rate tool itself became the income stream and shareholder distributions quintupled. That is a mechanical transmission failure, not economy-wide rent capture.

Wages are a channel inflation flows through, not the place where it starts.