Factors that impact on South African REITs trading at a discount or premium to their NAV
Master Thesis
2018
Permanent link to this Item
Authors
Supervisors
Journal Title
Link to Journal
Journal ISSN
Volume Title
Publisher
Publisher
University of Cape Town
License
Series
Abstract
This research investigates the factors that impact on South African Real Estate Investment Trusts (REITs) listed on the JSE trading at a discount or premium to their net asset value (NAV), known as the NAV spread. Globally, REITs have become an increasingly popular means for investing in real estate representing 41% of the global listed property industry and reaching a total market capitalisation of US$ 1.3 trillion as at June 2017 (EPRA, 2017). In South Africa, as at November 2017, there were 31 SA REITs listed on the JSE, representing a market capitalisation of over R422 billion (PropertyWheel, 2017). There are broadly two approaches to understanding and explaining the NAV spread, namely, the rational approach and the irrational approach (Morri et al., 2005; Mueller and Pfnuer, 2013). Factors under each approach were identified and discussed in the literature review. A quantitative research methodology was adopted in which 12 REITs were investigated between 1 July 2012 and 30 June 2017. A regression analysis using a linear mixed effects model was carried out in order to test whether the factors identified in the literature impacted on the NAV spread of SA REITs. Five rational and four irrational factors were regressed against two versions of NAV discount, namely, using a Traditional NAV Discount formula and an Unlevered NAV Discount formula which accounts for the effect of debt on the discount. Model diagnostics were carried out on two regressions and an outlying observation was found for each NAV Discount formula and hence each model was run twice, namely, the regression was run with the Traditional NAV Discount formula using the full set of observations (Model 1) and with one outlying observation removed (Model 2). Thereafter the regression was run with the Unlevered NAV Discount formula using the full set of observations (Model 3) and with one outlying observation removed (Model 4). The results of the regression analysis found only one factor, Dividend Yield (DY), to have a statistically significant impact on REITs trading at a discount or premium to their NAV in all four models employed. Sector Average Discount (SAD) was found to be statistically significant in Models 2, 3 and 4. Size (SIZE) was found to be statistically significant in Models 1 and 2. Return on Equity (ROE) was found to be statistically significant in Models 3 and 4. Consumer Confidence Index (CCI) was found to be statistically significant in Model 4. The SA Property Index (SAPI) was found to be marginally statistically significant in Model 1. Based on the model diagnostics Models 2 and 4 showed a better fit with the regression model and hence were the preferred models. The statistically significant variables in Model 2 were DY, SIZE and SAD; and in Model 4 were DY, SAD, ROE and CCI. This suggests that both rational factors (DY, ROE and SIZE) and irrational factors (SAD, CCI and SAPI) impact on South African REITs trading at a discount or premium to their NAV, however, the factors are dependent on whether debt is accounted for or not in the NAV Discount Formula. The results also suggest that the factors are partly influenced by the outlying observations, as was the case for SAD, CCI and SAPI.
Description
Keywords
Reference:
Gaszynska, D. 2018. Factors that impact on South African REITs trading at a discount or premium to their NAV. University of Cape Town.