Fee Structures in Private Equity Real

Estate

Author M a a rte n van d e r S p ek

Abstract Although fees in real estate are important for investors and fund
managers, they have received little attention in the finance
literature. In this paper, I examine fee structures for private
equity real estate funds from an investor’s perspective. Fee
structures, as proposed by fund managers in placing documents,
are used to calibrate the total fee load. The average total fee load
for closed-end funds equals 2.7%, which is substantially lower
than for private equity funds. Through regression and simulation,
I show that core and value-add funds charge significantly lower
performance fees compared with opportunistic funds, although
there is no difference in management fees. Moreover, larger
funds charge significantly less management fees. Investors can
substantially reduce fees by controlling the amount of leverage
and avoiding commitment fees and catch-ups.

One of the most important topics for investors in the private investment market is
also one of the least researched: fees. No standardization exists, yet the impact of
fees on the net performance of investors can be substantial. Moreover, because
we are in a low interest rate and economic growth environment as a result of the
global financial crisis, returns are expected to be modest. In this kind of
environment, every small return enhancement can make a difference. Institutional
investors are keen to reduce costs in order to add value, and since fees are an
important part of that, investors are likely to discuss and negotiate fees. However,
managers need to charge fees to compensate for their time and effort and to make
a profit. With this in mind, it would be reasonable to expect more insight into
what would be a fair amount of fees for each type of investment vehicle. Such an
expectation would also align with the most recent trend in regulations to improve
transparency and to protect investors. Unfortunately, transparency on fees is far
from market practice.
In this paper, I analyze the fee structure and fee load for private equity real estate
funds to help improve transparency. The central question is how investors can
compare different investment opportunities with different fee structures and how
investors can best reduce fee load. To achieve this, I introduce a single fee metric
and use regression to determine the main fee drivers.

JRER I Vo I. 39 I No. 3 - 2 0 1 7
3 2 0 van der Spek

Phalippou (2009) was one of the first to analyze private equity market fee
structures. He determined that the average private equity fund effectively charges
a 7% fee per annum and concluded that investors are fooled, as compensation
contracts are opaque and show lower fees at first sight. In a previous paper with
Gottschalg (Phalippou and Gottschalg, 2007), Phalippou even proved that private
equity underperformed the S&P 500 on an after-fee basis. Interestingly, Harris,
Jenkinson, and Kaplan (2014) found the opposite to be true, using actual cash
flow data from institutional investors, sourced by Burgiss; they showed that private
equity outperformed the S&P 500 by 3% annually. Unfortunately, they did not
analyze the impact of fees, as all data were net of fees, and only examined private
equity and not private real estate funds. While Fisher and Hartzell (2013) did
analyze the private real estate funds using the same database, they discovered that
the U.S. private real estate funds in the database on average underperform U.S.
private real estate and real estate investment trust (REIT) benchmarks. Again, the
data set consisted of net of fees return and fees were not investigated. Moreover,
the data they used were skewed towards value add and opportunistic funds in the
United States, while core and European and Asian funds were not sufficiently
covered.

Most of the other literature on fees focuses on the relationship between the
investor and the manager, the agent. The importance of agency theory was shown
by Eisenhardt (1989). Interestingly, Gompers and Lemer (1999) proved there is
no relation between incentive compensation and performance, while this would
normally be the reason to introduce performance fees. Furthermore, Robinson and
Sensoy (2011) found no relation between manager compensation and ownership
and the funds’ cash flow performance. In the mutual fund industry, Adams, Mansi,
and Nishikawa (2012) discovered that agency considerations and competition are
important determinants in the pricing of mutual funds. Using a cross-sectional
regression model, they found that disproportionately high fees are prevalent in
funds with multiple share classes and those with weak governance structures.
These findings are relevant for investors, as the information can be used to develop
fund selection criteria.
Besides the fund’s management and performance fees, managers have another
incentive. Chung, Sensoy, Stem, and Weisbach (2010) showed that there is a
reward stemming from the effect of current (good) performance on the ability to
raise larger funds in the future. Explicit fee structures of private equity funds thus
understate the actual incentive for the general partner. In a later paper, Chung,
Sensoy, Stern, and Weisbach (2012) explained that indirect pay for performance
from future fundraising is of the same order of magnitude as direct pay for
performance from carried interest. Capozza and Seguin (2000) demonstrated that
misalignment of the external manager results in underperformance of externally
managed REITs compared with their internally managed counterparts, thus
concluding that the right incentive structure is crucial to the performance of
REITs. As a result and due to their superior ability to resolve conflicts of interests
between REIT management and shareholders, internally advised REITs will
 Fee S t r u c t u r e s in P r i v a t e E q u i t y R e a l E s t a t e 3 2 1

dominate externally advised REITs (Ambrose and Linneman, 2001). While most
research is focused on investment products, private investment funds, Andonov,
Eichholtz, and Kok (2012) analyzed a database with pension fund allocations and
costs, and showed that larger pension funds are more likely to invest in real estate
internally and have lower costs and higher returns. Moreover, U.S. pension funds’
investment costs are twice as high as those of their foreign peers, while returns
are lower.
In this paper, I focus on the fee load of private equity real estate funds. These
private equity funds have several features that make them particularly attractive
for an empirical study. First, most of these funds are closed-end with a specific
finite lifetime, typically seven to ten years. Compensations are agreed upon
beforehand and are therefore fixed and rarely renegotiated during the funds’
lifetime. Second, it is uncommon that managers are removed and replaced from
managing the fund during this period, as removal is rather difficult and expensive
and usually only done as a last resort. Third, unlike for private equity, there is
scarce literature available on private real estate fund fees. Real estate industry
bodies [the European Association for Investors in Non-Listed Real Estate Vehicles
(INREV), the Asian Association for Investors in Non-listed Real Estate Vehicles
(ANREV), and the Pension Real Estate Association (PREA)] publish management
fee studies annually but report mostly on individual fees and not entire fee
structures. As a result, fee structures between funds remain relatively difficult to
compare. Finally, focusing on one industry, real estate, allows for better
comparability between different funds; as such, the results will be more
meaningful for the industry.
The analysis shows that private equity real estate funds have an average fee load
of 2.7%, which is substantially lower than private equity funds. A number of fund
characteristics impact this fee load. Core funds and club deals are generally
cheaper, but remarkably this bears little significance. Only the performance fee of
core funds is significantly lower than that of opportunistic funds; the same can be
concluded for value-add funds. Another important characteristic is size: larger
funds have significantly lower management fees. Funds investing in multiple
countries charge somewhat higher management fees, because of the necessary
additional work, resources, and complexity. The same is applicable to funds
investing in developing countries (+30 bps), while the opposite is true for
industrial funds (-2 7 bps). Additionally, investors need to be aware of the main
fee drivers that will have a substantial impact on fees but can be controlled for:
leverage, commitment fees, and catch-up clauses. The more leverage, the higher
the fee load on average. In particular, leverage has the highest impact when returns
are negative, which is the worst circumstance for investors. Meanwhile, funds with
fees on commitments are on average 46 bps more expensive, and a catch-up clause
can cost investors 27 bps on average but can be as high as 84 bps in market
scenarios in which target returns are met. During fund selection and negotiation,
investors should take these aspects into account to reduce their costs.
The paper is structured as follows: I start with a description of the most common
parts of a fee structure and introduce the database used for the analysis. The

JRER I V oI . 39 I No. 3 - 2 0 1 7
3 2 2 van der Spek

methodology used to investigate fee structures is explained next, followed by a

discussion of the results of the average fee load of private real estate funds, split
by fund type and between management and performance fees. A regression model
is then used to further examine the fee load and its characteristics and main
drivers. A simulation model is next used to analyze the impact of a volatile market
on the fee load and drivers. The paper closes with a discussion on the impact and
drivers of private real estate fund fees.

An Introduction to P r i v a t e E q u i t y Real Estate Fund Fees

A hitherto unexplored database was obtained from the Dutch institutional investor
PGGM. This database contains hundreds of private equity real estate placing
documents, including the terms and conditions of each investment proposal. As
PGGM is a large institutional investor and well known by investment managers,
most managers try to pitch their product to this investor. As a result, many placing
documents are provided to the investor in support of the proposition. Similar
databases have been used to analyze private equity markets. Metrick and Yasuda
(2010) used these investor contracts for their research and found that two-thirds
of expected revenue comes from fixed revenue components that are not sensitive
to performance. Litvak (2009) studied venture capital compensation based on
partnership agreements. One of her conclusions was that compensation is often
more complex and manipulable than it should be. More complex fee structures,
however, predict lower total compensation. Again, both studies focused on private
equity, but the database I use is for a specific private equity investment type: real
estate.
Most funds in the database were raising capital between 2005 and 2015, covered
a wide spectrum of styles, property types, and geographies, and were mostly
closed-end. All of these funds were offered to institutional investors as an
investment product, but whether these funds were eventually launched is unknown.
Other information that is unavailable include the actual fee structure and levels
after negotiations, and the actual performance. The possibility to negotiate differs
per investor and per market. The more capital there is chasing investment
opportunities with a limited number of fund managers, the more fund managers
can determine the fee height. On the other hand, investors are better positioned
to influence fees when markets are in decline, like during the years after the Global
Financial Crisis. In general, larger institutional investors can negotiate lower
management fees and even remove catch-ups, but smaller investors are rarely in
a position to influence fees. Core funds, mostly open-end, have specific fee
structures in place where fees are lower for investors investing over a certain
threshold. Negotiation would still be possible but only to make additional
arrangements, like tailor-made reporting. The fees for these funds are fixed.
Another way for investors to reduce fees is by investing more directly via JVs or
separate accounts; due to the large involvement and substantial commitment of
investors, fees tend to be lower. Only a few JVs have been captured by the
 Fee S t r u c t u r e s in P r i v a t e E q u i t y R e a l E s t a t e 3 2 3

Exhibit 1 | D e s c rip tiv e S ta tis tic s P riv a te R eal E state F u nd D a ta b a s e

Style # Structure # Target Countries # Catch-up #

Core 78 Closed-end 403 Single country 307 Yes 162

Value-add 161 Club deals 10 Multi country 106 No 251
Opportunistic 174

Region # Sector # Year # Target Leverage #

Mature Americas 80 Office 66 >2013 33 0% 15

Emerging Americas 15 Retail 62 2012 57 >0% & <40% 74
Mature Europe 137 Residential 58 2011 51 >40% & <50% 95
Emerging Europe 18 Industrial 30 2010 34 >50% & <60% no
Mature Asia Pacific 71 Mixed 175 2009 56 >60% 119
Emerging Asia Pacific 58 Debt 10 2008 75
Mixed Asia Pacific 30 Other 12 2007 42
Other 4 2006 30
<2005 35

database, but separate accounts are not included, as these are even more tailor-
made solutions. Fees can be avoided by investing in real estate directly. In that
situation, investors would bear all the costs involved and would have to cope with
finding and hiring the right local experts. This arrangement carries organizational
challenges and it is questionable if performance could keep up with the more
specialized managers. Unlike private equity, co-investments within private real
estate are not necessarily fee reducing. Consequently, most investors typically do
not target real estate co-investment opportunities when selecting a private real
estate fund. For some larger investors, however, this has become an important
consideration in order to create the opportunity to increase exposure to the best
deals.
The database provides a very good overview of what is available in the market,
what the average asking fee set by fund managers is, and what smaller investors,
who are unable to negotiate lower fees, should expect to pay. To have as much
consistency as possible between the different funds and to allow for the best
comparisons, the analysis only includes closed-end funds and club deals (typically
structured as a closed-end fund with only a few investors). Exhibit 1 provides an
overview of the type of funds in the database.
Unlike private equity, private real estate fee structures are not standardized, but
rather complex. There is a wide variation of fee structures, even by style.
Typically, there are three different styles for private real estate, namely core, value-

JRER I Vo I . 39 I No. 3 - 2 0 1 7
3 2 4 van der Spek

add, and opportunistic. These styles are related to the type of activities done within
the fund and therefore to the risk taken. Core is the least risky type where the
fund mainly invests in income-producing investments, like fully occupied
shopping center, uses low leverage, has no or very low development exposure and
generates a high proportion of return through income. A value-add fund may invest
in any property type and deliver returns from a balance of income return and
capital appreciation. The fund may allocate part of its investments in real estate
development. Typically, it will also invest in forms of active management, such
as active leasing risk, repositioning or redevelopment to generate returns through
adding value to the property. Usually the fund will use moderate leverage. An
opportunistic fund is the most risky style; typically it is able to use high amounts
of leverage, has a high exposure to development or other forms of active asset
management, and will deliver returns primarily in the form of capital appreciation.
The fund may invest in any market or sector and may be highly focused on
individual markets or property types.
The database consists of 78 core, 161 value-add, and 174 opportunistic funds.
The average fund has a target equity size of almost half a billion euros, and
combined with an average 50% loan-to-value target, it aims to invest a billion
euros in real estate. It is remarkable that the average LTV level for core funds is
over 40%. There are a number of reasons for this relatively high level of LTV.
While the average LTV for open-end funds is more likely to be in the range of
20%-30%, closed-end funds are using levels of leverage of even up to 50%. Some
Japanese funds are even presented as core funds with 70% leverage. However,
these high levels of leverage are mostly used for funds launched pre-crisis, and
are no longer in line with what is currently expected for core funds. That
opportunistic funds have lower leverage than value-add funds can be explained by
the substantial amount of opportunistic funds invested in emerging countries where
finance is not always readily available and therefore use lower LTVs. Value-add
funds, on the other hand, are more focused on mature economies where sufficient
leverage is available.
Usually core funds require less active management and should therefore charge
lower management fees. Value-add and opportunistic funds require more active
management and thus charge higher fees as risk and expected returns are higher.
Managers of these funds are more incentivized, making performance fees more
important. Exhibit 2 displays the types of fees most often used and the average
for each type. These averages are similar to the averages available in private real
estate industry management fee studies by INREV, ANREV, and PREA.
In general, there are three types of fees: fund management, active management,
and performance. Fund management fees are paid quarterly or annually to a fund’s
manager for his or her management services to the fund, covering services such
as the fund level structure management, arrangement of financing, fund
administration, fund reporting, and investor relations. Approximately 55% of all
funds apply different fees during the investment or commitment period (generally
the first three years) compared with the holding period. Fund management fees
 Fee Structures in Private Equity Real Estate I 325

E x h i b i t 2 | Summary Fee Statistics

J RER I Vo I . 39 I No. 3 - 2 0 1 7
 326

E x h i b i t 2 | (continued)
van

Summary Fee Statistics

der
Spek
 Fee S t r u c t u r e s in P r i v a t e E q u i t y R e a l E s t a t e 3 2 7

can be based on gross asset value (GAV, the value of the underlying real estate
portfolio), net asset value (NAV or equity), commitment (the amount for which
the investor has committed to invest), invested equity (the amount the investor has
actually invested in the fund), gross operating income (GOI), and net operating
income (NOI). Except from management fees charged on income, fund
management fees tend to be higher when style risk increases. Thus, fund
management fees for opportunistic funds are higher than those for core funds.
The most popular are management fees based on GAV, on commitments during
investment periods, and on invested equity during holding periods. When looking
at the regional differences (not shown in the exhibit), funds in the Americas are
less likely to charge management fees on GAV and NAV, while European funds
are more likely to do so.
The active management fee is paid for certain activities a manager needs to
perform to execute the strategy of the fund; these fees are typically paid only
once. Around 35% of all funds charge at least one active management fee, whereas
funds in the Americas are less likely to charge active management fees. The most
common fees are:1
■ Acquisition Fee: 82 bps of purchase price on average and used by 27%
of all funds. This fee is charged to a fund upon the acquisition of assets.
Acquisition fees are not typically charged if a property developer/
operator contributes assets to a fund. Unfortunately, it is not always clear
from the acquisition fee whether costs for external advisors (that is,
property agents) are charged to the fund or paid by the manager.
■ Disposition Fee: 80 bps of sale price on average and used by 17% of
all funds. This fee is charged to a fund upon the disposal of assets and
is similar to the acquisition fee.
■ Set-up Fee: 74 bps of commitments on average and used by 7% of all
funds. Set-up fees are charged to cover all costs directly related to the
structuring and establishment of a fund. These costs include, for example,
legal fees, tax advisory fees, structuring fees, and administration costs.
■ Financing Fee: Sometimes referred to as a debt arrangement fee, 50 bps
of par value of debt on average, and used by 2% of all funds. A financing
fee is paid to the manager for services in arranging debt for asset
purchases or refinancing. This fee would be in addition to any
arrangement fees paid to debt providers.
The last type of fee that can be charged to the investor is the performance fee.
This is mostly back-ended and is typically structured as a percentage of all cash
flows after the fund achieves a certain hurdle. The key terms used to structure the
performance fee are hurdle rate, carried interest, and catch-up. The hurdle rate is
the annualized percentage return beyond which the outperformance of net investor
returns is shared with the fund manager. The hurdle rate can be stated relative to
an absolute hurdle rate or to an index/benchmark. Within the database, a
performance fee, and therefore a hurdle, was used by 96% of all funds and the

JRER I V o I . 39 I No. 3 - 2 0 1 7
3 2 8 v a n d e r S p e k

average first hurdle was 9 .9 %. Up to the hurdle rate, no performance fee is paid,
but once the hurdle rate is achieved, profits are split as agreed to in the fund
documentation. This split is called the carried interest. A carried interest is
equivalent to the share of a fund’s profit that will accrue to the general partner.
The average carried interest is 20.7% (the most typical percentage is 20%). The
final key term is the catch-up fee. A catch-up takes effect when an investor’s
return reaches the defined hurdle rate, an agreed level of preferred return. The
fund manager then enters a catch-up period in which he or she may receive an
agreed percentage of the profits until the profit split determined by the carried
interest agreement is reached. This means the manager will be paid performance
fees for any return below the hurdle rate, but only after achieving this rate, hence
the name “catch-up.” Opportunistic funds are most likely to charge a catch-up
and core funds the least likely. Overall, 39% of the funds charge a catch-up fee
and the average split is 59% (that is, 59% of the return over the hurdle rate is
paid to the manager until he or she has achieved a percentage of the total return
equal to the carried interest rate). Once the catch-up is paid, the additional return
is then split according to the carried interest percentage. In some cases, there is
even a second (14.8% on average) or third hurdle (19.3% on average), and after
these hurdles, carried interest changes as well (not included in the exhibit) to 25%
and 27% on average, respectively. Funds in the Americas are more likely to charge
a catch-up (55% on average). Finally, a few managers are paid a performance fee
relative to a benchmark. Only four funds (1%) in the database used a relative
benchmark. This type of fee is very difficult to model; thus, for this study, it is
assumed that managers will not outperform the market. The impact of this
assumption is rather low given the number of funds using a relative benchmark.
Nevertheless, the performance fee might be slightly underestimated.

M odeling Fee S t r u c t u r e s

In order to compare fees among funds, each using different fee structures, one fee
metric must be introduced to cover the complete fee structure. This fee metric
will be the aggregation of all fees paid to the fund manager and will be called
the total fee load (TFL). To determine this metric, a cash flow model has been
developed in which all fees are included. Investors and fund managers in private
real estate typically use cash flow models to determine the IRR for each of their
investments. The TFL is calculated by taking the difference between the IRR after
costs and before fees, and the IRR after costs and fees. This reduction in IRR
equals the TFL, the loss of return due to fees paid to the manager. Once this
metric is available for each fund, given a certain real estate market scenario, it
can be determined which funds are more expensive than others and why.
Moreover, by assuming different market scenarios, it is possible to create a better
understanding of how sensitive fees are for different market circumstances.
For each fund, the lifetime, the length of the investment period, and the potential
extension period are known. Nevertheless, a number of assumptions must be made
 Fee S t r u c t u r e s in P r i v a t e E q u i t y R e a l E s t a t e 3 2 9

to build a cash flow model that is effective for all funds. First, the cash flow model
is on a quarterly basis and investments are done linearly during the investment or
commitment period at quarter end. Meanwhile, divestments are done linearly
during the last two years of the lifetime and during half of the potential extension
period. This period is taken based on the experience that most funds do not sell
all assets before the end of the lifetime. In fact, most of the time it is in the
manager’s interest to continue the fund due to income from management fees.
One hundred percent of the cash is returned to the investors each year.
Furthermore, fixed-rate debt is preferred by investors and fund managers, but
because most terms end before or at the end of the lifetime, it is possible that the
interest rate will be reset at a certain point during the lifetime. To avoid
complexity, debt is assumed to be fixed-rate debt during the full lifetime. This
assumption might underestimate the variability of the interest rate paid by the
funds. The percentage of development is 0% for core, 20% for value-add, and
60% for opportunistic. These numbers are based on research done by Fuerst
and Matysiak (2010), commissioned by INREV, and are necessary because
developments take longer to attribute to performance and during the development
(which is assumed to take 2.5 years) no income is generated. Value-add funds
tend to promise more additional value by reducing vacancies and redevelopment
activities. Most of this aims to achieve higher rental growth. For this reason, it is
assumed that value-add funds generate 1.5% more NOI growth compared with
core funds. This number is rather arbitrarily chosen, but it will generate a net IRR
in line with what value-add funds, on average, promise to investors. Opportunistic
funds, on the other hand, aim to provide more value growth by, for instance,
developing and buying distressed assets. For this reason, it is assumed that these
funds generate 20% additional value growth over the lifetime compared with core
funds. Again, this number is rather arbitrarily chosen, but it will generate a net
IRR in line with what opportunistic funds, on average, promise to investors. A
cash sweep is triggered when the NAV becomes negative. In reality this can
happen much sooner (or even later if the borrower is still able to pay his interest),
but this assumption is made for simplicity’s sake and because LTV and DSCR
(debt service coverage ratio) covenants are unknown for most funds. This
assumption might underestimate distress during poor market scenarios.
Finally, some assumptions have been made to create a real estate market scenario.
Net operating income (NOI) growth equals inflation, which equals 2%. For
emerging markets, NOI growth and inflation are increased by an additional 2%,
as these markets tend to generate higher economic growth, inflation, and thus NOI
growth. The interest rate is 3.1% as long as the LTV is less than or equal to 50%.
When the LTV is between 50% and 60%, the interest rate is increased by 40 bps;
between 60% and 70%, it is increased by 130 bps; between 70% and 80%, it is
increased by 270 bps; and finally between 80% and 90%, the interest rate is
increased by 550 bps. These spreads are in line with observations in today’s
market (see, for instance, Principal Capital Markets Insights, January 5, 2015).
The cap rate on a property portfolio level is 5.5%, which is similar to the year
end 2014 situation in the U.S. market based on National Council of Real Estate

JRER I Vo I . 39 I No. 3 - 2 0 1 7
3 3 0 van der Spek

Investment Fiduciaries (NCREIF) data, and the average vacancy is 5%. This
vacancy might be on the low side given the average market, but most institutional
funds target the better end of the market and relatively new assets, so it is fair to
assume a somewhat lower figure. No distinction has been made between property
types, so these general assumptions are used for all property types. This
assumption affects funds charging management fees on income; since these are
only a small portion of all funds, a significant impact on the results is unlikely.
Using IRR for analysis purposes can cause problems. Multiple IRRs are possible,
and because the process is iterative, it can take time to calculate the IRR. In
addition, Altshuler and Schneiderman (2011) highlighted the consequences of
using IRR-based instead of preferred return-based performance fees. They proved
that in some cases IRR-based incentive fees are much higher than preferred return-
based incentive fees. Such cases involve investments where capital and profit are
returned before the last capital is called and where an interim promote is paid as
well. Due to the assumptions, this situation cannot occur in this analysis. Even
though there are some issues with using IRR, it is important to use this metric
because it is the most significant metric used by investors to evaluate private fund
performance and by managers to determine performance fees.

The A v e r a g e Fee f o r P r i v a t e Re al E s t a t e F u n d s

It is important for investors to understand the total fee that is to be paid to the
fund manager. The difficulty is there is no real market average with which to
compare this information. Based on the database and modeling of the previous
section, however, it is possible to determine the average total fee load per fund
and establish a market average. Because all fees are calculated given a single
market environment, fee loads between funds are comparable. In reality, the
differences between markets are substantial and funds do not start at the same
time; as such, making a comparison between fees is actually harder. Exhibit 3
presents the average TFL overall and by style, and includes a split between
management and performance fees. The management fee (1.79%) is the biggest
proportion of the TFL, approximately 76%, which is somewhat higher than the
results Metrick and Yasuda (2010) found for private equity. For opportunistic
funds, this ratio is much more skewed towards performance fees, as the
management fee is 1.9% and the performance fee is 1.1%.
The average TFL is 2.36%, which is much lower than the average fee for private
equity, according to several aforementioned studies. Furthermore, the TFL, or even
the management fee load, is significantly higher than the average management
fees as shown in Exhibit 2. The main reason for this difference is that fees are
often not based on equity, while the IRR and the TFL are. Even if the fee on
commitments seems relatively low, on NAV the fee is a lot higher. Assuming that
investments are made gradually over time, fees on commitments can more or less
be doubled to obtain the fee on NAV. Additionally, costs charged in the first year
of an IRR calculation impact an IRR more negatively than when charged in the
 F e e S t r u c t u r e s in P r i v a t e E q u i t y R e a l E s t a t e 3 3 1

E x h i b i t 3 | Total Fee Load (TFL) per Style and Split between M anagem ent and Performance Fee

Core Value Add Opportunistic All funds

■ M anagem ent Fee Load ■ Performance Fee Load

fund’s later stage. Also, leverage will multiply fees on GAV, NOI, GOI, and some
of the active management fees. Lastly, TFL is a combination of multiple fees and
includes active management fees, like acquisition and disposition fees. These fees
should not be ignored by investors, as the impact can be significant. Therefore,
in line with Phalippou’s (2009) conclusions on private equity fees, investors must
look at aggregated figures, like TFL, to understand the full fee impact on return.
So long as it is not standard practice to include such information in the
documentation, investors might be fooled by seemingly lower fees.
As a robustness check to see whether the assumptions made in the cash flow
model are reasonable, a comparison can be made of the fee loads in Exhibit 3 by
style to the historical average difference between gross and net returns. This
information is, unfortunately, only available for the U.S. market through the
NCREIF Townsend Fund returns data series. These data series include open-end
core and value-add funds, which are excluded from the fee load analysis but are
known to charge lower management fees. In addition, the NCREIF data were
calculated on an annual basis, whereas the fee loads are calculated on an IRR—
that is, full lifetime—basis. The comparison is shown in Exhibit 4.
Given the time period and the returns in the final column of Exhibit 4, most core
and value-add funds are unlikely to have achieved their performance targets and
thus performance fees are likely to be low. As a consequence, the historical fee

JRER I Vo I . 39 I No. 3 - 2 0 1 7
3 3 2 van der Spek

E x h i b i t 4 | Total Fee Load |TFL) and Management Fee Load (MFL) for U.S. Fund Compared to Historical
Fee Load and Return in % Based on Rolling Annual NCREIF Townsend Data: 1 9 880:4-—201 2 :Q 4

NCREIF Townsend Gross — Net return NCREIF

TFL, MFL, Townsend
U.S. Only U.S. Only 5 % percentile Average 95% percentile Gross Return

Core 1.4 1.3 0 .8 1.0 1.1 6 .5

Value-add 1.9 1.6 0 .9 1.5 2.8 6 .5

Opportunistic 2 .9 1.7 0 .9 2 .9 8 .4 10 .7

N o t e : The returns are value weighted.

leakage should be compared to the management fee load (MFL). The MFL for
core funds is 1.3%, while the historical fee leakage is 1.0%. For value-add funds,
these numbers are 1.6% and 1.5%, respectively. Opportunistic funds, however, are
more likely to have achieved their targets, as the realized return is over 10%, and
thus some performance fee is likely to be paid. Given these observations, TFL
estimates seem to be close to the empirical results; therefore, the assumptions and
modeling appear very reasonable.
Also shown are the 5% and 95% percentiles to provide a better understanding of
the distribution of fees over the funds’ history. The variation between core funds
is minimal, while it is rather wide for the opportunistic funds. The difference
between gross and net return for value-add funds runs from 0.8% to 2.8%, which
is a significantly smaller range compared with the same range for opportunistic
funds. The variation in TFL between funds within the different styles for the
analyzed database is presented in Exhibit 5. Both core and value-add funds have
some lognormal distribution, while the spread of average TFLs for opportunistic
funds is much wider and can even be as expensive as a private equity fee, as
shown by the literature, with TFL levels of over 5%. Only the core funds
distribution is a lot wider than the presented interval for U.S. funds; this is because
the NCREIF database is dominated by open-end funds, which are rather large.
Besides the distribution, the variation between certain segments is substantial.
Exhibit 6 presents the TFL for different segments. This overview provides clear
indications of which type of fund is more expensive than others. Mixed sectors
and residential funds, for instance, are the most expensive, while industrial and
debt funds are relatively cheap. Funds investing in multiple countries have higher
fee loads than single-country funds. This seems fair, as managing a portfolio in
more than one country is considerably more difficult, complex, and time
consuming. The additional management cost of this is 10 bps.
One of the most common phenomena in the investment industry is the size effect,
and the opposite seems to be the case in fees within the private real estate fund
 Fee Structures in Private Equity Real Estate 3 3 3

E x h i b i t 5 | Distribution o f the Total Fee Load (TFL) by Style

30

E x h i b i t 6 | A verage and Standard Deviation o f the Total Fee Load (TFL), per Segment in %

Average Average
Segment TFL Std. Dev. Segment TFL Std. Dev.

Property Type Vintage Year

Mixed 2.69 1.20 £ 2005 2.11 1.12
Residential 2.41 1.19 2006 2.39 1.09
Office 2.21 1.06 2007 2.56 1.25
Other 2.13 0.67 2008 2.77 1.43
Retail 2.05 1.07 2009 2.44 1.12
industrial 1.76 0.75 2010 2.01 0.98
Debt 1.45 0.40 2011 2.24 0.85
Size in Target Equity 2012 2.29 1.09
£ 250 million 2.40 1.44 >2013 1.84 0.46
>250 and £ 500 2.24 1.16 Catch-up Clause
>500 and £ 1,000 2.48 1.21 Catch-up 2.98 1.23
>1,000 million 2.61 1.32 No catch-up 1.97 1.16
Structure Country Diversification
Closed-end 2.37 1.15 Multi Country 2.44 1.45
Club deal 1.96 1.22 Single Country 2.34 1.18

J RER I Vo I . 39 I No. 3 - 2 0 1 7
3 3 4 van der Spek

market. Large funds (measured by the target equity size) have higher fee loads
than small funds. In particular, funds exceeding 1 billion euros in equity have
higher fees, and the difference can be as much as 37 bps for funds between 250
million and 500 million euros. TFL does not seem stable over time. Based on the
results shown, old funds, funds launched just after the crisis, and the newest funds
are relatively inexpensive, while funds launched during the peak of the market,
2007 and 2008, are the most expensive. Club deals charge lower TFLs (41 bps
on average), which can again be explained by lower complexity of the structure.
Fund managers have less marketing to do, fewer investors to manage, and receive
higher amounts of capital to invest and therefore are satisfied with lower fees.
Finally, the most apparent fee booster based on these segmentations is the catch­
up clause. When a catch-up is included in the fund’s fee terms, the TFL is
increased by over 101 bps. If investors can negotiate for this clause to be excluded,
expenses drop significantly.
Although these segmentations provide some clear guidance in explaining TFL,
they do not provide the complete story, as some of these segments likely overlap.
The difference between small and large funds, for instance, might also be
explained by the investment style. For this reason, it is necessary to control for
these effects, which is done in the next section through a regression model.

Explaining the Private Real Estate Total Fee L oa d

I construct a cross-sectional regression model to analyze the relation between the

calculated TFL and the fund and fee structure characteristics. The explanatory
variables are discussed in the previous section. It is expected that higher
complexity will have an increased impact on fees. So funds investing in multiple
countries or developing economies and opportunistic and value-add funds should
have higher TFLs. In addition, a positive relation is expected between return and
leverage and the TFL and a negative relation with size. Lastly, funds with a catch­
up clause or fees on commitments during the holding period are expected to have
higher fee loads, as these terms generate higher fees for managers. Funds also
have been tested over vintage years, as pre-crisis funds demanded higher fees on
average. However, the result is disappointing: none of the years are significantly
different from the average. The difference between the years is clearly attributable
to, for instance, a change in style and structure rather than to vintage. For this
reason of insignificance and to keep the results accessible and readable, vintages
are excluded from the regression. Successful fund managers might be able to
charge higher fees on consecutive funds, as investors are more willing to provide
(more) capital. Chung, Sensoy, Stem, and Weisbach (2010), for instance, showed
that consecutive funds are likely to be larger and therefore generate more fees.
Unfortunately, the database does not provide sufficient information to include
this into the regression. The model used for the regression is shown in equation
(1).
 Fee S t r u c t u r e s in P r i v a t e E q u i t y R e a l E s t a t e 3 3 5

TFLj = a + /3, NetIRRi + Leverage, + /33Size

+ fi4Club; + Core, + /36ValueAdd:
+ /31MultiCountryi + Developing, + p9Retail,

+ p H)Office, + ft n Residential; + fi ^Industrial,

+ Pi3Debti + p i4OtherSectori + /31SCatchUp,
+ fi[6CommitmentFee t + £,. (1)

The dependent variable in the regression is TFL,, the average total fee load for
fund i. The independent variables are: NetIRR,, which is the net IRR for fund i;
Leverage;, which is the target debt-to-equity ratio, a rather common metric used
in corporate finance literature; and Sizet, which is the logarithm of the fund’s
target equity amount in millions. All other independent variables are dummy
variables. Club;, Core,, and ValueAddi take the value 1 if it is a club deal, core
or value-add fund, respectively. MultiCountry/f takes the value 1 if fund i invests
in more than one country. Developingi takes the value 1 if fund i invests in a
developing (or emerging) economy.2 Retail,, Office,, ResidentialIndustrial,,
Debtj, and OtherSectort take the value 1 if the fund is focused on investing in
retafi, office, residential, industrial, debt, and other sectors of real estate,
respectively. The dummy CatchUpj or CommitmentFeei is 1 if the fund uses one
of these types of fees. Finally, this model is used to estimate the impact of different
func characteristics on the TFL, as well as the management fee load (the
aggregation of fund and active management fees) and the performance fee load.
The results of these regressions are shown in Exhibit 7.
A number of features have a clear impact on the TFL, MFL, and PFL. The net
IRR has a clear positive impact on the PFL but an insignificant negative impact
on the MFL. As there is an effect on the fee load by design, it needs to be included
in the regression. The net effect on the TFL is significantly positive. Leverage is
a clear fee booster. Both the MFL as well as the PFL rise when leverage is
increased. If leverage is 2, which means that the debt is twice as much as the
equity stake, and LTV equals 67%, the TFL is increased by 68 bps (2 X 0.34%).
Over a third of this effect is included in the management fee load. Therefore,
leverage seems to be an easy way for fund managers to increase income. Although
size does not have a significant effect on the TFL, it does have a clear negative
effect on the management fee load. Hence, the bigger the fund, the less the
management fee is needed to cover fund management activities. Interesting to
note, however, is the positive impact on the PFL. Larger funds are apparently able
to demand higher performance fees. What is interesting to note is the contrast to
the previous exhibit, where fee loads seem to be higher for bigger funds. This

JRER I Vo I . 39 I No. 3 - 2 0 1 7
3 3 6 van der Spek

E x h i b i t 7 | Regression Results for the Total Fee Load, Management Fee Load, and Performance Fee Load

TFL MFL PFL

Constant 0.79' 1 .9 3 "' - 1 . 1 3 '"

(1.81) (6.28) (-3.49)
Net IRR 4 .3 2 " -0 .9 6 5 .2 7 '"
(2.26) (-0.72) (3.74)
Leverage 0 .3 4 '" 0 .1 2 "* 0 .2 3 '"
(5.46) (2.66) (4.88)
Log(Size) -0.03 - 0 .2 2 " 0.18*
(-0.26) (-2.36) (1.89)
Club -0 .1 0 -0 .3 0 0.20
(-0.37) (-1.51) (0.93)
Core - 0 .3 7 " -0 .1 4 -0 .2 3 '
(-2.09) (-1.16) (-1.73)
Value-add -0 .4 1 "' 0.06 - 0 . 4 8 '"
(-3.35) (0.75) (-5.25)
Multi country 0.09 0.14* -0.05
(0.84) (1.83) (-0.60)
Developing economy 0 .4 7 '" 0 .3 2 '" 0.15
(3.48) (3.39) (1.50)
Retail -0.11 -0.01 -0 .1 0
(-0.80) (-0.06) (-1.03)
Office 0.12 0.15 -0.03
(0.86) (1.58) (-0.33)
Residential -0.13 0.09 - 0 .2 2 "
(-0.90) (0.98) (-2.15)
Industrial - 0 .4 0 " -0 .1 9 -0.21
(-2.30) (-1.56) (-1.63)
Debt -0 .4 0 -0 .2 7 -0.13
(-1.39) (-1.35) (-0.61)
Other -0.36 -0 .0 7 -0 .3 0
(-1.40) (-0.38) (-1.54)
Catch-up 0 .4 7 '" -0.01 0 .4 8 '"
(4.23) (-0.10) (5.82)
Commitment Fee 0 .4 9 "* 0 .2 6 '*' 0 .2 3 '"
(4.77) (3.58) (3.07)
R2 0.47 0.18 0.50
Adj. R2 0.45 0.14 0.48

Notes: Regression results for the total fee load (TFL), management fee load (MFL), and
performance fee load (PFL), based on a flat market (flat net operating income growth and net
initial yields). Coefficients are in percentages and ^-statistics are in parentheses. Number of
observations is 458.
'Significant at the 10% level.
"Significant at the 5% level.
'"Significant at the 1% level.
 Fee S t r u c t u r e s in P r i v a t e E q u i t y R e a l E s t a t e 3 3 7

higher fee load for larger funds can therefore primarily be explained by factors
other than size. Club deals are structured as closed-end funds and do not show
significantly different fee loads, even though overall fee load is somewhat lower,
caused by a 30 bps lower MFL.
The results also reveal differences between styles. Core funds have a 37 bps lower
TFL than opportunistic funds, driven by a 14 bps lower MFL and a 23 bps lower
PFL. Value-add funds, however, are 41 bps less expensive, which is mainly driven
by a lower PFL. Management fees are more or less the same. Funds investing in
multiple countries or in developing economies are more expensive due to higher
management fees, which makes sense as costs and complexity are higher. For
bigger investors, this may be a good reason to focus on single-country funds, as
investors are able to diversify across countries themselves. When looking at the
different property types, there are a number of interesting observations to make.
Residential funds have a 22 bps lower PFL compared with mixed funds, which is
difficult to explain. Industrial funds on average are 40 bps cheaper. Management
for this type of property is less intensive and complex, and performance fees are
generally lower.
Perhaps the most remarkable result is the impact of a catch-up clause. If the
manager introduces a catch-up, the average TFL is increased by 47 bps, which
can be fully attributed to the performance fee. Though this effect is according to
expectations, the volume is staggering. Almost half a percent of return will be
lost due to a catch-up clause. Fees on commitments do not grow when the NAV
grows and are stable in absolute terms when values fall. Furthermore, this fee is
very high in the first couple of years when only a few investments are done.
Therefore, it makes sense that funds using fees on commitments are more
expensive (49 bps), although the management fee only accounts for half of this
increase. Funds using fees on commitments are also more likely to demand higher-
than-average performance fees.
With the number of variables used, multicollinearity may be a problem that
warrants further investigation. To analyze this potential issue, I use variance
inflation factors (VIF). Generally, a VIF exceeding 4 is seen as a sign that
problems of multicollinearity exist. In this research, the highest VIFs are around
2.7, which are for the core dummy variable and the net IRR. As such,
multicollinearity is not an issue with this regression.
All results are very plausible, and the goodness-of-fit of the regression model is
reasonably high, given R2 is around 50% for the TFL and PFL and almost 20%
for the MFL. The lower R2 for MFL can be explained by the fact that there are
many different types of management fee structures, while performance fees are
rather similar and more homogeneous. However, these regression results are based
on a very stable real estate market scenario in which NOI growth and yield
movements are the same for each year. This is unlikely to occur in reality, and
thus a Monte Carlo simulation has been used to better reflect reality.

JRER I Vo I . 39 I No. 3 - 2 0 1 7
3 3 8 van der Spek

Im pact of C h a nging M arkets on Total Fee Load

In real life, real estate markets are not stable, but instead have great variability.
Fund managers predominantly use IRR models assuming a flat market for their
fund’s placing documents. As this is not a fair reflection of investment markets,
variability must be incorporated into these models; thus, a Monte Carlo simulation
model is suitable for this research. Four market variables are used to simulate the
market: inflation, vacancy, NOI growth, and net initial yield. As this study focuses
on the impact certain fund characteristics have on the average fee load under
different market scenarios and not the exact level of return within a certain
scenario, it is sufficient to model each variable simply. Hence, all variables are
assumed to be normally distributed, as shown in equation (2).

x ~ M > , 2). (2)

Here, x is the vector of the aforementioned four market variables, /x is the mean
vector of these four variables, and 2 is the respective covariance matrix. Variables
of change in vacancy and change in net initial yield are assumed to follow a
random walk and therefore have an average of 0. Inflation and NOI growth are
assumed to have the same average of 2%, which means that rents will follow
inflation in the long run. Two percent inflation is more or less what is currently
expected. The standard deviations of the three real estate variables—change in
vacancy rate, NOI growth, and change in net initial yield—are based on NCREIF
historical data,3 but are increased by 30% since these portfolios are generally less
diversified than the market [see, for instance, Callender, Devaney, Sheahan, and
Key (2007a,b)]. The standard deviation of inflation is based on data retrieved from
Bloomberg. These data also show a negative and significant correlation of -0.45
between NOI growth and the change in vacancy rate, which makes sense as rents
are likely to rise when vacancy rate falls. As a result, a correlation coefficient
between NOI growth and the change in vacancy is included. As no other
correlation is significantly different from 0, the covariance matrix is built
according to equation (3).

2 =

17In f l 0 0 0
0 a D VAC p D V A C .D N O I a D VAC
‘ ' (t d n o i 0
2
0 P D V A C .D N O I ' a D VAC ' (J D N O I 17 r e n t 0
2
0 0 0 <JD N IY

(3)
 Fee S t r u c t u r e s in P r i v a t e E q u i t y R e a l E s t a t e 3 3 9

E x h i b i t 8 | Average M anagem ent Fee Load, Performance Fee Load, Total Fee Load, and N e t IRR

Simulation Model Linear Model

MFL PFL TFL IRR MFL PFL TFL IRR

Core 1.5 2 0.31 1.83 8 .6 1.47 0 .0 3 1.50 9 .2

Value-add 1.9 4 0.71 2 .6 5 10.6 1.86 0 .2 4 2 .1 0 12.0

Opportunistic 1.93 1.1 9 3 .1 3 14 .4 1.87 1.12 2 .9 9 15.2

All funds 1.86 0 .8 4 2 .7 0 11.8 1.79 0 .5 7 2 .3 6 12.8

Notes: Average management fee load (MFL), performance fee load (PFL), total fee load (TFL), and
net IRR in %, based on a Monte Carlo simulation model, reflecting dynamic real estate market
scenarios, and a linear model, reflecting a flat real estate market scenario.

Here, crlNFL is the standard deviation of inflation, aDVAC is the standard deviation
of the change in vacancy rate, crDNOI is the standard deviation of NOI growth,
a DN1Y is the standard deviation of the change in net initial yield, and p dvac,dnoi is
the correlation between the change in vacancy rate and NOI growth. Vacancy rate
and the net initial yield cannot be negative and thus some minimum are set. The
minimum value for the net initial yield is 3.5%, and although this number is set
arbitrarily, there is no historical observation of the average market yield below
these levels and thus this assumption is realistic. The minimum value for the
vacancy rate is set at 0%.
The Monte Carlo simulation generates 50,000 scenarios, in which 15 years are
modeled using the assumed distributions described above. So during this 15-year
period, the variables follow a random walk. Each scenario is applied to each fund’s
cash flow model. Because the main goal is to capture the variability of the fee
load in the market, the analysis focuses on the average impact and not on the
outliers or extreme scenarios. Based on these 50,000 scenarios, the average
management, performance and total fee load per fund were calculated. These
averages are shown in Exhibit 8.
Based on the simulation, the average TFL is 2.7%, 34 bps higher compared to
the linear model. The primary reason for this increase is a surge in the PFL. This
can be explained by option theory, as the performance fee is a kind of call option,
and an increase in volatility increases the value. Core funds are 33 bps more
expensive when adjusting the IRR for changing market circumstances, and this is
also mainly due to an increased PFL. Value-add funds are 55 bps more expensive.
Opportunistic funds, on the other hand, are only 14 bps more expensive. The
reason for this limited increase for opportunistic funds is that in the simulation
results the average IRR is lower for highly leveraged funds; this effect was also

JRER I Vo I . 3 9 I N o . 3 - 2 0 1 7
documented by Van der Spek and Hoorenman (2011). In fact, the average IRR
for opportunistic funds in the simulation study is 0.8% lower compared to the
linear model.

For a better understanding of the results, the average fee loads generated by the
simulation are used as inputs for the regression analysis. This provides a more
robust way to understand the different drivers of fees. The regression was done
in a similar manner as the one before, and the results are shown in Exhibit 9.
Again, no significant effect is found for vintages, so they are excluded from the
regression analysis.
The results are similar to the results based on a flat market, although there are
some clear differences. In the MFL regression, the IRR coefficient became positive
and significant. This means a strong return will have a positive impact on
management fees, which is in line with expectations as some of the management
fees are based on value and thus linked to performance. Another remarkable
difference is the effect of leverage on each type of fee load. This effect is much
stronger than in the linear scenario and is driven by the increased option value of
carried interest due to the increase in volatility. Leverage of 50% (which equals
an equity-to-debt multiple of 1) implies an increase in TFL of 54 bps and not the
34 bps in the linear scenario. Therefore, if investors solely rely on one scenario
provided by the placing document, they will underestimate the effect of leverage
on fees. The size effect is still present but only has an effect on the management
fees, which is in line with expectations; larger funds charge lower management
fees. Another interesting observation is that on average club deals place
management fees, while the performance fee is somewhat higher. This is not
surprising since more emphasis is placed alignment, but this effect is marginally
insignificant since the observation amount is low.
Styles only differ on the performance fee, where unmistakably opportunistic funds
charge the highest performance fees. Nevertheless, value-add funds charge the
highest, nearly significant, management fees. The most remarkable observation is
that it cannot be concluded that closed-end core funds have lower management
fees. This is contrary to the amount of work that is involved and the expectation,
as value-add and opportunistic funds require more work on the assets. Other fee
drivers, like lower leverage and return, make average fees lower, but the style
itself has no impact. The variable that became insignificant was the multi-country
dummy. Apparently the relation is not as strong as the initial results suggest,
although the effect is still positive. Another relation that is weaker than initially
indicated is the effect of investing in developing economies. The TFL coefficient
is reduced by a third to 30 bps. Across the different sectors, retail and residential
stand out. Retail has a significantly lower performance fee, which is
understandable since this sector is a more core and long-term investment compared
to other sectors. In contrast to initial findings, residential stands out, as none of
the fee loads is significantly lower, which is more in line with expectations. The
effect of a catch-up, based on simulated scenarios, is less severe, although the
 Fee S t r u ct ur e s in Pr ivate Equit y Real Estate 3 4 1

E x h i b i t 9 | Regression Results on the Average Total Fee Load, Management Fee Load, and Performance
Fee Load

TFL MFL PFL

Constant 0 .66 * 1.49 *** - 0 .80 ***

( 1.89 ) (4 .86 ) ( - 3 .85 )
Average Net IRR 6 .40 “ * 1.44 * 4 .99 ***
(6 .83 ) ( 1.75 ) (8 .94 )
Leverage 0 .54 *** 0 . 18*** 0 .36 “ *
( 11.29 ) (4 .22 ) ( 12.45 )
Log(Size) - 0.15 - 0 .20 ** 0.05
( - 1.40 ) ( - 2 . 13) (0 .73 )
Club - 0.15 - 0.33 0.18
( - 0 .63 ) ( 1.61 ) ( 1.29 )
Core - 0.16 - 0.01 - 0 . 15*
( - 1. 19) (0 .07 ) ( - 1.90 )
Value-add - 0.06 0.14 - 0 . 19***
( - 0 .64 ) ( 1.60 ) ( - 3 .32 )
Multi country 0.11 0.11 - 0.01
( 1.27 ) ( 1.47 ) ( - o. i l )
Developing economy 0 .30 *“ 0 .22 " 0.07
(2 .78 ) (2 .34 ) ( 1.04 )
Retail - 0.16 - 0.02 - 0 . 14*
( - 1.37 ) ( - 0 .21 ) ( - 1.98 )
Office 0.16 0.16 0.01
( 1.47 ) ( 1.60 ) (0 . 16)
Residential 0.08 0.12 - 0.03
(0 .72 ) ( 1. 17) ( - 0 . 39 )
Industrial - 0 .27 * - 0.20 - 0.06
( - 1.82 ) ( - 1.57 ) ( - 0 .69 )
Debt - 0.23 - 0.23 0.00
( - 0 .95 ) ( - 1.09 ) (0 .01 )
Other - 0.23 - 0.03 - 0.19
( - 1.04 ) ( - 0 . 14) ( - 1.46 )
Catch-up 0 .27 *** 0.04 0 .22 ***
(2 .97 ) (0 .52 ) (4 .00 )
Commitment Fee 0 .46 " * 0 .24 *** 0 .22 ***
(5 .44 ) (3 .23 ) (4 .39 )
R2 0.52 0.18 0.57
Adj. R2 0.50 0.15 0.55

Notes: Regression results on the average total fee load (TFL), management fee load (MFL) and
performance fee load (PFL), based on a Monte Carlo simulation of the real estate market
(simulating inflation, vacancy, NOI growth and net initial yields). Coefficients are in percentages
and t-statistics are shown in parentheses.
‘ Significant at the 10% level.
“ Significant at the 5% level.
“ ‘ Significant at the 1% level.

JRER I Vo I . 39 I No. 3 - 2 0 1 7
 E x h i b i t 1 0 | Average M anagem ent Fee Load, Performance Fee Load, and Total Fee Load

# of Obs. MFL PFL TFL

LTV = 0% 15 1.60 0.12 1.72

0% < LTV s 40% 74 1.54 0.43 1.97
40% < LTV < 50% 95 1.79 0.68 2.47
50% < LTV < 60% 110 1.86 0.97 2.84
LTV > 60% 119 2.14 1.19 3.34

Notes: Average management fee load (MFL), performance fee load (PFL), and total fee load (TFL),
based on a Monte Carlo simulation model per loan-to-value (LTV) cluster in %.

impact is an average increase of 22 bps on performance fees. One can argue that
top-quartile managers (i.e., managers who achieve higher returns on average
compared to their peers) are more likely to have a catch-up, as investors are more
likely to accept this type of arrangement with these managers. The additional
return they achieve could offset the additional 22 bps of fees. To analyze this, I
attempted to incorporate and link additional return data from the industry
platforms INREV and ANREV. Unfortunately, the combined data resulted in a
sample too small to properly link additional returns and catch-up fees.
Nevertheless, this is an interesting topic for further investigation.
Finally, it is interesting to note that the sum of the MFL and PFL coefficients is
more or less similar to the TFL coefficient, which would make perfect sense when
most of the TFL is explained. Furthermore, R2 slightly improved in the simulated
scenarios. For this regression, I check for a potential problem with
multicollinearity. The highest VIF in this regression is only 2.3 for the core
dummy variable, while all other VIFs are below 2. Therefore, there is no problem
with multicollinearity in this regression.
The impact of leverage on the TFL can also be shown by clustering each fund in
five ascending LTV clusters. As shown in Exhibit 10, a clear relation exists
between leverage and fee load for private real estate vehicles where management
typically is externalized. Therefore, increasing leverage creates a direct boost to
the total fee load, implying more profits for the manager. MFL and PFL increase
substantially as a result of an increase in leverage. Without leverage the average
fee load is 1.7%, but with a leverage of over 60%, it can be as high as 3.3%.
Because a fee is dependent on the return, it is interesting to group the results by
return and see whether the coefficients and parameters are stable for different
return levels. These return groups are shown in Exhibit 11, which includes the
 Fee S t r u c t u r e s in P r i v a t e E q u i t y R e a l E s t a t e 3 4 3

E x h i b i t 11 | Total Fee Load Simulation Results

G roup N e t IRR Interval % o f Obs. A verage TFL Std. Dev. TFL

G roup 1 N et IRR < 0% 6 .8 2 .4 3 1.3 2

G ro u p 2 0% s N e t IRR < 7.5% 19.0 1.91 0 .6 6

G ro u p 3 7.5% =s N et IRR < 15% 3 8 .2 2 .1 9 0.81

G ro u p 4 15% < N et IRR < 2 2 . 5 % 2 6 .6 3 .3 4 1.24

G ro u p 5 N e t IRR > 22.5% 9 .4 4 .2 2 1.72

A ll — 100 2 .7 0 1.00

Notes: Total fee load (TFL) sim ulation results grouped by return, including distribution, average TFL
and the standard deviation o f the TFL in %.

percentage of observations in each group, the average TFL per group, and the
standard deviation of the TFL within the group. It is remarkable that the TFL in
the lowest return group is higher than in the second group, meaning the loss of
IRR due to fees is higher when returns are negative than in the case when returns
are modestly positive. This can be explained by the fact that NAVs will be a lot
lower in these scenarios; as a result, fees on commitments or invested equity are
relatively high compared to NAV. Consequently, the variation in the TFL is a lot
higher as well, resulting in a standard deviation of 1.3. Furthermore, not
surprisingly, the variation in the TFL increases as the return increases.
Exhibit 12 shows the differences between styles within these IRR groups. It is
remarkable that the difference between opportunistic and value-add funds is rather
low. Core, on the other hand, is significantly lower for all returns. All styles have
higher TFLs in the lowest return group.
Also remarkable is the lower TFL for opportunistic funds compared with value-
add funds in the highest return group. The reason behind this is that value-add
fund fees are mostly not structured for these high levels of performance, while
opportunistic fund fees are. Therefore, fee loads seem very high, while in practice
they are unlikely to be achieved. Finally, the TFLs per group are used in the
regression model; the results are displayed in Exhibit 13. Most coefficients are
rather stable across return groups, although some clear differences exist in the
lowest and highest return groups. Leverage, for instance, slowly increases for each
higher return group but has the biggest impact on fees in a negative return
environment. Investors are thus confronted with a double negative impact with too
much leverage, as it amplifies the negative return and increases the fee load, yet
another reason for investors to be wary of using much leverage.

JRER I V o I . 39 I N o . 3 - 2 0 1 7
3 4 4 van der Spek

E x h i b i t 12 [ A v e ra g e Total Fee Load p e r Return G ro u p b y Style a n d the A v e ra g e TFL b y Style B ased on the

S im ulatio n Results

Given a certain return, there is little difference in the TFL between styles. Core
funds are consistently less expensive but only significantly lower when returns are
negative, while value-add fee structures additionally increase fees when returns
are high. In contrast, it is more likely that opportunistic fund returns are higher.
When the return exceeds the hurdle, the catch-up is operational and thus has a
large significantly positive impact on the TFL in groups 3 and 4. The additional
fee amount due to this catch-up can be as much as 84 bps, which is substantial.
Interestingly, the effect is strongly negative in group 5. This can only be explained
by the catch-up that is already fully paid out in group 5 and the average additional
carried interest that is higher for funds without a catch-up clause (average carried
interest of 22.8%) compared to those with a catch-up clause (20.4% on average).
Finally, funds with commitment fees are consistently more expensive.
 Fee S t r u c t u r e s in P r i v a t e E q u i t y Real Es t at e 3 4 5

E x h ib it 13 | Regression Results on the A verage Total Fee Load per Return G roup

Group 1 Group 2 Group 3 Group 4 Group 5

Constant 0.32 1.76“ * 1.62*** 2 .31*** 3.86***

(0.64) (6.58) (5.37) (5.14) (5.17)

Leverage 0 .7 6 “ * 0 .1 2 “ * 0 .1 8*** 0 .1 9*** 0.20*

(10.26) (3.14) (4.13) (2.87) (1.77)
Log(Size) 0.12 - 0 .1 7 * -0 .1 0 -0 .0 6 - 0 .1 4
(0.70) ( -1 .7 8 ) (-1 .0 0 ) (-0 .3 9 ) (-0 .5 4 )

Club - 0 .5 6 - 0 .3 0 -0 .2 7 -0 .0 9 - 0 .1 8
(-1 .4 9 ) (-1 .4 9 ) (-1 .1 7 ) (-0 .2 7 ) (-0 .3 3 )

Core - 0 .3 4 * -0 .1 7 - 0 .1 5 -0 .2 2 -0 .2 4
(-1 .7 3 ) (-1 .5 9 ) (-1 .2 8 ) (-1 .2 4 ) (-0 .7 9 )

Value-add -0 .1 7 0.05 -0 .0 0 0.03 0.12

(-1 .1 4 ) (0.62) (-0 .0 4 ) (0.21) (0.55)

Multi country 0.02 0.13* 0.13 0.22* 0.29

(0.13) (1.72) (1.59) (1.78) (1.33)
Developing economy 0.12 0 .29*** 0 .2 2 “ 0.08 0.43*
(0.71) (3.36) (2.23) (0.54) (1.78)
Retail -0 .1 7 0.00 -0 .0 7 -0 .1 3 - 0 .2 3
(-0 .9 1 ) (0.02) (-0 .6 1 ) (-0 .8 1 ) (-0 .8 4 )

Office 0.16 0.15 0.14 0.20 0.37

(0.91) (1.61) (1.34) (1.28) (1.41)
Residential -0 .0 1 0.11 -0 .0 8 -0 .1 2 0.27
(-0 .0 5 ) (1.16) (-0 .7 5 ) (-0 .7 0 ) (1.00)

Industrial -0 .3 0 - 0 .2 0 - 0 .2 8 * * -0 .1 4 0.05
(-1 .2 8 ) (-1 .5 7 ) (-1 .9 8 ) (-0 .6 5 ) (0.14)

Debt -0 .0 6 -0 .2 5 - 0 .3 5 - 0 .4 8 -0 .3 0
(-0 .1 5 ) (-1 .2 4 ) (-1 .5 2 ) (-1 .3 9 ) (-0 .4 3 )
Other 0.05 - 0 .0 2 - 0 .1 3 -0 .0 8 - 0 .6 3
(0.16) (-0 .1 2 ) (-0 .6 1 ) (-0 .2 6 ) (-1 .1 7 )
Catch-up 0.18 0.01 0 .4 2*** 0 .84*** - 1 .1 8 * * *
(1.27) (0.18) (4.93) (6.58) (-5 .6 5 )
Commitment Fee 0.17 0 .2 9 “ 0 .36*** 0 .6 6 “ 0 .83***
(1.29) (3.97) (4.45) (5.37) (4.11)
Rs 0.30 0.20 0.32 0.36 0.13
Adj. R2 0.27 0 .1 7 0.29 0.33 0.10

Notes: Regression results on the average total fee load (TFL) per return group, based on a Monte
Carlo simulation of the real estate market (simulating inflation, vacancy, NOI growth, and net
in tia l yields). Coefficients are in percentages and f-statisties are in parentheses. In Groups 1 -4 ,
there are 413 observations; in Group 5, there are 383 observations.
‘ Significant at the 10% level.
“ Significant at the 5% level.
‘ "S ig n ific a n t at the 1% level.
|
________________________ ____ ____ ____________________ ___________

JRER | V oI . 39 I No. 3 - 2 0 1 7
3 4 6 van der Spek

Conclusion

This paper aims to improve the transparency of the rather opaque world of private
equity real estate fees. The analysis is based on a unique investor’s private equity
real estate fund database containing ex ante fee structures, as recorded in
corresponding placing documents. Through IRR modeling and simulation, I
demonstrate that the average total fee load for private equity real estate funds is
2.7%, which is clearly lower than the average private equity fund fee load.
Furthermore, there are substantial differences between fund styles. These
differences, however, cannot be explained by a difference in the level of
management fees, which conflict with the impression that more work needs to be
done managing value-add and opportunistic assets. Rather, performance fees and
other fee drivers are responsible for such differences. Core and value-add funds
charge lower performance fees compared with opportunistic funds. As in most
other financial studies, size is an indicator that predominantly impacts
management fees: the bigger the fund, the lower the management fee on average.
Funds investing in the retail sector on average have a 14 bps lower performance
fee, reflecting the more long-term characteristic of this property type. The 27 bps
lower total fee load for industrial funds, on the other hand, is primarily a reflection
of the lower management intensiveness of this property type. Funds investing in
developing countries tend to have a 30 bps higher total fee load, mainly as a result
of a 22 bps higher management fee; this is a reflection of the relative amount of
work that needs to be done for this type of strategy. Complexity has its price.
The most striking features affecting fee loads are the commitment fee, leverage,
and catch-up clause. Funds applying a commitment fee are on average 46 bps
more expensive, as the fee is being paid over capital that is not yet invested and
thus not providing any return. Leverage increases management and performance
fees, even when controlling for the increase in return. Putting a 50% LTV in place
means that the total fee will increase by 54 bps on average. For fund managers,
this looks like an easy way to improve their profitability, while investors should
be aware of this effect and limit the amount of leverage. In negative market
scenarios, the leverage effect is even stronger. The other easy way for fund
managers to boost profits is to introduce a catch-up clause. Funds charging a catch­
up are on average 22 bps more expensive than other funds, although no additional
work is involved. In normal or somewhat more positive market scenarios, this
effect can be as high as 84 bps. Private equity real estate fund investors should
therefore take this impact into consideration during negotiations and only accept
this feature with top-quartile managers.
Finally, it must be mentioned that the total fee load does not tell the full story.
Fund expenses and other costs are excluded from this research, as these are not
well documented in placing documents. More transparency is needed to include
these costs in research and to determine and analyze the total expense ratio.
Moreover, the placing documents hold fee structures before negotiations, and
hence (larger) investors, in reality, will have lower fee loads due to negotiations.
 Fee S t r u c t u r e s in P r i v a t e E q u i t y Real Est at e 3 4 7

Finally, further research should be done on actual return data to analyze these fee
structures. Currently, there is not enough data available to perform such an
analysis. Nonetheless, fee structures should be a prominent topic on the agenda
when investors perform due diligence. This research reveals some features that
impact total fee load, making it easier for investors to negotiate fair terms.

Endnotes
1 Definitions are from INREV.
2 Definition according to IMF.
3 NCREIF Property Index (NPI) data are used for the period 1988-2012.

References

Adams, J.C., S.A. Mansi, and T. Nishikawa. Are Mutual Fund Fees Excessive? Journal of
Banking & Finance, 2012:38, 2245-59.
Altshuler, D. and R. Schneiderman. Overpayment of Manager Incentive Fees— When
Preferred Returns and IRR Hurdles Differ. Journal o f Real Estate Portfolio Management,
2011, 17:2, 181-89.
Ambrose, B.W. and P. Linneman. REIT Organizational Structure and Operating
Characteristics. Journal o f Real Estate Research, 2001, 21:3, 141-62.
Andonov, A., P. Eichholtz, and N. Kok. A Global Perspective on Pension Fund Investments
in Real Estate Funds Market. Journal o f Portfolio Management, 2013, 39:5, 32-42.
Callender, M., S. Devaney, A. Sheahan, and T. Key. Risk Reduction and Diversification in
U.K. Commercial Property Portfolios. Journal o f Property Research, 2007a, 24:4, 355-75.
------. Risk Reduction and Diversification in Property Portfolios. Investment Property
Forum, 2007b.
Capozza, D.R. and PJ. Seguin. Debt, Agency, and Management Contracts in REITs: The
External Advisor Puzzle. Journal o f Real Estate Finance and Economics, 2000, 20:2, 91-
116.
Chung, J-W., B.A. Sensoy, L.H. Stem, and M.S. Weisbach. Incentives of Private Equity
General Partners from Future Fundraising. NBER Working paper, 2010.
------. Pay for Performance from Future Fund Flows: The Case of Private Equity. Review
o f Financial Studies, 2012, 25:11, 3259-3304.
Eisenhardt, K.M. Agency Theory: An Assessment and Review. The Academy o f
Management Review, 1989, 14:1, 57-74.
Fisher. L.M. and D.J. Hartzell. Real Estate Private Equity Performance: A New Look.
Working paper, UNC Chapel Hill, 2013.
Fuerst. F. and G. Matysiak. Testing and Validating the INREV Fund Styles Framework.
Report for INREV, 2010.
Gompers, P. and J. Lemer. An Analysis of Compensation in the U.S. Venture Capital
Partnership. Journal o f Financial Economics, 1999, 51:1, 3-44.
Harris, R.S., T. Jenkinson, and S.N. Kaplan. Private Equity Performance: What Do We
Know? Journal o f Finance, 2014, 69:5, 1851-82.

JRER I Vo I . 39 I No. 3 - 2 0 1 7
348 v a n d e r S p e k

Litvak, K. Venture Capital Limited Partnership Agreements: Understanding Compensation

Arrangements. University of Chicago Law Review, 2009, 76:1.
Metrick, A. and A. Yasuda. The Economics of Private Equity Funds. Review of Financial
Studies, 2010, 23:6, 2303-41.
Phalippou, L. Beware of Venturing into Private Equity. Journal of Economic Perspectives,
2009, 23:1, 147-66.
Phalippou, L. and O. Gottschalg. The Performance of Private Equity Funds. Review of
Financial Studies, 2009, 22:4, 1747-76.
Principal Real Estate Investors. Capital Markets Weekly, March 2, 2015.
Robinson, D.T. and B.A. Sensoy. Do Private Equity Fund Managers Earn their Fees?
Compensation, Ownership, and Cash Flow Performance, Charles A. Dice Center Working
Paper No. 2011-14, 2011.
Spek, M.R. van der and C. Hoorenman. Leverage: Please Use Responsibly. Journal of Real
Estate Portfolio Management, 2011, 17:2, 75-88.

I would like to thank the participants at the Annual Conference o f the European Real
Estate Society in Vienna, J.L. Pagliari and P. Hayes fo r their useful comments, and
D. Brounen in particular fo r his useful feedback and constructive discussions.
Furthermore, / am especially grateful to PGGM fo r their continuous support.

Maarten van der Spek, PGGM Investments, Zeist, the Netherlands or mrvdspek@
gmail.com.
Copyright of Journal of Real Estate Research is the property of American Real Estate Society
and its content may not be copied or emailed to multiple sites or posted to a listserv without
the copyright holder's express written permission. However, users may print, download, or
email articles for individual use.