Other Studies looking at similar relationship:
(could be others)
The relationship between the indirect and direct real estate markets has attracted considerable research interest in recent years, particularly in the United States (e.g., Giliberto, 1990; Gyourko and Keim, 1992; Myer and Webb, 1993; and Barkham and Geltner, 1995) and the United Kingdom (e.g., Barkham and Geltner, 1995), with lags of up to one year evident in the U.K. and up to two years evident in the U.S. Similar studies concerning the major Asian markets of Hong Kong (e.g., Newell and Chau, 1996) and Singapore (e.g., Ong, 1994, 1995; and Liow, 2001) indicate significantly shorter lags of one quarter, reflecting their unique and dynamic real estate market characteristics, particularly for Hong Kong (Newell and Chau, 1996; Chau, MacGregor and Schwann, 2001; and Newell, Chau and Wong, 2004).
Aims and Objectives of the study:
To analyse the long-term relationship between direct and indirect real estate with a focus on the UK office market.
• Investigate the direct real estate market with a focus on the UK office market – Evaluate the market structure and understand the valuation and pricing of the asset.
• Investigate the indirect real estate market with a focus on the UK office market – Analyse the securitisation process with a focus on REIT’s, understand the pricing of the asset.
• Use a Vector Error Correction Model to analyse the long-term relationship between direct and indirect real estate market. (Using data on the UK office market)
• Draw conclusions on the relationship between direct and indirect UK office real estate market, discussing if cointegration exists.
Possible sub-headings taken from another study – these can be changed and added to
Direct real estate investments
Direct real estate investments from an investor perspective
Indirect real estate investments
The securitisation process
Mortgage backed securities
Equity based real estate securities
Real estate companies and funds
Real estate investment trusts -REIT’s (Background)
Indirect real estate investments from an investor perspective
Efficiency of the real estate asset market
Real estate return and risk
Valuation of property assets (DCF)
The issue with differences in asset valuation.
(About valuation of direct and indirect, (Arbitrage)
Differences in expected cash flows
Differences in the opportunity cost of capital
Impact of information efficiency
Linkages between real estate investments
Contemporaneous connections between real estate investments and other assets
Price creation and linkages within dual markets
Portfolio considerations (real estate for diversification)
Methodology to be employed
The researcher has chosen to use a multivariate time series regression model to investigate the relationship between the two time-series. The model to be used also highlights if a cointegrating relationship exists between the two time-series and if both variables adjust to a long-term equilibrium path. Firstly, we explain the basic vector auto regressive model and then how cointegration is tested for.
Vector Auto Regressive Model (VAR)
A vector auto regressive model (VAR) is a variation of an autoregressive model, it relies on two or more variables each of whose current values depend on different combinations of the previous K values of all variables and their error term (Brooks, 2014).
This model assumes the variables are stationary, however it is possible for two non-stationary variables to produce a linear combination that is stationary. This is defined as cointegration and it suggests two variables cannot move independently from one another and that they share a long-term equilibrium, much of the literature on this topic suggest cointegration exists between the two variables if this is found to be true instead of using a vector auto regressive model a vector error correction model (VECM) is required (Enders, 2004).
Cointegration and Vector Error Correction Model
As previously stated cointegration suggests that two variables share a long-term equilibrium path. There are two main methods to test for cointegration the Engle-Granger method and the Johansen method. This study will use the Johansen method, mainly due to the problem highlighted by Brooks and Tsolacos (2010) that it is not possible to perform any hypothesis tests about the actual cointegration relationship if using the Engle-Granger method.