Global Businesses ‘from Below’: Ethnic Entrepreneurs in Metropolitan Areas

The urban migration in the North American cities and many European cities displays a great diversity of urban ethnic landscapes. These territories do not just show to the concentration of resident foreign-born population in certain areas inside the cities or in the suburbs but also the relevance of the economic activities developed by the immigrants. The role of migrant entrepreneurs has been critical in the revitalization of certain derelict urban cores and peripheries which have been transformed sometimes into vibrant ethnic business spaces. Some of these spaces show the importance of concentration of ethnic businesses and coethnic residents; some others present different geographic patterns, such as concentration or dispersion of ethnic businesses among a non-coethnic population; other times it is the concentration of specialized businesses. The combination of the different factors: concentration vs dispersion, diversity vs specialization and the majority vs minority relationship between ethnic businesses and the residents (monoethnicity vs multiethnicity) creates a set of different types of ethnic business spaces. The paper will explain theses types of ethnic business spaces with examples, such as Petite Asie in Paris, or multiethnic Ravalistan in Barcelona’s old town, and the geographical factors that are shaping them. The methodology is based upon a review of contemporary literature and other research findings on ethnic landscapes mainly in metropolitan North America but also in European cities.


Introduction
Urban spaces with relevant immigrant entrepreneurs have received some attention by the sociology of migrations' literature. Instead few Geographers (and not that many Sociologists either) have attempted to explain how and why these entrepreneurs concentrate or scatter in the urban landscape and what are the spatiotemporal relationships that there are between the ethnic entrepreneurs and the residents in the neighbourhoods they locate. I try in this paper to establish different types of immigrant urban landscapes according to different factors: how ethnic businesses concentrate or spread out, the coincidence or not between the ethnicity or nationality of entrepreneurs and residents in the same neighbourhood and the diversity or the specialization of the activities carried out by the entrepreneurs in a given neighbourhood. The discussion starts in this section about some geographical implications of rather well known concepts in the sociology literature and it develops in the following sections to the ethnic urban landscapes typologies I mean to categorize in this paper.
Sociological studies on migrant ethnic entrepreneurs have dealt sometimes about the geographical distribution of the ethnic businesses. The ethnic centrality of Little Havana in Miami was studied by Wilson and Portes (1980), coined as 'economy of ethnic enclave', a concentration of immigrant or ethnic entrepreneurs in a delimited urban zone.
Other authors have established expressions such as 'middlemen minorities' (Bonacich, 1973) or have preferred to use the concept of 'ethnic economy' (Light et al., S98 2000). The debate around the use of these expressions has not yet finished nor has it reached a consensus among the researchers 1 (Portes et al., 2006).
According to Bonacich (1973) 'middlemen minorities' are entrepreneurs of an ethnic background which differs from the group or the groups they serve or sell their products to. 'Middlemen minorities' constitute a minority with respect to the predominant group of the zone where their businesses are located. So 'ethnic entrepreneurs' tend not to show off their own identity through geographic concentration of their business premises, but choose rather dispersion, like the Chinese grocery stores throughout the island of Jamaica before 1911. Bonacich underlines the tendency of economic 'concentration' (as domination and functional specialisation) of the 'middlemen minorities', but rather avoids the issue of the geographical distribution of businesses.
A different strategy is carried out in the 'economy of ethnic enclave' which follows the criterion of space concentration of ethnic businesses in a delimited zone (Wilson and Portes, 1980). Nevertheless spatial concentration is not the only requirement to qualify an urban zone as an 'economy of ethnic enclave'. For example, it must have an ethnic labour market, a pool of potential ethnic employees who otherwise would not easily find a job in the general market. The two authors insist that other cases contrasting with that of the exiled Cubans of the first generation, between the 1960s and the 1970s, in the metropolitan area of Miami (the later generation of the Marielitos is excluded) and that of New York's Chinatown, studied by Zhou (1992) should be excluded from the 'economy of ethnic enclave' definition. According to Portes and Shafer, the concept of 'economy of ethnic enclave' has been applied in an unsuitable way in other cases and lament the confusion surrounding the usage of this excessively popular concept. The same authors recognise that to validate those conditions empirically is not a simple task. Waldinger (1993) argues that it is indeed the condition of geographic concentration that excessively restricts the concept of 'economy of ethnic enclave'. Light and Gold also insist on the difficulty in finding cases that realistically fulfil the requirements of the 'enclave', so that they support a much more generic concept, that of 'ethnic ecomomies', "agnostic about clustering" (Light et al., 2000: 10) Waldinger (1993), like Light and Gold, simply proposes to leave the term of 'enclave' and to retain therefore only 'ethnic economy'. Thus the non-spatiality of the 'ethnic economies' seems to be an accomplished fact. Zhou (1992) verifies that the New York's Chinatown goes beyond its more visible boundaries in the Lower East Manhattan, with the establishment of 'satellite Chinatowns' (or 'quasi suburban Chinatowns', an expression of Waldinger and another collaborator) in the boroughs of Queens, in Flushing and of Brooklyn, in Sunset Park and Bay Ridge. So does the 'economy of ethnic enclave' in fact increase (becoming 'metropolitan' rather than just an enclave) or does it lose its character? According to Zhou (1992) the 'economy of ethnic enclave' has more to do with the social and economic organisation of an ethnic group than with a geographic concentration of businesses.
The little attention that the concept of 'middlemen minorities' renders to the geographical dispersion of the ethnic entrepreneurs, the restrictions against geographical concentration in the concept of the 'economy of ethnic enclave' and the supposed nonspatiality of the 'ethnic economies' do not prevent the researcher (specially the geographer) from asking: how do geographic processes of dispersion, concentration, centralisation and suburbanisation shape the landscapes where ethnic businesses locate?

'Ethnic centralities': derelict areas towards urban revitalisation
Many authors have argued that the appearance of concentrations of ethnic businesses, mainly small shops, provide rather derelict urban areas (central or suburban) with a sense of revivification 2 . These concentrations of immigrant or ethnic entrepreneurs along with a relevant number of residents of the same ethnic origin as the entrepreneurs may be called 'ethnic centralities' 3 . Portes and Stepick (1993) indicate that Miami was rather in decline throughout the 1950s, until the arrival of the Cubans in the 1960s. In Sydney, flourishing concentrations of ethnic businesses (Chinatown, Little Italy and Asiatown) have been promoted not only by ethnic entrepreneurs themselves but also through municipal policies and foreign investors. Some city councils have viewed the proliferation of ethnic businesses as a tourist attraction for some neighbourhoods, with restaurants, exotic products shops and exotic street decoration, as well as a means of reducing unemployment in such neighbourhoods (Collins, 2006;Shaw et al., 2004;Light and Gold, 2000;Teixeira, 2006).
Ethnic centralities are numerous in North America, with its Chinatowns, Hispanic barrios or colonias, Little Tokyo, Koreatown, Little Armenia, Thai Town, Little Saigon and Little Ethiopia of Los Angeles and Little Havana in Miami, etc.; in Western Europe, with a Banglatown in London (Shaw et al., 2004), the Curry Mile of Manchester, the Petite Asie in Paris; in Southern European cities, with Chinese centralities in quartiere Sarpi in Milan, the piazza Garibaldi in Naples, the quartiere di San Lorenzo in the historical centre of Florence, in the Dreta de l'Eixample, next to the Ciutat Vella district in Barcelona, in Lavapiés in Madrid, in the district of Russafa in Valencia, etc. 'Chinese centralities' also abound in cities in Asia, Latin America or Africa. Waldinger et al. (1990) and Rekers et al. (2000) expose the lack of attention given to the influence of geographic factors on the location of immigrant or ethnic group economic activities. Kaplan (1998) and Light et al. 4 (2000), among others, emphasise the spatial association between ethnic neighbourhoods (neighbourhoods of immigrants or minorities) and concentrations of ethnic businesses. Some examples given by Light and Gold from the United States, on this issue include Italian and Polish historical neighbourhoods in Pittsburgh, Poland in Detroit and Bronzeville in Chicago. The authors comment on the geographic coincidence or space correlation of ethnic residential zones and ethnic centralities leaning each other. According to Rekers and van Kempen "clusters of ethnic entrepreneurs often emerge when an ethnic population is concentrated in an urban area" (2000: 62).

The processes of concentration and urban centralisation in the ethnic centralities
In Toronto the Portuguese ethnic centrality coincides with the zone of Portuguese residents in Little Portugal. With the decrease of Portuguese immigrants, partly because of their suburbanisation, the entrepreneurs of Portuguese origin within Little Portugal are pessimists. The Portuguese ethnic businesses depend therefore 'excessively' on the coethnic demand (Teixeira, 2006). In Berlin businesses of Turkish S100 origin are concentrated in the neighbourhoods of Kreuzberg, Wedding and Tiergarten which have a high percentage of residents of the same nationality.
I found a high positive spatial correlation between the location of Pakistani residents and the presence of Pakistani-run businesses in some areas (census blocks) of Ciutat Vella district in Barcelona; and also between location of foreign residents and the presence of telephone call shops, run by foreigners in the same district (Serra, 2006).
The diversity of economic activities constitutes an important feature of the ethnic centralities. These concentrations represent activities attached to small convenience stores for a coethnic customer base in a neighbourhood with shops such as groceries, restaurants, laundries, hairdressing salons, clothing shops, telephone call shops, shops for the shipment of remittances, film and music shops. But these spaces may also incorporate more specialised businesses for a wider customer base, such as jewellery shops, workshops, supermarkets, shopping malls, bookstores, opticians, photography and mass media shops, doctors, lawyers and accountants, consultants, travel agencies, translators and real estate agencies, supplementary education academies, day-care centres, etc.
But what comes first, the immigrant residents or the ethnic businesses? According to Waldinger et al. (1990) the textile industry of the Manhattan's Lower East End (today's Chinatown) of New York attracted Jewish worker immigrants at the end of the nineteenth century, and these in turn installed their residences near the industries and warehouses, because of their low wages and the difficulty of transportation. The area then evolved as a flourishing Jewish neighbourhood, with an ethnic centrality of more than 1430 shops in 1890. Factories attracted immigrant population which in turn attracted coethnic entrepreneurs, who provided products and services to the residents.

Exceptions to the 'ethnic centralities'. The effects of suburbanisation in areas with ethnic businesses
The existence of ethnic residential concentrations does not always imply the location of coethnic businesses in the same area. In the United States there are numerous neighbourhoods of African-American population where there are few black entrepreneurs (let alone non-black ones). Yet sometimes the absence of a critical mass of ethnic residents or a decreasing ethnic population limits the presence of numerous of coethnic businesses, like in the aforementioned Little Portugal in Toronto. Indeed "there is little consensus about the degree to which ethnic businesses require residential concentrations to survive […] Eden Center, a thriving Vietnamese shopping district in suburban Washington, DC, functions successfully without the benefit of an ethnic residential concentration" (Gober, 2000: 86). Kaplan (1998) explains the evolution of ethnic businesses concentrations in North America that remain in the urban centre and the relationship with their former coethnic population, which increasingly 'suburbanised', switching from a monoethnic composition to a multiethnic one. In this case many ethnic businesses may be in a minority with respect to the new residents in the neighbourhood.
The space dissociation between residential zones of immigrants and concentrations of coethnic businesses and spaces of coethnic socialisation seems to respond to the so called phenomenon of heterolocalism: an alternative model of social and spatial behaviour to the assimilationist and the multicultural models (Zelinsky and Lee, 1998). According to Raulin (2000) this separation is not an ethnic exclusive feature. Rather, society in general has increasingly experienced this process since the nineteenth century, with the separation of functional zones to work, to reside, to make purchases and for leisure. Rekers et al. (2000) suggest that zones of immigrants' S101 residences and concentrations of coethnic entrepreneurs may be distant each other, 'as happened in the Chinatowns of New York and Los Angeles'.
In the 'ethnic centralities with few coethnics', unlike 'ethnic centralities', there are few conational residents of the predominant group of traders in the area, so they constitute concentrations of businesses run by an ethnic group in a neighbourhood not dominated by compatriot residents. In Sidney an active concentration of ethnic businesses in the suburbs, namely Little Italy, in Leichhardt, does not contain a majority of corresponding coethnic residents (Collins, 2006). I consider therefore Little Italy an 'ethnic centrality with few coethnics': the Italian entrepreneurs are a minority with respect to a majority of non-Italian residents in that neighbourhood. In the Los Angeles Koreatown, two thirds of the inhabitants are Hispanics and just a fifth of the residents are Koreans (most affluent Koreans there left for the LA suburbs). What characterises that Koreatown is not logically the ethnicity of most of its residents, but the economic activities of Korean entrepreneurs, no matter who the customers are or where they live. So 'ethnic centralities with few coethnics' often respond to business concentrations (be they in the urban core or in the suburbs) that are compatible with suburbanisation (and so a dispersion) of coethnic residents, such as the "apparent flight" of former coethnic residents from the old 'centrality'. The permanence of a central Chinatown in downtown Toronto (thus a traditional urban "ethnic centrality") is compatible with the suburbanisation of numerous shopping malls of Chinese origin in municipalities around the city. The influence of these 'ethnic centralities with few coethnics' attracts clients beyond the Greater Toronto Region. Most of the Chinese malls are indeed not located in neighbourhoods where the Chinese population constitutes the majority, but are somehow dispersed everywhere. Waldinger et al. (1990) outline the evolutionary and changing character of the ethnic businesses and neighbourhoods: what was previously an 'ethnic centrality' may later become an 'ethnic centrality with few coethnics'.
'Multiethnic centralities' consist of concentrations of entrepreneurs of diverse nationalities or ethnic origins in neighbourhoods where, generally speaking, a coethnic resident group does not clearly prevail either. Therefore, somehow there is a correspondence in the neighbourhood between the multiethnicity of the entrepreneurs and the multiethnicity of the residents. Li (1998) has characterised ethnoburbs as multiethnic suburbs.
In the Petite Asie, a 'multiethnic centrality' in Paris, some sizeable Asian supermarkets and many smaller Asian shops are concentrated, attracting Asians from the Paris banlieues. Yet the Petite Asie neighbourhood only hosts a small fraction of Asian residents (Raulin, 2000). However most of the shops are Asian (from Cambodia, Vietnam, Laos, Taiwan and China) in a more reduced zone of the neighbourhood, the so called Triangle de Choisy or Petite Asie, so a '(multiethnic) Asian centrality'.
Dearborn, in the suburbs of Detroit (Michigan) has a large Arab population, composed of Palestinians, Lebanese, Iraqis and Yemenis (amongst others). In the east zone of the city Arabs make up 45% of the population. There is an '(multiethnic) Arab centrality', with businesses representing all the Arab nationalities of the city. The Ciutat Vella, old town district, of Barcelona represents another example of 'multiethnic centrality'. In Ciutat Vella immigration is quite high with around 37% of the total population in the district in 2005, with Pakistanis, Moroccans, Dominicans, and Filipinos, amongst others. In that year Pakistani entrepreneurs represented approximately a third of the total ethnic traders. Next in line were Indian retailers (17%), then Chinese (16%), followed by Moroccans (9%), and Bangladeshi and Filipinos (3% respectively) (Serra, 2006). S102 'Specialised ethnic centralities' constitute concentrations of businesses of the same type run usually by entrepreneurs of a single ethnic group. The functional specialisation and geographical concentration are the main features of this type of centrality. Ethnic entrepreneurs in these 'specialised centralities' are, on the other hand, usually a minority regarding the majority of the residents in the neighbourhood. There are examples of concentrations of businesses of Chinese textile wholesalers in Lavapiés, in Madrid; in the Dreta de l'Eixample neighbourhood next to Ciutat Vella, in Barcelona; in the Popincourt and Sedaine streets, in Paris; in piazza Garibaldi, in Naples.

The processes of dispersion of the 'middlemen minorities'
The geographical dispersion of 'middlemen minorities' (Bonacich, 1973) constitutes the fifth type of landscape with ethnic businesses that I analise in this paper. Waldinger et al. (1990) describe 'middlemen minorities' as 'doing business outside the ethnic community'. The dispersion constitutes a spatial resource (just like concentration), it allows for a greater non-coethnic customer base (Gober, 2000), since coethnic demand may sometimes be insufficient.
The differences between 'ethnic centralities with few coethnics' (seen in the previous section) and the dispersion of 'middlemen minorities' are based upon the geographical distribution and the type of economic activities performed. In both cases most of the entrepreneurs share the same origin, but their origin is different from the majority of the resident ethnic groups. In the first case, nevertheless, a concentration of rather varied businesses in a particular zone or neighbourhood takes place; in the second instance, there is a dispersion of rather specialised ethnic businesses on a larger area. Examples of dispersion of 'middlemen minorities' are: the Chinese laundries in New York before 1965; the Chinese restaurants in Paris and in Barcelona; the Chinese bazaars in Madrid and in Barcelona; the Korean grocers in New York who monopolised this business in the 1980s; the restaurants of Bangladeshi Sylhetis in cities of the United Kingdom and the bars and restaurants and sale of alcoholic drinks by the Lebanese in Toledo (Ohio, USA), both in the 1980s (Waldinger et al., 1990).
The dispersion of ethnic entrepreneurs may be combined with concentration strategies. Thus, by the 1980s, a third of the Los Angeles county Korean entrepreneurs were concentrated in the Koreatown, while there were Korean traders dispersed in 87% of the 10,000 postal zones of the county (Waldinger et al., 1990). Therefore, there was an 'ethnic centrality with few coethnics' in Koreatown, compatible with a dispersion of 'middlemen minorities' on a county scale. Some authors emphasise the high level of competition between the Chinese restaurants. This is why the Chinese follow the strategy of geographic dispersion.
Pakistanis and to a lesser extent Bangladeshis are retaking the convenience and grocery stores in Barcelona. These small businesses are certainly concentrated (at a city scale) in the district of Ciutat Vella and especially in the El Raval neighbourhood in the western area of the Ciutat Vella district or in the Besòs i Maresme neighbourhood, with a remarkable Pakistani population, yet Pakistani grocers are also dispersed in other neighbourhoods inhabited mostly by autochthonous population.
The cartography of almost 900 Chinese restaurants in Paris in 1990(Ma Mung, 2000 reveals concentrations in some arrondissements or districts, in special in the 13e, in the 'multiethnic centrality' of the Petite Asie. Nevertheless, most Chinese restaurants are dispersed in all the arrondissements. Concentration and urban dispersion of Chinese restaurants are therefore pronounced in Paris in a compatible, simultaneous way. In El Raval in Barcelona two ethnic businesses spatial strategies coexist at the neighbourhood scale: on the one hand, 'dispersion', well represented by convenience stores (butcheries, grocers, remittance shops, video and telephone call shops, etc.) run by Pakistanis; on the other hand, 'concentration', whose examples constitute the fifteen souvenirs' shops of the Sindi Indians in la Rambla and the Chinese textile wholesalers in the Eixample.
Tables 1 and 2 respectively show the geographic processes that are more manifest in the five types of spaces of ethnic businesses and some examples of them in world cities.

Conclusions
Concentrations of immigrant residents and of ethnic businesses often display a spatial association. Yet the neighbourhoods of immigrants do not always hold a factor S104 of attraction for coethnic businesses. Each group of ethnic entrepreneurs may use the processes of concentration or dispersion in the location of their businesses as enterprise strategies.
I showed that concentration or dispersion, product and service specialisation or diversity, and the search of a coethnic or a non coethnic demand are processes of an eminently geographical character that take place in areas where ethnic businesses locate. I have set five typologies relating to ethnic business areas: Within 'ethnic centralities' there is both a concentration of ethnic businesses and of coethnic residents sharing the same space, like the Chinatowns of New York and Los Angeles or Banglatown, in Tower Hamlets, London.
In the 'ethnic centralities with few coethnics' there is a concentration of businesses of an ethnic group in an area where few coethnics live, like Koreatown in Los Angeles.
The 'multiethnic centralities', such as Ciutat Vella in Barcelona or the Petite Asie in Paris, include businesses of different ethnic groups in a neighbourhood which has also got residents of multiple ethnic groups. The diversity of the businesses characterises these three types of centralities.
Nevertheless in the 'specialised ethnic centralities' there is a high density of ethnic businesses destined to the same economic sector, like the Chinese textile wholesalers in the Roquette neighbourhood, in Paris or in the Curry Mile in Manchester. The functional specialisation distinguishes this type of centrality from the other three centralities.
Unlike the previous four centralities, the 'middlemen minorities' (Bonacich, 1973) represents the phenomenon of geographical dispersion of ethnic specialised entrepreneurs in a city or in an ample zone where most of the residents usually have a different ethnic origin than the entrepreneurs, such as the dispersed Chinese restaurants in so many European cities.
The role of geography in the configuration of neighbourhoods that include ethnic businesses with many or few immigrant or ethnic residents seems to have been rediscovered in recent years, as the recent publication of three collective books show Miyares and Airriess, 2007;Frazier and Tettey-fio, 2006).
A comparative study of the areas of ethnic businesses of diverse cities, in order to empirically apply the typologies and processes presented in this paper could be a next step. A research of this sort may imply some challenges or difficulties, like trying to use the right geographical scales and quantitative techniques to delimit ethnic businesses' areas in metropolitan areas with different extents and densities (of businesses and of people). For instance, at a metropolitan scale the geographical distribution of some of ethnic business may look concentrated, while at a higher (more local) scale the shops may show a rather dispersed location. The use of aggregated or non-aggregated ethnic groups for both residents as much as entrepreneurs also constitutes an aspect to consider when classifying a space of ethnic businesses.