{Accepted for publication by Transportation Research Board, 1999.}

TABLE 1. Empirical Studies of the Relationship of Nonwork Travel, Retail Structure and Urban Form

TABLE 2. Selected US Market Structure Trends with Travel Implications

FIGURE 1. Spatial Distribution of Planned Shopping Centers in the Seattle (Central Puget Sound) Metropolitan Region.

An essential tenet of the New Urbanism is that existing as well as new activity centers of metropolitan regions should be developed to higher densities and a greater mix of uses. New Urbanism suggests that if centers with different commercial functions are numerous and linked by high quality transit service, people will significantly reduce their use of automobiles for both commuting and nonwork travel. Working against this ideal of transit-oriented development (TOD) as an organizing concept for urban design are strong market and socioeconomic forces that have created the current highly variegated commercial structure characteristic of urbanized areas in the United States.

This paper reports the results of a preliminary exploration of retail industry decision making, consumer behavior, economic trends, and other key factors that have shaped the retail environment of metropolitan areas over the past fifty years to the present day. Well-known human motivations such as bargain hunting, comparison shopping, and variety-seeking are reflected in the ongoing executive decision making and resource allocation by chain stores and multinational service institutions that increasingly dominate consumer markets. Efforts to reshape existing land use patterns often meet strong opposition from current residents. This complex of consumer, organizational, and political behavior could well constrain the vehicle travel reductions sought by proponents of New Urbanism. Competitive retail industry practices and the associated consumer response are potential impediments to the success of TOD that should be addressed by metropolitan planners.

Proponents of New Urbanism believe that the number and length of automobile trips will be reduced if neighborhood development is more compact and if housing is within walking distance or a short bicycle ride of a core commercial center (1-3). They suggest this design concept, transit-oriented development (TOD), should be applied to existing as well as newly developed activity centers of metropolitan regions. Centers should be numerous and should be linked by high quality transit service.

New Urbanists think the proximity of retail stores to transit stations will induce more transit commuting. Because of the transit linkage between centers, TOD also has the potential of influencing the large number of automobile trips made to engage in nonwork activities; trips for purposes of shopping, eating out, and participating in recreational and cultural activities comprise about sixty percent of all person and household trips (4).

Efforts to reshape urban space following the principles of New Urbanism have gained a large base of support among planning professionals. Many metro and local areas in the United States have embraced TOD and incorporated it in their growth plans (3,5,6). Some areas have initiated construction of major new rail transit systems to link centers. Officials rationalize these investments in part on the assumption that high capacity transit service will induce future land use changes that will in turn increase per capita transit ridership.

Although there is an extensive body of work that addresses the retail environment and the decision-making process used to locate consumer-oriented businesses, regional planners appear to have given limited consideration to the many factors involved in causing retail establishments to locate in TOD zones. The prevalent planning view seems to be that public policies can leverage the desired spatial restructuring of commercial activities and travel patterns. For example, a transportation plan for a suburban county outside Portland, Oregon, recommends that all new retail space be "guided" to TOD sites through zoning mechanisms (7).

A systematic understanding of the market factors described in this paper would increase the realism in estimates of the extent to which the retail element of TOD can be accomplished, and how commercial activity centers might be configured to maximize transportation benefits. At the same time, comprehending these factors might also assist transportation policy makers in choosing transit investments that would likely prove cost-effective when integrated with compact development.

The issues raised here are timely from a nationwide perspective. The TRB Executive Committee has identified the impact of new urban development on travel behavior as a critical current question in transportation (8).

A central question for planners and decision makers is the magnitude of TOD’s travel impacts when translated to a regional scale in established metropolitan areas. Since new urbanism on the regional level is largely still a planner’s vision, the impact of TOD on travel demand, patterns and mode choice cannot be directly measured. Consequently, researchers have resorted to comparing older neighborhoods that approximate TOD and conventional suburban neighborhoods that do not. Other studies attempt to isolate the influence of specific design features and land use density and diversity. Also, metropolitan planning organizations and others have carried out limited modeling of TOD.

Previous empirical studies (see Table 1) suggest that compact and mixed-use development may produce localized transportation benefits. However, these investigations fall short of giving planners and decision makers confidence that the promised macro-scale transportation benefits of new urbanism can be achieved. In particular, they do not sustain the belief that the necessary restructuring of the urban landscape and retail marketplace can actually be accomplished. And even if major restructuring can be realized, they provide insufficient evidence that the large transit investments supporting the restructuring are likely to produce transportation system performance benefits that the traveling public seeks.

Dunphy suggests that if TOD is to make a meaningful difference in development patterns, there must be significant change on a regional scale (28). This change must be accomplished within the economic and political context of a particular region whose urban form has developed over a long period, the result of local zoning policies and myriad private investment decisions. Zoning, once established, is difficult to change, especially if the intent is to increase density. And because real estate is inherently a long-lived investment, a large majority of structures will still be standing at the end of the normal planning period.

Assuming that higher density centers linked by a quality transit service can be created, the scale of the transit investment required is an important consideration. Downs provides one estimate by calculating the number of TODs needed to accommodate the average population growth during the 1980’s of metropolitan areas with a 1990 population of one million or more (29). He concludes that TODs could handle the growth if their numbers were large, but that in each city this would require a regional transit system that might not be financially feasible.

Public support for such major investments requires that the potential benefits of TOD be clearly identified. Although localized benefits are important, the public will gauge benefits on a regional scale, since travel patterns usually extend beyond an individual’s home neighborhood. Most people will measure success by reduced congestion on major corridors and improved regional air quality, not by the more subjective goals of the proponents of New Urbanism such as less social segregation, a better quality of life, and a heightened sense of community.

TOD is more than a planning exercise; it involves major public investments. Sound public process dictates that officials estimate TOD’s benefits before making major policy decisions. Beyond building new regional transportation systems, governments will need to buy public services and infrastructure that support compact development -- streets, sidewalks, parks. To the extent that mode shift does not follow from the changes in land use, there will be additional public expense associated with the management of increased vehicle traffic within compact areas.

These investments will likely compete with other demands on the public purse. And, if intended benefits are not forthcoming, they will translate to lost opportunity costs for government and to wasteful expenditures of political capital required to achieve significant urban restructuring.

The primary purpose of this inquiry is to identify the forces that shape metropolitan area consumer market structure and provide insight into their possible significance to TOD. These are the same forces that are responsible for generating the current high levels of personal vehicle trips to engage in nonwork activities.

Our interest is the structural dynamics of the consumer marketplace on a metropolitan, trade area, and neighborhood scale. We are also interested in the personal and household travel behavior associated with consumer activities. Our ultimate goal is to determine whether auto usage will be significantly reduced by TOD, given the variety of activities afforded consumers, the freedom developers and business owners have to choose store locations, and the constraints on public investments. By how much can the spatial and accessibility relationships of residence, activity, and transit be restructured?

For purposes of simplicity, and to recognize the increasing integration of activities in the consumer marketplace, we use the term "retail" to refer to all consumer-related activities that fit the following categories: shopping for personal and household goods and services, eating out, engaging in recreation, and attending sports, entertainment and cultural events. The term "store" refers to the establishment or venue for each of these activities.

Our focus is on general market patterns as they have developed in North America since World War II. These patterns show a remarkable similarity from one metro area (30). We used the extensive literature on urban geography and retail distribution to develop a general understanding of these patterns. We have also developed a detailed analysis of the market structure of the Seattle metropolitan area for separate presentation (24).

The site selection processes that have produced these patterns are complex, as are the relationships between urban retail structure and trip generation and mode choice. Consequently, this was necessarily a preliminary investigation. However, we hope to stimulate more empirical research and more in-depth discussions, especially among members of the planning community, business people involved in commercial development, and decision-makers responsible for transportation policy and investments. A comprehensive exploration is important because policy makers often make expensive transportation investment decisions based on the assumption that store location decisions are malleable and that consumer activity patterns can be altered to accommodate and support those investments.

As Bookout (31) suggests, the challenge is to know the market that planning seeks to restructure, i.e., to gain more information on the "ever-changing needs, preferences, and aspirations of people who make up communities." Howe and Rabiega posed a similar question after finding that the attitudes of members of the Oregon planning profession were negative toward strip malls and positive about "urban village" forms of commercial structure: "What do consumer choices and travel patterns reveal about their relationship to the most elemental parts of the commercial urbanscape -- the stores?" (32). Calthorpe, a leading proponent of the New Urbanism, acknowledges this challenge: "Clearly much more research and analysis is needed to clarify and quantify the potential results of new land use patterns on our travel behavior. It is critical ... to effectively directing federal and state transportation dollars..." (1).

A detailed delineation of TOD has been provided by Calthorpe (1). He defines a TOD as a center with a mix of high-density residential, retail, office, public, and open space uses. Retail shops and services are in a commercial core within an easy walk of homes (600 meters, or about ten minutes). A transit station is at the center of the core. Uses in the core are "vertically integrated" -- apartments and offices rise above ground-floor stores.

Secondary areas for lower intensity uses surround the core to a distance of 1600 meters. These areas might be locations for single-family housing in a range of sizes, small parks, schools, and light industry. Housing design features include front porches, shallow setbacks from the street, and alley access to off-street parking. Streets largely conform to a grid pattern and link directly to the TOD for walking and biking access.

Calthorpe explains that the number and mix of commercial establishments in each TOD would vary depending on the size, location, and overall function of each center, whether servicing nearby residents or an entire community.

Calthorpe distinguishes two types of TODs -- urban and neighborhood -- depending on their articulation with the transit system and the intensity of their development (1). Urban TODs are located at stations on a trunk line of the regional system, which could be light rail, heavy rail, or express bus. Their locations are determined by station spacing, and are typically 0.8 to 1.6 kilometers apart. Urban TODs have high commercial intensities, employment clusters, and moderate to high residential densities.

Neighborhood TODs are located on a local or feeder bus line within three miles (no more than ten minutes) of a trunk line transit station. They are developed at moderate residential densities and provide for retail, service, entertainment, recreation, and civic uses. Neighborhood TODs can be closely spaced to form a "corridor" of activity nodes.

Proponents of TOD acknowledge that for meaningful regional transportation benefits to be achieved, centers must be linked by a high quality regional transit system, and the number of centers must be sufficient to allow for cost-effective transit operation. There must be a "transit metropolis" providing high accessibility to distant work centers from all areas (3). In other words, TOD will induce significant new transit commuting only if the work locations for many are within walking distance of stations. The regional linkage will also enable access to a range of goods, services, and recreation unavailable in one center.

Several market, consumer behavior, and political factors will play a role in determining whether the desired transportation benefits of TOD will be achieved. Our review of these factors finds that they are not nearly as well covered in the city and regional planning literature as the value-based theoretical justifications of TOD and New Urbanism. While some market features described here support TOD and the goal of reducing vehicle travel, others appear to work against public policy-based efforts to locate commercial activities in residential areas. And even if commercial activities can be located in centers, consumers may have rational reasons for not shifting trips from cars to walking and biking.

According to Jones and Simmons, market-oriented retail firms tend to locate in close proximity to other firms for two reasons: they offer goods that are similar but not identical, or they sell complementary goods (30). The former facilitates comparison shopping; the latter facilitates one-stop shopping. In both cases, clustering benefits both the retailers and their customers. Total travel distance and cost are reduced, and positive externalities are created, i.e., a total market that is greater than the sum of the individual markets when the same stores are not clustered. Agglomerations also offer retailers the benefit of reduced overhead. Parking is shared, as are other costs such as security and sometimes advertising.

Clustering occurs at several levels: in central business districts, regional malls, outlet malls, and smaller malls and retail strips along arterials. Clusters involve services as well as retail stores, e.g., post offices, libraries, banks and ATM machines in shopping centers and malls. A rising form of retail cluster is the "power center", which typically brings together a warehouse store, home improvement center, a major electronics retailer and other smaller stores offering goods and services. The mix of stores is usually subject to careful selection to maximize cumulative attraction and impulsive purchasing.

Fast food outlets cluster with department stores in regional malls to take advantage of high pedestrian flows generated by their neighbors. Restaurants also cluster. Pilsbury concluded, after an extensive study of the Atlanta-area restaurant industry, that clustering was the most important factor determining restaurant location (33). According to Pilsbury, this "competitive linkage" strategy has produced an almost total clustering of restaurants in most communities in Atlanta.

The retail marketplace continues to reinvent itself. A recent development is the "stacked entertainment zone," that may include restaurants, food courts, cinemas, ice rinks, video game arcades, art galleries, and spas. These highly diverse activity assemblages have replaced the department store as the destination anchor in some shopping centers.

Many kinds of retail firms are building larger stores that attract customers from a larger market area. Even stores that have been traditionally a part of neighborhood retail centers, and that remain so today -- groceries, barber shops, pharmacies, book stores -- have scaled up. Some retail formats have reached a market size that requires a store area, including parking, that would be a difficult fit in a core commercial center. "Big box" retailers -- discount department stores, warehouse club stores, home improvement centers, and other "category killers" -- generate large numbers of trips for many hours of both week days and weekends. Most customers arrive by car to be able conveniently to haul purchased items that are heavy, bulky, or numerous.

The growing size of market areas is obviously related to the greatly increased regional accessibility that personal vehicles and modern urban roadway systems provide. Some retail centers -- e.g., ethnic and lifestyle shopping districts, factory outlet malls, major recreation venues -- may be dependent on a market that extends across a metropolitan area. Regional shopping malls generate a considerable amount of "cross shopping," i.e., shoppers live close to one mall but also frequently shop at other regional malls.

Large stores and their lower prices are also facilitated by modern information technology, including bar-code price scanners to keep checkout lines moving, point of sale terminals wired to inventory management systems and credit/debit card networks, and global logistics management systems connecting stores to warehouses and factories worldwide. Another information technology issue driving the construction of some types of larger retail stores is the necessity of competition with the large inventories and low prices available in mail-order and on-line electronic shopping. This is true in books and computers, for example.

Even the convenience store, which has replaced the "mom and pop" neighborhood grocery as the nearest place to purchase food, bottled drinks, and tobacco, has achieved a market scale that creates a difficult fit except in locations at the edge of a residential neighborhood. Most convenience stores are on busy arterials, and their customers are drawn from a large trade area. Almost three-fourths of convenience stores sell gasoline (34). Their orientation is to drive-to and drive-by customers.

All stores seek visibility to attract customers and to provide convenient access to the site. Since the car is the dominant mode for shopping trips, many retail chains prefer stand alone sites on major roads and at key intersections. Such sites serve to project the image of the company and support its advertising, and they provide convenient access and competition-free parking. Compared to a shopping mall location, freestanding stores control their own business hours, including staying open around the clock. Stand alone sites also allow retailers to grow at faster rates than through traditional shopping center development.

Quite apart from the vehicle traffic consequences caused by large size, certain stores are difficult to site in residential areas because the activity generates noise and refuse, or just looks bad. These include modern grocery stores, auto repair services, and operations like brew-pubs that have manufacturing or processing operations just behind the customer counters.

In previously developed areas, current zoning is a central political issue for realization of restructured neighborhood retail. Residents resist rezones that allow more mixed-use development. Efforts to introduce commercial businesses into existing residential areas, even when not requiring zoning changes, often meet opposition. Expansion of commercial activities is more probable in commercial zones that have underutilized capacity. It will tend to take the form of the existing commercial center, which is most often an arterial strip.

Competition attracts price-conscious consumers to travel outside their neighborhood, trading higher travel and time costs for lower-cost merchandise. This is particularly the case when

consumers are familiar with the goods they seek and are responding to regional advertising. Newspapers typically carry advertising inserts describing low-priced goods available only in big-box stores and off-price retailers sited throughout a metropolitan region.

Although all stores generate some measure of non-local trips, some business activities (stores selling furniture, major appliances, or automobiles) generate higher levels of longer, "comparison" shopping trips, i.e., customers will bypass other similar stores to shop there. Competition for customers provides a mixture of price, quality, variety, and service choices.

People also travel farther to find variety or a unique shopping experience. For some, shopping is a recreational activity, and "satisfaction" is a large component. Malls that include food courts, multi-screen cinemas, amusement rides and electronic game parlors, concert stages, traveling festivals, and fashion, automobile, hobby, and crafts shows are playing to this preference.

Choice in the marketplace allows travelers to control the cost of a trip. For example, to avoid congestion or to combine several travel purposes in a chained trip involving one invariable destination (for example, grandma's house), consumers can access the same retail store at another location without increasing the time or direct cost of the trip.

Consumers exhibit considerable flexibility in the time scheduling of trips to retail centers, often made possible by extended store hours. Nonwork trips are combined with trips to and from work, and they originate from work sites. Tours involving several nonwork activities typically occur after work hours and on week ends.

Societal, behavioral, and market forces have combined to create major changes in retail structure and nonwork travel patterns over the last few decades. The changes they produce are ongoing and have important transportation policy implications. Consumers now have a large array of choices for their household needs, leisure-time pursuits, and other personal and household activities.

Table 2 summarizes some current national trends in the size, number, variety and spatial dispersion of stores, and provides selected examples or indicators. Equally important trends are apparent in the service and recreational sectors.