For the first time, data research has been used to chart the speculation in “hot land” in the Netherlands. Small pieces of agricultural land were sold, with the assumption to be allocated for building purposes. If that happens the value of the land will rise enormously. Last years 2300 hectares of agricultural land has been cut up in 27.000 small pieces. More than ten thousand people bought these pieces and invested an estimated € 700 million. But our research showed that not a single piece of “hot land” has been allocated for building. In fact, municipalities suffer from this fragmentation.
Besides media attention and parliamentary questions, the biggest impact of the publications are court cases. A lot of buyers of “hot land” think they are scammed by the traders who sold them the pieces of land. They sometimes paid €250 per m2 while the agricultural price is €8 per m2 or even less. Several sold pastures are not even fit for building purposes, since they are too close to highways or assigned as breeding grass fields for birds.
The need for trading licenses are part of these court cases. Previously, The Dutch Authority for Financial Markets (AFM) supervised the market for land speculation. Several traders were given fines since they did not have the right license. Nowadays, the AFM is not involved in supervising of land trading. Traders found a loophole in the trade laws and laws for the protection of retail investors. This research sheds light on the massive use and scope of this loophole that surprised several national experts on land ownership.
For researching land speculation we used the online database of the Dutch Land Registry (Kadaster) in combination with maps of Esri (ArcGIS program). Esri used algorithms and machine learning for detection of possible land speculation locations in The Netherlands.
While checking the ownership files of split up pastures, we discovered new traders in the Dutch company register (Kamer van Koophandel). Those traders offered new, unknown locations of land speculation.That’s how we build up our data set in Excel and in the end made a map of all locations. On this map you can find all 27000 individual pieces of land sold by traders to retail investors.
Because of the combination of our research and the maps of Esri, we have an almost 100% total coverage of all locations of land speculation in the Netherlands.
After this we surveyed all 355 municipalities in The Netherlands for comment on our results. We asked them if they were aware of the land speculation in their municipality and how this could effect their real estate policy. The response was very high (60+%).
What was the hardest part of this project?
The hardest part was the Dutch Land Registry. They couldn’t help us with our research, because it’s not part of their official regulatory tasks. So we had to build a database ourselves, this took some months.
After the first interviews with specialists and scientist, we discoverd no one has ever looked at the scale of speculation in the whole country. The research we did is unique.
The Netherlands are a small country and land for building purposes is extremely scarce. The scale of speculation with agricultural land as we found doesn’t solve the problem of the shortage of housing we are facing at this moment in the Netherlands. Because one hectare (10.000 m2) can be divided into 100 pieces (100 m2 each). A municipality has to negociate with 100+ different owners if they have plans for building on that location.
This means 10.000+ retail investors will probably never see any return of their investment. Since this land is hard to sell, even without profit, heirs will probably inherit these small pieces, which makes the number of land owners even larger.
What can others learn from this project?
We worked with three journalist on this project. One datajournalist, one journalist specialised in financial markets en one journalist specialised in real estate.
We were lucky a trainee from a high school could work with the maps of Esri, this could help us with our research and to visualize the results in the end.
If there’s no dataset available, you have to build one yourselves. In this way you get really into your subject.