Introduction

It is argued by Wall et al. (1985, 604) that “by definition tourism involves the movement of people away from their place of permanent residence and their temporary sojourn in other locations”. This form of mobility triggers a demand for lodging and other services (car hire, restaurants, tours) in order to satisfy tourist needs. As tourism expands, the provision of commercial accommodation emerges as a business sector, including hotels, guest houses, bed and breakfasts, backpacker hostels and homestays (Timothy and Teye 2009). The accommodation sector offers “temporary homes away from home for tourists, and, as such, are the focal points from which most tourism activity emanates in a destination” (Shoval et al. 2011, 1594). In the USA Raitz and Jones (1988) consider hotels as historical cornerstones of urban development, as landscape artefacts which marked the early business and social core of cities symbolizing community progress and achievement as well as investment opportunity.

In modern cities the structures built to supply tourism markets “have become increasingly visible in the landscape” (Wall et al. 1985, 604). Pearce (1979) shows that various changes in city form in western cities can be attributed to the development of high rise hotels in central business districts and of other hotels in suburban areas. Hotel development is at the heart of the growth of urban tourism and “without proper hotel facilities even the most richly attractive city could not become a major tourist centre” (Arbel and Pizam 1977, 18). Overall, Arbel and Pizam (1977, 18) can assert “the primary factor in the development of urban tourism is the accommodation industry”. Tourism accommodation is viewed as affecting urban form and function; in addition, the image of cities can be influenced by the types of hotels that are established (Wall et al. 1985). Arguably, therefore, “hotels are the most prominent and representative expression of tourism” (Shoval and Cohen-Hattab 2001, 910). Within the contemporary cityscape, hotels are considered as perhaps “the only facet of the urban tourism product of which tourists are virtually the sole consumers” (Shoval 2006, 59). Despite the central role played by hotels within city landscapes as well as for the making of modern urban tourism, the hotel sector is under-researched both in urban studies and tourism scholarship. In the extensive international literature on urban restructuring and spatial change, the accommodation sector in general and hotels in particular are overlooked especially in comparison to other urban functions such as housing, retailing or office development (Shoval and Cohen-Hattab 2001, 911). It is against such a backcloth that this paper seeks to offer a modest contribution to the sparse scholarship on spatial patterns of hotel development and location change within the urban developing world.

As a whole, issues around hotel development and the location of hotel accommodation are so far little represented on the African tourism research agenda (Rogerson and Rogerson 2011; Rogerson and Visser 2011a). The focus here is on urban South Africa where during the past decade, there occurred a substantial upturn of scholarship, mainly contributed by urban geographers and planners, on issues of economic restructuring and spatial change associated with post-apartheid transformation and the growth of service-oriented cities (e.g. Turok and Watson 2001; Beavon 2004; Lemanski 2006; Visser and Kotze 2008; Borel-Saladin and Crankshaw 2009; Tredoux and Dixon 2009; Pirie 2010; South African Cities Network 2011). Notwithstanding this surge of interest around urban restructuring and spatial change, it is remarkable to observe that almost no attention has been given to the roles and impacts of the hotel sector. For example, in Pirie’s (2005) detailed overview of 15 years of urban writings on Cape Town, no mention is made of the hotel sector. Even within the expansion of scholarship around urban tourism in South Africa (see e.g. Rogerson 2002; Rogerson and Visser 2007), it is acknowledged that the hotel sector has been neglected (Rogerson and Visser 2011a, b). In a national analysis of research trends and directions, Visser and Hoogendoorn (2011) identify hotels as one of the critical “investigatory blindspots” of South African tourism research. Among issues scrutinized in recent South African work on hotels are inter alia, the changing historical nature of the hotel (Rogerson 2011), property development (Venter and Cloete 2007), the segmentation of the hotel sector (Rogerson 2010, 2011a, b; Rogerson and Kotze 2011) and the adoption of responsible tourism practices (Kohler 2007; Van der Merwe and Wöcke 2007; Frey and George 2010). The topic of the spatial distribution of hotel accommodation in cities has not been investigated.

Urban tourism in South Africa has recorded a phase of considerable expansion in recent decades with some significant variations noted between the country’s major cities (Rogerson and Visser 2007). This article focuses at the intra-urban scale of analysis on the locational distribution of hotels in three urban tourism destinations, Cape Town, Durban and Port Elizabeth. All three cities benefited from the post-apartheid expansion of tourism which has occurred in South Africa, with Cape Town in particular boosted as a popular focus for long haul international tourism (Rogerson and Visser 2004, 2006, 2007; Ferreira and Visser 2007). Specifically, this research seeks to investigate the shifting patterns of hotels during the period 1990–2010 as one facet of wider processes of urban spatial restructuring taking place in these leading South African coastal cities. Although for each of these cities there is useful work which documents the early evolution of coastal tourism (Bickford-Smith 2009) as well as various aspects of recent urban tourism growth (Grant and Butler-Adam 1992; Visser 2002; Heath 2004; Kotze 2006; Maharaj et al. 2006; Ferreira and Visser 2007; Pirie 2007a; Preston-Whyte and Scott 2007; Ferreira 2011), the local geographies of hotels remain a research void. Methodologically, this paper draws upon material sourced from a national audit and two comprehensive databases which were prepared on the South African hotel sector for 1990 and 2010. The research databases for 1990 and 2010 contain information for each individual hotel establishment concerning size (number of rooms), quality (star rating) and location. The study follows Timothy and Wall (1995) in applying cartographic and qualitative tools to interpreting this developing world case of urban hotel location.

International Research on Hotel Location

Among others Egan and Nield (2000, 613) point to the limited range of research on the intra-urban location of hotels. Currently, most of the existing empirical literature centres on hotel location in tourist historic cities of Western Europe (Ashworth 1989: Ashworth and Tunbridge 1990; Urtasun and Gutiérrez 2006) and Asia (Timothy and Wall 1995). Other works on hotel location are spread across a range of cities in Canada (Wall et al. 1985), Israel (Arbel and Pizam 1977; Shoval and Cohen-Hattab 2001; Shoval 2006), New Zealand (Pearce 1987), Turkey (Dökmeci and Balta 1999) and USA (Baum and Mezias 1992; De Bres 1994; Baum and Haveman 1997). Timothy and Wall (1995, 64) contend that much work on urban tourism has “focused on the cities of the western world and there is a dearth of studies of urban tourism, especially the spatial distribution of tourism functions, in other locations”. Importantly, in respect of a research agenda for developing countries, Oppermann et al. (1996, 55) make clear that “analyses of hotel location and the evolution of the urban tourism landscape are almost non-existent”. Within the past two decades, the only contributions on the geography of hotels in modern urban tourism destinations of the global South are those authored by Oppermann et al. (1996) for Kuala Lumpur, Malaysia and Bégin (2000) for Xiamen, China.

The current international literature considers the question of urban hotel location from a variety of perspectives. Shoval (2006) applies what is considered to be the first model of the intra-urban location of hotels utilising a concentric land use model which is based on the principle of land rent curves and representing a derivative of the classic von Thünen agricultural land use model (von Thünen 1826). This particular model locates the hotel district in the city centre and more specifically between the central business district (CBD) and a zone of commerce. Several other studies on hotel location utilise the seminal works by Ashworth (1989) and Ashworth and Tunbridge (1990) on accommodation in historic cities as the basis for their empirical investigations. Ashworth (1989) and Ashworth and Tunbridge (1990) produced a model of urban hotel location which is anchored on the tourist historic city. This influential study of urban tourism acknowledged that “those interested in the study of tourism have tended to neglect the urban context in which much of it is set, while those interested in urban studies…have been equally neglectful of the importance of the tourist function of cities” (cited in Hall and Page 1999, 162).

The Ashworth and Tunbridge (1990) model divides the tourist historic city into distinct zones where specific hotel profiles cluster. Based upon an examination of urban tourism in European cities, they stated that hotel location evolves with urban development. Six types of hotel locations are differentiated namely: (1) historic city locations, (2) railway station locations, (3) along main access routes, (4) clusters of small and medium hotels in ‘nice’ locations, (5) large modern hotels in a transition zone between the CBD and historic city, and (6) the urban periphery including along motorways and airport transport interchanges. Each of these zones corresponds to a particular hotel profile. Of special significance for location analysis is the transitional zone which is observed in many tourist cities as the prime location for new hotels and tourist-oriented activities, situated between an increasingly specialized tourist historic core and a new commercial centre, “integrating tourism and hotels into the renewed urban structure” (Bégin 2000, 452).

Following the appearance of the Ashworth and Tunbridge (1990) model, a number of tourist scholars applied it to specific case studies of hotel location and tourist functions. The Timothy and Wall (1995) study of Yogyakarta in Indonesia notes that unlike many other historical cities where the attractions are central, the tourist attractions in this case study were situated some distance from the city centre. Their study applied the historic city hotel location model and concluded that the spatial distribution of tourist land uses in Yogyakarta “have a considerable resemblance to those portrayed by the models of Ashworth and Tunbridge” (Timothy and Wall 1995, 72). Moreover, it was considered the “locational classification of commercial accommodation proposed by the former authors can also be applied to Yogyakarta” (Timothy and Wall 1995, 72). In another application Bégin (2000) interrogates changing patterns of hotel location in the Chinese city of Xiamen. It was noted that “along with urban development and the rapid growth of tourism, hotels have become a ubiquitous component of China’s urban landscape” (Bégin 2000, 449). The aim was to analyse how the distribution of hotels evolved amidst the radical transformations which have affected Chinese urban and tourism developments. The results confirm that whilst tourism is a relatively recent activity in urban China, the pattern of hotel location corresponds fairly closely with that described by Ashworth and Tunbridge (Bégin 2000). Analysis of hotel distributions in this Chinese city indicated that as tourism expanded and new hotels opened, the centre of gravity gradually shifted away from the old core towards more recent neighbourhoods. This spatial shift occurred for similar reasons as recorded in European tourist historic cities, namely the lack of available land for new hotels, congestion in the old town, restrictions on building size and accessibility (Bégin 2000).

In an earlier work conducted on Malaysia’s capital city of Kuala Lumpur, Oppermann et al. (1996) suggest that although many hotel location zones in Kuala Lumpur can be likened to those in western cities, certain distinct differences are evidenced in South East Asian tourist cities. This particular study identified seven different types of hotel locations in Kuala Lumpur. Several of these hotel location zones, such as the railway station, historic city, access roads and airport locations, coincide with the Ashworth and Tunbridge (1990) typology for developed countries. Significantly, however, others deviate. The most notable is the emergence of a new CBD in what was once a prestigious residential suburb. Oppermann et al. (1996) suggest that the city’s ‘Golden Triangle’, Kuala Lumpur’s foremost hotel location, represents a new location type which is commonly found across urban South East Asia where there are combinations of large modern hotels and deluxe shopping centres. Other related examples are Orchard Road, Singapore and similar hotel clusters in Jakarta and Kuching. Accordingly, it is argued that “different factors may be at work in developing countries as compared to the developed world” with the added caution that the transference of urban tourism concepts “may be inappropriate” (Oppermann et al. 1996, 62).

Hotel location is viewed as an important indicator of changing inner city space. Often, however, rapid changes may take place in hotel location “in relation to urban changes” (Bégin 2000, 463). In addition, it is evident that shifts in hotel location frequently evolve in relation to technological changes; in particular, “location is linked to the dominant form of transportation technology” (Shoval and Cohen-Hattab 2001, 911). Indeed, “with the widespread introduction of air travel, hotels developed in the vicinity of airports, partly re-introducing a focal point of new hotel construction not unsimilar to the railway age” (Oppermann et al. 1996, 56). The evolutionary locational trajectories of hotels have been tracked for different historical time periods across several cities including Madrid (Urtasun and Gutiérrez 2006), Istanbul (Dökmeci and Balta 1999) and Toronto (Wall et al. 1985). The complexities of political upheavals and social and cultural differences over a 150-year study period are additional considerations for interpreting the location of hotels and tourist accommodation in Jerusalem (Shoval and Cohen-Hattab 2001). Research on tourist Jerusalem, however, suggests the evolution and transformation of the city resulted in a pattern of hotel location “directly in line with the model proposed by Ashworth” (Shoval and Cohen-Hattab 2001, 921). Nonetheless, whilst Ashworth and Tunbridge (1990) utilised factors such as advances in transportation, access, land values, historical continuity and land use policies as explanations for the changing location of accommodation, in Jerusalem other contributing factors relate to its unique history as an occupied or divided city and the many “social and cultural differences between the two largest population groups in the city: Jews and Muslims” (Shoval and Cohen-Hattab 2001, 922).

Further analysis of urban hotel development has been undertaken by modelling the spatial hierarchy of hotels applying bid rent or land rent models which are founded once again upon the classic works of von Thünen (1826) and Alonso (1964). Egan and Nield (2000) maintain their study is not targeted “to explain the location of hotels in general, but rather that we are seeking to develop a model that will explain the observable spatial hierarchy of hotels within the major cities of the UK”. It is proposed that “within cities it is possible to identify a hierarchy of hotels based on location” (Egan and Nield 2000, 613). By using Alonso’s model of bid rent analysis, they predict the intra-urban location of three different hotel types, namely of luxury hotels, budget hotels and business hotels. It is considered that luxury hotels generally locate in the inner city and as they compete with other inner city users, potential locations are rare and carry high land values. Business hotels mainly are located in secondary locations at the edge of the city centre, and budget hotels are forced to the edge of city locations unable to outbid the other users closer to the city. This kind of location is, however, favoured by the mass tourist market where low room rates are important considerations for package tour operators. A parallel examination of differentiation of hotel locations by types of hotel was carried out in Jerusalem where average room rates in the centre were measured against the urban periphery using a concentric model of land use founded on the principle of land rent curves (Shoval 2006). The findings revealed that hotels in the centre are preferred by independent tourists and secure higher average room rates than hotels away from the centre. In addition, the more distant hotels are frequented mainly by tour groups which pay lower prices due to bulk bookings. Overall, it is concluded that such results “underline the fundamental importance of a hotel’s location and how vital it is for there to be an accord between a hotel’s category, its potential customers and its location” (Shoval 2006, 71).

Within often relatively compact and well-defined geographic areas or ‘hotel districts’, there exist high levels of localized competition as documented in New York’s Manhattan hotel industry (Baum and Mezias 1992; Baum and Haveman 1997). Hotel entrepreneurs confront important location choices for new establishments, whether to locate a new venture close to competitors, both in geographic and product space, or to position far from competitors in respect of geographic and product space (Urtasun and Gutiérrez 2006, 385). Certain authors recognize negative aspects of the impact of hotel development in inner city areas, especially within historic cities. Hotels divert scarce land from business and housing in inner cities, and the increasing flow of tourists can intensify pressure on the general state of urban infrastructure. The development of hotels in the inner core of historical cities can threaten their unique and famous skylines (Dökmeci and Balta 1999). In the example of Istanbul it was disclosed that, despite town planning regulations and traffic restrictions, the group of one-star to four-star hotels still dominate and cluster in the city’s old centre among its many historic buildings. Here it was recognized that future sites for business and tourist hotels need to be found “without destroying the world-famous skyline of this historically unique city” (Dökmeci and Balta 1999, 108). Accordingly, it has been concluded that whilst urban tourism—including the development of clusters of hotels—has many economic benefits, it also creates serious social and environmental problems (Arbel and Pizam 1977, 18).

Overall, it is now evident from international case studies that the development of tourism accommodation is not evenly or randomly distributed throughout cities, rather it is confined in clusters to certain specific locations. Factors impacting upon the urban geography of hotels will encompass “the morphology, history and various economic functions of the particular city in question, as well as the type of tourism it attracts” (Shoval 2006, 62). In Toronto the two strategic locations for hotels are the city downtown and airport strip (Wall et al. 1985, 614). In US cities Getz (1993) pointed to the growth of so-termed tourism business districts to designate areas of a city where tourist-oriented attractions and services (including hotels) are located together with CBD functions. Outside the city centre, beyond airport locations (see McNeill 2009), other potential bases for hotel clusters can be in the vicinity of convention centres, trade fair sites or beach strips in coastal cities such as Rio de Janeiro or Tel Aviv (Shoval 2006). Furthermore, it is clear that “the spatial pattern of hotels in world cities, such as New York, London and Paris, will be different from that of a small historic city which, in turn, will differ from that of a city without a tradition of tourism, whose visitors are mostly there on business” (Shoval 2006, 62).

The Changing Patterns of Hotels in South Africa’s Coastal Cities

South Africa’s cities have been major beneficiaries from the expansion of both international and domestic tourism which has taken place in the country over the past two decades, and most especially since the 1994 democratic transition (Rogerson and Visser 2004, 2007). Although all three coastal cities under investigation—Cape Town, Durban and Port Elizabeth—represent long-established leisure tourism destinations, local stakeholders have recognized tourism’s importance for economic expansion and made energetic efforts to promote and diversify tourism as one of the leading sectors for driving local economic development (Rogerson and Visser 2007). In all three cities, urban tourism development has been encouraged by investment in new or improved tourism products and infrastructure, including waterfront developments, beachfront improvements, leisure entertainment centres, casinos, convention centres, events promotion, and airport upgrading. Considerable impetus for new product innovation and tourism product upgrading was derived from the FIFA 2010 Soccer World Cup during which Cape Town, Durban and Port Elizabeth were all host cities. This section analyses the profile of hotel stock in the three cities in terms of numbers of hotel establishments, quality of hotel stock (indexed by star grading) and size of hotel establishment. Subsequently the discussion moves to analyse the shifting spatial patterns of hotels as a whole and as differentiated by quality standards of hotels.

Growth, Upgrading and Hotel Size

The heightened tempo of tourism within the three coastal cities is reflected in the expansion of the local accommodation sector, including hotels. Table 1 shows the growth in numbers of hotels that occurred in the three cities. An uneven performance is evident in the net growth of hotels between the three centres with Cape Town recording an extraordinary boom in hotel numbers which is explained by the city’s emergence a popular international tourism destination during the post-apartheid period (Visser 2002; Ferreira and Visser 2007). Between 1990 and 2010, Port Elizabeth records also a notable increase in the net growth of hotels with a near doubling of the city’s hotel stock. This expansion in hotel establishment numbers accompanied the city’s consolidation as a mainly domestic leisure destination (Heath 2004). By contrast, a much slower pace of hotel growth is apparent in Durban. Historically, Durban was South Africa’s premier traditional domestic leisure tourism destination. The city experienced, however, a major shift in its tourism market during the 1990s because of a relative decline in its popularity for the predominantly white middle class clientele. As a consequence, Maharaj et al. (2006) and Preston-Whyte and Scott (2007) point out that Durban faced the challenge of reinventing itself as a tourism destination.

Table 1 Growth in number of hotel establishments 1990–2010
Full size table

Further insight into the changing stock of hotel accommodation in the three cities is provided by Table 2 which shows the shifts in the quality of hotels in each of the cities. What is evident in all three cities is that there has been an upward shift in the quality of hotel stock with a much higher proportion of hotels graded as three- to five-star accommodation as opposed to the lower one- to two-star stock of hotels. For example, in Cape Town, there is a dramatic growth in the group of upmarket four- to five-star-graded hotels from a total of 21 establishments in 1990 to 145 establishments by 2010. Indeed, by 2010, the largest share of all five-star hotel accommodation in South Africa was situated in the city of Cape Town. The upgrading of the hotel stock in Cape Town is further shown by the finding that in 1990 the largest number of hotel rooms in the city was in the group of three-star hotels (1,030 rooms); by 2010, this had changed to the largest number of available hotel rooms being in the category of four-star accommodation (6,152 rooms). One niche segment of the growth in four- to five-star accommodation in Cape Town is the emergence of boutique hotels (Rogerson 2010). In Durban a parallel upgrading of hotel stock is apparent with the greatest expansion taking place in the tier of three- to four-star accommodation. As indexed by numbers of available rooms, in both 1990 and 2010, the largest share of hotel accommodation in Durban was provided in the segment of three-star accommodation; in 1990, Durban had a total of 1,892 three-star rooms whereas in 2010 the number was 3,117 in total. In Port Elizabeth considerable upgrading of hotel stock is again recorded; in 1990, the city had only four hotels within the three- to five-star grading; by 2010, 19 hotels in the city were positioned in the category of three- to five-star accommodation. Alongside the trend towards a growth in three- to five-star-graded hotels in all the coastal cities, Table 2 records the net decline in the stock of one- to two-star-graded hotels. In total for 1990, the three cities had 30 two-star hotels and 41 one-star establishments; by 2010, this had dwindled to 6 two-star hotels and 19 one-star establishments.

Table 2 Number of hotels by star grading 1990–2010
Full size table

Another element of the changing hotel stock in South Africa’s coastal cities between 1990 and 2010 relates to the size of hotels, which is indexed by number of rooms. Table 3 provides the profile of hotel accommodation in the three cities as differentiated by size of establishment. Some differences are evident between the three coastal cities in respect of the size distribution of their hotel stock. In the case of Cape Town in 1990, the hotel stock was dominated by hotels with 50 rooms or smaller which represented 61% of the city’s hotel establishments. By 2010 there is a much wider spread of available hotel accommodation in the city in terms of different sizes of hotels. In terms of average size of hotel, in Cape Town in 1990 this was 66 rooms and in 2010 73 rooms. In Durban the city’s hotel accommodation was dominated in 1990 by its group of large hotels (more than 250 rooms) which represented 58% of the city’s stock of hotel accommodation; by 2010, this had declined to only 36% of the city’s hotel rooms. During the study period, the largest expansion is recorded in the group of hotels offering between 101 and 250 rooms; these grew from a 21% share of city hotel rooms in 1990 to 41% by 2010. Of the three coastal cities, Durban records the highest average size of hotel; in 1990, the average size of hotel was 95 rooms, by 2010 the average size was 116 rooms. Arguably, the healthy growth of hotels in Port Elizabeth during the period 1990–2010 has been led by establishments in the size range of 51–100 rooms; whereas in 1990, this size class of hotel contained only 25% of the city’s stock of hotel rooms, by 2010 this had risen to 43% of the total. In common with both Durban and Cape Town, the average size of hotel establishment in Port Elizabeth grew from 61 rooms to 87 rooms between 1990 and 2010.

Table 3 Numbers of hotels by size class as determined by rooms
Full size table

Location Patterns

The changing spatial patterns of hotel accommodation are examined here separately for each city in order to identify local trends in relation to wider processes of urban restructuring. A discussion of observed macrotrends in relation to international debates and patterns of hotel location is reserved for the concluding section.

Figure 1 maps the location patterns of hotels in the Cape Town metropolitan area in 1990 and 2010. It is evident that hotel accommodation in Cape Town is clustered in particular parts of the city. In 1990 the largest clusters of hotels are recorded in the CBD and the proximate coastal area of Sea Point. The rapid growth of new hotels in the city between 1990 and 2010 is mainly concentrated in and around the CBD, Foreshore and the Victoria and Alfred Waterfront, and the adjacent coastal areas, notably Sea Point, Green Point and Bantry Bay. By 2010 distinct clusters of hotels can be observed away from this inner core of Cape Town in the upmarket southern suburbs of the city, at Claremont and Constantia, as well as coastal areas of Bloubergstrand, Hout Bay, Simon’s Town and Fish Hoek. On the edge of the Cape Town metropolitan area, other clusters can be found at the decentralized nodes around Milnerton, Bellville and Somerset West.

Fig. 1
figure 1

The spatial distribution of hotels in Cape Town 1990 and 2010

Full size image

What is most striking about the location of hotels in the Cape Town metropolitan area in 1990 and 2010 is the continued dominance and strengthening of the inner city core and adjacent coastal suburbs. This pattern of hotel location can be accounted for by initiatives launched by the city council in partnership with the private sector to renew and upgrade the inner city economy. The priorities of the Cape Town Partnership were those of “economic growth and job creation, building strong trading communities, delivering equitable and effective services and fostering a creative city” (Pirie 2007b, 128). As part of wider initiatives to physically upgrade Cape Town’s inner core, since 2000 there has been the transformation of derelict buildings, deserted historic squares, period office blocks and disused working spaces into a myriad of different types of apartment buildings, hotels, hostels, restaurants, clubs, bars and cafes. This infrastructural upgrading served to transform the inner city from merely a place of work to a 24-h city to live, work and play (Pirie 2007b). Coupled with the re-emergence of the inner city as a desirable place to work, live and visit has been the growth and continued success of the adjacent V & A Waterfront, the most popular tourist attraction in South Africa, which was transformed “from port to playground” (Ferreira and Visser 2007, 227). This successful landmark development serves to physically link the waterfront with the inner city, resulting in positive spillovers for new commercial developments in these and surrounding areas.

Another important basis for continued growth of hotels in the inner core of Cape Town relates to business tourism as opposed to leisure tourism. In 2003 the Cape Town International Convention Centre opened within the inner city. In addition, conference facilities available at other hotels in the city core as well as the waterfront confirmed the inner core as a major business tourism hub as well as leisure tourism hub. The decentralized clusters of hotels in the Cape Town metropolitan areas are commercial and retail nodes [Bellville, Milnerton (Century City), Claremont] or at what Ashworth and Tunbridge (1990) would describe as ‘nice’ locations; examples would include Constantia, Hout Bay and Simon’s Town.

Analysis of the geography of hotels in Port Elizabeth provides a contrast to the Cape Town experience for the period 1990–2010. It is shown on Fig. 2 that in 1990, the city’s small existing base of hotels was focused mainly in the CBD cluster and at Humewood proximate to the coastal strip. By 2010 the relative importance of the CBD of Port Elizabeth for hotel development had diminished markedly. The new geography of hotels in the city is dominated by the cluster of new hotel developments that had taken root at Summerstrand and the further consolidation of the Humewood area. Between 1990 and 2010, therefore, the geographical distribution of the city hotels shifted away from the inner city and CBD in particular to the outer suburbs and more especially to the Port Elizabeth beachfront. The demise of the 1990 cluster of CBD hotels is part of the broader decline of the inner cities of many large South African metropolitan areas. Although initiatives are ongoing to renew the inner core of Port Elizabeth, these initiatives have not retained the hotels in the CBD. By 2010 for both Port Elizabeth’s leisure and business tourists, the major location for hotel development is the beachfront and close environs.

Fig. 2
figure 2

The spatial distribution of hotels in Port Elizabeth 1990 and 2010

Full size image

Examination of the geography of hotels in Durban shows a combination of the spatial trends that were discernible in Cape Town and Port Elizabeth. Figure 3 reveals that in 1990, Durban hotel accommodation was concentrated in three major clusters at the beachfront, the CBD and the northern coastal node of Umhlanga. Outside of these three areas, minor hotel clusters are evidenced at the two decentralized suburban nodes of Berea and Morningside and along the arterial roads to the north, south and west of Durban. By 2010 the cluster of hotels at the CBD has strengthened, the beachfront largely has retained its hotel stock, albeit with little additional growth, and the Umhlanga node greatly expanded. Limited hotel development has taken place also at the two decentralized suburban nodes. In common with hotel development patterns in Cape Town, the Durban experience is of the continued importance of the CBD area, which adjoins the beachfront. In contrast to Cape Town, however, other hotel clusters are more significant than the CBD, a result which reflects some physical decline in the inner core of Durban which spills over to the deterioration of the beachfront during the 1990s (Preston-Whyte and Scott 2007). Although partnership initiatives have taken place to reinvigorate Durban’s inner city and beachfront, including a casino development, upgrading of the city waterfront, and establishment of a new entertainment and aquarium complex, these have not achieved the degree of success of central Cape Town in attracting major hotel clusters. One common experience between the two cities is the establishment in 1997 of an international conference centre in Durban, a natural focus for business tourism which was allied to area infrastructural upgrading and the development of additional tourism facilities. The Durban case also has a parallel with Port Elizabeth in the strengthening of newer clusters of beachfront hotels at Umhlanga, which is now a substantial decentralized retail and commercial hub. In respect of all three cities, it is observed that both in 1990 and 2010, the airports of these cities have not emerged as a focal point for any significant concentration of hotel developments.

Fig. 3
figure 3

The spatial distribution of hotels in Durban 1990 and 2010

Full size image

Finally, a brief analysis is merited of the spatial patterns of different quality standards of hotels in the three coastal cities. Figures 4, 5 and 6 show the geographical distribution of hotels of different star grading for 1990 and 2010, respectively, for Cape Town, Durban and Port Elizabeth. Across all three cities, one common observed trend is for differences to exist in the distributions of different quality standards of hotel accommodation as indexed by star grading. For example, for 1990 in Cape Town, the geography of one-star accommodation exhibits a dispersed pattern with such accommodation spread through the inner city, some coastal locations as well in parts of the city’s industrial areas. Two-star accommodation geared to budget leisure travellers is concentrated mainly in the Sea Point and CBD clusters. Three-star accommodation targeted at business as well as midrange leisure travellers is again mainly focused in the CBD, Sea Point as well as at Bellville and at the False Bay coastal strip. The upmarket four- and five-star accommodation in Cape Town is spread largely through ‘nice’ suburban areas as well as the CBD and proximate coastal areas (Fig. 4). Likewise, the spatial differentiation of hotels of different quality standards is evident from an examination of hotel patterns in Durban in 1990. Once again, the pattern of one-star accommodation is more spatially dispersed than for any other quality level of hotel accommodation. The cheaper two-star grade accommodation highlights the importance of beachfront locations both close to the CBD and at Umhlanga. The beachfront close to the CBD is the geographical axis for both three- and four-star hotel accommodation targeted at business travellers and midupper range leisure travellers. The luxury five-star accommodation is limited and occurs at the CBD, the beachfront and Umhlanga (Fig. 5).

Fig. 4
figure 4

Spatial patterns of hotels in Cape Town 1990 and 2010 by star grading

Full size image
Fig. 5
figure 5

Spatial patterns of hotels in Durban 1990 and 2010 by star grading

Full size image
Fig. 6
figure 6

Spatial patterns of hotels in Port Elizabeth 1990 and 2010 by star grading

Full size image

Over the period 1990–2010, there is evidence of both the continued spatial differentiation of hotels of varying quality standards and of certain changes in the geographical distribution of hotels of particular standards. In the case of Port Elizabeth for both 1990 and 2010, different spatial trends can be observed across hotels of different quality (Fig. 6). The Durban example discloses different geographies emerge for those hotels of lower standards (one to two star) which are declining in numbers as opposed to hotels in the mid- to luxury quality range (three to five star) which have been expanding in numbers (Fig. 5). Between 1990 and 2010, the declining group of lower standard hotels shifts from a dispersed distribution to a concentration in the CBD and adjacent beachfront. The two-star cluster existing at Umhlanga in 1990 has disappeared by 2010 as a result of upgrading of the hotel stock in this area to midrange and higher accommodation standards. By 2010 the geographical axis of the growing group of midrange (three star) to luxury accommodation has moved both northwards along the coast and away from the CBD and inland to ‘nice’ locations such as Berea, Morningside, Westville and Kloof (Fig. 5).

The Cape Town example confirms again the tendency for continued differences in the location patterns of hotels of varying quality standards in this city (Fig. 4). Once again, the lower quality levels of accommodation exhibit distributions which are clearly different to the city’s luxury end accommodation. Whereas the surge of new hotel developments which occurred in Cape Town (in particular post 2000) cluster in and around the city’s CBD, the waterfront and coastal suburbs such as Sea Point, Green Point and Bantry Bay, the declining rump of one- and two-star hotel accommodation is to be found only in cheaper properties in the CBD, industrial areas and the decentralized commercial node of Bellville. The most distinctive feature of the Cape Town maps of hotels of different standards is of the role of the CBD as premier location for hotels of all quality standards. This is as a consequence, as noted earlier, of the highly successful inner city revitalization initiatives coupled with the natural beauty of the historic inner core, seafront and the iconic views of Table Mountain. Another notable feature of the Cape Town hotel geography is the strengthening of upmarket hotel clusters in ‘nice’ locations scattered through the metropolitan area, with good examples being those at Bloubergstrand, Claremont, Constantia, Simon’s Town and Somerset West (Fig. 6).

Conclusion

Timothy and Wall (1995, 65) observe “hotels are the purest and most visible manifestations of tourism in the city”. Accompanying as well as facilitating the expansion of urban tourism are the growth and consolidation of the commercial lodging or accommodation sector, the most important element of which is the hotel (Wall et al. 1985; Timothy and Teye 2009). This article addressed an under-researched topic in scholarship around both urban tourism and urban restructuring, namely the changing spatial patterns of hotel accommodation. Arguably, most of the existing empirical work and debate around hotel location concentrates upon ‘tourist historic’ cities with the spatial balance of research weighted towards urban tourism destinations in developed countries or the global north. Although Asian cities have attracted a small amount of attention, the current literature is silent on considerations of hotel location in tourist cities of sub-Saharan Africa. Here a preliminary investigation was pursued of hotel location within the context of the recent expansion of South African tourism focusing upon the country’s three leading coastal cities.

The results of this analysis of hotel location patterns in Cape Town, Durban and Port Elizabeth show both certain common trends as well as differences with previous international research. The most obvious commonality relates to the clustering of hotels in specific parts of these urban tourism destinations. In common with the observed patterns in both developed and developing world case studies, the CBDs of South Africa’s three coastal cities have been an important locational focus for the hotel sector as a whole. Nonetheless, with the changing morphology and restructuring of cities, geographical change in hotel patterns was recorded. South Africa’s coastal cities do not evidence a parallel with the new CBD locations which were identified in the Oppermann et al. (1996) investigation. There is strong evidence, however, of what Ashworth and Tunbridge (1990) style as ‘nice’ locations in suburban areas and which in all three cities would extend to include certain beachfront locales. The most distinctive South African coastal city is Cape Town, an international tourism destination in which the CBD and its surrounds have strengthened as a locus for the hotel sector because of the waterfront redevelopments and successful partnership initiatives undertaken for the physical revitalization of the city’s downtown core. A further trend in South African cities confirming the international experience is of the existence of a marked spatial differentiation and a hierarchy of locations for hotels of different quality standards. In the final analysis, an understanding of hotel change in urban South Africa is inseparable from broader processes of local economic restructuring and the growth of service-oriented cities which are driving changes in the spatial patterns of the country’s urban centres as a whole.