Have you thought about buying a flat recently, only to discard the idea after a sneak peak at the prices? You are certainly not alone: my recent deep dive into the data shows that prices on the Danish real estate market have indeed skyrocketed over the last three decades, while income levels have not managed to keep up with that growth. In this article, I present the main findings of my study, and touch briefly upon the way I arrived at my conclusions.

There, you will find aggregate data on sales price (measured per m²) and two metrics of affordability, which are both derived by relating sales price to personal income. The latest data in the app is from 2024 and I intend to update the app once a year, as soon as new annual data becomes available.

Introduction

In a recent publication, the European Parliament reported that rising housing prices and rents were a major concern for many Europeans. In fact, EU citizens were so disgruntled by the rising cost of living that it was the primary driver behind their decision to vote in the 2024 EU parliamentary elections. This is not surprising, seeing how at the EU level, housing has seen an average increase of 48% in the period between 2015-2023. Meanwhile, GDP per capita has only increased by 34% in the same period, highlighting a growing gap between income and housing prices. And that is without even considering the distinction between GDP per capita and actual disposable income at the individual level.

To see how Denmark has fared relative to the EU as a whole, I took a deep dive not only into sales prices of flats but also in how they related to personal income. This allowed me to assess how housing affordability has changed over time.

Flats in Denmark cost 4.6 times more than they did in 1992; in Copenhagen, the increase is as high as 8.5

In little more than 30 years, the average sales price of flats in Denmark has increased a staggering 4.6 times (measured per m²). Naturally, this development is reflected differently across the country, with some municipalities showing higher or lower rates of change over the period. As Figure 1 shows, between 1992-2024, flat prices have increased the most in Frederiksberg, Copenhagen, and curiously, Vejen:

Figure 1. The 15 municipalities with the highest price increase between 1992-2024

The lowest price hikes were recorded in Brønderslev, Faxe and Randers; however, there are no municipalities where prices have decreased relative to their 1992 level.

Looking at the most recent data, Figure 2 offers little surprise: in 2024, flats were most expensive to buy in Denmark’s biggest cities as well as their surrounding municipalities:

Figure 2. The 15 most expensive municipalities for buying a flat in 2024

Buying in provincial areas was cheaper, but then again, income levels there were also lower, which brings us to the next topic: that of housing affordability.

Buying a flat in Denmark has become 33% less affordable over the last three decades

During the past three decades, the average sales price of flats has increased 4.6 times (measured per m²). Meanwhile, the national average disposable income has only increased 3 times, growing at a rate more or less identical to that of GDP:

Figure 3. Indexed sales price and national average disposable income + national GDP

These diverging trends have naturally led to lower overall housing affordability. Specifically, while in 1992, the average national disposable income could buy you 18.6 m², in 2024, that number had decreased to just short of 12.5 m², which corresponds to a 33% decline in housing affordability:

Figure 4. Number of m² that can be bought with annual disposable income (national average)

The national trend is observed across most municipalities and certainly among Denmark’s four largest cities, as the chart below indicates:

Figure 5. Number of m² that can be bought with average annual disposable income in Denmark’s four largest cities

Even though housing in Denmark has generally become less affordable (that is the case in 81 out of 91 municipalities, or 89%), there are a few municipalities which have followed the opposite trend. As the chart below indicates, that is precisely the case for Brønderslev, Randers and Faxe:

Figure 6. Number of m² that can be bought with average annual disposable income in municipalities that do not follow the general trend

Still, the conclusion is clear: housing in Denmark has become way less affordable over the past 33 years. Whereas 2.7 annual after-tax salaries were enough to buy a 50 m² flat in 1992, today you would need 4 annual after-tax salaries to do the same.

Larger municipalities have seen stable price increases, while smaller municipalities have had much more volatile prices

Denmark’s four largest cities, represented by the five municipalities shown on the chart below, have been and continue to be the country’s most expensive areas to buy in. Frederiksberg municipality is by far the most expensive in the whole of Denmark, with the gap between it and the second most expensive municipality Copenhagen increasing over time, particularly in the period following the global financial crisis of 2008:

Figure 7. Average price per m² for Denmark’s four largest cities over time

Looking at the data, a clear trend of increase is seen across most municipalities, with a slight drop in the immediate aftermath of the financial crisis of 2008, and then again in 2023, a year marked by higher interest rates than what people had become accustomed to in the preceding decade. These are but flukes in an otherwise consistent development, and prices in 2024 have already picked up from their temporary slump.

The case is not as clear-cut when it comes to the cheapest municipalities to buy flats in, which includes places such as Holstebro, Randers and Slagelse. On the lower end of the price spectrum, the trend is a lot less pronounced and prices are less stable, with much larger year on year changes, as Figure 6 shows. Therefore, while the price in Copenhagen dropped by 1.4% between 2021-2023, the price in Favrskov decreased by 21.4%, while in Holstebro the drop was likely as high as 41.3%.

In summary, we see that between 1992-2024, flat prices have increased substantially in most municipalities, with the price increase being stronger in places where it was already more expensive to buy. Cheaper municipalities, on the other hand, seem to have experienced much more price volatility during the last three decades, though some of that volatility may be due to having fewer sales compared to larger municipalities, which can therefore more easily skew the average price for the area.

It’s unlikely that flats will get cheaper in the upcoming five years

Assuming historical trends continue into the future, it is unlikely that there will be any significant decrease in housing prices over the course of the next 5 years. This is the conclusion from a series of forecasting models built at the municipality level, where housing prices were modelled as a function of different macroeconomic factors. In practice, this means that municipalities where prices have been historically increasing will continue to be on the rise, while areas where prices have been declining will continue to do so, as the chart below illustrates:

Figure 8. Historical data and predictions for average price per m² for select Danish municipalities

In the models, average sales price for each municipality was specified as a function of Denmark’s national GDP, annual inflation and the annual median interest rate as reported by Denmark’s national bank. Doing the modelling separately for each municipality meant that the local price differences were preserved, even though predictions were generated for many different areas. The future values of GDP and inflation were sourced from the IMF’s World Economic Outlook (WEO, last updated in April 2025), while for interest rate, the all-time historical average was used.

Temporary price drops can be expected if interest rates go up or in the event of an economic recession (which will result in a lower GDP). But unless we are dealing with some sort of black swan event, any future price decreases will likely not last more than a couple of years. If you are waiting for “the bubble to burst”, you may get lucky, but you would also have to get lucky with being spared the potential negative effects of the economic downturn that might trigger falling flat prices.

Concluding remarks: Less affordable housing in the context of growing inequality

In a recent article, Danish public broadcaster DR discussed the growing issue of overpriced housing, with an emphasis on how the dominance of private companies on the rental market has made housing less affordable. In that article, we find politicians from Denmark’s two largest cities (Copenhagen and Aarhus) calling for the construction of more flats for sale (ejerlejligheder) as a way of addressing the issue of unaffordable housing. However, my analysis points out that flat sales prices have already increased too much relative to the general income level, which does not exactly make them an affordable option for many.

But what is affordable housing and why does it matter? The United Nations (UN) classifies adequate housing as a fundamental human right, and defines housing as inadequate if its cost “threatens or compromises the occupants’ enjoyment of other human rights”. There is a general consensus that housing costs must not exceed 30-40% of a household’s disposable income in order for the housing to qualify as affordable. With housing prices increasing at a much faster level than disposable income, things in Denmark are not exactly going in the right direction.

The hard reality is that, thanks to an unfortunate mixture of macroeconomic factors and government policies, housing prices in Denmark have increased at a much faster pace than personal income, leaving more and more people out of the market.

According to Eurostat, only about 61% of people residing in Denmark owned their home in 2024, which is 5 percentage points down compared to two decades ago. The data used in this study shows that housing affordability in the same period (2003-2024) has decreased by 6.5% at the national level. Meanwhile, income inequality in Denmark has increased by 27%, as measured through the Gini coefficient (where the latest available data is for 2023).

In the context of growing inequality and progressively less affordable housing, it may be difficult to stay optimistic. Indeed, for the average person, navigating the widening gap between personal income and housing costs will likely continue to be a challenge in the upcoming years. Prospective buyers will be hoping for “the bubble to burst”, while prospective sellers will naturally hope for further increases in housing prices.

Meanwhile, for policy makers, the current situation also represents a challenge: faced with multiple crises at the same time, will they have the capacity to do anything about the rising cost of living, of which housing constitutes a significant part? It remains to be seen whether politicians will be able to balance the interests of private investors with protecting the average citizen’s right to adequate housing, or whether we will see a rise in populism due to parts of the population feeling left behind.

Appendix: Data and method

Data sources

Data on sales prices was sourced from Finans Danmark and captured the development in the average price of “owned flats” (called ejerlejligheder in Danish) in the period 1992-onward. This data was further supplied with population and income data from Danmarks Statistik (DST), as well as with macroeconomic data from both DST and the IMF. For a comprehensive list of the input data, please consult this GitHub page.

Please note that both housing prices and disposable income are nominal and represent the actual value measured in each year, which allows for direct comparisons between the different metrics.

Why I focused on flat prices only

I did not look at the prices of houses (huse and rækkehuse), rental flats (lejeboliger) and partly-owned flats (andelsboliger) due to the more limited availability of such data, though I expect them to follow a similar trend to that of flats. In addition, I focused on the prices of flats rather than houses in order to get better transferability of the conclusions to the private rental market, which is dominated by flats, at least in urban areas.

Augmenting the data quality of flat prices

In the source, average housing prices were available for each post code on a quarterly basis, with some missing values in cases where too few sales were recorded. For the purposes of this analysis, the input data was aggregated into annual data at the municipality level so that it could be linked to important background data (such as local disposable income data and national macroeconomic data).

As a part of the data preparation, imputation was used to ensure that all municipalities have a complete historical record of housing prices. The imputation was carried out using a random forest algorithm, which took factors such as the average national sales price, number of sales and local income into account, and had an accuracy of about 67%. In the accompanying app, historical prices are shown including the imputation (called “estimates” in the app), but you can filter those out if you only wish to see actual prices.

Measuring housing affordability

Housing affordability was calculated by relating sales price to local disposable income in each municipality, resulting in the following two metrics, which are presented on the Affordability by municipality and Affordability by gender pages in the app:

• Number of m² that the average person can buy with their annual disposable income
• Number of years’ worth of annual disposable income needed to buy a 50 m² flat

Those separate measures were created for better interpretability and are perfectly correlated with each other.

Please note that due to the way the data is aggregated from post codes to municipalities and due to the imputation algorithm, some rounding off may slightly impact the final numbers (e.g. the municipal or national averages) from what may be reported in other sources.

Predicting future flat prices

As mentioned in the article, I also built predictive models for housing prices for all municipalities included in the data. The models were simple OLS-based regressions, each one fitted on a subsample of the data (one municipality at a time). The models do not differ in their specification, and future flat prices are always modelled as a function of the same three input variables: national GDP, inflation rate and median interest rate.

The accuracy and goodness-of-fit of all models are evaluated through the mean absolute percentage error (MAPE) and R² scores. The first one is indirectly presented as “accuracy” in the app, where accuracy is defined as 1 - MAPE. The forecasts are generated one municipality at a time and then concatenated in a single dataset in an automated manner. The app does not show forecasted prices by default, but they can be enabled via the Selected price type(s) filter. As a general rule of thumb, flat prices for the current year and all future years are based on forecasts rather than historical data.

📊🦉 Explore the free interactive data app that serves as a companion to this article here.

🔎📈 Explore the R and Python code used to prepare the data and create the app on the project’s GitHub page.

---

Cover image source: Own production