Sebastian Gl

1
The Influence Of Ownership,Sector Allocation And
Investment Style On Portfolio Volatility

• Sebastian Gläsner
• IPD Investment Property Databank GmbH, Wiesbaden

2
Research motivation

• The German property market described by the IPD
  Total Return index DIX showed low volatility over
  the last two decades
• Even in times of worldwide financial crisis in
  2008 and 2009, the Total Return remained
  (slightly) positive
• The underlying property appraisals have
  frequently been criticised for smoothing and
  lagging tendencies as a consequence of
  sustainable factors that enter the valuation
  process
• Although the polycentric German property market
  does not necessarily need to show as much
  volatility as for example the London-focused UK
  market, we want to deepen our understanding of
  the index performance

3
German properties of foreign investors

• As IPD offers specialised benchmarks for groups
  of clients, such as e.g. insurances, we compared
  the group foreign investors to the German
  investors
• Data refers to standing investments only (no
  transactions, no developments)
• Index calculation is capital weighted (IPD
  standard)

4
Investor performance Total return

• Sharpe outperformance of foreign investors from
  2003 to 2007, peak in 2006
• Strong underperformance in 2008 and 2009

5
Investor performance Income return

• Constantly higher income returns of foreign
  investors

6
Sector allocation

• Foreign investors are strongly over-allocated in
  sector industry (32.4 vs. 2.3)
• Both investor groups have a focus on office and
  retail investments

7
The German index by investor groups

• We assigned each portfolio to an investor type
• Asset manager
• Foreign investor
• Insurance
• Open-end fund (retail and institutional)
• As we apply statistical analysis, each property
  has the same weight, the index is not capital
  weighted like IPD standard
• We excluded properties with a CV of less than
  500.000 EUR from the analysis

8
DataInvestors in the German DIX
9
Properties by investor group
Total returns Total returns
InvType 2006 2007 2008 2009 analysis possible?
asset manager 5
foreign investor 39 51 78 144 gtgtgt analysis for 2006-2009
insurance 1 1 1 1 gtgtgt analysis for 2006-2009
open-end fund 17 15 13 17
industry 62 67 92 162  
asset manager 16 23 36 89
foreign investor 5 17 42 46 gtgtgt analysis for 2007, 2008 and 2009
insurance 626 536 428 442 gtgtgt analysis for 2007, 2008 and 2009
open-end fund 542 373 370 247
office 1,189 949 876 824  
asset manager 9 5 3
foreign investor 5 7 10 9 gtgtgt no analysis for sector "other"
insurance 128 71 62 65 gtgtgt no analysis for sector "other"
open-end fund 95 55 52 37
other 237 138 124 114  
asset manager 672 717
foreign investor 2 1 136 150 gtgtgt analysis for 2008 and 2009
insurance 559 459 427 414 gtgtgt analysis for 2008 and 2009
open-end fund 17 15 7 10
residential 578 1,147 570 1,291  
asset manager 1 1 1 19
foreign investor 105 140 234 298 gtgtgt analysis for 2006-2009
insurance 249 172 163 183 gtgtgt analysis for 2006-2009
open-end fund 230 160 142 99
retail 585 473 540 599  
total 2,651 2,774 2,202 2,990  
10
Investors in the German DIXMeans of Total Returns
mean TR  
sector InvType 2006 2007 2008 2009 analysis possible?
industry asset manager 5.7
industry foreign investor 9.4 7.2 -0.1 -3.3 gtgtgt analysis for 2006-2009
industry insurance -4.0 -2.5 -32.1 -40.0 gtgtgt analysis for 2006-2009
industry open-end fund -2.9 -0.2 5.9 4.3
total   5.5 5.4 0.4 -2.7  
office asset manager 7.2 5.4 4.9 1.3
office foreign investor -10.5 6.7 0.9 -5.2 gtgtgt analysis for 2007, 2008 and 2009
office insurance 1.4 4.7 4.1 3.0 gtgtgt analysis for 2007, 2008 and 2009
office open-end fund -1.3 1.9 3.2 3.6
total   0.2 3.7 3.6 2.5  
other asset manager 2.8 -4.1   3.9
other foreign investor 2.1 -0.8 -2.3 -9.8 gtgtgt no analysis for sector "other"
other insurance 4.0 4.1 2.0 0.7 gtgtgt no analysis for sector "other"
other open-end fund 2.3 4.4 5.1 2.5
total   3.2 3.6 2.9 0.6  
residential asset manager   7.3   6.6
residential foreign investor -15.8 -4.9 1.5 1.4 gtgtgt analysis for 2008 and 2009
residential insurance 7.1 5.8 4.8 4.9 gtgtgt analysis for 2008 and 2009
residential open-end fund 7.5 4.5 3.9 8.3
total   7.0 6.6 4.0 5.4  
retail asset manager -22.1 4.8 3.5 -0.3
retail foreign investor 8.8 8.4 0.1 1.5 gtgtgt analysis for 2006-2009
retail insurance 2.8 3.3 4.2 1.9 gtgtgt analysis for 2006-2009
retail open-end fund -0.2 4.2 5.1 3.5
total   2.7 5.1 2.7 1.9  
total   2.6 5.2 3.3 3.3  
 
  higher mean total return higher mean total return  
  lower mean total return lower mean total return  
Foreign investors achieved higher total returns
in 2006 and 2007 and lower total returns in
2008 and 2009 in all analyzed sectors
11
Investors in the German DIXStandard deviation of
Total Returns
sd TR
sector InvType 2006 2007 2008 2009
industry asset manager 6.8
industry foreign investor 16.1 9.1 6.6 10.6
industry insurance        
industry open-end fund 18.4 14.7 10.2 3.9
total   16.8 10.9 8.2 10.8
office asset manager 13.7 11.4 5.5 6.4
office foreign investor 18.4 7.4 7.4 12.2
office insurance 11.9 8.9 9.7 10.0
office open-end fund 10.6 6.7 5.8 4.9
total   11.5 8.3 8.0 8.8
other asset manager 12.9 27.8   6.2
other foreign investor 20.2 13.4 22.5 17.8
other insurance 11.7 7.1 10.4 12.0
other open-end fund 6.5 5.4 5.5 5.9
total   10.2 8.4 10.4 11.2
residential asset manager   16.2   11.2
residential foreign investor 19.2   7.2 9.3
residential insurance 11.6 7.5 8.5 9.0
residential open-end fund 11.7 3.1 1.5 4.9
total   11.7 13.3 8.3 10.4
retail asset manager      
retail foreign investor 13.4 8.6 9.0 10.7
retail insurance 12.5 8.1 8.8 12.8
retail open-end fund 11.3 6.9 5.6 5.9
total   12.6 8.2 8.5 10.7
total   12.1 10.8 8.4 10.4

  lower cross spread lower cross spread
  higher cross spread higher cross spread
The spread (sd) between singleproperty returns
within the groups does not vary significantly
12
Foreign investments are more volatile

• Foreign investorsMean total returns of 7.7 in
  2006/2007
• Returns dropped to 0.1 in the 2008/2009 period
• Domestic investors
• Mean TR of 3.7 in the 2006/2007 period
• Even higher returns (4.2) in 2008/2009

Investor / period mean TR median TR
domestic 08/09 4.2 4.8
foreign 08/09 0.1 1.0
domestic 06/07 3.7 4.4
foreign 06/07 7.7 8.2
-30 -20 -10 0 10 20 30
13
Industry investments are more volatile

• DIX (excl. industry)
• Mean total returns of 3.9 in 2006/2007
• Similar returns of 3.3 in the 2008/2009 period
  
• Industry sector
• Mean TR of 5.5 in the 2006/2007 period
• Sharpe decline to -1.6 in 2008/2009
• gtgtgt As foreign investors are stronglyover-allocat
  ed in the industry segment, both effects interfere

Investor / period mean TR median TR
DIX 08/09 3.3 4.3
industry 08/09 -1.6 -1.2
DIX 06/07 3.9 4.7
industry 06/07 5.5 7.3
-30 -20 -10 0 10 20 30
14
Retail properties Domestic vs. foreign investors

• Analysis of retail sector, as both investor
  groups have sufficient assets for comparison
• Domestic retail investors
• Low total returns of 2.3 in 2006/2007
• Higher returns of 3.5 in the 2008/2009 period
• Foreign retail investors
• Mean TR of 8.6 in the 2006/2007 period
• Sharpe decline to 0.9 in 2008/2009
• gtgtgt As each group has more than 100 retail
  properties in every analyzed year, the conclusion
  is that foreign investors had their assets
  appreciated in the 2006/2007 period and
  depreciated in the 2008/2009 period. For German
  investors there was only a minor change in the
  market values of their retail investments

Investor / period mean TR median TR
retail DIX 08/09 3.5 5.0
retail foreign 08/09 0.9 1.1
retail DIX 06/07 2.3 4.4
retail foreign 06/07 8.6 8.2
-30 -20 -10 0 10 20 30
15
Accounting for investment risk

• To account for different risk levels of
  investments, analysis focuses on capital growth
  rather than Total Return
• The risk of a property is measured by the maximum
  income return the property realized between 2000
  and 2009
• The terciles of the resulting distribution of
  income returns define the dummy regression groups
  IRlow, IRmed and IRhigh

16
Dummy regression Explaining capital growth 2009
by Investor type, sector and income return
lm(formula CG09 InvForeign InvFund InvInsurance SecIndustry SecOffice SecResidential SecRetail Irlow IRhigh) lm(formula CG09 InvForeign InvFund InvInsurance SecIndustry SecOffice SecResidential SecRetail Irlow IRhigh) lm(formula CG09 InvForeign InvFund InvInsurance SecIndustry SecOffice SecResidential SecRetail Irlow IRhigh) lm(formula CG09 InvForeign InvFund InvInsurance SecIndustry SecOffice SecResidential SecRetail Irlow IRhigh) lm(formula CG09 InvForeign InvFund InvInsurance SecIndustry SecOffice SecResidential SecRetail Irlow IRhigh) lm(formula CG09 InvForeign InvFund InvInsurance SecIndustry SecOffice SecResidential SecRetail Irlow IRhigh)
Coefficients
Estimate Error t value Pr(gtt)
(Intercept) -3.6 1.0 -3.539 0.000
InvForeign -4.9 0.6 -7.981 0.000
InvFund 1.1 0.7 1.597 0.110
InvInsurance -0.5 0.5 -0.945 0.345
SecIndustry -2.5 1.2 -2.069 0.039
SecOffice 1.2 0.9 1.317 0.188
SecResidential 5.3 1.0 5.538 0.000
SecRetail 1.8 1.0 1.803 0.071 .
IRlow 0.1 0.4 0.250 0.803
IRhigh 0.4 0.4 1.102 0.270
---
Signif. codes 0 '' 0.001 '' 0.01 '' 0.05 '.' 0.1 ' ' 1 Signif. codes 0 '' 0.001 '' 0.01 '' 0.05 '.' 0.1 ' ' 1 Signif. codes 0 '' 0.001 '' 0.01 '' 0.05 '.' 0.1 ' ' 1 Signif. codes 0 '' 0.001 '' 0.01 '' 0.05 '.' 0.1 ' ' 1 Signif. codes 0 '' 0.001 '' 0.01 '' 0.05 '.' 0.1 ' ' 1

Residual standard error 0.09332 on 2980 degrees of freedom Residual standard error 0.09332 on 2980 degrees of freedom Residual standard error 0.09332 on 2980 degrees of freedom Residual standard error 0.09332 on 2980 degrees of freedom Residual standard error 0.09332 on 2980 degrees of freedom Residual standard error 0.09332 on 2980 degrees of freedom
(2956 observations deleted due to missingness) (2956 observations deleted due to missingness) (2956 observations deleted due to missingness) (2956 observations deleted due to missingness) (2956 observations deleted due to missingness)
Multiple R-squared 0.1222, Adjusted R-squared 0.1195 Multiple R-squared 0.1222, Adjusted R-squared 0.1195 Multiple R-squared 0.1222, Adjusted R-squared 0.1195 Multiple R-squared 0.1222, Adjusted R-squared 0.1195 Multiple R-squared 0.1222, Adjusted R-squared 0.1195 Multiple R-squared 0.1222, Adjusted R-squared 0.1195
F-statistic 46.09 on 9 and 2980 DF, p-value lt 2.2e-16 F-statistic 46.09 on 9 and 2980 DF, p-value lt 2.2e-16 F-statistic 46.09 on 9 and 2980 DF, p-value lt 2.2e-16 F-statistic 46.09 on 9 and 2980 DF, p-value lt 2.2e-16 F-statistic 46.09 on 9 and 2980 DF, p-value lt 2.2e-16 F-statistic 46.09 on 9 and 2980 DF, p-value lt 2.2e-16

• Foreign investments have significantly lower
  capital growth

17
Dummy regression Explaining capital growth 2008
by Investor type, sector and income return
lm(formula CG08 InvForeign InvFund InvInsurance SecIndustry SecOffice SecResidential SecRetail IRlow IRhigh) lm(formula CG08 InvForeign InvFund InvInsurance SecIndustry SecOffice SecResidential SecRetail IRlow IRhigh) lm(formula CG08 InvForeign InvFund InvInsurance SecIndustry SecOffice SecResidential SecRetail IRlow IRhigh) lm(formula CG08 InvForeign InvFund InvInsurance SecIndustry SecOffice SecResidential SecRetail IRlow IRhigh) lm(formula CG08 InvForeign InvFund InvInsurance SecIndustry SecOffice SecResidential SecRetail IRlow IRhigh) lm(formula CG08 InvForeign InvFund InvInsurance SecIndustry SecOffice SecResidential SecRetail IRlow IRhigh)
Coefficients
Estimate Error t value Pr(gtt)
(Intercept) -0.2 1.4 -0.129 0.897
InvForeign -5.5 1.3 -4.152 0.000
InvFund -0.7 1.3 -0.538 0.591
InvInsurance -0.6 1.3 -0.502 0.616
SecIndustry -0.7 1.1 -0.634 0.526
SecOffice 0.4 0.7 0.563 0.574
SecResidential 2.9 0.8 3.877 0.000
SecRetail 0.7 0.8 0.978 0.328
IRlow -1.2 0.4 -2.730 0.006
IRhigh -1.2 0.4 -3.166 0.002
---
Signif. codes 0 '' 0.001 '' 0.01 '' 0.05 '.' 0.1 ' ' 1 Signif. codes 0 '' 0.001 '' 0.01 '' 0.05 '.' 0.1 ' ' 1 Signif. codes 0 '' 0.001 '' 0.01 '' 0.05 '.' 0.1 ' ' 1 Signif. codes 0 '' 0.001 '' 0.01 '' 0.05 '.' 0.1 ' ' 1 Signif. codes 0 '' 0.001 '' 0.01 '' 0.05 '.' 0.1 ' ' 1

Residual standard error 0.07451 on 2192 degrees of freedom Residual standard error 0.07451 on 2192 degrees of freedom Residual standard error 0.07451 on 2192 degrees of freedom Residual standard error 0.07451 on 2192 degrees of freedom Residual standard error 0.07451 on 2192 degrees of freedom Residual standard error 0.07451 on 2192 degrees of freedom
(3744 observations deleted due to missingness) (3744 observations deleted due to missingness) (3744 observations deleted due to missingness) (3744 observations deleted due to missingness) (3744 observations deleted due to missingness)
Multiple R-squared 0.09674, Adjusted R-squared 0.09303 Multiple R-squared 0.09674, Adjusted R-squared 0.09303 Multiple R-squared 0.09674, Adjusted R-squared 0.09303 Multiple R-squared 0.09674, Adjusted R-squared 0.09303 Multiple R-squared 0.09674, Adjusted R-squared 0.09303 Multiple R-squared 0.09674, Adjusted R-squared 0.09303
F-statistic 26.08 on 9 and 2192 DF, p-value lt 2.2e-16 F-statistic 26.08 on 9 and 2192 DF, p-value lt 2.2e-16 F-statistic 26.08 on 9 and 2192 DF, p-value lt 2.2e-16 F-statistic 26.08 on 9 and 2192 DF, p-value lt 2.2e-16 F-statistic 26.08 on 9 and 2192 DF, p-value lt 2.2e-16 F-statistic 26.08 on 9 and 2192 DF, p-value lt 2.2e-16

• Foreign investments have significantly lower
  capital growth

18
Dummy regression Explaining capital growth 2006
by Investor type, sector and income return
lm(formula CG06 InvForeign InvFund InvInsurance SecIndustry SecOffice SecResidential SecRetail IRlow IRhigh) lm(formula CG06 InvForeign InvFund InvInsurance SecIndustry SecOffice SecResidential SecRetail IRlow IRhigh) lm(formula CG06 InvForeign InvFund InvInsurance SecIndustry SecOffice SecResidential SecRetail IRlow IRhigh) lm(formula CG06 InvForeign InvFund InvInsurance SecIndustry SecOffice SecResidential SecRetail IRlow IRhigh) lm(formula CG06 InvForeign InvFund InvInsurance SecIndustry SecOffice SecResidential SecRetail IRlow IRhigh) lm(formula CG06 InvForeign InvFund InvInsurance SecIndustry SecOffice SecResidential SecRetail IRlow IRhigh)
Coefficients
Estimate Error t value Pr(gtt)
(Intercept) 0.2 2.0 0.111 0.912
InvForeign 2.6 2.1 1.218 0.223
InvFund -2.8 1.9 -1.436 0.151
InvInsurance -0.2 1.9 -0.128 0.898
SecIndustry -1.8 1.6 -1.090 0.276
SecOffice -2.5 0.7 -3.394 0.001
SecResidential 3.9 0.8 4.606 0.000
SecRetail -1.4 0.8 -1.650 0.099 .
IRlow -0.8 0.5 -1.495 0.135
IRhigh -0.5 0.5 -1.014 0.311
---
Signif. codes 0 '' 0.001 '' 0.01 '' 0.05 '.' 0.1 ' ' 1 Signif. codes 0 '' 0.001 '' 0.01 '' 0.05 '.' 0.1 ' ' 1 Signif. codes 0 '' 0.001 '' 0.01 '' 0.05 '.' 0.1 ' ' 1 Signif. codes 0 '' 0.001 '' 0.01 '' 0.05 '.' 0.1 ' ' 1 Signif. codes 0 '' 0.001 '' 0.01 '' 0.05 '.' 0.1 ' ' 1

Residual standard error 0.07451 on 2192 degrees of freedom Residual standard error 0.07451 on 2192 degrees of freedom Residual standard error 0.07451 on 2192 degrees of freedom Residual standard error 0.07451 on 2192 degrees of freedom Residual standard error 0.07451 on 2192 degrees of freedom Residual standard error 0.07451 on 2192 degrees of freedom
(3744 observations deleted due to missingness) (3744 observations deleted due to missingness) (3744 observations deleted due to missingness) (3744 observations deleted due to missingness) (3744 observations deleted due to missingness)
Multiple R-squared 0.09674, Adjusted R-squared 0.09303 Multiple R-squared 0.09674, Adjusted R-squared 0.09303 Multiple R-squared 0.09674, Adjusted R-squared 0.09303 Multiple R-squared 0.09674, Adjusted R-squared 0.09303 Multiple R-squared 0.09674, Adjusted R-squared 0.09303 Multiple R-squared 0.09674, Adjusted R-squared 0.09303
F-statistic 26.08 on 9 and 2192 DF, p-value lt 2.2e-16 F-statistic 26.08 on 9 and 2192 DF, p-value lt 2.2e-16 F-statistic 26.08 on 9 and 2192 DF, p-value lt 2.2e-16 F-statistic 26.08 on 9 and 2192 DF, p-value lt 2.2e-16 F-statistic 26.08 on 9 and 2192 DF, p-value lt 2.2e-16 F-statistic 26.08 on 9 and 2192 DF, p-value lt 2.2e-16

• Foreign investments with higher growth rates, but
  not significant due to few observations
• Neither significant results for 2007 and 2005

19
Overview Significant capital growth differences
2005-2009
CGInvType sectorIR 2005   2006   2007   2008   2009  
CGInvType sectorIR Sig. sign Sig. sign Sig. sign Sig. sign Sig. sign
foreign owned prop. 20 53 102 214 327
InvForeign ()  () (-) - -
InvFund . - -
InvInsurance - -
SecIndustry               -
SecOffice . - -        
SecResidential
SecRetail  . -          .
IRlow - - -
IRhigh -
Signif. codes 0 '' 0.001 '' 0.01 '' 0.05 '.' 0.1 ' ' 1 Signif. codes 0 '' 0.001 '' 0.01 '' 0.05 '.' 0.1 ' ' 1 Signif. codes 0 '' 0.001 '' 0.01 '' 0.05 '.' 0.1 ' ' 1 Signif. codes 0 '' 0.001 '' 0.01 '' 0.05 '.' 0.1 ' ' 1 Signif. codes 0 '' 0.001 '' 0.01 '' 0.05 '.' 0.1 ' ' 1 Signif. codes 0 '' 0.001 '' 0.01 '' 0.05 '.' 0.1 ' ' 1 Signif. codes 0 '' 0.001 '' 0.01 '' 0.05 '.' 0.1 ' ' 1 Signif. codes 0 '' 0.001 '' 0.01 '' 0.05 '.' 0.1 ' ' 1 Signif. codes 0 '' 0.001 '' 0.01 '' 0.05 '.' 0.1 ' ' 1 Signif. codes 0 '' 0.001 '' 0.01 '' 0.05 '.' 0.1 ' ' 1 Signif. codes 0 '' 0.001 '' 0.01 '' 0.05 '.' 0.1 ' ' 1

• Stars indicate significance level, sign refers
  to the sign of the estimate, i.e. whether the
  deviation is positive or negative
• Foreign investments deviate highly significant
  negative in 2008 and 2009
• Objects with high income returns were appreciated
  in 2005 and depreciated in 2008

20
Dummy regression Significant capital growth
differences retail 2005-2009
CG InvForeign IRlow IRhigh 2005   2006   2007   2008   2009  
CG InvForeign IRlow IRhigh Est. Sig. Sig. sign Sig. sign Sig. sign Sig. sign
foreign owned retail prop. 105 140 234 298
InvForeign 6.7 5.7 3.0 -5.4 -4.0
IRlow -2.5 0.1 -0.4 -0.2 -1.1
IRhigh 0.7 -0.8 -1.7 -0.9 1.5
Signif. codes 0 '' 0.001 '' 0.01 '' 0.05 '.' 0.1 ' ' 1 Signif. codes 0 '' 0.001 '' 0.01 '' 0.05 '.' 0.1 ' ' 1 Signif. codes 0 '' 0.001 '' 0.01 '' 0.05 '.' 0.1 ' ' 1 Signif. codes 0 '' 0.001 '' 0.01 '' 0.05 '.' 0.1 ' ' 1 Signif. codes 0 '' 0.001 '' 0.01 '' 0.05 '.' 0.1 ' ' 1 Signif. codes 0 '' 0.001 '' 0.01 '' 0.05 '.' 0.1 ' ' 1 Signif. codes 0 '' 0.001 '' 0.01 '' 0.05 '.' 0.1 ' ' 1 Signif. codes 0 '' 0.001 '' 0.01 '' 0.05 '.' 0.1 ' ' 1 Signif. codes 0 '' 0.001 '' 0.01 '' 0.05 '.' 0.1 ' ' 1 Signif. codes 0 '' 0.001 '' 0.01 '' 0.05 '.' 0.1 ' ' 1 Signif. codes 0 '' 0.001 '' 0.01 '' 0.05 '.' 0.1 ' ' 1

• Capital growth of foreign investments deviates
  highly significant negative in 2008 and 2009 and
  positive between 2005 and 2007
• Objects with high income returns were depreciated
  in 2008

21
Sum up

• Volatility of total return differs between
  domestic and foreign investors in Germany
• Foreign investors are strongly over-allocated in
  the industry segment
• Asset allocation is a necessary control variable
  when comparing domestic and foreign investments
  on the German market
• Retail investments of foreign investors showed
  significantly more volatility between 2005 and
  2009
• Open research questions
• Are foreign investors more opportunistic than
  domestic investors (by the fact that they went
  abroad)?
• Differences between foreign and domestic
  investors in other countries?
• Differences in the selection of properties within
  a sector (core office versus opportunistic office
  within Frankfurt)?

22
Thank you for your attention!
IPD Investment Property Databank
GmbH Kirchgasse 2 65185 Wiesbaden Dr. Sebastian
Gläsner Telefon [phone]
89 sebastian.glaesner_at_ipd.com