Influence of population demographics on real estate prices in Zuid-Holland
Summary
Because the Dutch population is aging, a change in housing preferences can be expected.
According to the lifecycle theory older people have accumulated more wealth, and therefore should
be able to spend more on their housing. This thesis aims to model the real estate market between
2009 and 2016 in Zuid-Holland, comparing the results of linear regression to random forest
regression while trying to incorporate the development of local population change in people over 65.
Data from the Dutch national real estate broker association (NVM) is being used, enriched with
publicly available neighborhood statistics. Analyses have been performed for each year, for both
models. Results show the relations of structural, locational, and neighborhood variables on the
recorded transaction price per square meter. From these results, conclusions have been drawn on
the effectiveness of the linear regression in relation to the forest regression, as well as the
performance and impact of the inclusion of age dynamics into hedonic price modeling. It was shown
that the inclusion of age dynamics added a very slight value to the adjusted R-squared, however this
variable was not consistently significant. Furthermore, the random forest regression has shown to
consistently outperform the linear regression.