Public v. Private Real Estate Equities:
A More Refined Comparison
Copyright By PowCoder代写 加微信 powcoder
. Pagliari, Jr.
. Monopoli
NAREIT Institutional Investor Forum
June 3, 2003 – ,
The Objective:
Understand the Similarities and Dissimilarities
between Public and Private Real Estate
Investment Vehicles
Presentation Agenda
• Background
• Evidence & Methodology
• Implications
Background
• Returns:
NAREIT – NCREIF = 500 bps per annum
14.7% – 9.7% = 500 bps (1978-2002)
• Is there some structural reason (e.g., optimal contracting)?
• Is it happenstance? Does the platform/vehicle matter?
• The answer matters in terms of portfolio implications
– If one platform dominates, then that vehicle will dominate over time
– If neither platform dominates, then the vehicles can be viewed as substitutes for
one another
• At present, institutional investors (e.g., pension funds & LICs) invest relatively
little in private RE & even less in REITs
Circumstantial Evidence: Popularity of NAREIT v. NCREIF
Ended NAREIT NCREIF Total NAREIT NCREIF Total
1981 $758 $3,545 $4,302 17.6% 82.4% 100.0%
1982 788 4,731 5,518 14.3% 85.7% 100.0%
1983 893 8,782 9,676 9.2% 90.8% 100.0%
1984 1,684 11,216 12,900 13.1% 86.9% 100.0%
1985 2,174 15,040 17,214 12.6% 87.4% 100.0%
1986 3,017 17,573 20,590 14.7% 85.3% 100.0%
1987 5,894 21,868 27,762 21.2% 78.8% 100.0%
1988 6,814 27,858 34,672 19.7% 80.3% 100.0%
1989 7,641 32,076 39,717 19.2% 80.8% 100.0%
1990 9,386 37,328 46,714 20.1% 79.9% 100.0%
1991 8,606 36,197 44,803 19.2% 80.8% 100.0%
1992 13,691 38,803 52,494 26.1% 73.9% 100.0%
1993 26,319 40,216 66,535 39.6% 60.4% 100.0%
1994 50,663 40,147 90,810 55.8% 44.2% 100.0%
1995 70,414 47,472 117,886 59.7% 40.3% 100.0%
1996 100,424 53,793 154,217 65.1% 34.9% 100.0%
1997 173,927 64,984 238,911 72.8% 27.2% 100.0%
1998 213,232 65,900 279,133 76.4% 23.6% 100.0%
1999 205,821 79,744 285,564 72.1% 27.9% 100.0%
2000 232,711 94,631 327,343 71.1% 28.9% 100.0%
2001 $284,510 $109,911 $394,421 72.1% 27.9% 100.0%
Exhibit 1: Year-end Total Market Capitalization of NAREIT and NCREIF
Total Market Capitalization Market Share
The year-end market capitalization of the NCREIF Property Index is reported on a gross-
asset basis.
(millions of $)
The year-end market capitalization of NAREIT equity index represents the market value of
each firm’s common stock plus the book value of its liabilities and its preferred stock (where
the latter two items were available). It, however, excludes the value o
Nearly ten-
fold increase
Reversal of
Circumstantial Evidence (continued):
Big Pension Funds Favor Private Real Estate
Private Allocation Public Allocation
Real Estate Fund Total to Private REIT Fund Total to Public
Defined-benefit Fund Investment Ranking Assets Real Estate Defined-benefit Fund Investment Ranking Assets REITs
1 . California Public Employees $13,000 1 $143,887 9.03% General Motors $1,100 6 $82,500 1.33%
2 . Ohio State Teachers 6,024 22 47,336 12.73% State Teachers 991 8 74,915 1.32%
3 . California State Teachers 5,296 3 95,553 5.54% Ohio Public Employees 967 18 50,459 1.92%
4 . General Motors 5,000 6 82,500 6.06% California Public Employees 946 1 143,887 0.66%
5 . Michigan State Retirement 4,042 20 49,266 8.20% Maryland State Retirement 674 36 26,608 2.53%
6 . State Teachers 3,840 8 74,915 5.13% Oregon Employees 613 29 35,051 1.75%
7 . Florida State Board 3,621 5 88,514 4.09% SBC Communications 604 23 46,405 1.30%
8 . Washington State Board 3,592 25 41,916 8.57% Teamsters Central States 545 73 11,889 4.58%
9 . Los Angeles County Employees 3,505 37 25,910 13.53% Kentucky Retirement 453 75 11,784 3.84%
10 . State Common 3,439 2 106,091 3.24% Texas County & District 441 108 8,315 5.30%
11 . Ohio Public Employees 3,325 18 50,459 6.59% Ohio State Teachers 425 22 47,336 0.90%
12 . Illinois Teachers 2,378 44 21,712 10.95% 395 30 34,744 1.14%
13 . Verizon 2,193 11 65,936 3.33% BellSouth 388 38 25,046 1.55%
14 . Wisconsin Investment Board 2,118 14 55,473 3.82% Pennsylvania Employees 376 39 24,576 1.53%
15 . Colorado Employees 2,090 34 26,928 7.76% Florida State Board 345 5 88,514 0.39%
16 . Pennsylvania School Employees 2,064 21 48,000 4.30% Division 331 10 66,691 0.50%
17 . Oregon Employees 1,970 29 35,051 5.62% Bakery & Confectionery Union 319 174 5,000 6.38%
18 . General Electric 1,868 9 68,769 2.72% Kansas Public Employees 298 102 8,717 3.42%
19 . AT&T 1,804 32 27,601 6.54% Illinois Municipal Retirement 296 63 13,928 2.13%
20 . IBM 1,725 13 56,500 3.05% Colorado Employees 258 34 26,928 0.96%
21 . North Carolina 1,573 17 52,575 2.99% Illinois State Universities 240 91 9,892 2.43%
22 . Massachusetts PRIM 1,520 35 26,802 5.67% I.A.M. National 223 52 19,000 1.17%
23 . Western Conference Teamsters 1,517 42 22,604 6.71% Chicago Public School Teachers 203 98 9,252 2.19%
24 . Nevada Public Employees 1,299 70 12,825 10.13% Michigan State Retirement 200 20 49,266 0.41%
25 . Pennsylvania Employees 1,277 39 24,576 5.20% University of California 194 24 41,974 0.46%
Total $80,080 $1,351,699 Total $11,825 $962,677
Average $3,203 22 $54,068 6.46% Average $473 47 $38,507 2.00%
Amounts (shown in $ millions) as reported in the January 21, 2002 issue of Pensions & Investments , for the period ended September 30, 2001.
Invested in Private Real Estate Equities Invested in Public Real Estate Investment Trusts (REITs)
Exhibit 2: Top 200 Pension/Endowment/Union Funds with Defined-benefit Assets Invested in Real Estate Equities
Circumstantial Evidence (continued)
• Big pension fund investors favor private real estate:
– 90% of pension fund capital devoted to total RE is in private RE
– Top 25 Pension Funds allocation to RE: $80 billion
– Top 25 Pension Funds allocation to REITs: $12 billion
– Ten funds common to both lists: RE:REITs ≈ 6:1
– Clientele Effect
Big funds favor private RE, and
Small funds & individuals favor REITs
Circumstantial Evidence (continued)
• Average premium /discount to NAV ≈ 0 (1990 – 2001)
• Shifting investor sentiment (about REITs relative to RE)
Competing Theories
• RE = Capital-Intensive Business:
– Economies of scale
– Large firms will dominate
• Well-Grounded Economic Principle:
– Increases in marginal costs declines with growing firm size
• Dissenting Points of View:
– RE = Local Business, Diseconomies of Scale, etc.
• Is Size Really Much of a Differentiator?
– Ten largest REITs = $107 billion in assets (not all RE; excludes OP units)
– Ten largest advisory firms = $75 billion in assets
Competing Theories (continued)
• Classical Financial Theory Suggests Wrapper Doesn’t Matter:
– It’s what’s under the wrapper that counts
• Modern Financial Theory Takes a More Nuanced View:
– Optimal contracting issues
– Manager /shareholder conflicts
– Corporations suffer from a series of agency conflicts (even before Enron,
WorldCom, etc.)
• Similar Concerns Voiced About the Sometimes Too-Cozy Relationship
between Pension Funds & Their Advisors:
– Not limited to RE advisors
Background (continued)
• No Conclusive Argument Emerges:
– Circumstantial evidence is mixed
– Theory is inconclusive
Presentation Agenda
• Background
• Evidence & Methodology
• Implications
Examining the Empirical Evidence
• Need to Control for:
– Leverage
– Core v. non-core property types
– Varying mix of core property types
– Appraisal smoothing
• Additional Differences (which are
difficult to control):
– Geography (NCREIF tends to cluster
in top-tier cities)
– Development activities
– Building franchise value
– G&A v. investment management fees
(depends on how the assets are held)
Examined in Paper
Not Examined in
(Lack of Data)
Controlling for Differences: Leverage
Controlling for Differences: Leverage
• NCREIF ← Unlevered Index
• NAREIT ← Levered Index
• Increasing LTV & Declining Debt Cost
Controlling for Differences: NAREIT Leverage
• Recognize that asset-level returns (ka) represent a blend of returns to
indebtedness (kd) and returns to equity (ke)
• Weighted average cost of capital (“WACC”):
• Need to solve for ka → an apples-to-apples comparison
Book Interest Expense
Book Value of Debtd
Market Return on Common Equity
Market Value of Common Equitye
k k (LTV) k ( LTV)= + −
Controlling for Differences: Data Sources
• NAREIT, NCREIF, CRSP, COMPUSTAT
• Complete Data Covers About 80% of Population
– Those missing tend to be smaller firms
– More coverage recently
• Coverage as Shown in Exhibit 7
Controlling for Differences: Data Sources
Firm-Years Available in: Number Percent Number Percent Number Percent
CRSP and COMPUSTAT 1,618 79.5% 322 66.5% 1,296 83.6%
CRSP but not COMPUSTAT 254 12.5% 145 30.0% 109 7.0%
COMPUSTAT but not CRSP 59 2.9% 9 1.9% 50 3.2%
Neither CRSP nor COMPUSTAT 103 5.1% 8 1.7% 95 6.1%
NAREIT Total 2,034 100.0% 484 100.0% 1,550 100.0%
Average Number of Firms per Year 97 40 172
This exhibit summarizes the data coverage by firm-years (i.e. , combination of total number of years and total
number of firms) for which sufficient data was available. The return data was provided by CRSP (and, in some
cases, supplemented by SNL Financial
1981-2001 1981-1992 1993-2001
Exhibit 7: Data Coverage based on Firm-Years
Controlling for Differences: Core v. Non-Core
Controlling for Differences: Core v. Non-Core
– Stabilized core only
– Predominantly an office & retail index
Controlling for Differences: Core v. Non-Core
Controlling for Differences: Core v. Non-Core
– Stabilized & non-stabilized
– Core & non-core
– Predominantly a retail & residential index
Controlling for Differences: Property Types
• NAREIT – Eliminate non-core property types:
Controlling for Differences: Property Types
Controlling for Differences: Property Types
• For “diversified” segment, divide equally among:
– Apartments
– Industrial
• For “mixed” segment, divide equally among:
– Industrial
• With the IPO explosion in 1993, greater emphasis on property-type
orientation
Controlling for Differences: Property Types
Controlling for Differences: Appraisal Smoothing
Appreciation Returns – NAREIT v. NCREIF
1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002
NAREIT Equity
NCREIF Property Index
Controlling for Differences: Appraisal Smoothing
– An exchange-traded security
– Valuations are not an issue
– Not an exchange-traded security
– Valuations are primarily done by appraisal
– Appraisal “smoothing” problems
⇒ Econometric/academic solution:
Inject more volatility into RE time series
Controlling for Differences: Appraisal Smoothing
0.40, for this studyα =
• Because of relatively few sales, appraisers use “comps”
• Appraisers are thought to partially “anchor” on past appraisals (kt-1)
• Econometric technique to recover “true” values (kt*):
1(1 )* t tt
1* 2.5 1.5t t tk k k −= −
Presentation Agenda
• Background
• Evidence & Methodology
• Implications
The Empirical Questions
• Ho : REITs – RE = 0 v. Ha : REITs – RE ≠ 0
• Ho : σ2REITs = σ2RE v. Ha: σ2REITs ≠ σ2RE
• It is not enough to observe a difference!
• The difference must be statistically meaningful
(i.e., unlikely to have observed the difference by chance)
Results: Restated Returns
Results (continued): The Crux of Our Comparison
Results (continued)
Restating Returns: NAREIT NCREIF
Average Std. Dev Average Std. Dev
As Reported 13.47 14.66 8.43 5.91
Data Availability 0.27 <0.85>
Eliminate Non-Core 0.96 <0.01>
Eliminate Leverage <2.36> <5.51>
Restate for NAREIT Mix 0.91 <0.71>
Appraisal Smoothing 0.00 3.39
Restated Returns 12.34% 8.29% 9.34% 8.59%
Results (continued): Statistically Testing Return Differences
• Competing Hypothesis:
– Ho : REITs – RE = 0
– Ha : REITs – RE ≠ 0
• Statistical Test:
t ≈ 2.00 in order to reject the null
hypothesis (Ho) at “statistically
significant” levels
Cannot reject Ho at conventional
confidence levels
REITs RE∴ =
Results (continued): Statistically Testing Return Differences
Illustration of Difference Test
-9.27% -7.95% -6.62% -5.30% -3.97% -2.65% -1.32% 0.00% 1.32% 2.65% 3.97% 5.30% 6.62% 7.95% 9.27%
Return Differential
Observed, Empircal Difference
Difference under Null Hypothesis
Critical Rejection RegionCritical Rejection Region
Results (continued): Two Related Points
• If fees & costs of private RE exceeded fees & costs of public REITs by
250 bps per year, then REITs > RE:
• Most recent period, change is smaller (62 bps)
Results (continued): Statistical Tests of Volatility
2 2:o REITs REH σ σ=
• Competing Hypotheses:
• Statistical Test:
2 2:a REITs REH σ σ≠
2 2Cannot Reject o REITs REH σ σ∴ ⇒ =
Reinforces the econometric approach of injecting
volatility into private RE return series
Results (continued)
• Statistical tests indicate there is no meaningful difference between the
two restated series.
• This conclusion is largely driven by the “noisy” nature of the
(differential) return series.
• The “new” REIT era shows a considerably smaller spread in returns (60
Presentation Agenda
• Background
• Empirical Evidence & Methodology
• Implications
Implications
• Greater long-run synchronicity between public and private-market real
estate vehicles, probably fed by:
Improved market efficiency
Increased market capitalization
Better data availability
Implications (continued):
• Portfolio Management Implications:
– Public- and private-market vehicles should be viewed somewhat
interchangeably (since investors can alter the property-type mix and
leverage of their real estate holdings)
– While the “platform” may not matter in a statistical sense, the platform
may matter with regard to:
Transparency
Executive Compensation
Clearly a “clientele” effect has developed among
(big and small) pension fund investors, which
seems to indicate that they value these
characteristics differently
Public v. Private Real Estate Equities:
A More Refined Comparison
. Pagliari, Jr.
. Monopoli
NAREIT Institutional Investor Forum
For a copy of the complete paper, visit:
http://www1.kellogg.nwu.edu/fps/DisplayPub.asp?pub_sequo=2119
Public v. Private Real Estate Equities:A More Refined Comparison
The Objective:
Presentation Agenda
Background
Circumstantial Evidence: Popularity of NAREIT v. NCREIF
Circumstantial Evidence (continued): Big Pension Funds Favor Private Real Estate
Circumstantial Evidence (continued)
Circumstantial Evidence (continued)
Competing Theories
Competing Theories (continued)
Background (continued)
Presentation Agenda
Examining the Empirical Evidence
Controlling for Differences: Leverage
Controlling for Differences: Leverage
Controlling for Differences: NAREIT Leverage
Controlling for Differences: Data Sources
Controlling for Differences: Data Sources
Controlling for Differences: Core v. Non-Core
Controlling for Differences: Core v. Non-Core
Controlling for Differences: Core v. Non-Core
Controlling for Differences: Core v. Non-Core
Controlling for Differences: Property Types
Controlling for Differences: Property Types
Controlling for Differences: Property Types
Controlling for Differences: Property Types
Controlling for Differences: Appraisal Smoothing
Controlling for Differences: Appraisal Smoothing
Controlling for Differences: Appraisal Smoothing
Presentation Agenda
The Empirical Questions
Results: Restated Returns
Results (continued): The Crux of Our Comparison
Results (continued)
Results (continued): Statistically Testing Return Differences
Results (continued): Statistically Testing Return Differences
Results (continued): Two Related Points
Results (continued): Statistical Tests of Volatility
Results (continued)
Presentation Agenda
Implications
Implications (continued):
Public v. Private Real Estate Equities:A More Refined Comparison
程序代写 CS代考 加微信: powcoder QQ: 1823890830 Email: powcoder@163.com