CEM OCCASIONAL PAPER SERIES

Explaining rental variance

within Singapore’s suburban
retail supply
 IP 2/11

EXPLAINING RENTAL VARIANCE WITHIN

SINGAPORE’S SUBURBAN RETAIL SUPPLY

An information paper by Dougie Crichton

CEM OCCASIONAL PAPER SERIES May 2011

abouT The auThor

Dougie Crichton joined the Pamfleet group in 2008 and currently heads their Singapore office.
Dougie graduated from the University of the West of england, Bristol, in 1995 and is a member of
the royal Institution of Chartered Surveyors. He spent five years in the UK working for gvA grimley
and King Sturge before relocating to Hong Kong in 2000.

Dougie worked as a Senior Asset manager and general manager for Hongkong Land for over seven
years in Hong Kong, Bangkok and Jakarta. He obtained a rICS Postgraduate Diploma in Property
Investment in 2005 and an mBA in Construction and real estate in 2009. His dissertation was
awarded a British Council of Shopping Centres educational Trust retail Award 2009.

inTroduCTion
Singapore is one of the world’s more prosperous and dynamic economies. In 2009, its gDP per capita
of S$53,143 exceeded that of its Asian peers and, according to the International monetary fund, ranked
higher than the UK.

Asian retail markets have benefited from the regional economic upswing and continued to strengthen
through rising consumer confidence, an improved employment market and growth in regional tourism.
Luxury and business fashion is a clear growth story in Asia, but what about the less attractive but more
practical ‘necessity spending’ for comparatively wealthy Singaporeans?

1
IP 2/11
Explaining rental variance within Singapore’s suburban retail supply

The investigation
The author’s initial analysis of sampled neighbourhood centre (NC) tenancy schedules identified that
national food and beverage (f&b) and retail operators are paying over 100% rental differential in high-
density suburban Housing Development Board (HDB) areas. This research paper attempts to answer
why there is such variance in rentals.

Singapura – The Lion City

Singapore became an independent island state in 1965. Presently it has a land size of 274 square miles
(710km2) and a population 4.8 million people. With 17,650 people per square mile (6,814 per km2), it is
one of the most densely populated islands in the world. London has a population of 7.2 million people
living within 610 square miles (1,580km2): 11,803 people per square mile (4,557 per km2); a population
density approximately 30% lower than Singapore.

Since the 1950s, a steady increase in Singapore’s population has required the adoption of a new
town planning concept overseen by the HDB, the country’s public housing authority. Today, 84% of
Singaporeans live in HDB flats.

The island has a tropical rainforest climate with an average temperature range of 22°C to 34°C, high
humidity and a daily average rainfall of 2.36 inches (60mm). These characteristics have formed a unique
environment for citizens and created specific demands on the provision of retail, including f&b.

2
 IP 2/11

May 2011

Sample neighbourhood centres

The focus of this investigation is on NCs, 10,750ft² to 150,000ft² (1,000–14,000m2). The International
Council of Shopping Centres (ICSC) describes NCs as ‘designed to provide convenience shopping for
the day to day needs of consumers in the immediate neighbourhood’ (ICSC 2004: 2).

The author identified four NC assets located in three HDB estates. All four NCs are privately owned,
with a family concern owning both Beta and Charlie, and the other two under individual ownership.

Name HDB estate Net lettable area ft²

Alpha Plaza* Ang Mo Kio 31,123

Beta Plaza* Ang Mo Kio 36,956

Charlie Plaza* Clementi 26,581

Delta Plaza* Toa Payoh 30,429

* Project names changed for confidentiality reasons Source: Author’s compilation

Media and misinformation

Jones Lang LaSalle estimates island-wide retail space per capita at 6.28ft², compared to Hong Kong’s
ratio of 13.2ft² per capita. Singapore has room to expand its retail stock further, based on its current
population.

The local media regularly highlight retailers’ concerns over rising retail rentals increasing their
operating costs and squeezing their profit margins. Comparatively, Singapore’s prime retail rentals are
substantially lower than in Hong Kong. CBRE’s Asia Market Review Q4 2010 highlights Singapore prime
retail rents at US$23.60 per ft² per month, with prime Hong Kong retail at US$66.10 per ft² per month.
A comparison between Singapore and Hong Kong’s suburbs again emphasises the trend, with rents at
US$22.70 per ft² per month (Singapore) and US$38.80 per ft² per month (Hong Kong).

Market practice: Typical market practice in Singapore is for a retail lease term of two to three years.
The structure of the preferred lease is ‘triple net’, which means the tenant is liable for all property taxes,
building insurance and maintenance costs. Depending on the influence and strength of the landlord, the
preference is for the tenant to pay a fixed base rent plus a percentage of sales turnover. The flagship
retail REIT, CapitaMall Trust, stated in its 2009 Annual Report that occupancy costs remained ‘healthy’
at 16.4%.

3
IP 2/11
Explaining rental variance within Singapore’s suburban retail supply

Singapore – local facts

Micro: The HDB Sample Household Survey 2003 highlighted HDB home ownership at 97.1% with low
housing cost, estate facilities, location and transportation as the most liked aspects. The proportion
of the HDB population patronising facilities (market/supermarket, shops, hawker centre, eating house/
coffee shop/food court) on a weekly basis ranged from 60% for the various eating establishments to
64% for shops and over 85% for markets/supermarkets.

Shopping choice behaviour: Ibrahim (2002) summarised research which found that the majority of
shoppers (81.9%) patronised the shopping centres at NCs for their routine major grocery shopping,
with shoppers preferring to use a regional centre rather than the central, traditional Orchard area. Travel
factors such as effort, comfort and length of shopping trips, are more important than shopping centre
characteristics when choosing their shopping destination for major grocery shopping.

4
 IP 2/11

May 2011

South East Asia focused retail studies

Similarities between Hong Kong and Singapore suburban retail were identified; namely higher
population density, reliance on public transport (rather than private car) and differences in shop size.
Chau and Yu highlighted the importance of street level pedestrian flows and the fact that locational
characteristics are more important than physical characteristics, identifying a new genre of retailers, ‘an
independent, small scale retailer selecting a property with micro location and physical characteristics to
suit its business’ (2000: 2).

Quantitative analysis of rental evidence

The author enjoyed full disclosure of the sample NCs’ tenancy schedules. The classification of tenants
was broken down into two groups: food and non-food. In describing the NC tenants, the f&b offering
was primarily for food catering through to restaurants. Common non-food subgroups in the sample
were clothing, shoes, hairdressers, nail salons, massage/spas, mobile phone accessories and opticians.

A simple observation was that food operators, by average, occupy less than 50% of lettable space and
pay a lower average rent than non-food operators. CapitaMall Trust (CMT)’s 2010 Annual Report states
that, within its own portfolio, ‘F&B remained the largest contributor to gross rental income at 25.4% of
the total portfolio’. F&B was the second largest in terms of net lettable area at 16.3% (CMT 2010: 27).
It is interesting to note that CMT’s evidence conflicts with the NC data analysis.

Summary of initial tenancy schedule analysis

• Average ground floor units range from 426ft² to 818ft².

• Average gross ground floor rents range from S$9.65 per ft² to S$40.55 per ft².

• Food retailers occupy 44.56% of lettable space and contribute 43.75% of income.

• Non-food retailers occupy 55.45% of lettable space and contribute 56.25% of income.

• Non-food rents are 18.88% higher than food rents.

Source: Author’s compilation

Three advanced methods of analysis were used to reduce the bias from mean averaging. The
correlation coefficient, which determines the degree to which the movements of two variables are
associated, in this case size and rent, was selected. Basic averaging is vulnerable to distortion owing to
outliers and, by adopting the interquartile range, the outliers are removed. Finally, the coefficient of the
multiple of determination (R²), which is the amount of variability in rent explained by size, was used.

5
IP 2/11
Explaining rental variance within Singapore’s suburban retail supply

Advanced quantitative analysis

Average Average 2nd 2nd

NC unit gross Correlation quartile quartile R² %
size ft² rent coefficient unit size rent per ft²
S$ ft² S$

Alpha 794 9.65 0.427 654 11 29.16

Beta 426 40.55 0.581 341 42 33.84

Charlie 818 18.04 0.792 388 21 74.55

Delta 480 15.69 0.678 367 17 46.31

Source: Author’s compilation

Focusing on rental by product group, the table below supports the trends identified earlier.

Median rentals by product group

2nd quartile 2nd quartile

Difference
NC Classifiction rent per ft² rent per ft²
%
S$ S$

Food 11 11.50 4.5

Alpha
Non-food 10.50 (4.5)

Food 42 35.00 (16.67)

Beta
Non-food 42.00 0

Food 21 21.00 0
Charlie
Non-food 22.00 4.76

Food 17 15.50 (8.82)

Delta
Non-food 18.00 5.55
Source: Author’s compilation

Analysis of the rental schedules highlighted an interquartile range from 341ft² to 654ft² per ground
floor tenant. Monthly rents vary from S$11 per ft² to S$42 per ft², with a noticeable difference between
product groups in terms of occupancy and rental contribution.

6
 IP 2/11

May 2011

Qualitative secondary research

The initial analysis of the rental schedules provided a good balance of qualitative observations that
would form the basis of the semi-structure interviews with the stakeholders: landlords, retailers and
agents.

Landlords: Three general areas were of genuine interest. First, the internal decision making within the
landlord stakeholder company regarding leasing matters. Was the motivation to maximise rents or a
belief in optimising tenant mix for the benefit of the whole NC?

Second, the typical landlord approach to retailers, from the author’s experience, is usually one of
conflict. How does a landlord agree on the rental that a retailer is prepared to pay? Does a landlord
understand the retailer’s operational business model or exploit an operator’s expansion strategy?

Third, the role of agents in Singapore is perhaps misunderstood. What are the expectations of agents
from private landlords? Does a landlord look towards an agent in terms of providing comparable rental
information or does the landlord motivate the agent through commission or through a longer-term
relationship?

All three landlords declined to be interviewed. Undeterred, the author successfully approached and
interviewed senior management individuals from the HDB, CapitaLand and Hongkong Land.

Stakeholder: Landlords

Interview observations:

• Priorities are increasing footfall and avoiding vacancy.

• Branded tenants are regarded as anchors.

• Basic awareness of rent affordability by product group.

• Poor intelligence on comparable rental evidence.

• Passive attitude towards leasing agents.

Source: Author’s compilation

Retailers: Although only two candidates were secured, these particular interviews were rewarding as
their businesses operated five branded food and non-food group businesses. The candidates were
familiar with the author’s sample of NCs and, in fact, were tenants in two of the four assets (Beta and
Charlie).

7
IP 2/11
Explaining rental variance within Singapore’s suburban retail supply

The main themes of the interview touched upon selection of store location, determining rental,
awareness of behaviour towards shopping centre/rental theory, decision making regarding operating
margins, and the use of retail agents.

STAKeHOLDer: reTAILerS

interview observations:

• A fixed rent preferred by retailers.

• Store location dependent on existing stores and competitors.

• Decision making based on comparable sales and expected ‘hit’ rate.

• Increase number of stores to increase sales.

• Six-year period allowed to write off fit-out costs.

• Agents used for introductory purposes only.

Source: Author’s compilation

agents: The role of the agent was broken down into three subgroups. The first group, which specialised
in leasing, was easy to identify. The author approached the leasing agent, Joanna Lee Property
Consultants, of the Beta project in Clementi for interview. It was essential to know how the agent
managed the landlord’s expectations of rental compared with those proposed by the potential tenant.

The second area for discussion was the research work carried out by DTz Debenham Thorpe, which
releases regular market Watch updates that discuss retail trends in Singapore. It was felt that a sector
overview from a research expert would be a positive contribution.

Alpha Beta

8
 IP 2/11

May 2011

finally, it was essential to interview a valuer who is instrumental in advising banks for lending purposes
as to the value of suburban retail assets. Colliers was selected, as it is used by lenders and the HDB for
the setting of rental rates for vacant units.

STAKeHOLDer: AgeNTS

interview observations:

• generally undervalued by landlords and retailers.

• Preference for landlords and retailers to negotiate directly.

• value add potential for market research and retail mix positioning.

• resistance by landlords to pay commission for securing new tenants.

• Subjective approach to market value valuations.

• Industry professionals would prefer to work for a landlord than an agency firm.

Source: Author’s compilation

summary: The behaviour of a typical NC landlord appears to be largely independent of other landlords
in the surrounding area. retailers tend to exploit the NC landlords’ lack of information and secure units
with different levels of rent. The role of the agent is an informal and irregular part of the leasing process
owing to the internal conflicts which result from landlords employing in-house leasing managers.

Charlie Delta

9
IP 2/11
Explaining rental variance within Singapore’s suburban retail supply

Conclusions
Singapore has characteristics of land scarcity, high population density, dependency on public transport,
comparative underprovision of retail space per capita and a local business culture. These factors create
a unique environment where, owing to limited supply, there will always be excess demand by retailers.
Quantitative and qualitative research led to the following conclusions.

Macroeconomic factors: Three key factors were identified: GDP, inflation and interest rates. All three
factors have an effect on consumer confidence and behaviour. However, the author is in agreement with
the opinion of Hardin and Wolverton (2000: 319) that, at the NC level, macroeconomic variables had
‘little effect on aggregate retail returns’.

Micro-market characteristics: The relative size of Singapore and town planning policy allow high
accuracy on household information, such as household size, age and ethnic groups, income and
savings ratios. This data is used extensively by retailers to identify their target customers, who show a
strong demand for staple goods, creating an efficient local retail mix with a loyal customer base.

Demand, supply, locational and physical characteristics: Singapore retail stock per capita is 6.28ft²,
Hong Kong is 13.2ft². Supply will continue to be absorbed owing to the underprovision of stock;
however, the location in relationship to public transport is of paramount importance. The retailers
interviewed repeatedly stressed the importance of proximity to major public transport hubs, while
securing a unit with high footfall is a major factor in decision making.

Theory on rent analysis and forecasting: On a practical level, multiple regression analysis is the key
to understanding the variables which affect rental values on HDB estates. A basic regression analysis
was carried out, but a more advanced multiple regression analysis would have been useful, particularly
quantifying locational characteristics.

Retailer business model: Gerbich (1998) states that the rental of retail space is sensitive to sales
volume. Several problems face the retailer; in a relatively small market there is familiarity of retail stock,
so the criteria of selection, location and hit rate can force greater risks to be taken with new stock. Such
trends do push up rents, and this is evident with the Beta NC rental analysis. The retailer is also left with
the problem of increasing the number of stores to increase sales turnover, but at some stage the brand
will reach saturation and the cannibalisation of neighbouring stores will affect turnover.

Typical business practice: In securing a retail unit, the preference is for direct dialogue between
landlord and tenant. This is a business culture issue, as there is a desire to avoid paying an agency
commission fee to a third party. The mechanism for rental will be a fixed S$ per ft² because of the
difficulty of trusting monthly sales figures provided by retailers and the lack of a cost-effective electronic
point-of-sale system.

10
 IP 2/11

May 2011

Stakeholders: The author was declined interviews with the sample NC owners and instead relied
on the feedback of branded landlords. These were primarily concerned with generating footfall and
avoiding the negative consequences of visible vacant units. All the sample NCs have vacant units
(either ground or second floors). It appears that vacant units are not treated as a priority unless a
satisfactory rental is achievable.

The next observation is the inherent conflict between landlord and tenant. The landlord wants a fair,
market rent for the unit, while the retailer wants to minimise the rent to maximise margins. The landlord
has a basic awareness of affordability by product groups, but is uncertain as to the exact rent, owing to
poor intelligence from the rental market. The reluctance of independent landlords to communicate and
share rental information creates a negative effect on achieving current market rentals.

There is some agreement on what a product group’s percentage of sales should be paying, and this
can be translated into a higher fixed rent. A landlord is of the opinion that a retailer should pay rent
equivalent to 12% to 15% of turnover, and an f&b outlet 7% to 8%. Retailers will accept 10% to 12%
of expected turnover and f&b operators from 12% to 15%. The disparity with perceived f&b turnover
margins could be explained by the fact that interviewed landlords were talking about mid-market
restaurants, whereas the retailers were discussing fast food and lower-end f&b operators.

According to agents, their role in the leasing process is undervalued. Landlords would benefit from the
leasing agent’s market knowledge; however, they tend to neglect the value of agency advice. It was
observed that agents do not provide any quantitative analysis to smaller landlords regarding rental
variance in their assets. In summary, landlords and retailers rely on agents for introductory purposes
only.

Summary: The rental variance identified is primarily due to two main factors: location and the landlord’s
business culture. The maxim, ‘location, location, location’, is very relevant to Singapore’s high-density
suburbs, particularly when in close proximity to public transport or suburban shopping malls.

There appears to be minimal sharing of rental information between smaller, private landlords. This
problem is compounded by the fact that in-house leasing managers are not expected to use the
services of agents. This behaviour is common within hierarchical local businesses where a single,
independent approach can create an inward-looking culture. The refusal to engage agents passes the
competitive advantage on to the retailer.

The observed variance benefits the retailer. With such poor communication among smaller landlords,
retailers can negotiate the fixed rent that they want to pay rather than accept a rent that they can afford
to pay. Therefore, typical landlord behaviour has a negative impact on the value of NCs as an asset
class.

11
IP 2/11
Explaining rental variance within Singapore’s suburban retail supply

Recommendations for further research

Further investigation into regression analysis would allow more variables, e.g. the impact of GDP, real
disposable income, household savings ratio and interest rates, to be considered. These could reveal
other factors that affect rental variance. Quantifying the importance of location would be extremely
valuable.

The above-average rental rates in the Beta NC are remarkable. Retailers are of the opinion that the Beta
NC is overrented. It would be useful to monitor the lease renewal phase when negotiations begin in
2011 and to observe whether the current rents are sustainable.

References and further reading

CBRE (2010) Asia Market View Q4 2010, CBRE Research, Asia.

CapitaMall Trust (2010) Consistent Performance, Sustainable Growth, Report to Unitholders.

Chau K W and Yu C K (2000) ‘The determination of street level retail shop prices in Hong Kong’, paper
delivered at the Pacific Rim Real Estate Society Conference, Sydney, 23–27 January.

Competition Commission Singapore (2008) ‘Market study on retail rental space in Singapore; Summary
of Study and CCS’, Views, 31 October.

DTZ Research (2009) ‘Temporary oversupply, long term sustainability in the Singapore retail market’,
Research Alert, May.

Gerbich M (1998) ‘Shopping center rentals: An empirical analysis of the retail tenant mix’, Journal of
Real Estate Research, 15(3): 283–296.

Hardin G H and Wolverton M L (2000) ‘Micro-market determinants of neighbourhood center rental

rates’, Journal of Real Estate Research, 20(3): 299–322.

Housing and Development Board (2003) Public Housing in Singapore: Resident’s Profile and Physical
Aspects, HDB Sample Household Survey, 2003.

Ibrahim M F (2002) ‘Shopping choice behaviour in Singapore: Suburban malls rule’, Centre for Real
Estate Studies (CRES), Department of Real Estate, National University of Singapore, 2(22): 8.

International Council of Shopping Centers, ICSC Shopping Center Definitions, 2004.

Photographs – acknowledgements
Cover and header image © Wolfgang Sladkowski 2009 – www.aseanmobile.mobi

Pages 8–9 © Dougie Crichton 2011

12
© College of Estate Management 2011

All rights reserved by the College of Estate Management. No part of this

publication may be reproduced, stored or transmitted in any form or by any
means without prior written permission from the College of Estate Management.

CEM warrants that reasonable skill and care has been used in preparing this
report. Notwithstanding this warranty, CEM shall not be under liability for any
loss of profit, business, revenues or any special indirect or consequential damage
of any nature whatsoever or loss of anticipated saving or for any increased
costs sustained by the client or his or her servants or agents arising in any way,
whether directly or indirectly, as a result of reliance on this publication or of
any error or defect in this publication. CEM makes no warranty, either express
or implied, as to the accuracy of any data used by CEM in preparing this report
nor as to any projections contained in this report which are necessarily of any
subjective nature and subject to uncertainty and which constitute only CEM’s
opinion as to likely future trends or events based on information known to CEM
at the date of this publication. CEM shall not in any circumstances be under any
liability whatsoever to any other person for any loss or damage arising in any way
as a result of reliance on this publication.
 CEM is the leading provider of education, training
and research for the real estate and construction
industries. No other institution offers the same range
and quality of specialist expertise to the property
profession.

Over the past 90 years, we have helped more than

150,000 people, at all levels of the profession, with a
wide range of business and academic backgrounds, to
gain the skills they need to enhance their careers.

While we are an independent organisation, we have

a close relationship with the University of Reading,
strong links with a range of professional bodies and
with major property firms. The College is increasingly
global in outlook.

Drawing on our extensive knowledge base,

professional contacts and independent standpoint,
research is a core area of CEM’s activities, both to
ensure the quality and relevance of our education
programme and to offer a vital service to the property
profession.

the logo is intended to

bleed off the top left hand
corner of the page.
the dots should line up with
the page edge, allow 3mm
bleed and pull in the picture
box to hide the dots.

The College of Estate Management

Whiteknights, Reading,
RG6 6AW, United Kingdom
Tel: +44 (0) 118 921 4696
Email: research@cem.ac.uk
www.cem.ac.uk Patron: HRH The Prince of Wales