A vast amount of literature on hedonic price modelling has been formulated on overseas
property markets. Very little currently exists in South Africa and this poses a risk for
sellers and estate agents of a residential property when listing it on the open market, as
this could result in an extended list period, reducing the original asking price. This paper
seeks to examine Gauteng’s West Rand residential property market and formulate a
multi-variate regression model to best predict property prices, determined by a property’s
structural characteristic. The research tracks residential sales from 1996 to 2009, a
thirteen-year sample period from which a composite property index, to account for
inflation and real house price growth, has been formalised. Correlation and regression
analysis was used to interpret the data at the relevant significance level. In order to
account for locational attributes present in property values, the data set was divided into
locational quadrants and run as dummy variables. A further regression was run on a
screened data set to create an ordinary least squares equation that could be used to show
the relationship between property values and structural characteristics. The results
indicated a good fit with an R2 of 69.5%. This regression was then applied practically to
predict property prices for houses that have transacted in the West Rand property market,
and plotted along a value/price graph using the 45-degree true value frontier line. The
relevant results were then interpreted, and recommendations given.
| Identifer | oai:union.ndltd.org:netd.ac.za/oai:union.ndltd.org:wits/oai:wiredspace.wits.ac.za:10539/9640 |
| Date | 05 May 2011 |
| Creators | Dodds, Robert Scott |
| Source Sets | South African National ETD Portal |
| Language | English |
| Detected Language | English |
| Type | Thesis |
| Format | application/pdf, application/pdf, application/pdf |
Page generated in 0.0019 seconds