An empirical analysis of the Norwegian housing market : what drives the house price?
Abstract
According to OECD (2018), Norwegian real house prices increased 44 percent from the end of
2008 until 2017, compared to an OECD average of 8 percent. Meanwhile, there has been a
substantial change in several important aspects of the Norwegian households’ financial base.
In order to understand the house price development this thesis develops a house price model
based on fundamental drivers of the housing market. Including debt in the model improves the
in-sample fit, especially around the Financial Crisis. However, the model with debt is not
robust. As a result the households’ debt is not a part of the preferred model. In the out-of-sample
period the real interest rate was close to zero, which made credit cheaper and increased
expectations of future house price growth. As the preferred model emphasize the interest rate,
it is able to accurately forecast the development of house prices in the out-of-sample period
from 2015 to 2017.
In contrast to earlier studies, such as Jacobsen and Naug (2004) and Anundsen and Jansen
(2013), the analysis finds that both the growth of house prices one year ago and one quarter ago
impact the house price development. The importance of backdated growth implies that people
observe the historic house price when they shape expectations to the future house price. This
mechanism can cause a self-reinforcing house price spiral, which can lead to severe house price
bubbles. Especially the effect of the growth in the preceding quarter is worrying, as it implies
that people react faster to changes in house price growth, than what was found in the earlier
studies of the Norwegian housing market.
The main drivers of the Norwegian house price are found to be the households’ real income,
the real interest rate, the housing supply, backdated growth in house prices and the households’
expectations. According to the model, the real house prices in Norway were above their
fundamental value from mid-2016 until the end of the sample in 2017.