Publié par : Published by : Publicación de la :

Faculté des sciences de l’administration Université Laval Québec (Québec) Canada G1K 7P4 Tél. Ph. Tel. : (418) 656-3644 Fax : (418) 656-7047

Édition électronique : Electronic publishing : Edición electrónica :

Aline Guimont Vice-décanat - Recherche et partenariats Faculté des sciences de l’administration

Disponible sur Internet : Available on Internet Disponible por Internet :

http ://www.fsa.ulaval.ca/rd rd@fsa.ulaval.ca

DOCUMENT DE TRAVAIL 2002-008

SHOPPING CENTER RENTS AND AGGLOMERATION ECONOMIES : PRELIMINARY FINDINGS FROM EMPIRICAL EVIDENCE François Des Rosiers Marius Thériault Ünsal Özdilek

Version originale : Original manuscript : Version original :

ISBN – 2-89524-145-7

Série électronique mise à jour : On-line publication updated : Seria electrónica, puesta al dia

06-2002

SSHHOOPPPPIINNGG CCEENNTTEERR RREENNTTSS AANNDD AAGGGGLLOOMMEERRAATTIIOONN EECCOONNOOMMIIEESS::

PPRREELLIIMMIINNAARRYY FFIINNDDIINNGGSS FFRROOMM EEMMPPIIRRIICCAALL EEVVIIDDEENNCCEE

Paper presented

at the 9th European Real Estate Society Conference, Glasgow, Scotland, June 4-7, 2002

by

François Des Rosiers, Ph.D., Faculty of Business Administration, Marius Thériault, Ph.D., Director, Land Planning Research Center,

and Ünsal Özdilek, Ph.D. Candidate,

Laval University, Quebec City, Canada

ABSTRACT This study is an attempt to model shopping centre rents and to investigate whether agglomeration economies are a driving force in the rent determination process. It is part of a research program based on physical and financial information obtained from seven super-regional, regional and community shopping centre managers in Quebec City for the 1998-2000 period. At this point, some 784 shops are covered in the study with gross base rent, operating expenses and gross leasable area being available for all of them. Partial information on lease duration as well as yearly sales is also available. While still preliminary, findings clearly suggest that larger centres generate substantial agglomeration economies for several categories of shops. Retail structure internal to shopping centres as well as cyclical factors also play a significant role in the rent determination process.

KEY WORDS:

Shopping Centres, Rents, Agglomeration Economies, Commercial Mix

____________________________________

1. OBJECTIVE AND CONTEXT OF RESEARCH

This study is an attempt to model shopping centre rents and to investigate whether

agglomeration economies are a driving force in the rent determination process. It is part of

a research program based on physical and financial information obtained from seven super-

regional, regional and community shopping centre managers in Quebec City for the 1998-

2000 period. At this point, some 784 shops are covered in the study with gross base rent,

operating expenses and gross leasable area being available for all of them. Partial

information on store frontage, lease duration as well as yearly sales is also available. When

2

completed, the data bank will include a series of shopping centre amenities and design

features (physical configuration, parking facilities, landscaping, etc.) as well as

neighbourhood characteristics (household composition and income) and urban externalities

(accessibility to services, proximity of competitors, etc.) in order to account for all major

internal and external determinants of shopping centre rents.

2. LITERATURE REVIEW

The academic literature on shopping malls has evolved around various theories of urban

spatial structure (Hotelling, 1929; Christaller, 1933; Lösh, 1940 and Alonso, 1964) with

strategies relating to mall configuration and store location within shopping centres

replicating those observed at the urban level (Vandell and Lane, 1987; Pearson, 1991;

Brueckner, 1993; Roulac, 1996; Brown, 1999). In contrast to what prevails in the

residential market (Follain and Malpezzi, 1980; Noland, 1980; Sirmans and Benjamin,

1991; Benjamin and Sirmans, 1994; Jud et al., 1996; Des Rosiers and Thériault, 1994 &

1996; Chinloy and Maribojoc, 1998) and office sector (Rosen, 1984; Hekman, 1985;

Gabriel and Nothaft, 1988; Wheaton and Torto, 1995; Sivitanides, 1997) where the rent

issue has been widely investigated, studies on the dynamics of commercial rent structuring

remain embryonic because of the confidential nature of the required information.

The mechanics underlying additional, or overage rents - expressed as a percentage of yearly

sales over and above a given, pre-negotiated threshold - are among the issues raised by

authors (Hartzell et al., 1987; Benjamin et al., 1990; Bruecker, 1993; Colwell et al., 1998;

Wheaton, 2000; Chun et al, 2001). Benjamin et al. (1990) were the first to apply hedonics

to the analysis of commercial rent. In their study, base rents derived from 103 commercial

leases pertaining to national, local and independent stores are regressed against sales,

discount rates, overage rents, lease terms, lease provisions, etc. Results suggest that while

base rents are lower where higher overage rent rates apply, they rise with higher sales

thresholds. In another similar study, Sirmans and Guidry (1993) point out that higher

consumer traffic levels are a prerequisite for the success of a store. In a totally different

urban context, Tay et al. (1991) investigate the Hong Kong commercial market. Their data

base includes 405 stores distributed among nine high-rise shopping centres. In contrast with

the literature, their study namely reveals that rent level is positively related to the age of a

shopping centre due to both customers’ fidelity - which tends to grow with time - and

3

continuous improvements to buildings. It also suggests that while the unit rent of a store is

positively correlated with the size of a centre, it is inversely related to its own size. This

latter finding brings us to the core of the current paper, that is agglomeration economies.

With location theories as the conceptual background (Weber, 1929), sales potential in

shopping centres are looked upon through the concepts of agglomeration economies and

externalities derived from the presence of anchor tenants (Eaton and Lipsey, 1983;

Mulligan, 1983; West et al., 1985; Ghosh, 1986; Ingene and Ghosh, 1990; Fisher and

Yezer, 1993; Eppli and Benjamin, 1993; Mejia and Benjamin, 2002) as well as from tenant

mix and product diversity (Konrad, 1984; Pashigan and Gould, 1998). Behind the concept

of agglomeration economies lies the reduction of consumer search and of uncertainty costs.

Such advantages allow major tenants to negotiate lower rents with shopping centres’

owners (Anderson, 1985), the fact that their departure may cause rental income to drop by

as much as 25% (Gatzlaff et al., 1994) greatly enhancing their bargaining power.

According to Nelson (1958) and Eppli and Shilling (1996), the clustering of similar stores

leads to an increase in their total sales level, thereby contributing to the success of the

shopping centre.

The image of a shopping centre may also impact upon sales level (Brown, 1992; Kirkup

and Rafiq, 1994; Anikeeff, 1996). It stems from consumers’ perception of major occupants

(Nevin and Houston, 1980), shopping centre size and configuration as well as the quality of

goods and services offered. In this respect, image is increasingly dependent upon fashion

(James et al, 1976; Jain and Etgar, 1976; Mazursky and Jacoby, 1986; Grewal, 1998).

Similarly, it affects tenants in their negotiation for an optimal location (Mejia et al., 2001).

Finally, accounting for all these features raises the spatial autocorrelation issue, recently

addressed by Carter and Haloupek (2000) on the grounds of previous work performed

mainly on the residential market (Griffith, 1987; Pace and Guilley, 1998; Dubin et al.,

1999).

4

3. DATA BANK AND ANALYTICAL APPROACH

3.1 Original Data Bank

As mentioned earlier, this study is based on physical and financial information obtained for

seven super-regional (2), regional (2) and community (3) shopping centres in Quebec City.

While financial data apply to the 1998-2000 period, leases were negotiated over a much

longer period that extends from 1976 to 2000. At this point, 784 stores are accounted for in

the study. Whereas net rent – defined as total yearly rent minus operating expenses - and

gross leasable area (GLA) is available for all retail units, partial information only is

available with respect to yearly sales (526 stores), store frontage (231) and lease duration

(621). Eight retail categories are distinguished in the analysis, namely jewellery stores (58),

clothing stores (264), shoeshops (57), restaurants (128), personal services (74) specialty

stores (151), kiosks (36) and anchor stores (16), all of which totalling nearly 2,4 million

square feet of GLA. As for shopping centre size, it is accounted for via dummy variables

used in interaction with other descriptors to generate hedonic coefficients that are specific

to each shopping centre category. Community (COM – 79 stores and less), regional (REG –

80 to 149 stores) and super-regional (SREG – 150 stores and over) centres’ rented space

roughly amount to 283 000 sq. ft., 612 000 sq. ft. and 1 489 000 sq. ft., respectively. GLA

distribution by retail category and shopping centre size is displayed in Exhibit 1 while

Exhibit 2 provides an operational definition for each variable used in the modelling process.

3.2 Structural and Cyclical Attributes

In addition to basic financial and physical characteristics, a series of five market structure

and cyclical attributes are developed in order to capture the effect on rent of both

commercial dynamics within shopping centres and overall economic situation. These are:

Lease duration (LEASEDURN): It might be expected that risk reduction deriving

from a longer lease term should result in a rent discount for the tenant, provided that

regional commercial rents at the time of negotiation are relatively stable or

declining; otherwise, a longer than average term may penalize the landlord.

An agglomeration index (AGGLOMINDX): Defined as the ratio of GLA for a

given retail category within a shopping centre to total GLA in that centre, this index

5

measures agglomeration economies resulting from the amalgamation of a large

number of tenants in a given field of retail activity and, therefore, from the

increased consumer traffic it implies. It should be expected that the larger the index,

the higher the rent that the landlord can extract from tenants in that category.

A concentration index (CONCNINDX): While the landlord can take advantage of

the abundance of retailers in a given commercial category, he will in contrast be

prone to grant a rent discount where relatively few tenants are in control of the field,

since any defection might induce a severe drop in yearly revenues. The Herfindahl

Index, designed at measuring industrial concentration, is used here. Since it

accounts for both the number of stores in a retail category and their relative share of

the total GLA, it mirrors the level of competition in a field and the bargaining power

of tenants. It can be assumed that the larger the index, the lower the rent paid.

The GLA open to negotiation (OPENLEASE) in any given year is also a good

indication of the bargaining power of tenants in a shopping centre. Indeed, where a

substantial proportion of total GLA is under negotiation at the same time, the

landlord should be expected to grant rent discounts in order to avoid massive

departures to competitors. Therefore, the higher the index, the lower the rent.

Finally, economic cycles (CYCLE91_92) should theoretically affect the negotiation

process. An analysis of the consumption expenses of households and retail sales in

the Quebec metropolitan area between 1981 and 2001 (expressed in 1992 constant

dollars) indicates that the 1991-92 period is characterized by a drop in commercial

activity. Thus, the coefficient of this dummy variable should display a negative

sign.

3.3 Analytical approach

Two functional forms are used in the study, namely the linear and log-linear ones; a

logarithmic transformation is also applied to the GLA variable. Exhibit 3, displaying basic

descriptive statistics for all variables, clearly indicates that neither NETRENT nor GLA are

normally distributed; Figure 1 provides further evidence of this. Hence the rationale for

such transformations. While included in the data base and available for some 526 stores,

unit sales (SALES/SQFT) are not included in the analysis at this stage of the research. The

6

final calibration of the equations involves the removal of 15 extreme residuals considered

to be highly detrimental to the model performances; these represents less than 2% of the

initial data base. Finally, the regression procedure adopted is a combination of both

standard and stepwise approaches.

4. MAJOR FINDINGS

4.1 The Linear Form – Model A

As can be seen from Exhibit 4, the linear model yields interesting results in spite of the non-

linearity of the dependent variable. While the explanatory performance of Model A, which

displays an adjusted R-Square of 0,714 and a F value of 84,2, is quite fair, its predictive

performance remains rather weak: considering the global average rent ($64), the relative

SEE stands at 38,4%, which was to be expected in the light of the highly skewed

distribution of the NETRENT/SQFT variable. Examination of the regression coefficients

suggests that no inconsistency can be detected with respect to either their magnitude or

sign, with the exception of the ANCHOR parameter estimate which is unexpectedly

positively signed whereas it should be negative. Safe for a few coefficients whose statistical

significance do not meet the 0,10 (CYCLE91_92; SPECIALTY*SREG) or the 0,05

(SHOESTORE*COM; AGGLOMINDX*REG) threshold, all other descriptors exhibit

strong t values. Most interestingly, the two structural attributes indicative of tenants’

bargaining power (CONCNINDX*SREG and OPENLEASE*REG) emerge as significant

and with the right sign.

Considering that the linear model may not provide the best indications on the hedonic

shopping centre rent function, let us turn to Model B for a more reliable interpretation of

the coefficients obtained via the log-linear form.

4.2 The Log-Linear Form – Model B

Regression results for Model B are shown in Exhibit 5. While overall performances

obtained with the log-linear form are quite similar to those derived from the linear one1, the

1 It should be kept in mind that performance tests obtained with a linear form cannot be compared with those derived from a non-linear function without first applying the test suggested by Box and Cox (1964). The latter consists in computing the linear model using the transformed dependent variable Yi / YG , with YG being the geometric mean of the Y.

7

statistical significance of parameter estimates is substantially improved for most

descriptors. Only one coefficient (KIOSK*REG) is not significant at the 0,05 level while

all coefficient signs – including that of the ANCHOR variable - are in line with theoretical

expectations. Most important though, four out of five structural and cyclical variables

emerge as statistically significant, lease duration (LEASEDURN) alone being rejected from

the model. As for excessive multicollinearity, VIFs suggest that there is none, the highest

VIF value standing at 2,89, which is well below the critical threshold of 10. In spite of that,

it is interesting to note that there might be a structural link between store agglomeration

(AGGLOMINDX*REG) and the clothing retail sector (CLOTHING*REG) in regional

shopping centres.

In order to make an adequate interpretation of the implicit rents of commercial attributes

derived from Model B, it is appropriate to keep in mind the retail structure of shopping

centres. In that respect, Exhibit 1 may prove very helpful.

4.3 Interpreting the coefficients of basic physical and retail category attributes

As expected, the prominent role of gross leasable area in the retail rent determination

process clearly emerges from the model, the negative sign of the Ln_GLA coefficient

indicating that unit rent decreases with store size. More precisely, increasing average store

size (3 000 sq. ft.) by 10% will result in a 2,4% drop in unit rent2. As shown by the Beta,

standardized coefficient, frontage is also a major determinant3 and exerts a positive

influence on rent due to the enhanced visibility and attractiveness it confers to a store.

Thus, raising the mean frontage (34 ft.) by 10% increases unit rent by roughly 1,5%.

Quite interestingly, the presence of a jewellery impacts differently on rents depending on

the size category of the shopping centre it belongs to. In community shopping centres

(JEWELRY*COM), the effect is negative whereas it is positive in both regional

(JEWELRY*REG) and super-regional (JEWELRY*SREG) structures. A twofold

explanation can be brought forward: on the one hand, the relative scarcity of this retail

activity in community centres (6,8% of stores and 1,6% of GLA) as opposed to regional

2 This marginal impact is obtained by applying mean values to all model variables and then simulating a 10% increase in GLA, from 3 000 to 3 300 sq. ft. As a result, unit rent drops by 2,4%, from $38,53 to $37,59. 3 In the current state of the data base, this attribute is only available for super-regional centres.

8

centres where jewellers are in excess (25,7% of stores and 3,0% of GLA) translates into a

rent discount in the former case and in a location premium in the latter case. On the other

hand, jewellery stores located in community centres typically offer cheaper products than

those found in more glamorous, either regional4 or super-regional ones that also benefit

from agglomeration economies; hence the positive sign of the latter coefficients. The same

rationale, combining relative scarcity, prestige location and agglomeration economies,

applies – with even greater relevance – to clothing and shoe stores located in community

(negative sign) and regional (positive sign) shopping centres.

Agglomeration economies are clearly at stake when it comes to interpreting the marginal

contribution of restaurants to unit rents, with regional and super-regional centres

commanding substantially higher rents than community ones. Rent patterns differ though

with respect to personal services and specialty stores. In spite of their relative abundance,

the negative coefficients assigned to the SERVICES*COM and SPECIALTY*COM

estimates reflect the bargaining power of local retailers that are at the very core of

community centres’ mission, namely proximity services. In contrast to all other statistically

significant retail categories, specialty stores located in super-regional centres

(SPECIALTY*SREG) also command lower unit rents: according to the “combined” and

“compared” purchase concept, consumers are attracted by high-tech specialty stores

(computer, audio and video stores) which, for that reason, benefit from rent discounts;

hence the negative sign attached to the related coefficient.

Kiosks are found only in regional and super-regional shopping centres. Since they occupy

the central alley of shopping malls where consumer traffic is optimal, they generate very

high sales per square foot and, consequently, command the highest unit rents of all retail

categories. Their contribution to the retail rent determination process is among the

interesting findings of this study. While the KIOSK*REG coefficient is only significant at

the 0,10 level, the parameter estimate of the KIOSK*SREG descriptor (Beta coefficient of

0,286) undoubtedly suggests that kiosks located in super-regional centres substantially

increase their overall profitability; actually, the unit rent they pay represents a 144%

4 It is worth noting that among the two regional shopping centres included in the study, one is located in a rather low-class, popular environment while the other attracts high-income, most sophisticated customers. Considering that bias, the magnitude of the regional centres-related coefficients would project a different picture if the “low-profile” centre were distinguished from the “high-profile” one.

9

premium over the base (intercept) rent. Finally, anchor tenants do benefit, as expected,

from a major rent discount; while it amounts to roughly 24% on average, it can be much

higher in some cases.

4.4 Interpreting the coefficients of structural and cyclical descriptors

Central to this study, findings pertaining to agglomeration economies in shopping centres

corroborate the literature on the subject. They clearly suggest that while larger size

shopping centres involve greater competition among tenants, store clustering also generate

substantially higher sales volumes, which translates into higher rents. When applied to

regional shopping centres (AGGLOMINDX*REG), a 10% increase in the mean

agglomeration index (0,14) results in a 1,3% rise in unit rent. In contrast, and according to

theoretical expectations, concentration levels in retail activity, as measured by the

Herfindahl Index, impact negatively on rents. In as much as super-regional shopping

centres are concerned (CONCNINDX*SREG), a 10% rise in the average concentration

index (0,11) translates into a 1,5% drop in rent. Tenants who negotiate their lease while a

large number of retailers do the same do experience, as expected, a strengthening of their

bargaining power. In regional shopping centres, every 10% increase in the proportion of

open leases (OPENLEASE*REG) results in a 0,8% reduction in unit rent. This raises the

importance of lease renewal strategy for shopping centre owners. Finally, a recessionary

economic cycle affects retail rents negatively. Over the 1991-92 period, overall shopping

centre rents in the Quebec metropolitan region dropped by 12%.

5. SUMMARY OF FINDINGS AND SUGGESTIONS FOR FURTHER RESEARCH

5.1 Summary of findings

This study is an attempt to model shopping centre rents and to investigate whether

agglomeration economies are a driving force in the rent determination process. It is based

on physical and financial information obtained for seven super-regional, regional and

community shopping centres in Quebec City. While financial data apply to the 1998-2000

period, leases were negotiated over a longer period that extends from 1976 to 2000. At this

point, 784 stores are accounted for in the study. In addition to basic financial and physical

characteristics, a series of five market structure and cyclical attributes are developed in

10

order to capture the effect on rent of both commercial dynamics within shopping centres

and overall economic situation. Main findings can be summarized as follows:

As expected, the prominent role of gross leasable area in the retail rent determination process clearly emerges from the model, the negative sign of the Ln_GLA coefficient indicating that unit rent decreases with store size;

Frontage is also a major determinant and exerts a positive influence on rent due to the enhanced visibility and attractiveness it confers to a store;

Quite interestingly, the presence of a jewellery impacts differently on rents depending on the size category of the shopping centre it belongs to. In community shopping centres, the effect is negative whereas it is positive in both regional and super-regional structures. A combination of relative scarcity, prestige location and agglomeration economies provides a sensible explanation to such findings;

A similar rationale applies – with even greater relevance – to clothing and shoe stores located in community and regional shopping centres;

Agglomeration economies are clearly at stake when it comes to interpreting the marginal contribution of restaurants to unit rents, with regional and super-regional centres commanding substantially higher rents than community ones;

Rent patterns differ though with respect to personal services and specialty stores. In spite of their relative abundance, the negative contribution of this retail category in community shopping centres reflects the bargaining power of local retailers that are at the very core of community centres’ mission, namely proximity services;

Specialty stores located in super-regional centres also command lower unit rents due to the attraction they exert on high-tech goods consumers;

Since they occupy the central alley of shopping malls where consumer traffic is optimal, kiosks command the highest unit rents of all retail categories;

Anchor tenants do benefit, as expected, from a major rent discount which amounts, on average, to 24% of base rent;

Findings clearly suggest that while larger size shopping centres involve greater competition among tenants, store clustering generates substantial agglomeration economies that translate into higher rents;

Concentration levels in retail activity, as measured by the Herfindahl Index, impact negatively on rents;

The proportion of leasable space under negotiation in a given year strengthens the bargaining power of tenants, who consequently benefit from a rent discount;

Finally, a recessionary economic cycle affects retail rents negatively.

11

In short, this study corroborates previous research findings about the critical role played by

anchor tenants in the generation of agglomeration economies as well as the importance of

tenant mix, product diversity, quality of goods and image in the retail rent structuring

process. While increased attractiveness has an overall positive effect on both sales and unit

rents for most stores - namely in the case of kiosks - , they allow major tenants and other

specialty goods retailers to negotiate lower rents with shopping centres’ owners. Such

factors however impact differently on rents depending on shopping centre size and profile.

Finally, retail structure within shopping centres and lease renewal strategies by owners will

affect tenants’ bargaining power and, therefore, rent levels.

5.2 Suggestion for further research

While this study is but a preliminary investigation into the retail rent dynamics, it raises

several interesting issues namely with respect to the measurement of externalities and

agglomeration, as opposed to dispersion, effects. While shopping centre configuration and

design deserve further attention, resorting to concepts and analytical tools found in

geography and spatial economics to measure such phenomena may lead to major

developments in the field.

12

REFERENCES Alonso, W., (1964). Location and Land Use. Toward a General Theory of Land Rent, Cambridge, Ma.,

Harvard University Press, 205 pages. Anderson, P. M. (1985). Association of Shopping Center Anchors with Performance of Non-Anchor

Speciality Chain’s Stores, Journal of Retailing, 61, pp. 61-74. Anikeeff, M. (1996). Shopping Center Tenant Selection and Mix: a Review, in J. Benjamin, ed., Research

Issues in Real Estate, Vol. 3 : Megatrends in Retail Real Estate, Norwell : Kluwer Academic Publishers. Benjamin, J. D. and Sirmans, G. S. (1994). Apartment Rent : Rent Control and Other Determinants, Journal

of Property Research, 11, pp. 27-50. Benjamin, J. D., Boyle, G. W. and Sirmans, C. F. (1990). Retail Leasing : Determinants of Shopping Center

Rents, Journal of the American Real Estate and Urban Economics Association (AREUEA), 18 :3, pp. 302-312.

Box, G. E. P., and D. R. Cox (1964) An analysis of transformations, Journal of the Royal Statistical Society, 26, 211-243.

Brown, S. (1992). Tenant Mix, Tenant Placement, and Shopper Behavior in a Planned Shopping Centre, The Services Industries Journal, 12 :3, pp. 384-403.

Brown, G. M. (1999). Design and Value : Spatial Form and the Economic Failure of a Mall, Journal of Real Estate Research, 17:2, pp. 189-225.

Brueckner, J. (1993). Inter Store Externalities and Space Allocation in Shopping Centers, Journal of Real Estate Finance and Economics, 7:1, pp. 5-17.

Carter, C. and Haloupek, W. J. (2000). Spatial Autocorrelation in a Retail Context, International Real Estate Review, 3:1, pp. 34-48.

Chinloy, P. and Maribojoc, E. (1998). Expense and Rent Strategies in Real Estate Management, Journal of Real Estate Research, 15:3, pp. 267-282.

Christaller, W. (1933). Die Zentralen Orte in Suddeutschland, Verlag. Traduit par Baskin, C.W., (1966), Central Places in Southern Germany, Englewood Cliffs, N. J., Prentice-Hall, 230 pages.

Chun, G. H., Epli, M. J. and Shilling, J. D. (2001). A simulation analysis of the relationship between retail sales and shopping center rents, The Journal of Real Estate Research, 21:3, pp. 163-186.

Colwell, Peter F. and Munneke, H. J. (1998). Percentage Leases and the Advantages of Regional Malls, Journal of Real Estate Research, 15 :3, pp. 239-252.

Des Rosiers, F. and Thériault, M. (1994). Implicit Prices of Rental Services : Modeling the Quebec Market, Assessment Journal, 1: 4, pp. 47-60.

Des Rosiers, F. and Thériault, M. (1996). Rental Amenities and the Stability of Hedonic Prices : A Comparative Analysis of Five Market Segments, The Journal of Real Estate Research, 12 :1, pp. 17-36.

Dubin, R. A., Pace, R. K. and Thibodeau, T. G. (1999). Spatial Autoregression Techniques for Real Estate Data, Journal of Real Estate Literature, 7: 2, pp. 79-95.

Eaton, B. C. and Lipsey, R. G. (1983). An Economic Theory of Central Places, Economic Journal, 92, pp. 56-72.

Eppli, M. J. and Benjamin, J. D. (1993). The Evolution of Shopping Center Research : A Review and Analysis, Journal of Real Estate Research, 9 :1, pp. 5-32.

Eppli, M. J. and Shilling, J.D. (1996). How Critical is a Good Location to a Regional Shopping Centre ?, Journal of Real Estate Research, 12:3, pp. 459-469.

Fisher, J. D. and Yezer, A. (1993). Spatial Structure and Rents in Shopping Centers, Working paper, Indiana University) et de Losh (Losch, A. (1938). The Nature of Economic Regions, Southern Economic Journal, 5, pp. 71-78.

Follain, J. R. and Malpezzi, S. (1980). Dissecting Housing Value and Rent, Washington, D. C. : Urban Institute, Working paper UI249-17.

Gabriel, S. and Nothaft, F. (1988). Rental Housing Markets and the Natural Vacancy Rate, AREUEA Journal, 16, pp. 419-429.

13

Gatzlaff, D. H., Sirmans, G. S. and Diskin, B. A. (1994). The Effect of Anchor Tenant Loss on Shopping Center Rents, The Journal of Real Estate Research, 9 :1, pp. 99-110.

Ghosh, A. (1986). The Value of a Mall and Other Insights from a Revised Central Place Model, Journal of Retailing, 62, pp. 244-249.

Grewal, D. R., Krishnan, R., Baker, J. and Borin, N. (1998). The Effect of Store Name, Brand Name and Price Discounts on Consumers’ Evaluations and Purchase Intentions, Journal of Retailing, 74 :3, pp. 331-352.

Griffith, D. A. (1987). Spatial Autocorrelation : A Primer, Washington D. C. : Association of American Geographers.

Hartzell, D. J., Shulman D. G. and Wurtzebach, C. H. (1987). Refining the Analysis of Regional Diversification for Income-Producing Real estate. The Journal of Real Estate Research 2 :2, pp. 85-95.

Heckman, J. (1985). Rental Price Adjustment and Investment in the Office Market, AREUEA Journal, 13, pp. 23-42.

Hotelling, H., (1929). Stability in Competition, The Economic Journal, 39, pp. 41-57. Ingene, C. A. and Ghosh, A. (1990). Consumer and Producer Behaviour in a Multipurpose Shopping

Environment, Geographic Analysis, 22, pp. 70-91. Jain, A. J. and Etgar, M. (1976). Measuring Store Image Through Multidimensional Scaling of Free Response

Data, Journal of Retailing, 52 :4, pp. 61-70. James, D. L., Durand, R. M. and Dreves, R. A. (1976). The Use of a Multi-Attribute Attitude Model in a Store

Image Study, Journal of Retailing, 52 :2, pp. 23-31. Jud, D. G., Benjamin, J. D. and Sirmans, G. S. (1996). What Do We Know about Apartments and Their

Markets ? Journal of Real Estate Research, 11, pp. 243-258. Kirkup, M. and Rafiq, M. (1994). Managing Tenant Mix in New Shopping Centers, International Journal of

Retail and Distribution Management, 22: 6, pp. 29-37. Konrad, S. (1984). Location and Spatial Pricing with non-convex Transportation Schedules, Bell Journal of

Economics, 12. Lösch, A. (1940). Die räumliche Ordnung der Wirtschaft, Jena. Traduit par Woglom, W.H. et Stolper, W.F.,

(1954), The Economics of Location, New York, Yale University. Mejia, Luis C. and Benjamin, John, D. (2002). What Do We Know About the Determinants of Shopping

Center Sales ? Spatial vs. Non-Spatial Factors, Journal of Real Estate Literature, 10:1, pp. 3-26. Mazursky, D. and Jacoby, J. (1986). Exploring the Development of Store Images, Journal of Retailing, 62 :3,

pp. 145-163. Mejia, Luis C., Eppli, M. J. and Benjamin, J., D. (2001). Retail Tenant Rents in a Market with Shopping Mall

Image Differentiation, Working Paper. Mulligan, G. F. (1983) Consumer Demand and Multipurpose Shopping Behavior, Geographical Analysis, 15,

pp. 76-81. Nelson, R. L. (1958). The Selection of Retail Locations, New York : Dodge. Nevin, J. R. and Houstan, M. J. (1980). Image as a Component of Attraction to Intraurban Shopping Areas,

Journal of Retailing, 56, pp. 77-93. Noland, C. W. (1980). Assessing Hedonic Indexes for Housing, Santa Monica, Calif. : Rand Corporation,

Working paperN-1505-HUD. Pace, R. K. and Gilley, O. W. (1998). Generalizing the OLS and the Grid Estimator, Real Estate Economics,

26: 1, pp. 331-347. Pashigan, B. and Gould, E. (1998). Internalizing Externalities : The Pricing of Space in Shopping Centers,

Journal of Law and Economics, XLI :1, pp. 115-142. Pearson, T. D., (1991). Location ! Location ! Location ! What is Location, The Appraisal Journal, 59 :1, pp.

7-20. Rosen, K. T. (1984). Towards a Model for the Office Building Sector, AREUEA Journal, 12, pp. 261-269. Roulac, S. E. (1996). Foreword State of the Discipline : Malaise or Renaissance ? Journal of Real Estate

Research, 12 :2, pp. 111-123.

14

Sirmans, G. S. and Benjamin, J. D. (1991). Determinants of Market Rent, Journal of Real Estate Research, 6, pp. 357-379.

Sirmans, C. F. and Guidry, K. A. (1993). The determinants of shopping centre rents, The Journal of Real Estate Research, 8 :1, pp. 107-115.

Sivitanides, P. S. (1997). The Rent Adjustment Process and the Structural Vacancy Rate in the Commercial Real Estate Market, Journal of Real Estate Research, 13:2, pp. 195-209.

Tay, R. S., Lau, C. K. and Leung, M. S. (1991) The Determination of Rent in Shopping Centers : Some Evidence from Hong Kong. Journal of Real Estate Literature, Cleveland, 7:2, pp. 183-196.

Vandell, K. D. and Lane, J. S. (1987). The Economics of Architecture and Urban Design : Some Preliminary Findings, Journal of the American Real Estate and Urban Economics Association, 17:2, pp. 235-265.

Weber, A. (1929). Theory of the Location of Industries, Chicago, University of Chicago Press, 259 pages. West, D. S., Von Hohenbalken, B. et Kroner, K. (1985). Tests of Intra urban Central Place Theories,

Economic Journal, 95, pp. 101-117. Wheaton, W. C. and Torto, R., G. (1995). Office Rent Indices and their Behavior Over Time, Journal of

Urban Economics, 35:2, pp. 121-139. Wheaton, W. C. (2000). Percentage Rent in Retail Leasing : The Alignment of Landlord – Tenant Interests,

Real Estate Economics, 28: 2, pp. 185-204.

________________________________________________

15

Exhi

bit 1

: Str

uctu

ring

Dat

a B

ase

by R

etai

l Cat

egor

y an

d Ty

pe o

f Sho

ppin

g C

entr

e

Nb.

%G

LA (s

f)%

Nb.

%G

LA (s

f)%

JEW

ELR

Y5

6,8%

4 63

71,

6%19

8,6%

18 4

573,

0%C

LOTH

ING

1317

,6%

23 7

148,

4%82

37,1

%21

8 12

835

,6%

SHO

ES2

2,7%

4 99

51,

8%22

10,0

%39

796

6,5%

RES

TAU

RAN

T12

16,2

%12

983

4,6%

3415

,4%

33 6

985,

5%SE

RVI

CES

2027

,0%

41 8

2614

,8%

2210

,0%

34 6

205,

7%SP

ECIA

LTY

1925

,7%

120

397

42,6

%36

16,3

%69

749

11,4

%AN

CH

OR

34,

1%74

148

26,2

%5

2,3%

197

699

32,3

%K

IOSK

00,

0%0

0,0%

10,

5%23

80,

0%To

tal

7410

0,0%

282

700

100,

0%22

110

0,0%

612

385

100,

0%

Nb.

%G

LA (s

f)%

Nb.

%G

LA (s

f)%

JEW

ELR

Y34

7,0%

25 4

221,

7%58

7,4%

48 5

162,

0%C

LOTH

ING

169

34,6

%39

7 03

726

,7%

264

33,7

%63

8 87

926

,8%

SHO

ES33

6,7%

58 3

423,

9%57

7,3%

103

133

4,3%

RES

TAU

RAN

T82

16,8

%56

923

3,8%

128

16,3

%10

3 60

44,

3%SE

RVI

CES

326,

5%38

048

2,6%

749,

4%11

4 49

44,

8%SP

ECIA

LTY

9619

,6%

307

394

20,6

%15

119

,3%

497

540

20,9

%AN

CH

OR

81,

6%60

0 53

940

,3%

162,

0%87

2 38

636

,6%

KIO

SK35

7,2%

5 61

00,

4%36

4,6%

5 84

80,

2%To

tal

489

100,

0%1

489

315

100,

0%78

410

0,0%

2 38

4 40

010

0,0%

SU

PER

_REG

ION

AL (2

)

A

ll C

ateg

orie

s (7

)St

ore

Cat

egor

y

Stor

e C

ateg

ory

CO

MM

UN

ITY

(3)

R

EGIO

NAL

(2)

16

Exhibit 2: Operational Definition of Variables

N.B.: M = Metric variable; D = Dummy variable

V A R IA B L E

O PE R A T IO N A L D E FIN IT IO N T Y PE

N E T R E N T /SQ F T N et rent / sq . foot (D ependent variable); expressed as total gross rent (base + overage) m inus transferable operating expenses.

M

B asic A ttributes G L A G ross Leasable A rea (sq. feet). M

SA L E S/SQ F T Store yearly realized sales / G LA . M

F R O N TA G E Store frontage (feet) M

JEW E L R Y Jew elry retailer. D

C L O T H IN G C lothing retailer; includes w om en 's, m en 's and children 's read-to -w ear clothes and accessories.

D

SH O ESTO R E Shoe retailer; includes fam illy shoes, w om en 's, m en 's & boy's shoes, athletic footw ear, unisex / jean store.

D

R E ST A U R A N T Food services; includes restaurant & fast food (w ith/out liquor), sandw ich & pizzas stores, candy & nuts stores.

D

SE R V IC E S Personal and financial services; includes banks & insurance, m edical & dental offices, beauty salons, cleaners & dryers, unisex hair, barbershops, travel agents.

D

SP E C IA L TY Specialty and gifts retailer; includes radio, video & m usic centres, cards & gifts, books, decorative accessories, eyeglasses-optician stores.

D

A N C H O R A nchor tenant; large, chain stores having betw een 20 000 et 200 000 sq. feet of G LA - U sually seen as providing both stability and custom er draw ing pow er to the shopping centre (W al-M art, S im ons, T oys' R ' U S, The Bay, Sears, etc.).

D

K IO SK K iosk store; sm all, light structure w ith open sides, usually located at a central position of the m all, close to high pedestrian traffic and having a G LA of betw een 70 and 300 sq. feet (cellular phone, candy and lottery stores).

D

Structural and cyclical attributes LE A SED U R N Lease duration (years) M

A G G L O M IN D X A gglom eration index, expressed as the ratio of G LA for a given retail category w ithin a shopping center to total G LA in that center.

M

C O N C N IN D X H erfindahl Index, used as a m easure of the concentration of a retail category w ithin a given shopping center; it is expressed as the sum of the squared proportion of each store’s G LA in that category.

M

O P EN L EA SE Leasable area open to negotiation during a given year, in a given shopping center, as a proportion of total G LA in that shopping centre; it is used as a m easure of the bargaining pow er of tenants.

M

C Y C L E91_92 Lease w as negotiated over the 1991-1992 recessionary period. D

17

Exhi

bit 3

: Des

crip

tive

Stat

istic

s

VARI

ABLE

SNb

. Com

pute

dTy

peM

ean

Med

ian

Mod

eSt

d. D

ev.

Min

.M

ax.

NETR

ENT/

SQFT

($)

784

M64

5430

460,

430

8G

LA (S

q.ft)

784

M3

041

1 23

32

000

10 8

2922

163

034

SALE

S/SQ

FT1 (

$)52

6M

431

339

214

335

42

627

FRO

NTAG

E*SR

EG2 (F

t.)23

1M

3428

3320

913

6JE

WEL

RY58

D0,

070

1CL

OTH

ING

264

D0,

340

1SH

OES

TORE

57D

0,07

01

REST

AURA

NT12

8D

0,16

01

SERV

ICES

74D

0,09

01

SPEC

IALT

Y15

1D

0,19

01

ANCH

OR

16D

0,02

01

KIO

SK36

D0,

050

1LE

ASED

URN3

(Yea

rs)

621

M10

1010

51

47AG

GLO

MIN

DX78

4M

0,14

0,08

0,22

0,13

0,00

0,86

CONC

NIND

X78

4M

0,11

0,08

0,02

0,13

0,02

1O

PENL

EASE

362

1M

0,08

0,08

0,00

0,06

0,00

0,50

CYCL

E91_

9237

D0,

050

1

1. S

ales

wer

e no

t ava

ilabl

e fo

r one

sup

er-re

gion

al s

hopp

ing

cent

re (2

58 c

ases

mis

sing

).2.

Fro

ntag

e w

as a

vaila

ble

for o

nly

one

supe

r-reg

iona

l sho

ppin

g ce

ntre

(FR

ON

TAG

E*SR

EG, 2

31 c

ases

).3.

Sin

ce y

ear o

f lea

se n

egoc

iatio

n w

as a

vaila

ble

for 6

21 s

tore

s on

ly (1

63 c

ases

mis

sing

), bo

th le

ase

dura

tion

(LEA

SED

UR

N)

an d

pro

porti

on o

f ope

n le

ases

(OPE

NLE

ASE)

wer

e co

mpu

ted

on th

is g

roun

d.

18

Fig

ure

1: N

et U

nit R

ent a

nd G

LA D

istr

ibut

ions

NET

REN

T/SQ

FT

300,0

280,0

260,0

240,0

220,0

200,0

180,0

160,0

140,0

120,0

100,0

80,0

60,0

40,0

20,0

0,0NET

REN

T/SQ

FT

Frequency

160

140

120

100 80 60 40 20 0

Std.

Dev

= 4

5,92

M

ean

= 63

,7

N =

784

,00

GLA

1600

00,0

1500

00,0

1400

00,0

1300

00,0

1200

00,0

1100

00,0

1000

00,0

9000

0,0

8000

0,0

7000

0,0

6000

0,0

5000

0,0

4000

0,0

3000

0,0

2000

0,0

1000

0,0

0,0GLA

Frequency

700

600

500

400

300

200

100 0

Std.

Dev

= 1

0829

,43

M

ean

= 30

41,3

N =

784

,00

19

Exhibit 4: Model A - Linear Form / N=769 / K=23

Model Summaryb

,850 ,722 ,714 24,60Model1

R R SquareAdjusted R

SquareStd. Error of the

Estimate

Dependent Variable: NETRENT/SQFTb.

ANOVAb

1171410 23 50931 84,162 ,000450838 745 605

1622248 768

RegressionResidualTotal

Model1

Sum ofSquares df Mean Square F Sig.

Dependent Variable: NETRENT/SQFTb.

Coefficientsa

157,98 8,44 18,71 ,000-14,88 1,12 -,371 -13,25 ,000 2,10

,32 ,05 ,131 6,03 ,000 1,25-29,40 11,21 -,051 -2,62 ,009 1,0325,64 6,20 ,087 4,13 ,000 1,1826,48 4,70 ,119 5,63 ,000 1,19

-30,21 7,37 -,082 -4,10 ,000 1,0615,34 4,89 ,103 3,14 ,002 2,89

-29,64 17,51 -,033 -1,69 ,091 1,0116,67 5,85 ,061 2,85 ,005 1,21

-22,04 7,72 -,057 -2,85 ,004 1,0735,51 5,11 ,159 6,94 ,000 1,4029,69 3,52 ,195 8,42 ,000 1,44

-25,82 5,98 -,087 -4,32 ,000 1,10-25,64 6,29 -,082 -4,08 ,000 1,0915,07 5,54 ,068 2,72 ,007 1,69-4,65 3,16 -,033 -1,47 ,141 1,3657,34 24,77 ,045 2,32 ,021 1,01

122,24 5,14 ,555 23,76 ,000 1,4622,15 8,67 ,067 2,55 ,011 1,8330,62 16,65 ,058 1,84 ,066 2,63

-60,07 19,51 -,081 -3,08 ,002 1,87

-78,57 25,94 -,080 -3,03 ,003 1,86

-4,48 4,30 -,021 -1,04 ,298 1,05

(Constant)

Ln_GLAFRONTAGE*SREGJEWELRY*COMJEWELRY*REGJEWELRY*SREGCLOTHING*COMCLOTHING*REGSHOESTORE*COMSHOESTORE*REGRESTAURANT*COMRESTAURANT*REGRESTAURANT*SREGSERVICES*COMSPECIALTY*COMSPECIALTY*REGSPECIALTY*SREGKIOSK*REGKIOSK*SREGANCHORAGGLOMINDX*REGCONCNINDX*SREGOPENLEASE*REGCYCLE91_92

B Std. Error

Unstand'zd Coeff.

Beta

Stand'zdCoeff.

t Sig. VIF

CollinearityStatistics

Dependent Variable: NETRENT/SQFTa.

20

Exhibit 5: Semi-Log Model / N=769 / K=23

Model Summaryb

,843 ,711 ,702 ,3543Model1

R R SquareAdjusted R

SquareStd. Error of the

Estimate

Dependent Variable: Ln_NETRENT/SQFTb.

ANOVAb

231 23 10,024 79,845 ,00094 745 ,126

324 768

RegressionResidualTotal

Model1

Sum ofSquares df Mean Square F Sig.

Dependent Variable: Ln_NETRENT/SQFTb.

Coefficientsa

5,725 ,122 47,07 ,000-,259 ,016 -,457 -16,04 ,000 2,10,006 ,001 ,177 8,02 ,000 1,25

-,780 ,161 -,097 -4,83 ,000 1,03,345 ,089 ,083 3,86 ,000 1,18,291 ,068 ,092 4,29 ,000 1,19

-,928 ,106 -,177 -8,74 ,000 1,06,206 ,070 ,098 2,92 ,004 2,89

-1,189 ,252 -,093 -4,71 ,000 1,01,295 ,084 ,076 3,50 ,000 1,21

-,518 ,111 -,095 -4,66 ,000 1,07,429 ,074 ,136 5,82 ,000 1,40,361 ,051 ,168 7,12 ,000 1,44

-,800 ,086 -,191 -9,28 ,000 1,10-,819 ,091 -,186 -9,05 ,000 1,09,191 ,080 ,061 2,40 ,017 1,69

-,114 ,045 -,057 -2,50 ,013 1,36,591 ,357 ,033 1,66 ,098 1,01,891 ,074 ,286 12,02 ,000 1,46

-,269 ,125 -,057 -2,15 ,032 1,83,940 ,240 ,125 3,92 ,000 2,63

-1,350 ,281 -,129 -4,80 ,000 1,87

-1,028 ,374 -,074 -2,75 ,006 1,86

-,129 ,062 -,042 -2,08 ,038 1,05

(Constant)

Ln_GLAFRONTAGE*SREGJEWELRY*COMJEWELRY*REGJEWELRY*SREGCLOTHING*COMCLOTHING*REGSHOESTORE*COMSHOESTORE*REGRESTAURANT*COMRESTAURANT*REGRESTAURANT*SREGSERVICES*COMSPECIALTY*COMSPECIALTY*REGSPECIALTY*SREGKIOSK*REGKIOSK*SREGANCHORAGGLOMINDX*REGCONCNINDX*SREGOPENLEASE*REGCYCLE91_92

BStd.Error

Unstand'zd Coeff.

Beta

Stand'zdCoeff.

t Sig. VIF

CollinearityStatistics

Dependent Variable: Ln_NETRENT/SQFTa.