Smart Beta for Global REITs. An initial study.

1. SMART BETA STRATEGIES FOR REIT MUTUAL FUNDS
Alex Moss, Consilia Capital
Kieran Farrelly, The Townsend Group
American Real Estate Society Conference
Fort Myers , Florida
April 2015

2. CONTENTS
• Background
• Purpose of the Study
• Data
• Methodology
• Results
• Conclusions and Next Steps
2

3. BACKGROUND – KEY POINTS
• Post GFC, there has been a change in emphasis on the factors which influence
investment decisions, affect performance, and determine asset allocation mixes,
and product design, which are particularly relevant for real estate. Namely;
• Focus on income based assets in a low interest rate environment (real
estate)
• Increased emphasis placed on liquidity (REITs*)
• Interest in combining asset types for specific solutions ( listed/unlisted for DC
schemes)
• Emphasis on diversifying away equity and bond market risk (low correlation
“alternative” buckets)
• Greater use of maximum drawdown as a key risk measure (DC funds )
• Growing acceptance of certain Smart Beta strategies (active management at
passive cost)
*For simplicity the term REITs is used to describe all global listed real estate markets in this paper.
3

4. BACKGROUND –KEY QUESTIONS
• Against this background, what Smart Beta strategies can be developed to provide the
investment solutions and risk/return profiles currently required by asset allocators?
• Is it possible to devise automated trading strategies (with a low turnover) which will
enhance performance?
• Are there likely to be more Smart Beta products for REITs ? Currently we are aware of
the Kempen Fundamental Index strategy and the Dow Jones Townsend Core REIT
Index.
4

5. BACKGROUND - DOW JONES TOWNSEND CORE US REIT INDEX
5
• This Index aims to measure the performance of a basket of securities that could serve
as a public-market analog for privately-held institutional “core” real estate investments
• To be eligible , a company must be both an equity owner and operator of commercial
and/or residential real estate and meet minimum requirements for size and liquidity
• Specifically excluded are companies invested in the following property types: Factory
Outlets, Hotels, Manufactured Homes, Mixed Industrial/Office and Suburban Office
26.0%
14.6%
18.0%
9.9%
22.7%
15.4%
17.7%
9.3%
1-year 3-year 5-year 10-year
Annualized Total Return - Gross of fees as of end of February 2015
Dow Jones Townsend Core U.S. REIT Index MSCI US REIT Index Total Return

6. PURPOSE OF THE STUDY
In this study we are interested in discovering whether the free float market
capitalisation weighted global benchmark would have consistently underperformed a
Smart Beta strategy utilising the following factors :
1) Gross Assets
2) Equal Weighting
3) Gearing - Loan to Value ( Low and High) - EW
4) Valuation - Price to Book Value (Low and High) - EW
5) Size – Gross Assets (Small and Large) - EW
6

7. CAVEATS
• No transaction costs are taken into account
• Portfolios are only rebalanced at calendar year ends and then held for the next 12
month period
• No constraints such as minimum liquidity , maximum number of portfolio constituents
etc have been applied
• No account has been taken of resultant regional weightings
7

8. DATA
• EPRA Global Developed Index constituents
• COMPUSTAT for fundamental data
• Bloomberg and CRSP for share price and total returns data
• Frequency: Annual
• Currency: US$ (Unhedged)
• Return: Total Return
• Period: 2004-2014
8

9. METHODOLOGY
• Key metrics used: Loan to Value (LTV) , Gross Assets (GA), Price to Book Value
(PBV)
• Establish benchmark constituents on an annual basis – this is the initial selection
criteria
• Determine annual returns for all constituents
• Determine fundamental data (LTV, GA, PBV) for all benchmark constituents
• Sort by quartile (if appropriate)
• Apply weighting criteria (EW, Gross Assets)
• Calibrate portfolio annual return
9

10. RESULTS – GROSS ASSETS
2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014
FTSE EPRA/NAREIT Developed Index TR 32.13% 15.25% 41.93% -7.20% -47.39% 37.55% 20.08% -5.65% 28.99% 4.31% 15.72%
Total Asset Weighted 38.04% 17.21% 43.70% -9.12% -47.61% 80.66% 24.02% -9.37% 30.94% 13.56% 16.97%
Smart Beta Relative 5.91% 1.96% 1.77% -1.92% -0.22% 43.11% 3.95% -3.72% 1.95% 9.26% 1.25%
10

11. RESULTS – EW
2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014
FTSE EPRA/NAREIT Developed Index TR 32.13% 15.25% 41.93% -7.20% -47.39% 37.55% 20.08% -5.65% 28.99% 4.31% 15.72%
Equal Weight 39.20% 17.27% 44.24% -8.49% -44.10% 60.57% 27.47% -7.90% 30.02% 11.91% 16.40%
Smart Beta Relative 7.08% 2.02% 2.31% -1.30% 3.29% 23.02% 7.39% -2.24% 1.03% 7.60% 0.68%
11

12. RESULTS – GEARING
2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014
FTSE EPRA/NAREIT Developed Index TR 32.13% 15.25% 41.93% -7.20% -47.39% 37.55% 20.08% -5.65% 28.99% 4.31% 15.72%
Equal Weight Low LTV Quartile 40.56% 26.59% 47.75% -7.01% -47.59% 83.98% 13.12% -13.62% 28.07% 11.30% 11.07%
Smart Beta Relative 8.43% 11.34% 5.82% 0.19% -0.20% 46.43% -6.96% -7.97% -0.91% 6.99% -4.65%
2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014
FTSE EPRA/NAREIT Developed Index TR 32.13% 15.25% 41.93% -7.20% -47.39% 37.55% 20.08% -5.65% 28.99% 4.31% 15.72%
Equal Weight High LTV Quartile 37.72% 16.70% 46.62% -17.48% -42.07% 56.59% 42.95% -2.82% 30.12% 7.40% 23.09%
Smart Beta Relative 5.59% 1.45% 4.69% -10.28% 5.32% 19.05% 22.88% 2.83% 1.13% 3.10% 7.37%
12

13. RESULTS- VALUATION
2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014
FTSE EPRA/NAREIT Developed Index TR 32.13% 15.25% 41.93% -7.20% -47.39% 37.55% 20.08% -5.65% 28.99% 4.31% 15.72%
Equal Weight Low BTM Quartile 35.21% 20.72% 39.48% -14.59% -34.29% 28.39% 29.85% -0.85% 25.05% 5.31% 25.94%
Smart Beta Relative 3.08% 5.47% -2.46% -7.39% 13.10% -9.15% 9.78% 4.80% -3.93% 1.01% 10.22%
2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014
FTSE EPRA/NAREIT Developed Index TR 32.13% 15.25% 41.93% -7.20% -47.39% 37.55% 20.08% -5.65% 28.99% 4.31% 15.72%
Equal Weight High BTM Quartile 44.41% 14.46% 46.29% -2.85% -46.06% 117.51% 37.63% -7.41% 28.06% 16.45% 9.91%
Smart Beta Relative 12.28% -0.79% 4.36% 4.35% 1.32% 79.96% 17.56% -1.76% -0.93% 12.14% -5.81%
13

14. RESULTS – GROSS SIZE
2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014
FTSE EPRA/NAREIT Developed Index TR 32.13% 15.25% 41.93% -7.20% -47.39% 37.55% 20.08% -5.65% 28.99% 4.31% 15.72%
Equal Weight Low Total Assets Quartile 37.97% 26.16% 42.30% -3.66% -40.23% 50.27% 23.95% -8.17% 30.95% 10.17% 16.85%
Smart Beta Relative 5.84% 10.91% 0.37% 3.54% 7.16% 12.73% 3.88% -2.52% 1.96% 5.86% 1.13%
2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014
FTSE EPRA/NAREIT Developed Index TR 32.13% 15.25% 41.93% -7.20% -47.39% 37.55% 20.08% -5.65% 28.99% 4.31% 15.72%
Equal Weight Hightotal Assets Quartile 38.39% 16.62% 40.68% -13.84% -49.77% 76.34% 22.49% -7.80% 28.20% 12.73% 17.16%
Smart Beta Relative 6.27% 1.37% -1.26% -6.65% -2.38% 38.79% 2.41% -2.15% -0.79% 8.42% 1.44%
14

15. RESULTS – OVER TOTAL PERIOD
Mean Geo Mean
FTSE EPRA/NAREIT Developed Index TR 12.34% 8.92%
Equal Weight 16.96% 13.03%
Total Asset Weighted 18.09% 13.20%
Equal Weight Low LTV Quartile 17.66% 12.48%
Equal Weight High LTV Quartile 18.08% 13.97%
Equal Weight Low BTM Quartile 14.57% 12.07%
Equal Weight High BTM Quartile 23.49% 17.25%
Equal Weight Low Total Assets Quartile 16.96% 13.72%
Equal Weight Hightotal Assets Quartile 16.47% 11.50%
15

16. CONCLUSIONS AND NEXT STEPS
• Promising initial results
• Simple Smart Beta strategies can create material performance differentials vs the
index
• Next Steps:
• Make use of higher frequency and longer time series data
• Explore additional strategies – fundamental and technical
• Incorporate additional filters such as liquidity and regional constraints
• Include transaction costs – measure ‘real’ investor level returns
• Assess regional level strategies
• Factor model, risk and diversification potential (within real estate and multi-asset
levels) analysis
• Explore whether there is a cyclical dimension to the various strategies which is
predictable
16