Energy Policy 46 (2012) 489–497

Contents lists available at SciVerse ScienceDirect

Energy Policy
journal homepage: www.elsevier.com/locate/enpol

The impact of energy labels and accessibility on office rents

Nils Kok a,n, Maarten Jennen b
a
Maastricht University, The Netherlands
b
RSM Erasmus, CBRE Global Investors, The Netherlands

a r t i c l e i n f o abstract

Article history: Energy consumption in the commercial property sector offers an important opportunity for conserving
Received 23 May 2011 resources. In this study, we evaluate the financial implications of two elements of ‘‘sustainability’’
Accepted 10 April 2012 – energy efficiency and accessibility – in the market for commercial real estate. An empirical analysis of
Available online 19 April 2012
some 1100 leasing transactions in the Netherlands over the 2005–2010 period shows that buildings
Keywords: designated as inefficient (with an EU energy performance certificate D or worse) command rental levels
Energy efficiency that are some 6.5 percent lower as compared to energy efficient, but otherwise similar buildings
Commercial real estate (labeled A, B or C). Furthermore, this study shows that office buildings in multi-functional areas, with
Valuation access to public transport and facilities, achieve rental premiums over mono-functional office districts.
For policymakers, the results documented in this paper provide an indication on the effectiveness of the
EU energy performance certificate as a market signal in the commercial property sector. The findings
documented here are also relevant for investors in European office markets, as the importance of
energy efficiency and locational diversification is bound to increase following stricter environmental
regulation and changing tenant preferences.
& 2012 Elsevier Ltd. All rights reserved.

1. Introduction Part of the return to energy efficiency improvements consists

of relatively predictable energy savings, but under standard lease
Energy efficiency in the built environment can play an impor- contracts and in multi-tenant buildings, these typically flow to
tant role in the reduction of global carbon emissions (Stern, 2008). the occupants. For investors, the return is thus uncertain, con-
This has been recognized by the introduction of the Energy sisting of better marketability of properties (e.g., lower vacancy
Performance of Buildings Directive (EPBD) in January 2003, risks, higher rents while keeping total housing costs fixed, shorter
leading to the implementation of energy performance certificates rent-free periods) and higher valuations (following lower cap
for residential and commercial buildings across the European rates and reduced depreciation). The implementation of energy
Union (see Andaloro et al., 2010, for a comparative analysis of performance certificates can be regarded as an additional step
progress towards implementation of the EPBD). towards transparency of energy consumption in buildings,
Even though the global financial crisis and subsequent enabling private and corporate occupiers to take energy efficiency
downturn in commercial property markets have not dented the into account when making housing decisions. Indeed, recent
ever-increasing interest in the ‘‘sustainability’’ of the built envir- evidence shows that the EU energy label is effective as a signaling
onment, the impact of energy efficiency and sustainability on real device in the residential housing market (Brounen and Kok, 2011).
estate investment performance remains a heavily debated sub- When it comes to commercial real estate, evidence on the
ject, and many (institutional) investors are reluctant to invest in adoption and evaluation of energy labels is scant. Existing ‘‘green’’
energy efficiency measures and retrofitting of existing properties. labels, such as BREEAM in the UK and the US Green Building
Of course, financing and liquidity constraints also play a role, Council’s LEED, offer the opportunity to assess the financial
especially given current macroeconomic conditions, but the lack impact of ‘‘sustainability’’ on real estate assets, but most research
of evidence on the returns to energy efficiency improvements is US-based. Studies of the European property market have thus
continues to be one of the most important barriers to energy far been hindered by the slow diffusion of heterogeneous labeling
efficiency investments. schemes and a lack of centralized transaction data.
This paper is the first to offer systematic insight in the effects
of the main components of sustainability – energy efficiency and
accessibility – on realized rents in the European office market,
n
Corresponding author. Tel.: þ31 433883838; fax: þ31 433884875. using the Netherlands as a case study. We create a unique dataset,
E-mail address: [email protected] (N. Kok). combining the databases of the three largest real estate agents in

0301-4215/$ - see front matter & 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.enpol.2012.04.015
490 N. Kok, M. Jennen / Energy Policy 46 (2012) 489–497

the Netherlands, CBRE, DTZ Zadelhoff and Jones Lang LaSalle, of energy efficiency investments in real estate. The costs of the
thereby overcoming the problem of lack of centralization in data various services are bundled and paid out of the energy savings,
gathering. The real estate transaction data is combined with the with the actual units of energy saved often guaranteed by the
database of energy performance certificates (EPCs) for buildings, ESCO. Studies of the Lawrence Berkeley National Lab show energy
maintained by the Dutch Ministry of Economic Affairs. EPCs are savings of about 15–20 percent, with payback periods varying
based on energy performance assessments and mandatory for between 5 and 15 years (depending on the implemented technol-
every real estate transaction. Labels vary from G, for inefficient ogy, see Goldman et al., 2005, for a discussion). However, the
properties, to Aþþ for highly efficient buildings. ESCO market is mostly focused on semi-public real estate
Based on an empirical analysis, correcting for the most (e.g., schools, universities, and hospitals), which has the advan-
important value drivers in real estate – location, age and size of tage of being owner-occupied, thereby reducing concerns about
office properties – we document for a sample of some 1100 rental the split-incentive between tenants and building owners in the
transactions in the Netherlands that properties designated as less reduction of energy consumption. Furthermore, academic evi-
efficient (with energy labels D or lower) yield significantly lower dence on the risk-return characteristics of ESCO-investments in
rents. We also find that the rental growth of energy-efficient the energy efficiency of commercial real estate is limited.
offices deviates strongly from rental growth of inefficient offices, The literature on the implications of energy efficiency invest-
a development starting in 2009. Furthermore, the analysis shows ments for building owners is more developed, even though most
that both the distance to train stations and the ‘‘walkability’’ of studies focus on the US commercial property market. Recent
offices (i.e., the location of an office relative to facilities such as research documents a positive effect of sustainability and/or
restaurants and shops) have an important impact on rental prices energy efficiency labels on rents and prices of office properties:
in the office market. this holds pre-crisis (Eichholtz et al., 2010; Fuerst and McAllister,
The results documented in this research provide the first 2011) as well as during the financial crisis (Eichholtz et al., in
credible evidence that ‘‘sustainability’’ (or: energy efficiency and press). An important finding is the quite precise relation between
accessibility) matters for corporate tenants in the European real the extent of energy efficiency and rents and prices: tenants pay
estate market, corroborating with existing evidence on the US about 95 cents for a dollar in energy savings and investors
office market (Eichholtz et al., 2010) and the Dutch residential capitalize lower energy costs with a discount rate of 8 percent
market (Brounen and Kok, 2011). These findings have implica- (just slightly higher as compared to current cap rates in some US
tions for the portfolios of real estate investors: energy efficiency office markets).
affects rental levels and rental growth. Accessibility is priced as The market capitalization of energy efficiency and sustain-
well. Both components of sustainability have an impact on the ability in Europe has been assessed empirically for the residential
value of ‘‘green’’ as well as ‘‘non-green’’ commercial properties. market. A recent study by Brounen and Kok (2011) evaluates the
For policymakers, the results documented in this paper provide dissemination and market valuation of EU energy labels (EPCs) in
an indication on the effectiveness of the EU energy performance the Dutch housing market and finds that, even though the
certificate as a market signal in the commercial property sector. adoption rate of energy labels is low, the relative energy efficiency
The remainder of this paper is organized as follows: Section 2 of private dwellings has a significant impact on selling prices.
provides some background on the literature examining the Market evidence on willingness to pay for ‘‘green’’ real estate
financial implications of ‘‘sustainability’’ in buildings. Section 3 in the European commercial property market is mostly anecdotal:
discusses the EU energy performance certificate and its diffusion Jones Lang LaSalle and CoreNet Global frequently survey corpo-
in the Dutch office market. Section 4 describes the financial data rate tenants and investors regarding their desire for more sustain-
and presents the analysis, relating the ‘‘sustainability’’ of a large able office space. According to the March 2010 survey ‘‘y83
sample of office buildings to realized rents. Section 5 is a brief percent [of the tenants] is willing to pay a rental premium, if this
conclusion and provides some implications for policy makers. is reflected in tangible benefits’’ (Jones Lang LaSalle, 2010). The
willingness to pay for ‘‘green’’ is partly determined by social
factors but of course also by the cost reduction potential.
2. Background literature In order to reduce total housing costs, i.e., the combined cost of
rent and service and energy charges, the tenant will, ceteris
The influence of sustainability and energy efficiency on the paribus, prefer energy efficient buildings over non-efficient alter-
financial performance of commercial property investments is natives (the tenant pays either directly or indirectly for energy
mostly a topic of speculation, rather than being subject to consumption). However, it has been argued that the structure of
rigorous empirical evaluation. One can make a distinction lease contracts may influence energy-saving behavior. The CBRE
between the direct implications of investments in the energy Global Office Occupier Guide shows that in the United States
performance of buildings and the indirect effect that energy there are three ways in which the tenant generally pays for
efficiency may have on building rents and prices. The expected electricity usage. In some, mostly single-tenant buildings, a meter
return from investments in building energy efficiency, following is installed and the tenant pays directly to the utility company. A
from cost savings, is often thought off as clear-cut: it is claimed second possibility is that the landlord pays the utility company,
that the real estate sector represents the largest opportunity to but that the tenant then compensates the landlord based on
reduce carbon emissions at attractive returns (Enkvist et al., actual usage measured by a submeter. A third option is that
2007). It is also assumed that building improvements are gen- electricity expenses are included in the service charges on a price
erally NPV positive, with varying discount rates for different types per square foot basis. Some electricity-intensive tenants, e.g., data
of consumers (Jaffe and Stavins, 1994). The question remains why centers, would pay an additional amount to cover the above-
private investments in building energy efficiency have thus far average usage of electricity. The last type of energy expense
not taken off at a larger scale (this is also known as the ‘‘energy billing may induce moral hazard issues in multi-tenant buildings,
efficiency paradox,’’ see Kok et al., 2011, for a discussion). as additional usage is spread over several tenants.
Recent micro-economic studies investigating the return on On the European continent the options for energy cost charges
building retrofit investments are mostly focused on the US are similar to the United States and the nature of compensating
‘‘ESCO’’ market—Energy Service Companies (ESCOs) are specia- the landlord for energy costs does not differ substantially across
lized in the development, installation, financing and maintenance countries. In France, Germany, Spain and the United Kingdom it is
 N. Kok, M. Jennen / Energy Policy 46 (2012) 489–497 491

most common that the tenant signs a contract with an electricity building characteristics of all commercial buildings with an
supplier directly for the floorspace occupied by the tenant and energy performance certificate. As of December 2011, more than
pays for water and electricity charges for common areas through 10,500 buildings (including retail, office, sports facilities, schools,
the service charges. In the Netherlands, it is more common that etc.) had been certified. The total square footage of these build-
energy and water costs are covered in the service charges per ings represents some 30 million square meters, with offices
square meter with annual settlement based on actual usage by all accounting for 68 percent (20.6 million square meters). The
tenants combined at the end of the year. If available the tenant energy label database covers about 44 percent of the total Dutch
pays for actual usage based on a meter. A sign of technological office market (estimated at 47 million square meters).
advances and more energy usage awareness is that newer build- Information available for each property includes address, year
ings in a warm country like Spain are equipped with air- of construction, year of renovation, surface area and energy
conditioning equipment separately for each tenant, whereas characteristics (i.e., energy index and energy label).
cooling charges in older buildings are still based on per square
meter service charges. Technological advances and an increase in 3.1. Energy performance certificates: age and size
individual metering will limit moral hazard issues and allow
energy conscious tenants to reduce total housing costs by actively Fig. 1A shows the composition of energy labels for different
monitoring heating and cooling settings. construction periods. Mainly due to the fast growth in office space
Of course, in market equilibrium, all types of rental contracts during the 1991–2000 period, this era has the highest representa-
will ultimately lead to similar total housing costs, despite differ- tion in the database. Quite clearly, technological developments and
ences in the way that energy costs are paid. Tenants focusing on increasingly strict building codes have lead to more efficient con-
total housing costs will thus favor energy efficient buildings over struction over the past decades. The number of least efficient office
similar non-efficient buildings, but the question remains how buildings (energy label G) is the largest in the category ‘‘Constructed
survey-based evidence on willingness to pay for energy efficiency before 1940.’’ The absolute number of G-labeled properties remains
translates into occupancy rates, rents and prices of commercial real relatively constant until 1990, but the share of inefficient office
estate in practice. Recent evidence of DTZ Zadelhoff documents buildings decreases due to the rise of newly constructed offices.
that, based on appraisals of 150 offices in the Netherlands, energy Even though A-labeled buildings represent about 50 percent of all
labels are positively correlated with building value (DTZ Zadelhoff, construction during the 2001–2010 period, Fig. 1A once again
2011). A similar study of Troostwijk Real Estate, a consultancy, reinforces the challenge facing society—dealing with the heritage
compares rents and prices of office properties with ‘‘green’’ labels of the past. The existing building stock consists mostly of inefficient,
and ‘‘non-green’’ labels and also documents a positive relation
between energy efficiency and values (Troostwijk Real Estate,
2011). The results of both studies show substantial differences in
rental levels and discount rates between efficient and inefficient
office buildings. We note, however, that these studies do not fully
account for crucial determinants of commercial building value, such
as location and year of construction. Omitting these factors poten-
tially affects the results: energy efficient buildings are often
recently constructed, which affects their marketability.

3. The EU Energy Performance Certificate (EPC)

The European Union implemented the Energy Performance of

Buildings Directive (EPBD) in January 2003 with the explicit goal
of promoting energy performance improvements in buildings. The
Directive, which was recently recast, includes an explicit element
on the disclosure of energy performance in buildings: ‘‘yMember
states shall ensure that, when buildings are constructed, sold or
rented out, an energy performance certificate is made available to
the owner or by the owner to the prospective buyer or tenant
(Article 7, Energy Performance of Buildings Directive, EU, 2009).’’
The Directive has lead to the implementation of national energy
performance certificates (EPCs) for residential dwellings as well
as utility buildings (e.g., office, retail, schools, and healthcare
facilities) across the European Union. Despite the intentions of a
swift and uniform introduction of EPCs across Europe it is
apparent that not all member states have implemented similar
policies. Andaloro et al. (2010) examined the uniformity and
excellence of the EPBD across Europe and found that there are
still large variations across Europe, and most countries are just
halfway towards achieving ‘‘excellence’’.
Agentschap NL, an agency of the Dutch Ministry of Economic
Affairs, exerts quality control and maintains registration of the
energy performance certificates in the Netherlands. For the
purpose of this study we were granted access to the database of
this government agency, which provides information on the Fig. 1. (A) Label composition per construction period (number of buildings).
energy performance rating, the address, and some physical (B) Average building size (m2).
492 N. Kok, M. Jennen / Energy Policy 46 (2012) 489–497

‘‘non-green’’ buildings. This phenomenon not only holds for very old, reduce carbon emissions by about 16 percent in 2020 (as compared
historic office properties (pre-1940), but for all buildings con- to 2010 levels). This explicit demand-shock, created by the largest
structed until the 1980s. The improvement of energy efficiency tenant of office space in the Dutch market, may give an impulse to
partially coincides with the implementation of energy efficiency improving the energy efficiency of the existing office stock in the
requirements, which first became active in 1996. A similar relation Netherlands. However, the green procurement policy may also
between vintage and energy consumption has been documented for have financial implications for the owners of office properties that
the residential housing market. Brounen et al. (in press) document do not live up to the leasing criteria of the government.
that dwellings constructed before 1980 are at least 50 percent less Fig. 2 analyses the fraction of the office market in each of the
efficient as compared to dwellings constructed after 2000. 12 provinces in the Netherlands that does not meet the green
Fig. 1B shows the relation between surface area (i.e., building procurement criteria (red bar). In each province, the share con-
size) and the energy performance certificate. With the exception stitutes at least half of the building stock. Owners of less energy
of A þ and Aþþ, the average size of properties with labels A efficient office space will be directly affected, and corporate
through G is quite similar. Buildings with label A are largest, on tenants following the example set by the public authority may
average, and buildings with label G are smallest, on average. reinforce this trend.
However, differences are quite small, which may indicate that a
minimum building size is not a prerequisite for engaging in
energy efficiency investments. This contrasts the oft-invoked 4. The financial implications of energy efficiency
thesis on ‘‘economies of scale’’ in building retrofits (Kok et al.,
2011). 4.1. Data
The most efficient office properties in our sample (Aþ and
Aþþ) are very small: the size of these buildings is 450 m2 and Transaction data of commercial real estate are notoriously
1900 m2, on average, which is significantly smaller than office difficult to obtain in most countries. Lacking consistent data like
properties with an A-label (with an average size of 5000 m2). We for instance those provided by the CoStar Group (which covers
note that the number of observations rated A þ or Aþþ is quite about 80 percent of all transaction in the US and has reasonable
limited. coverage in the UK), we combine the proprietary transaction
databases of the largest real estate agents in the Netherlands:
CBRE, DTZ Zadelhoff and Jones Lang LaSalle. For each transaction,
3.2. Green public procurement we collect information on the realized rental price per square
meter, the transaction size, the type of lease (new lease, sublease
Green public procurement has now been implemented by or lease extension), the transaction date, and building character-
most European member states. This implies that national govern- istics (including size and age).
ments choose for a ‘‘sustainable’’ alternative when making pur- Using building address, we merge the transaction databases
chasing decisions (this includes everything ranging from office with information on the energy characteristics of the building
supplies to office buildings). The Dutch government started their (i.e., energy performance certificate, modeled energy consump-
green procurement program in January 2010. The program has tion and the energy index), as collected by the Ministry of
major implications for the real estate market: the National Economic Affairs. This results in a unique dataset of 1072 rental
General Services Administration (the government institution transactions during the 2005–2010 period.
responsible for providing office space for government employees) Besides energy efficiency, the ‘‘sustainability’’ of an office build-
just considers buildings with energy label C or higher for new ing is also, and to a large extent, determined by its location relative
leases (alternatively, the landlord can improve the energy label of to public transport and amenities. (Of course, ‘‘location, location,
the property by at least two steps). From an environmental location’’ is also the most important determinant of rental prices.)
perspective, the policy seems to be effective: a recent research We use the longitude and latitude of each transaction and calculate
report documents that, as a result of the green procurement the crow fly distance to the nearest highway entrance and exit, and
criteria, office buildings used by the Dutch government will to the nearest railway station. Obviously, the type of railway station
(central station versus peripheral station) makes a substantial
difference in accessibility. To approximate the service quality of
railway stations, we use the ‘‘Rail Service Quality Index’’ (RSQI),
which takes into account the number of routes, frequency and
number of stops (see Debrezion et al., 2009).
To map the location of a property relative to amenities in the
direct vicinity, such as restaurants and retail facilities, we use the
‘‘Walk Score’’.
The Walk Score algorithm is based on the crow fly distance of a
given address to a varied set of neighborhood amenities. Certain
categories receive a higher weight, and scores are distance
weighted (see http://www.walkscore.com/ for more information
about the ‘‘Walk Score,’’ underlying parameters, and algorithms).
The rationale for including ‘‘Walk Scores’’ is embedded in the
urbanization of economies. Clustering of services in a confined area
is often found to increase the efficiency of labor and, ceteris paribus,
increases the rents that companies can afford to pay (see Melo et al.,
2009, for an extensive review on urban agglomeration economies).
The economic impact of the Walk Score on real estate values
Fig. 2. Green public procurement in the office market. (For interpretation of the
has been studied previously by Pivo and Fisher (2011). The
references to color in this figure, the reader is referred to the web version of this authors document that the ‘‘walkability’’ of locations is value
article.) enhancing for office, retail and industrial properties in the US.
 N. Kok, M. Jennen / Energy Policy 46 (2012) 489–497 493

Table 1
Descriptive Statistics.

Average Median St. Dev. Maximum Minimum

Panel A: Buildings Labeled A–C

Rent (h per m2) 166.25 160 45.43 350 48
Size of transaction (m2) 1785.54 950 2351.14 20000 60
Property size (m2) 10118.19 5710 9497.19 37904 123
Energy efficiency index 1.1 1.15 0.16 1.3 0.49
Distance to train station (m) 1178.71 755 1153.19 4700 66
Station score (RSQI) 1.56 1.4 0.72 3.47 0.29
Distance to highway junction (m) 1261.1 870 1119.61 7749 158
Google walk score 61.59 62 12.29 88 22
Sublet (percent) 1.93 0 13.77 100 0
Lease extension (percent) 1.65 0 12.77 100 0
Renovated (percent) 17.36 0 37.92 100 0
Age (years) 13.68 10 12.27 96 1
Panel B: Buildings Labeled D–G
Rent (h per m2) 152.28 146 39.23 360 14.63
Size of transaction (m2) 1473.86 925 1757.36 27047 60
Property size (m2) 7293.84 4852.07 7035.13 37053 121.37
Energy efficiency index 1.75 1.62 0.36 3.05 1.31
Distance to train station (m) 1327.88 936 1552.78 15561 36
Station Score (RSQI) 1.65 1.46 0.74 3.47 0.25
Distance to highway junction (m) 1554.53 1281 1601.23 1603.9 102
Google walk score 62.53 63 13.67 98 7
Sublet (percent) 1.27 0 11.20 100 0
Lease extension (percent) 2.96 0 16.97 100 0
Renovated (percent) 24.68 0 43.15 100 0
Age (years) 27.77 22 20.56 160 5

4.2. Descriptive statistics

Table 1 provides descriptive statistics for the dataset of 1072

rental transactions, separately for energy efficient (‘‘green’’)
buildings with energy performance certificates A, B or C, and for
less efficient (‘‘non-green’’) buildings with energy performance
certificates D, E, F or G.
The parameter of interest, the realized rent per square meter,
is h152, on average for offices with label D or lower. This
compares to an average rent of h166 for office properties with
label C or higher. Note that this simple (though convenient)
comparison does not control for other important quality differ-
ences between the two samples: energy-efficient offices are 14
years old, on average, whereas the average ‘‘energy hog’’ is twice
as old. Furthermore, the former are substantially larger and closer
to public transport and highways.
Fig. 3A and B illustrate the simple correlation between the
energy index and rents and the Walk Score and rents, respec-
tively. Clearly, there is a negative relation between (modeled)
energy consumption and the realized rents in Dutch offices.
Lower energy efficiency implies a lower rent. (But of course, this
relation is not necessarily causal.) Besides energy efficiency, the
accessibility of facilities in the direct vicinity of a property has a
direct impact on rents as well: a higher Walk Score is correlated
with higher realized rents.

4.3. Analysis and results

In line with Eichholtz et al. (2010, in press) and Brounen and

Kok (2011) we analyze the effect of energy efficiency, approxi-
mated by energy performance certificates, using the following
hedonic specification:
zX
¼1
log Ri ¼ a þ bi X i þ dz zz þ gi Li þ rGi þ ei ð1Þ Fig. 3. (A) Energy index and rental prices (1,076 Observations, 2005–2010). (B)
Z Google walk score and rental prices (1,076 Observations, 2005–2010).

where the dependent variable is the natural logarithm of the coefficients). Xi is a vector of quality characteristics, such as age
realized rental price per square meter in building i (the logarith- and size of a building. It has been shown that locational density is
mic transformation facilitates an easy interpretation of the an important determinant of commercial office rents in the
494 N. Kok, M. Jennen / Energy Policy 46 (2012) 489–497

Table 2
The value of location (dependent variable: logarithm of rent per square meter).

(1) (2) (3) (4)

Age (years)  0.001 [0.001]  0.001nn [0.001]  0.001nn [0.001]  0.001nn [0.001]
Renovated (1¼ yes)  0.010 [0.019]  0.012 [0.019]  0.013 [0.019]  0.013 [0.019]
Distance to highway junction (km) 0.026 [0.027]  0.042 [0.029]  0.041 [0.029]
Distance to train station (km)  0.132nnn [0.031]  0.134nnn [0.031]
Station Score (RSQI)  0.015 [0.039]
Google Walk Score 0.004nnn [0.001]
Size of transaction (m2, log) 0.006 [0.008] 0.006 [0.008] 0.005 [0.008] 0.005 [0.008]
Property size (m2, log) 0.042nnn [0.010] 0.043nnn [0.011] 0.035nnn [0.010] 0.035nnn [0.010]
Sublet (1 ¼yes)  0.037 [0.045]  0.039 [0.045]  0.047 [0.044]  0.046 [0.044]
Lease extension (1 ¼yes) 0.014 [0.035] 0.014 [0.035] 0.025 [0.035] 0.026 [0.035]

Constant 4.771nnn [0.171] 4.531nnn [0.178] 5.157nnn [0.260] 5.091nnn [0.256]

Broker dummy? Yes Yes Yes Yes

Time dummy? Yes Yes Yes Yes
Location dummy? Yes Yes Yes Yes

Observations 1071 1071 1057 1057

R2 0.714 0.714 0.720 0.720
Adj. R2 0.645 0.645 0.653 0.653

Standard errors are in brackets. Significance at the 0.10, 0.05, and 0.01 levels are indicated by *, **, and ***, respectively.

Netherlands (Jennen and Brounen, 2009). To control for density public transport matters more than accessibility by car in traffic-
(a proxy for the well-known adage ‘‘location, location, y’’) we clogged Holland.
include zz, a binary variable that is unique for each four-digit ZIP The application of the Walk Score to approximate density in a
code. These ZIP code fixed-effects control very precisely for real estate pricing framework is relatively new in the literature.
location-specific determinants of office rents, including (but not The results show that the presence of amenities in the direct
limited to) density. In addition, we include Li, a vector of location neighborhood of an office building is positively and significantly
characteristics, such as the distance to railway stations and the related to rents. This finding is in line with recent evidence for the
Walk Score. Gi is a binary variable with a value of 1 if an office US and important for real estate investors. There is currently
building has an energy performance certificate of D or lower (i.e., much talk about new work formats, which changes the traditional
an inefficient, ‘‘non-green’’ building), and zero otherwise. In mono-functional use of an office building in order to facilitate
alternative specifications, we use the specific label category (with flexible working hours and flexible office space. An attractive
label D as the reference group), or the energy index. neighborhood and multifunctional locations are important for
Table 2 provides the results of the regression analysis as ‘‘the office of the future,’’ and our analysis shows that tenants are
represented by Model (1), focusing on the control variables and already paying higher rents for locations with a more extensive
the financial implications of accessibility. The explanatory power set of facilities in the direct vicinity.
of the model (Adj. R2) is quite strong: about two-third of the The coefficients for the remainder of the control variables are
variation in rental prices in the Dutch office market can be in line with expectations: compared to new leases, subleases yield
explained with a rudimentary set of building characteristics and lower rents and lease extensions yield higher rents, on average
location controls. (although the effects are not statistically significant).
In line with expectations, we document in Column (1) that Table 3 extends the analysis with variables that reflect the
older office buildings realize lower rents, as compared to younger energy characteristics of the office properties in our sample.
buildings of comparable size and at similar locations. Controlling for year of construction (negative effect on rents),
However, this effect of building vintage on rents is quite weak: building size (positive effect) and location (four-digit ZIP code),
a decade of decay results in rents that are just one percent lower, the results in Column (1) show that the energy index is negatively
on average. In alternative specifications, we control for the effect related to rental prices. A one-point increase in the energy index
of building vintage using ‘‘year of renovation’’ (rather than ‘‘year results in a rental decrease of about 5 percent. The rental
of construction’’ in combination with a ‘‘renovation’’ dummy). The difference between the most efficient building in our sample
results are comparable to those reported here and available from (energy index of 0.49) and the least efficient building (energy
the authors upon request. Furthermore, leasing a square meter is index of 3.05) is more than 12 percent.
more expensive in larger buildings, ceteris paribus: an increase in In Column (2), we address the effect of energy efficiency on
building size of 1 percent leads to an increase of about 4 percent rental prices in the Dutch office market. The analysis indicates
in rent per square meter. that office properties labeled as ‘‘inefficient,’’ ‘‘non-green’’ labels
Accessibility, as measured by the distance to the nearest (D and lower) have realized rents that are some 6 percent lower
highway entrance or exit, does not have a significant influence than rents in comparable offices with labels reflecting higher
on realized rents (Column 2). However, we document that levels of energy efficiency. This provides the first evidence on the
distance to the nearest railway station matters for tenants (and negative effect of higher energy consumption on rental levels in
thus for investors), with higher rents for buildings located closer commercial office space, even when controlling rigorously for the
to public transport. These results hold while controlling for most important determinants of rents (location, size and vintage).
location using four-digit ZIP code areas and corroborate with In Column (3), we add an ‘‘Amsterdam-effect’’. The office
earlier findings for the Amsterdam office market, in a study on the market of Amsterdam, like Frankfurt, London and Paris, is quite
effect of clustering on office rents (Jennen and Brounen, 2009). For distinct from the remainder of the national property market and it
every kilometer increase in distance to the nearest railway may well be that the importance of energy efficiency is affected in
station, rents decrease with about 13 percent. Accessibility of these major hotspots with international allure. (For example, the
 N. Kok, M. Jennen / Energy Policy 46 (2012) 489–497 495

Table 3
The value of energy efficiency (dependent variable: logarithm of rent per square meter).

(1) (2) (3) (4)

Energy Efficiency Index  0.047nn [0.022]

‘‘Non-green’’ label (D or lower)  0.065nnn [0.016]  0.075nnn [0.018]

‘‘Non-green’’ labelnAmsterdam  0.032 [0.030]

Label Category
A 0.042 [0.030]
B 0.054n [0.032]
C 0.097nnn [0.026]
E  0.008 [0.029]
F  0.005 [0.029]
G 0.023 [0.025]
Size of transaction (m2, log) 0.005 [0.008] 0.005 [0.008] 0.006 [0.008] 0.005 [0.008]
Age (years)  0.001 [0.001]  0.000 [0.001]  0.000 [0.001]  0.000 [0.001]
Renovated (1 ¼yes)  0.014 [0.019]  0.019 [0.019]  0.018 [0.019]  0.012 [0.020]
Distance to train station (km)  0.132nnn [0.031]  0.136nnn [0.031]  0.135nnn [0.031]  0.139nnn [0.031]
Station Score (RSQI)  0.018 [0.039]  0.027 [0.039]  0.031 [0.039]  0.033 [0.040]
Distance to highway junction (km)  0.046 [0.029]  0.046 [0.029]  0.047 [0.029]  0.048 [0.030]
Google walk score 0.004nnn [0.001] 0.004nnn [0.001] 0.004nnn [0.001] 0.003nnn [0.001]
Property size (m2, log) 0.033nnn [0.010] 0.032nnn [0.010] 0.031nnn [0.010] 0.029nnn [0.010]
Sublet (1¼yes)  0.052 [0.044]  0.056 [0.044]  0.060 [0.044]  0.060 [0.044]
Lease extension (1 ¼yes) 0.030 [0.035] 0.031 [0.035] 0.030 [0.035] 0.028 [0.035]
Constant 5.194nnn [0.260] 5.096nnn [0.254] 5.096nnn [0.254] 5.170nnn [0.259]

Broker dummy? Yes Yes Yes Yes

Time dummy? Yes Yes Yes Yes
Location dummy? Yes Yes Yes Yes

Observations 1057 1057 1057 1057

R2 0.722 0.726 0.726 0.727
Adj. R2 0.654 0.659 0.660 0.660

Standard errors are in brackets. Significance at the 0.10, 0.05, and 0.01 levels are indicated by *, **, and ***, respectively.

tenant mix and investor demand for office space may be different ‘‘green’’ buildings and a portfolio of ‘‘non-green’’ buildings, using
in these markets.) However, we do not find significant evidence exactly the same model as in Table 3, Column (4). These indices
that a ‘‘green’’ energy label has a different value in Amsterdam as control for location and quality characteristics of the building.
compared to the remainder of the country. Until the start of 2009, the rental developments were quite
We also include the individual label categories, rather than similar for both groups. The portfolio of ‘‘green’’ office properties
just the ‘‘non-green’’ variable. Column (4) shows that the differ- had slightly stronger rental growth pre-crisis, but this growth
ence between labels A–C and labels D–G is not just driven by the disappeared soon after the start of the financial crisis, as vacancy
most efficient buildings: on the contrary, buildings with labels B rates started to increase following higher unemployment levels.
and C achieve a rental premium over less efficient office buildings As of 2009, however, there is a marked difference in the economic
in the neighborhood. Offices with energy performance certificate effect of energy efficiency. Controlling for location and year of
A realize a higher rental rate, but this effect is not significant. construction, Fig. 4 shows that ‘‘energy hogs’’ are currently facing
These results can be explained in several ways: first, our results relatively strong declines in rent, while more efficient office
quite possibly reflect a ‘‘crisis effect’’: the high-end of the market buildings show rental growth concomitantly. The aggregated
rises disproportionally fast under favorable macro-economic effect of these opposite effects is large and provides an indication
conditions, while rents also decrease faster in more challenging that sustainability (or at least energy efficiency) is capitalized by
economic periods. A similar result is documented in Eichholtz tenants in commercial property markets. Portfolios of real estate
et al. (2010). Second, we have a limited number of buildings with investors will likely be affected by this trend.
an A-label in our sample, which may result in a ‘‘small sample’’
bias. Third, the threshold label requirement for the government
(and some corporate tenants) is the C-label, and some building 5. Conclusions and discussion
owners may have invested just enough to improve the energy
efficiency of their inefficient buildings (i.e., with labels in lower The interest of the real estate sector in energy efficiency and
categories) to the C or B level. These (unobservable) investments sustainability has been growing steadily over the past few years,
may be reflected in the higher premiums for buildings with B and notwithstanding the global economic downturn. There is a
C labels. putative discussion on the costs and benefits of ‘‘greening’’ real
estate, but building improvements in the commercial property
4.4. Dynamics sector have not taken off on a large scale yet. The apparent lack of
investments in energy efficiency and sustainability is partially
We also analyze the value of energy efficiency through time. due to the substantial upfront, often on-balance, capital cost of
Fig. 4 shows a rental index for energy efficient office properties building retrofits, which is reinforced by the current credit
(green line) and for inefficient office properties (red line). This constraints among real estate investors (a direct effect of the
index is based on the quarterly change in rents for a portfolio of financial crisis). In addition, the lack of systematic evidence on the
496 N. Kok, M. Jennen / Energy Policy 46 (2012) 489–497

Fig. 4. Rent index efficient office buildings (labels A, B and C) and non-efficient office buildings (labels D and lower). (For interpretation of the references to color in this
figure, the reader is referred to the web version of this article.)

returns to ‘‘greening’’ existing properties forms an important changes that enable office workers to be more efficient and save
barrier to further inflow of private capital into building retrofits. real estate related costs for the employer. Besides limiting the
This paper adds to understanding better the economic value of square meters per employee, this also changes the view on the
energy efficiency and offers the first systematic, rigorous evidence design and surroundings of a building, with buildings increasingly
on the market implications of EU energy performance certificates viewed as meeting point rather than purely a place to work. The
in the commercial property market. However, energy efficiency ‘‘life’’ around office buildings becomes more and more important.
alone is just one of the variables comprised under sustainability, Our results show that tenants already pay higher rents for space
which also includes elements such as accessibility, water and in offices with a broad palette of amenities in the direct vicinity,
waste management, indoor air quality and building management. as compared to space in offices at mono-functional locations. If
This paper includes another tangible factor of sustainability: the the trend of new ways of working continues, the discount for
location of a building relative to public transport, and accessibility more traditional office locations, without facilities in the direct
of amenities, such as restaurants, retail and fitness facilities. neighborhood, may further increase. Moreover, data on the
European regulation is focused on the built environment, but location of offices in our sample shows that buildings with less
there is no regulation with respect to urban transport as of yet. efficient labels are generally located further away from railway
However, cities in for example Italy, the Netherlands, Germany stations. As location is fixed in time, these offices will to a lesser
and the UK are currently experimenting with charges for vehicles extent be able to profit from the broadening of the sustainability
(see Ekins and Lees, 2008, for a broader discussion of the impact theme and the corresponding valuation of accessibility by public
of EU policies). Although speculative, it is not unimaginable that transport.
regulations targeted at urban transport become more widespread The results in this study provide a clear market indication that
in the near future, thereby not only changing the outlook for the sustainability matters for real estate users, which is in line with
built environment but also increasing the importance of accessi- recent evidence for the US office market (Eichholtz et al., 2010, in
bility for tenants and building owners. press). Our findings have important implications for the portfolios
Our analysis of some 1100 recent rental transactions in the of investors in the European office market and beyond. Rental
Netherlands provides evidence that, on average, a less efficient, growth in efficient and less efficient buildings differs markedly.
‘‘non-green’’ office building achieves a 6.5 percent lower rent as Accessibility pays as well. Both components of sustainability have
compared to similar buildings with a ‘‘green’’ energy label. a direct impact on the valuation of ‘‘non-green,’’ inefficient
Importantly, we control for the most important determinants of buildings as well as offices at traditional, mono-functional
rental values of office buildings, such as location, building vintage locations.
and size. Besides the ‘‘discount’’ for office buildings with labels Real estate appraisers can use the tangible results in this paper to
indicating lower levels of energy efficiency, the results show that integrate the most important elements of sustainability (energy
train stations represent a positive externality for corporate efficiency, accessibility and availability of amenities) in the valuation
tenants—for every kilometer increase in distance to the nearest of properties: less sustainable implies lower income (and quite
railway station, rents decrease with some 13 percent. Facilities in possibly higher risk). Banks and other real estate financiers can
the direct vicinity of buildings have a positive effect on rents exploit the measurable elements of sustainability in the evaluation
as well. of existing and prospective lending agreements. For less efficient
Many property markets currently face historically high office properties that are not adequately improved, situated in areas
vacancy rates. In the Netherlands, for example, it is estimated without direct access to railway stations and amenities, the credit risk
that about 14 percent of the building stock is vacant. The faced by banks may be affected through lower cash flows, which may
poisonous combination of low economic growth, a shrinking put the debt service coverage ratio (DSCR) or the interest coverage
labor force and a changing perspective on the use of office space ratio (ICR) in jeopardy. The credit risk may also be affected by lower
will negatively affect the outlook for some office markets. In values of financed properties, leading to a higher loan to value
addition, corporate tenants are increasingly making the transition ratio (LTV).
to more innovative ways of working. The ‘‘New World of Work’’ Importantly, this research offers insight into the profit oppor-
represents a range of physical, technological and behavioral tunities of building retrofits as well. Sustainability is here to stay
 N. Kok, M. Jennen / Energy Policy 46 (2012) 489–497 497

for real estate investors. Innovative financing mechanisms, such Brounen, D., Kok, N., 2011. On the economics of energy efficiency in the housing
as ‘‘retrofit funds,’’ ESCO’s or ‘‘on bill’’ financing through utility market. Journal of Environmental Economics and Management 62, 166–179.
Brounen, D., Kok, N., Quigley, J.M. Residential energy use and conservation:
companies, enable large-scale inflows of private capital to invest economics and demographics. European Economic Review, http://dx.doi.org/
in energy efficiency. Using third-party capital, real estate inves- 10.1016/j.euroecorev.2012.02.007, in press.
tors can improve the quality of their real estate portfolio, profit Debrezion, G., Pels, E., Rietveld, P., 2009. Modelling the joint access mode and
from lower operational costs, enjoy an improvement in the railway station choice. Transportation Research Part E 45, 270–283.
DTZ Zadelhoff, 2011. Het Effect van Verduurzaming op de Marktwaarde van
marketability of properties and, ultimately, are hedged against Bestaande Kantoren. DTZ Zadelhoff, Amsterdam.
market and macroeconomic trends that will affect the value of Eichholtz, P.M.A., Kok, N., Quigley, J.M., 2010. Doing well by doing good: green
their property portfolio. This way, the transition to a more office buildings. American Economic Review 100, 2494–2511.
Eichholtz, P.M.A., Kok, N., Quigley, J.M. The economics of green building. Review of
sustainable built environment not only contributes to lower CO2
Economics and Statistics, in press.
levels (ultimately leading to a low-carbon economy), but in Ekins, P., Lees, E., 2008. The impact of EU policies on energy use in and the
tandem, it will yield the opportunity for the creation of share- evolution of the UK built environment. Energy Policy 36, 4580–4583.
holder value through energy efficiency and a decrease in sustain- Enkvist, P.-A., Naucler, T., Rosander, J., 2007. A cost curve for greenhouse gas
reduction. The McKinsey Quarterley 1, 35–45.
ability-related financial risks.
Fuerst, F., McAllister, P., 2011. Green noise or green value? Measuring the effects of
environmental certification on office values. Real Estate Economics 39, 45–69.
Goldman, C.A., Hopper, N.C., Osborn, J.G., 2005. Review of US ESCO industry
Acknowledgment market trends: an empirical analysis of project data. Energy Policy 33,
387–405.
Jaffe, A.B., Stavins, R.N., 1994. The energy paradox and the diffusion of conserva-
Financial support for this study has been provided by tion technology. Resource and Energy Economics 16, 91–122.
Agentschap NL and DTZ Zadelhoff. Kok is supported by a VENI Jennen, M.G.J., Brounen, D., 2009. The effect of clustering on office rents: evidence
grant from the Dutch Organization for Scientific Research (NWO). from the Amsterdam market. Real Estate Economics 37, 185–208.
We are grateful to CBRE, DTZ Zadelhoff, Jones Lang LaSalle and Jones Lang LaSalle, 2010. Gebruikersvisie Op Duurzame Huisvesting, Amsterdam.
Jones Lang LaSalle.
Agentschap NL for help in assembling, interpreting, and verifying Kok, N., McGraw, M., Quigley, J.M., 2011. The diffusion of energy efficiency in
the data used in this analysis. We also acknowledge the helpful building. American Economic Review 101, 77–82.
comments of two anonymous referees, as well as Dirk Brounen, Melo, P., Graham, D., Noland, R., 2009. A meta-analysis of estimates of urban
agglomeration economies. Regional Science and Urban Economics 39,
Albert Hulshoff, Bob Kesseler, Ronald van Ouwerkerk and Jan
332–342.
Vermaas. Pivo, G., Fisher, J.D., 2011. The walkability premium in commercial real estate
investments. Real Estate Economics, 39.
References Stern, N., 2008. The economics of climate change. American Economic Review 98,
1–37.
Troostwijk Real Estate, 2011. Verduurzaming van Kantoorpanden Loont
Andaloro, A.P.F., Salomone, R., Ioppolo, G., Andaloro, L., 2010. Energy certification Vastgoedmarkt.
of buildings: a comparative analysis of progress towards implementation in
European countries. Energy Policy 38, 5840–5866.