Demystifying the Swedish Housing Market
I understand you… Just like most of those living in Sweden, I had to study (a.k.a suffer from) the housing market. My love-hate relationship with real estate started with me renting a student room in Gothenburg and being forced to move three times within 9 months. A year later, I moved to Stockholm and again struggled to find a place to rent. Eventually, my boss knew someone, who knew someone, who saved me from the streets. Three years later I made the big decision to buy my own flat. Fast-forward four years, I realized that my real estate-investing bravado only led me to loose money when i sold the flat. Following the lessons learned, I have decided to become a savvier investor and better research the market in order to spot the right opportunity.
I realize I am not alone here. Housing is one of the hottest topics in Sweden, specially in Stockholm, where property prices have been rising roughly 15% year over year. In other words, even the uninitiated can understand that purchasing your own property makes more sense then renting (also, mortgage rates have been super low in the past 10 years). Furthermore, if you really need to rent (say you lack the minimum deposit requirements to get a mortgage loan), you will have to battle the second-hand rental market to find a place to live.
In this post I am proposing a solution which I hope will help ease the pain in understanding the housing market and devising a property purchase strategy.
Välkomen till Sverige and good luck finding an investable property
Here is the drill: Potential property buyers only have 2 opportunities to see the property. Typically one on Saturday and another on Sunday, each visit lasting around 30 min. Can you feel the anticipation building up? Imagine how many people will be visiting the property together with you. Fear of missing out feelings take over as you realize you may miss your dream-home to any of those tens (or hundreds) of other couples investigating (measuring and photographing) your next crib.
If you are like me, you not only want to find a good place to live, but you also want to maximize the chances of making a good investment. Thus, you may have additional questions, such as:
- When is the best time to buy?
- What is the best location?
- What defines an investment grade property?
People say that one should spend 1 to 2 years researching the Real Estate Market before devising a purchase strategy. I tend to agree with this statement, provided there is limited historical data available and no good tools to help one make such important decision in a short amount of time. Today most buyers use two websites to search for properties: Hemnet and Booli. Despite these websites allow historical analysis of sale prices, the information is only presented in list-format (spread across several pages), not offering any easy way to download the underlying data in bulk to perform a more comprehensive analysis. As a consequence, following one’s gut becomes the default route.
Introducing the Booli Web Data Connector for Tableau
To address this gap I’ve decided to build my first Tableau Web Data Connector. This will allow anyone to download data not only for properties currently on sale, but most importantly historical sales data, enabling users to easily and quickly demystify the market and get answers to the questions that matter to them. This was only possible thanks to Booli’s API, allowing a user to download housing data when integrated to third party applications.
I’ve fetched some data with the connector and created a few vizes that quickly helped me better understand the market as a whole within a few hours, rather than months. I’ve focused on central Stockholm area, as this is where I wanted my next crib to be located.
The dashboard above provides interesting first insights about the market, while displaying summary data of over 50000 properties sold in Stockholm between 2011 and early March 2016:
- In average, the properties are listed for sale for 3.1MSek and are eventually sold for 3.5MSek. The Margin of Negotiation is over 10%, i.e., people end up paying around 10% higher than the listing price. Actually, if we focus on Q1–2016, the margin is 15%! This is big red alert for me, indicating that either estate agents unwillingly underestimate the market value of the property, or they do it on purpose (thus providing a false sense of affordability, which will ultimately lead to a bidding race among buyers).
- It takes approximately 18 days for a property to sell (from when it’s listed until it is taken off the market). As discussed earlier in this article, considering each property is shown up to 2 times a week, lasting 30 minutes each time, 18 days indicate that sellers organize only two weekends of showings. For buyers this means that in theory they would only have the chance to spend a maximum of 2 hours in the property before making a purchase decision (in reality we know people see the property once or twice, and only for a few minutes before they start bidding).
- We can also see the monthly fees are around 3150 sek. Not bad, considering these are recurring costs which will never decrease over time. Outside central Stockholm, the fees are significantly higher, mainly due to unpaid debts and the fact that in central areas, buildings typically rent out commercial spaces to local businesses (helping reduce residents’ monthly running costs).
- Fastighetsbyrån, Svenskt Fastighetsförmedling and Notar dominate the real estate market in this region, with over 40% market share.
On that last point, I decided to create another viz comparing the evolution of these broker firms market shares over time. Svenskt Fastighetsförmedling and Notar are fighting for the second place year after year since 2014. Länsförsäkringar is expanding quickly since 2013, at the expense of Erik Olsson and Svenska Mäklarhuset.
Spotted some amazing vizes further down… Before you continue, how can I build my own vizes?
For anyone reading this article, here is how my first “quick and dirty”, but fully functional version of the connector looks like:
To use it (for free!), simply start up your copy of Tableau Desktop or Tableau Public, choose the “Web Data Connector” and paste this link (updated on 13/April/2018):
Type the city name (e.g. Stockholm, Nacka, Malmö), select between Live or Historical data and you are good to start deriving your own insights. If you are curious like me, you can dive deeper in the analysis. Let’s look into some further insights:
When should one buy a property in Stockholm?
I wanted to know if there is a better time to buy a property, so I came up with this viz that shows me the seasonality pattern for 50000 properties sold since 2011 in Stockholm:
I instantly fell in love this chart, as it clearly indicates that the supply of properties during the summer months (June, July, August) is much lower than the yearly average and the average sale prices are also lower. In other words, that’s shopping time!! And please, never EVER try to sell your house during Summer or in December, as you will have to settle for much less. I actually made this mistake myself. I sold my first flat in July and lost a lot of money. As my estate agent claimed: “Supply is lower during summer, so prices go up. That’s because, those buying homes during these months of low supply are typically desperate to find a place to live”. In reality, desperate are the ones selling during summer, not the buyers. My 2 cents: Do not trust others, trust your data ;)
What is an investment grade property?
From now on, I want to make sure any property I invest in Sweden will not devaluate (at least not as much as my first investment did :-). While having Fika with some friends, I asked if such thing as “investment grade” properties existed in Sweden. I’ve then learnt that the value of a property is directly related to its location, as well as being identified as “Sekelskifte” . Sekelskifte means “turn of the century” in Swedish (basically, anything constructed before between roughly 1850 and 1920). As I wanted to validate those claims of location+Sekelskifte driving prices with my own data, I firstly identified such properties on the map by colouring them using the following simple formula:
I didn’t want to live in a one-bedroom flat or pay too high-maintenance costs, so I needed to narrow the search down further with my personal requirements:
- 3 rooms
- at least 75 square meters
- low Avgift (i.e. Monthly maintenance fees < 2500 Sek)
- Maximum Price 6M Sek
This is what I got:
This chart was a good confirmation that in order to secure a good investment and satisfy my personal requirements, I should limit my search to a specific locations, such as Vasastan, Norrmalm and Östermalm. This came with the added benefit that I wouldn’t have to rush from one side of town to another to attend property showings.
Furthermore, I also wanted to confirm that people prefer older type of properties (Sekelskifte). Therefore, I’ve built this viz showing price trends vs. construction year (each bubble is a property sold, red ones are Sekelskifte):
Boom! Downward trend-line! My friends were right, the newer the property, the lower the price. I would definitely recommend investing on a “Sekelskifte”, rather than a “Funkis” (functional style, modern-type built after the 1920s). Here are examples of each:
Patience is a virtue
Say that you've followed my steps and have now devised your own purchase strategy. Next you will be trying to apply this strategy on the live property data that you have just collected with the Booli Web Data Connector. However, applying your strategy on your dashboard ends up filtering out every item from the view, so you are left with no potential properties to buy. What do you do now? Should you change your strategy?
I would say, if you can afford waiting, be patient and stick to your original strategy. Refresh the dataset every new week and see if a property that interests you shows up in your dashboard. Much patience, right? Yes, this is no different to stock trading and arguably the best traders will not trade based on their gut, but rather on facts.
With Tableau you can start being a little creative and adventurous with your strategy and try spotting unique opportunities. For example, I’ve decided to analyse the impact of having a property listed for too long on the market, i.e. those that didn’t sell quickly enough. My hypothesis was that these properties would sell for less, potentially even lower than the listing price (translating into a negative Margin of Negotiation). My hope was that this would even hold true for Sekelskifte properties.
Earlier in this article we saw that a property takes in average two weekends to sell (18 days). Here is how to compute how many days it takes in average to sell a house:
I combined this metric with the Margin of Negotiation to create a viz based on historical data that could potentially confirm we can really pay less than the listing sale price:
These two trend lines pointing down don’t lie. It doesn’t really matter if it’s a Funkis or a Sekelskifte, the longer it takes to sell, the higher the discounts one could achieve. There are even some Sekelskifte-type properties between 1 and 2 months to sell and offering 25% discount.
There could be a number of reasons why these properties took so long to sell. Perhaps they were in really bad shape, presenting water leaks and mold, located on noisy streets, etc. But maybe there wasn’t anything wrong those properties. Other things may have contributed to a non-sale, such as poor quality ads, building door got blocked and very few buyers attended the showing. Perhaps these are very large properties, say 200 square meters, but the it holds only 2 (enormous) rooms and 1 bathroom. One would expect at least 3 bedrooms and 2 bathrooms for a property of this size, so if you are into property reprogramming, these could be one-of-a-kind chances.
Great opportunities are available out there for budding property investors, as well as to regular families looking to buy their dream homes. And no, two years of market research are not necessary.
I would love to hear from you what other insights and visualizations you have created with the Booli Web Data Connector for Tableau. Get dirty with data!
Note: Interactive versions of the visualizations presented in this article are available in Tableau Public at:
Disclaimer: Important legal information
This article may include market analysis. All ideas, opinions, and/or forecasts, expressed or implied herein, information, charts or examples are for informational and educational purposes only and should not be construed as a recommendation to invest, trade, and/or speculate in the markets. Any investments, trades, and/or speculations made in light of the ideas, opinions, and/or forecasts, expressed or implied herein, are committed at your own risk, financial or otherwise.
