- Количество слайдов: 22
European Real Estate Society Annual Conference 2010 (ERES 2010) Milan, Italy Measurement of Retail Concentration and Variety in Vertically-used Large-scale Retail Properties Dr. Tony Shun-Te Yuo Department of Real Estate and Built Environment National Taipei University, Taiwan
Introduction • Micro-scale location and allocation problems within retail properties (Shopping Centres/Malls) • Multi-unit large-scale retail property. • Seeking the maximum retail agglomeration economies-enhance property income value • Optimizing the mall configuration, predicting shopping behaviours, and allocating total floorspace according to the previously planned tenant mix strategy • However, some modern micro-scale location theory generated from the observation of out-of-town shopping centres (horizontally-used) cannot be used in some cases in Asia (vertically-used).
Principle 1: for shopping mall • The floor plan configuration should allow maximum number of customers pass the maximum number of shops (Morgan and Walker, 1988) • Using simple geometric configuration to achieve this goal
Principle 2: for shopping mall • Dumbbell shaped or its extension to I, L, Y, X, Z. Anchor stores at the mall ends and standard/smaller tenants along the single corridors that connecting the anchors. (ULI, 1999; Fong, 2003; Carter and Vandell, 2005) Standard/small stores Anchor stores Pedestrian flow plans Anchor stores Pedestrian flows
Principle 3: for shopping mall • Dispersion of non-anchor stores of the same type. (Carter and Haloupek, 2002) • We agree with this observation, However. . . “if stores of the same type…will be where transportation costs between the stores are minimized. ” (Carter and Haloupek, 2002)
Three principles • The floor plan configuration should allow maximum number of customers pass the maximum number of shops (Morgan and Walker, 1988) • Dumbbell shaped or its extension to I, L, Y, X, Z. Anchor stores at the mall ends and standard/smaller tenants along the single corridors that connecting the anchors. (ULI, 1999; Fong, 2003; Carter and Vandell, 2005) • Dispersion of non-anchor stores of the same type. (Carter and Haloupek, 2002)
Arguments • Although: – generally widely agreed by researchers for (suburban or out of town) shopping centres. • However, – most of these principles were difficult applying to the cases in Taiwan and other urban/metropolitan or city-based countries with high population density, such as Hong Kong, Singapore Japan, and China.
Arguments • 1. Vertically-used retail property-against the mall configuration derived from Anchor-store based gravity model-need some other strategy to increase the space efficiency. • 2. Dispersion of non-anchor store (Carter and Haloupek, 2002) is not because of P-median- minimizing the total distance, but maximizing the total spill-over effect. And… In Vertically-used retail property. It’s departmentalization of non-anchor store of the same type.
Located in dispersion
Example for Case A
Allocated in Departmentalization
Example of case B
Comparison of the two cases
Dispersion and departmentalization of tenants for the same type
Hypotheses for empirical study • H 1: The lower the total floor levels, the larger single floor area, and the less complex of pedestrian routes, retail tenants of the same type has to be located in dispersion: • so as to stimulate shoppers circulation, and generate higher inter-store externalities among tenants. The goal is to achieve the maximum rental values. In contrast to this hypothesis, the opposite concept should also stands, • H 2: The higher the total floor levels, the smaller single floor area, and the more complex of pedestrian routes, retail tenants of the same type has to be located in departmentalization: • so as to increase the vertical space usage efficiency. Reducing shopping costs (Searching costs and comparison costs) of customers. The goal is also to achieve the maximum rental values.
Measurement of departmentalization The index measuring the degree of departmentalization of retail categories (A 5 ij): measuring the departmentalization of retail categories (more then 3 units) clustering within 5 metres in each floor level:  Where A 5 ij: The index to measure the degree of same retail categories i agglomerate within 5 metres in floor j : departmentalized floorspaces (5 metres) within floor j. ：Total floorspace for floor j
Measurement of spatial complexity • In this paper we refined the index into a Space. Weighted Inter. Connection Density (SWICD), so as to measure the complexity under various scale of space weighted. The measurement is defined as: • Where, SWICDi: is the space-weighted interconnection density of floor i; Di: is the total number of links in floor i; Pi: is the total decision points in floor i; and Si: is the size of floor i.
Results Table 1: multi-regression results of degree of departmentalization β SE p 0. 182 0. 141 0. 2050 Y=A 5 ij α VIF Totalev 7. 43 E-02 0. 016 0. 0000*** 2. 366 SWICD 2. 82 E-03 0. 002 0. 0720* 1. 425 Units 4. 44 E-04 0. 000 0. 0130** 2. 528 R square Adj r square F test Sample size 0. 401 0. 364 10. 929*** 53 Note: *: 90% of significance, ** for 95% of significance, *** for 99% of significance Table 3 Average degree of departmentalization on the 4 level (above and under) Class N Average degree of departmentalization Sd SE Under 4 levels 8 0. 1088 0. 1052 Above 4 levels 46 0. 7740 Total 54 0. 6755 Means 95% range Min Max 0. 1967 0. 0000 0. 2779 0. 7057 0. 8424 0. 0755 1. 0000 0. 5878 0. 7632 0. 0000 1. 0000 lower upper 0. 0372 0. 0208 0. 2300 0. 0339 0. 3214 0. 0437
Conclusion • This research agrees that these principles are suitable for a planned shopping centre with no physical restriction in lot size and shape, low total floor levels and simple geometrical pedestrian flow. • Nevertheless, the first principle is difficult to sustain, while the case with characterises of multi-level, high complexity of floor plans. The second principle also is not possible to stand under a vertical structured building. • The anchors would not be a simple “in the ends” location, but have to find the locations for their original roles. That is to enhance the total spillover customer drawing power to other small tenants, but the locations for anchors may not be as simple as in the ends of main corridors anymore.
Conclusion • Cater and Haloupek (2002) suggested that, under P-median problem, nonanchor retail stores of the same type should located in dispersion, so as to enjoy the minimized total distance from supply points to demand points. • However, although agree with the dispersion concept in suburban shopping centres, we argue that the main reason for this dispersion is not for minimizing total distance but to enhance inter-store externalities of the whole centre. • It would be an opposite situation for the third principle, • While: u Vertical structure (more than 4 levels in total), u Complexity for shoppers circulation is high, • Centre managers has to reduce this complexity by departmentalizing the stores of the same type to transform the floor plan into purposive zones. • Hence the shoppers could identify the place without wayfinding difficulties.