The effects of property specific style-based variables on the returns of South African listed property equities

Abstract

Contrary to the Capital Asset Pricing Model (CAPM) and the Efficient Market Hypothesis (EMH), styles-based variables of value, size and momentum have been well documented to show strong associations with cross-sectional equity returns on the Johannesburg Stock Exchange (Hoffman, 2012; C. Muller & Ward, 2013) and on international stock exchanges (Carhart, 1997; Fama & French, 1992). Yet little research has exclusively focused on style-based variables and the JSE listed property sector. It is worth noting the total market value of property stocks or real estate investment trusts (REITs) has grown rapidly both in South African and Internationally. In 1995 the market capitalisation of the JSE listed property sector was R15 billion, but by November 2015 this had grown to a market capitalisation of R665 billion (iNET Bridge, 2015). This research study examined the effects of 22 style-based variables on all JSE listed property companies over the period December 1995 to December 2015. The style-based variables included property specific variables, such as geography location, sector allocation, and vacancy percentage. Since such data was not readily available from financial data libraries, a database was created directly from the financial statements reported by each company. The same methodology and graphical time series approach as C. Muller and Ward (2013) was used to evaluate the performance of the examined styles. The study found that portfolios constructed on the basis of ranked style-variables exhibited significant effects over the period. Most notably, both the size (market capitalisation) and value (earnings yield) effects showed significant excess returns and linear relationships, but persistence has not been evident after 2010. Other style-variables that exhibited effects over the period were: total geographic concentration, price to NAV, vacancy percentage of portfolio, value traded, value traded as a percentage of market capitalisation, interest cover ratio, loan to value and geographic international percentage. Furthermore, the study found no evidence of the momentum effect, which was found to be the best performing strategy by C. Muller and Ward (2013) for general equities.