Ethnic economies in the age of retail chains: comparing the presence of chain-affiliated and independently owned ethnic restaurants in ethnic neighbourhoods

ABSTRACT

Many entrepreneurs in ethnic economies start out by serving fellow community members in their local ethnic neighbourhood. Large corporations, however, are creating retail chains that target the ethnic market and expand into ethnic neighbourhoods. Little research has determined the extent to which chain retailers are co-existing or overtaking independently owned businesses in such places. This paper uses negative binomial regressions on over 50,000 ethnic restaurants in the New York metropolitan area taken from Yelp.com in 2015 to demonstrate the pervasiveness of the ethnic chain restaurant industry in ethnic neighbourhoods. For ethnic populations whose cuisines are very popular with mainstream consumers, such as Italians, Chinese, and Mexicans, results follow the traditional theory in which chains are more likely to locate away from ethnic neighbourhoods. For groups whose cuisines are less popular, such as Caribbeans, Indians, and Koreans, however, a new pattern emerges in which chains exist alongside independent businesses in ethnic neighbourhoods. In some cases, chains are more likely than independently owned restaurants to locate in ethnic areas, which may complicate the ability of local communities to form ethnic economies and confound the ways in which ethnic economies can help immigrants achieve socioeconomic mobility.

KEYWORDS:

• Ethnic economies
• immigrant adaptation
• ethnic neighbourhoods
• entrepreneurship
• retail chains

Acknowledgements

The author would like to thank Mike Benediktsson, Paul DiMaggio, René Flores, Steven Gold, Charles Hirschman, Balázs Kovács, Douglas Massey, Martin Ruef, Seema Shah, Edward Telles, Stewart Tolnay, and Emilio Zagheni for their helpful suggestions and comments on earlier drafts. The author would also like to thank two anonymous reviewers for their feedback.

Disclosure statement

No potential conflict of interest was reported by the author.

ORCID

Mahesh Somashekhar http://orcid.org/0000-0003-2447-0317

Notes

1 Successful ethnic economies sometimes become ethnic enclave economies (Gold 2015). In this paper, I discuss successful ethnic economies more generally, although the discussion can extend to ethnic enclave economies as well.

2 http://www.paradisebiryanipointe.com/franchise.php.

3 I could have formulated Hypothesis 1a, along with Hypothesis 1b, using a proportional outcome. As the proportion of a neighbourhood’s population of a given ethnicity increases, in other words, there may be an increase in the proportion of the neighbourhood’s ethnic restaurants comprised of chains. This formulation would have been ineffective, however, because ethnic chain restaurants are not found in large numbers across all neighbourhoods. The small Ns lead to highly misleading proportions and result in unstable linear and logistic regression models, so I avoid using proportional dependent variables in the analysis.

4 http://www.goldenkrustbakery.com/content/index.php?do=company&sect=overview.

5 http://www.goldenkrustbakery.com/content/index.php?sect=viewArticle&do=pressReleases&categoryId=3&contentTypeCode=article&pressReleaseId=56&am.

6 Beyond New York City, there are smaller towns that have outright banned chain stores from entering their communities. Only one exists in the New York metro area, Port Jefferson, NY (Institute for Local Self Reliance 2006). Results remain unchanged with this place taken out of the analysis.

7 http://a816-restaurantinspection.nyc.gov/RestaurantInspection/SearchBrowse.do.

8 Given the arbitrary boundaries of Census tracts, it is possible that results would differ if I operationalised neighbourhoods using different geographical units. In robustness checks, I buffered neighbourhood boundaries one kilometre and five kilometres outside of Census tract boundaries, and findings changed very little. Results of these robustness checks are available from the author.

9 Operationalising the Creative Class is difficult to do and, in some cases, may not have much explanatory power (Nathan 2015). Nonetheless, when I replace the Creative Class variable with something similar such as the college education rate, or I take the variable out of the analysis altogether, results are virtually identical to those that are reported.

10 If a Census tract was in a relevant Neighbourhood Tabulation Area (NTA), as defined by the New York City Department of Planning, then I treat the Census tract as part of an ethnic district. The relevant NTAs are: SoHo-TriBeCa-Civic Center-Little Italy, Bensonhurst, and Belmont for Italians; Chinatown, Sunset Park, and Flushing for Chinese; Astoria and Brooklyn Heights-Cobble Hill for Mediterraneans; Murray Hill-Kips Bay and Jackson Heights for Indians; Washington Heights and Flatbush for Caribbeans; and Flushing for Koreans.