The Unbalanced Loan: Ending Indexation by Building a Fairer Mortgage

This is the second piece in the series on the problems Iceland now has to solve at home. The first was about inflation — the two levers, monetary and fiscal, that bring it down. This one is about the machine that keeps it coming back, and about a quiet injustice sitting at the centre of Icelandic household finance. It builds on an argument I made last November, after the Supreme Court’s mortgage ruling, that Iceland has a once-in-a-generation chance to redesign its mortgage market. The referendum is behind us; the redesign is ahead of us. Here is how I think we should approach it.

Let me begin not with economics but with fairness, because that is where this problem really lives. And to see the unfairness clearly, we have to do something simple that is almost never done: take an ordinary Icelandic mortgage apart, name every risk inside it, and ask, for each one, two plain questions. Who carries this risk today? And who is actually equipped to carry it?

There are five real risks in a home loan. Let us walk through them one at a time.

The first is inflation risk — the risk that the cost of living rises over the decades you are repaying. In an Icelandic indexed loan, the borrower carries this entirely. That is what indexation is: every rise in the consumer price index is added to the amount you owe, so inflation inflates your debt. Now ask who is equipped to carry it. A household has one income and one home; it cannot hedge inflation, cannot diversify it, cannot lay it off to anyone. A bank can. A bank has a treasury, access to financial markets, the ability to issue its own inflation-linked bonds to investors who want them, and thousands of loans to spread the risk across. The party with every tool to manage inflation risk hands it to the party with none.

The second is interest-rate risk — the risk that the price of money itself changes over the life of the loan. On a variable-rate Icelandic mortgage, the borrower carries this too: when rates rise, the payment rises, and there is nothing the household can do about it. And again, this is a risk that banks and bond markets are built to manage — it is their daily business, priced and traded in markets designed for exactly that purpose. The household has no such market and no such tools.

The third is house-price risk — the risk that the property loses value. The borrower carries this, and here I want to be completely fair: this risk belongs to the borrower. You own the asset; you enjoy it when it rises and you bear it when it falls. That is ownership, and there is nothing unjust about it. I include it precisely so you can see that the argument is not that borrowers should carry no risk. Some risks are rightly theirs.

The fourth is income risk — the risk that you lose your job or your earnings fall and can no longer make the payment. The borrower carries this, and this one is largely theirs too, in the nature of things. But notice that the Icelandic system offers the household no shock absorber against it — no built-in flexibility for the moment life goes wrong. Hold that thought, because the fairer systems do offer one.

The fifth is credit risk — the risk that the borrower fails to pay, which is the lender’s risk. This is the one risk the bank carries. And even this is heavily softened: the loan is secured against the house, and the borrower is personally liable on top of that. The single risk the lender holds is the best-protected risk in the entire contract.

Now stand back and look at the whole picture. Of the five risks in a mortgage, two are legitimately the borrower’s — the value of their house and their own income. But the two risks that a household is least able to manage — inflation and interest rates — have also been placed on the household, even though the bank is the party built to handle them. And the one risk the bank does carry, credit risk, is the one most thoroughly secured.

The borrower carries almost every risk in the transaction, including the ones they cannot manage. The lender carries one, and it is collateralised. The party with the balance sheet, the treasury, and the professional risk managers has handed the two hardest risks to the party with one income and one home. That is the asymmetry, and once you have seen it you cannot unsee it.

This matters because a loan is supposed to be a sharing of risk. The lender provides capital and takes on some of the uncertainty of the future, in exchange for the return it earns. Indexation breaks that bargain. It guarantees the lender a real return, protected against inflation, while handing the single largest uncertainty of the next forty years — what inflation will do — to the household, the party least able to absorb it and least able to see it coming. The risk has been placed with exactly the wrong party.

And here is the point that legislation can act on directly. Banks do not need to transfer inflation risk to borrowers. They have the means to hedge it themselves — through inflation-linked bonds sold to willing investors, through the covered-bond structures used elsewhere in the world, through diversification across their whole loan book, through the ordinary tools of a modern treasury. Indexation is not the only way to handle inflation; it is merely the cheapest way for the bank, because it costs the bank nothing to write the risk onto someone else’s principal. Hedging costs a little; offloading costs nothing. So the bank offloads. Legislation can change that calculation — it can require that inflation risk sit with the party equipped to manage it, and let the modest, transparent cost of hedging be priced openly rather than dumped invisibly onto families. That is not anti-bank. It is simply insisting that the institution built to carry a risk should carry it, instead of handing it to the household that cannot.

I want to be fair about how we got here, because it was not a conspiracy. Indexation was introduced in 1979, after a decade in which inflation ran wild and destroyed the value of savings and loans alike. In that context it had a logic: it made long-term lending possible again when money could not be trusted to hold its value. It was a painkiller for a real disease. But the painkiller was never put down. Forty-seven years later, the emergency measure has calcified into the permanent architecture of Icelandic home finance, and the injustice inside it is now structural and inherited rather than intended. That does not make it any less unjust. It just means the fix is a matter of design, not blame.

The proof: the lender never loses

You do not have to take the risk breakdown on faith. It shows up, plainly, in the banks’ own audited accounts.

If the borrower truly bears almost all the risk in a mortgage, then the lender’s one risk — that the borrower fails to pay — should almost never cost it anything. And it almost never does. Across Iceland’s three banks, loan losses have run near zero for years, and in several years the banks actually released provisions they had set aside — booking a gain, because the losses they had braced for never came. The only real blip in a decade was the 2020 pandemic, and even that was modest. The banks’ own risk models assume roughly 30 basis points of loss “over a lifetime”; in practice they have come in far below even that.

And the deeper number is the one in the second panel. Since 2011, through the pandemic, the inflation spike, and the sharpest interest-rate shock in a generation, not one of Iceland’s three banks has had a single losing year. Fifteen straight years of profit, all three, without exception. The years that were hardest on Icelandic borrowers were among the most profitable for the institutions that lent to them.

I want to be careful here, because this is not an accusation of wrongdoing. The banks did nothing improper; they operated a system exactly as it was built to work. That is the whole point. The asymmetry is not a matter of anyone behaving badly — it is designed into the architecture of the indexed loan. A structure in which one party carries every serious risk and the other has not had a losing year in fifteen is not a fair sharing of the future. It is a machine for making one side of the contract lose-proof.

Where the inflation lands

Now to the consequence, because this asymmetry does not just sit quietly on the banks’ balance sheets. It reshapes the whole housing market — and here the story connects to everything else I have written in this series.

Iceland has indexed almost everything to inflation. Mortgages are indexed. Rents are indexed. And wages, through the mechanism of collective agreements tied to the cost-of-living index, effectively chase inflation too. Each of these protections is individually reasonable — no household wants to fall behind rising prices, and no union would be doing its job if it let its members’ wages be eroded. But add them all together and you get an economy in which nothing absorbs inflation. Every flow is protected. And a cost that everyone is protected from does not disappear — it recirculates, and it compounds, and it goes looking for the one place in the system that is not indexed to anything: the price of the asset at the centre of it all. The house.

Look at what actually happened over the last decade, and the loop is visible in the data.

Since 2015, Icelandic wages rose about 109%, and household mortgage debt rose about 111% — almost exactly in step. That near-identical climb is not a coincidence; it is the fingerprint of indexation. Wages chase the cost-of-living index; the indexed mortgage is tied to the same index; so the two move together, both tracking inflation by design. But house prices rose 162% — pulling decisively away from both. The gap between the red line and the other two is the distance by which an ordinary household, running as fast as its indexed wages allow, has been priced out of its own market.

Why do prices pull ahead rather than simply keeping pace? Partly because of the indexation loop itself. The indexed mortgage lowers the entry payment — the first instalment is small, because the inflation is loaded onto the principal instead — which lets a given income support a larger loan. That extra borrowing capacity does not make housing more affordable; it gets capitalised straight into the price. The buyer can bid more, so the house costs more, so the next buyer needs an even larger indexed loan, and around it goes. Housing is also the heaviest single component of the consumer price index, so when house prices rise they push the index up, which lifts wages and indexed loan balances, which supports still higher prices. The instrument sold as making housing affordable is one of the engines making it unaffordable.

And indexation is not the only force lifting the price — it would be dishonest to claim it were. Demand has grown from outside the wage economy entirely. Iceland now draws well over two million tourists a year, and the short-term-rental market has converted a meaningful slice of central Reykjavík housing into tourist accommodation — which both adds demand priced off nightly yield rather than local salaries, and removes homes from the residential stock. At the same time the workforce has been transformed: immigrants now make up around a quarter of it, drawn by work, and every one of them also needs somewhere to live. Many, priced out of the centre like everyone else, end up in the towns around the capital — Mosfellsbær, Hafnarfjörður, Selfoss — with the congestion and commuting pressures that follow. None of this is anyone’s fault: tourists, new arrivals, and young Icelandic families are all simply trying to live and work. But it means the price of a home is being pushed up by forces that have nothing to do with Icelandic wages — which is precisely why wages, even indexed ones, can never catch it. You cannot win a race against a target that is being lifted by demand your wages do not touch.

Put the two together and the trap is complete. From the inside, indexation amplifies borrowing capacity and recirculates inflation into the asset. From the outside, tourism and a growing population push demand off a base that has nothing to do with local pay. And through all of it, whichever force moves the price, the lender is untouched — holding inflation-protected, collateralised debt, and, as its own accounts show, not losing money in any of it. The household absorbs every shock. The house runs away. The lender never loses.

Indexation is a substitute for something better

Here is the insight that changes how to think about the solution. Indexation is not a feature Iceland chose because it is good. It is a workaround — a crude substitute for a piece of financial machinery that Iceland never fully built.

In a country with a deep, functioning mortgage-bond market, lenders do not need to shove inflation risk onto borrowers, because they can fund each loan by selling a bond to an investor who is willing to hold that risk and is paid to. The risk goes to someone who wants it. Iceland, lacking that machinery at scale, reached instead for the blunt tool: index the loan, and make the household the risk-holder of last resort. Indexation is what you use when you do not have the real thing.

Which means the goal is not simply to ban indexation. A ban, on its own, would just remove a tool without replacing what it does, and the market would seize up. The goal is to build the real thing — the machinery that makes indexation unnecessary — so that indexation withers because nobody needs it any more, not because it was outlawed. And here is the good news: we do not have to invent that machinery. It exists, it has worked for two centuries, and it is just across the water.

The door the Supreme Court already opened

We are not starting from a standing stop. In October 2025, the Supreme Court of Iceland ruled that Íslandsbanki’s variable-rate mortgage terms — which let the bank change rates on vague, discretionary grounds — were unlawful. Rates, the Court held, must be tied to transparent, external benchmarks, not to a bank’s internal discretion. (A later ruling, in the Arion case, confirmed that indexation itself remains lawful — so this is not a court striking down indexation; it is a court striking down opacity and discretion.)

That distinction matters, and it is precisely the opening. The Court has already knocked out one half of the unbalanced bargain — the lender’s freedom to move the goalposts. The government has responded by proposing a Central Bank–administered benchmark rate, which would replace discretionary pricing with something transparent and comparable across lenders. That is the right instinct, and it is the first plank of a redesign. The question is what we build on top of it — and my answer, which I argued in more detail last November, is the Danish model.

Why Denmark, specifically

Denmark solved this problem two hundred years ago, and its mortgage system is widely regarded as the best in the world. It rests on a principle that speaks directly to the asymmetry I described at the start.

The core of it is match funding — the “balance principle.” Every mortgage is funded, one-to-one, by a covered bond with identical cash flows and maturity. The mortgage bank is not a risk-warehouse; it is a pass-through conduit between the borrower and a bond investor, and it keeps only a transparent margin for its trouble. The interest-rate risk does not land on the household and it does not sit on the bank’s balance sheet. It goes to bond investors who choose to hold it and are paid to hold it. Risk is placed with the party that wants it. That is the fair arrangement that indexation replaced with an unfair one.

And there is a feature of the Danish system that is almost the exact mirror-image of Icelandic indexation, and worth dwelling on. A Danish borrower can always repay the loan by buying back the bonds that funded it, at their market price. So when interest rates rise, those bonds trade below face value — and the borrower can retire the debt at a discount. Rising rates improve the Danish household’s position. Compare that to Iceland, where rising inflation grows the borrower’s principal. In Denmark, the machinery hands the household a hedge. In Iceland, it hands them a ratchet. Same life event — rates move against you — and the two systems do precisely opposite things to the family’s balance sheet.

The Danish system carried this through the 2008 crisis without a single government bailout, despite house prices falling by an amount comparable to the United States. It is not magic — it clears partly through credit rationing rather than pure risk-based pricing, and borrowers pay administrative and prepayment fees for what they get. But it is transparent, fair, and proven, and it is the destination Iceland should be building toward.

How to get there without hurting people

The sequence matters as much as the destination, because done in the wrong order, this reform crushes the very households it is meant to help.

First — and this is why the inflation piece came before this one — win the inflation fight. You cannot move people out of indexed loans and into transparent, market-rate ones while inflation is high, because at high rates the non-indexed payment is brutal. The transition is only humane once inflation is falling and rates are coming down. Encouragingly, this migration has already begun on its own: as non-indexed rates have eased, record numbers of Icelandic households have voluntarily refinanced out of indexed loans. People leave indexation the moment the alternative becomes affordable. Our job is to make that voluntary drift the deliberate direction of the whole system.

Second, turn off the flow. New lending should default to transparent, benchmark-based, and — over time — match-funded terms, not indexed ones. The Central Bank already holds the regulatory levers to steer this: the loan-to-value and debt-service rules that govern new mortgages can be used to make indexed lending the exception rather than the default. Combined with the benchmark the government is already building and the legal shift the Court has already forced, the flow of new indexed debt can be wound down within a few years. Fast on the flow.

Third, convert the stock patiently. The existing pile of indexed mortgages cannot be force-converted overnight without inflicting exactly the payment shock we are trying to avoid. It should be migrated over a decade or more, as inflation falls and as the covered-bond market deepens, giving households the room to move when the numbers work for them. Patient on the stock.

Fourth, build the machinery — and open it up. Iceland needs to develop a genuine covered-bond market on the Danish pattern: match-funded, transparent, with the buy-back prepayment right that gives households their hedge. And here is where this piece hands off to the next one. Today, Icelandic mortgages are financed overwhelmingly by the pension funds, whose required real return — the 3.5% actuarial floor I have written about before — effectively sets a floor under the cost of capital for the entire country. A deep covered-bond market that draws in other investors, including foreign ones, does more than fund fairer mortgages. It begins to erode that floor, by ending the pension funds’ role as the captive, sole financier of Icelandic housing. That is the bridge to the cost of capital, which is where this series goes next.

The fairness at the end of it

Strip away the machinery and this is, in the end, a question about what kind of bargain we think is just.

For nearly half a century, Icelandic households have signed loan agreements in which they carry almost all the risk and the lender carries almost none — in which the family bears the inflation that the whole system, through indexation, is designed to pass to them. It was born of a real crisis and it made sense in its moment. But it has long since stopped being a fair deal, and the country now has, for the first time in a generation, both the legal opening and the proven blueprint to replace it with one that is.

A fair loan shares its risks with the party best able to bear them. Denmark has shown, for two centuries, that this is not idealism — it is engineering. The Supreme Court has cracked the door. The inflation fight, won properly, walks us through it. What remains is the will to build — and building, as I have argued through this whole series, is the thing we do here when we decide something is ours to fix.

This one is ours to fix.

---

Sources and notes

• Indexation mechanics and history. Verðtryggð lán — negatively-amortizing loans whose principal rises with the consumer price index — were introduced in 1979 following the inflation of the 1970s, and remain the majority of Icelandic mortgages (estimates range from about half to roughly 60% of the stock). Indexed borrowers pay a real interest rate (in the region of 3.5%) with inflation added to principal. Non-indexed variable mortgage rates rose from about 3.7% (2021) to 5.8% (2022) to about 9.5% (2023). (Central Bank of Iceland; S&P; EMF Hypostat; academic literature on verðtrygging.)
• The refinancing shift. During the 2023–25 rate rises, households moved into indexed loans to lower monthly payments; more recently, as non-indexed rates have eased, Íslandsbanki reports a record number of homeowners refinancing from indexed to non-indexed loans. (Íslandsbanki; Iceland Review; Central Bank of Iceland.)
• The Supreme Court rulings. On 14 October 2025 the Supreme Court of Iceland ruled that Íslandsbanki’s variable-rate mortgage terms, which allowed discretionary rate changes without clear criteria, were unlawful, requiring variable rates to be tied to transparent benchmarks; banks paused products and the Financial Stability Council issued warnings. On 10 December 2025, in a case against Arion Bank, the Court found the terms of a specific indexed mortgage lawful. Indexed lending itself remains legal; Íslandsbanki restarted indexed loans as five-year fixed products. (Supreme Court of Iceland; Landsbankinn; Iceland Review; Arion Bank statements.)
• The government benchmark proposal and the pension floor. The government proposed a Central Bank–administered benchmark reference rate to replace discretionary pricing. Icelandic pension funds finance a large share of household mortgages (estimates around two-thirds) and, by actuarial practice, discount liabilities at a long-run real return near 3.5%, which functions as a structural floor under the domestic cost of capital. These points, and the risks of hard-linking mortgages to government bond yields (the Housing Finance Fund / IL Fund experience), are developed in the author’s November 2025 piece, “Iceland’s Mortgage Market After the Supreme Court: From Discretion to Discipline.” (Central Bank of Iceland; Ministry of Finance; startupiceland.com, November 2025.)
• The Danish mortgage model. Danish mortgages operate on the “balance principle”: each loan is match-funded one-to-one by a covered bond with identical cash flows and maturity, with the mortgage bank acting as a pass-through conduit that retains neither interest-rate nor prepayment risk and earns a transparent margin. Bond prices are public; borrowers obtain close-to-capital-markets pricing. Borrowers may prepay by buying back the underlying bonds at market price, so that when rates rise and bonds trade below par, debt can be retired at a discount — a natural hedge for the household. The system, over 200 years old, came through the 2007–09 crisis without government bailouts despite a US-magnitude house-price fall. It clears partly through credit rationing rather than pure risk-based pricing, and borrowers pay front-end, administrative (bidragssats, typically 0.5–1% a year), and prepayment fees. (Finans Danmark; Nykredit; New York Fed Staff Report 848; IMF Country Report 07/123; academic and industry sources.)
• Borrower-based rules. The Central Bank of Iceland maintains loan-to-value and debt-service-to-income limits on new mortgages (rules no. 217/2024), which are available as levers to steer the composition of new lending. (Central Bank of Iceland.)
• The banks’ returns and losses (Figure: “The lender never loses”). Net interest margin, return on equity, cost-to-income, and cost of risk were computed by the author from the audited consolidated financial statements of Íslandsbanki, Arion Bank, and Landsbankinn, 2011–2025 (income statements and balance sheets as published in the banks’ factbooks; figures in ISK). Return on equity is calculated on average equity (opening plus closing, divided by two), so it differs by a few tenths of a percentage point from each bank’s own reported ROE, which some compute on a period-end or slightly different equity base. “Cost of risk” is net loan impairment as a share of average loans to customers, in basis points; a negative value denotes a net release of provisions (a gain). Post-2021 average ROE: Íslandsbanki ~11.3%, Arion ~13.9%, Landsbankinn ~10.3%. None of the three reported a loss in any year from 2011 to 2025. The banks would reasonably attribute part of these returns to the high interest-rate environment rather than to market structure; the response is that genuine competition tends to compress margins over time regardless of the rate level, and net interest margins here were stable rather than compressed. Loan impairments in some recent years relate to the Grindavík eruptions rather than to mortgage defaults, and several banks waived indexation or interest for affected borrowers.
• House prices, wages, and mortgage credit (Figure: “Everything rose together”). House price index: Eurostat / Statistics Iceland residential property price index (2015 = 100), quarterly, averaged to annual. Wages: Statistics Iceland wage index (launavísitala), monthly, averaged to annual and rebased to 2015 = 100. Household mortgage debt: Central Bank of Iceland, loans to households, year-end stock, rebased to 2015 = 100. Change 2015 to 2025: house prices approximately +162%, mortgage debt approximately +111%, wages approximately +109%.
• Demand-side context. Iceland receives well over two million foreign visitors a year (Icelandic Tourist Board; Statistics Iceland). Immigrants made up roughly a quarter of the employed workforce in 2025 (about 24.6%, up from about 12% in 2015), per Statistics Iceland labour-force register data. The short-term-rental and housing-supply effects are described qualitatively; this piece does not attempt to quantify their separate contribution to house prices, which is contested and beyond its scope. The claim is not that indexation is the sole driver of house prices — supply constraints and external demand matter materially — but that indexation systematically tilts the market toward higher prices and more leverage on top of those forces.
• Caveat. The risk-asymmetry argument describes the standard structure of indexed variable-rate lending; individual contracts vary, and lenders do bear credit risk and, as the October 2025 ruling demonstrated, legal and conduct risk. The sequencing and institutional design proposed here are directional; the detailed transition — pace of stock conversion, covered-bond market design, and treatment of the pension sector’s role — would require careful technical work beyond the scope of this piece. Figures should be checked against the latest Central Bank of Iceland and Statistics Iceland data at the time of reading.