Are BIDs Good for Business? The Impact of BIDs on Neighborhood Retailers in New York City

Abstract

The business improvement district (BID) model has been widely adopted as a place-based strategy for commercial renewal and sustainability. This study uses the National Establishment Time-Series (NETS) Database to estimate the impact of New York City BIDs on change in retail sales and change employment between 2000 and 2008. Since the self-financing mechanism of BIDs becomes a compulsory tax on retail tenants, this study examines whether the model improves retail performance relative to comparable areas of the city that never adopted BIDs. Although effects vary by BID size, retail sales and employment are depressed by small community BIDs.

Keywords

business improvement districts community BIDs small businesses program impact

During the 1970s a handful of cities, starting with Toronto then New Orleans and New York City, adopted the business improvement district (BID) model as a novel mechanism for financing and managing the revitalization of forlorn commercial corridors and downtown shopping districts. By 2009 more than fifteen hundred BIDs had been established in hundreds of urban and suburban municipalities across the United States and internationally (Hoffman and Houstoun 2010). It is widely agreed that transnational adoption of BIDs is principally a function of the model’s promise to “deliver increased economic and employment activity at little or no direct cost to taxpayers” (Stokes 2006, 174). That is to say, in an era of urban entrepreneurialism, fiscal austerity, and market-based solutions to redevelopment, BIDs not only provide a variety of place-based services—such as street cleaning, security, streetscape enhancement, façade improvement and beautification, capital projects, district marketing and business support—intended to enhance overall commercial life, but BID operating expenses are largely covered by a compulsory tax assessment. All commercial and industrial property-owners within BID boundaries agree to pay additional tax to pay for services, rather than drawing on city coffers. In New York City (NYC), for instance, the overall property tax assessment constitutes approximately 75 percent of BID operating revenue. From the perspective of property owners, BID managers, and other local stakeholders, the self-financing structure of BIDs fosters local autonomy by providing a reliable stream of revenue, thereby mitigating numerous challenges that have historically bedeviled urban commercial corridors (Mitchell 2008). Despite the importance of the varied “safe and clean” BID services, as they are commonly described, an outstanding empirical question remains: Does the model yield discernible benefits for street-level BID retailers? Since the compulsory property tax assessment is typically treated as a “pass-through” to retail tenants, it is reasonable to expect that, all else equal, BID retail establishments¹ derive economic benefit subsequent to BID formation.

After decades of decline, a renaissance in urban retailing is manifest in “strong market” cities like Washington, DC, NYC, and Chicago. Unlike traditional patterns of urban development that concentrated commercial life downtown in central business districts and pedestrian malls, the contemporary variant is marked by polycentric urban retailing, or the dispersal of retailing in outlying or previously underretailed urban neighborhoods. City managers, civic leaders, and property owners increasingly champion BIDs as an effective tool for enhancing commercial corridor aesthetics, maintaining social order, and attracting new retailers, investors, and consumers (Briffault 1999, 2010; Hoyt 2005; Hoyt and Gopal-Agge 2007; Levy 2001; Mitchell 2008).

In NYC, municipal agents and civic leaders underscore the potential for BIDs to “increase property values, improve sales and decrease commercial vacancy rates” among other purported benefits (NYC Department of Small Business Services 2009, 1). According to Hoyt and Gopal-Agge’s (2007, 954) synthesis of scholarship on the functioning of BIDs globally, there is “considerable consensus around the notion that BIDs provide benefits to property owners, businesses, residents, and visitors within their jurisdiction.” Some BID proponents contend that the model helps improve business profitability and property values (Hoffman and Houstoun 2010), strategically advance retail districts (Gopal-Agge and Hoyt 2008), and helps local businesses enrich the commercial environment (Frug 2010). According to Caruso and Weber (2006, 192), other proponents claim that the model boosts “a municipality’s vitality through increased property values, higher sales tax revenues, and a better quality of life for its residents.” The self-financing structure and autonomous governance of BIDs coupled with targeted place-based investment have been characterized by some scholars as a local economic development innovation (Gross 2005; Mitchell 2008). Despite widespread support for the BID model, the utility of BIDs for retail tenants has not been systematically examined. Although BIDs purportedly produce local economic development effects through sales tax revenue and job creation, a dearth of research examines the direct effects of BIDs for neighborhood retail² performance. In all fairness, reliable tests of the effect of BIDs on firm performance has become more feasible only recently with more readily available parcel-level firm panel data necessary for constructing quasi-experimental designs that measure economic growth, before and after BID formation, relative to reasonable non-BID comparison areas.

This study relies on a novel balanced panel of establishment-level retail sales and employment data joined to BID polygons and non-BID parcels throughout NYC. These fine-grained panel data allow for making reliable estimates of retail performance, before and after BID formation, relative to observationally similar commercial districts in NYC that never established BIDs. Since a disproportionate number of BIDs established in NYC since 2000 occupy a relatively small footprint, provide a narrow scope of services, have limited organizational capacity, and are generally located in weaker economic environments of outlying neighborhoods (Gross 2005; Mitchell 2001; Stokes 2006), this study focuses on the impact of these smaller BIDs, referred to in this study as “Community” BIDs, on conventional measures of business performance, namely, sales and employment. This study addresses three primary questions: First, how do NYC BIDs affect retail performance? Second, how does retail performance within BIDs vary by BID size and retail structure (independently owned vs. chain store)? And third, what does this suggest for the trajectory of BID formation?

The next section of this article situates BIDs within two distinct literatures: one on retail theory and the other on BID impact with regard to public safety and commercial property values. The impact literature provides the methodological basis for examining retail performance. Following the literature review, I present a brief history of BIDs in NYC and summary statistics for the BIDs included in this study and non-BID areas of NYC with regard to retail attraction and retention. The fourth section presents the research design and description of the data, followed by study findings that are organized by research question. In the final section of the article, I discuss implications for planning and policy, as well as areas for further research. I confront the inevitable question of whether BIDs are good policy. One of the primary findings show that both sales and employment decline for existing independent neighborhood retailers within BIDs relative to comparable non-BID areas. This corroborates cautionary claims of qualitative studies that find disparate capacity of BIDs mediated by BID resources. That is to say, the impact of BIDs cannot be homogenized; it must be analyzed by BID type and retail structure.

How Does Retail Theory Help Us Understand BIDs?

Scholars of business, psychology, geography, and economics have developed a voluminous literature on retailing. Though much of the literature employs individual-level analyses, and focuses on consumer choice and retail location effects (Fox, Postrel, and McLaughlin 2007), there is also a robust literature concerned with firm agglomeration; central place theory or the distribution of goods and services within market areas; and spatial interaction theory, which offers a normative framework for connecting consumer shopping choices and retail environments. Spatial interaction theory assumes a trade-off between the allure of shopping district attractiveness and the deterrent of travel distance (Brown 1993; Huff 1962). The probability that consumers would choose a district within a competitive environment of alternatives is directly proportional to the district’s relative attractiveness and inversely proportion to travel distance (Brown 1993). In other words, consumers bypass proximate shopping districts for shopping options generally considered more attractive on the basis of an array of perceived factors. Numerous studies advance Huff’s (1962) early two-dimensional spatial interaction models by assuming consumers shop in multiple districts, and by also incorporating a mix of attraction and deterrence variables or “pull factors” that vary for different categories of goods and services, and socioeconomic and cultural groups (Brown 1993; Dawson and Kirby 1980; Nevin and Houston 1980; Teller and Reutterer 2008). Though retail theory is not explicitly concerned with BIDs, one can infer that in crowded or dense retail markets, BIDs serve as a mechanism for differentiating shopping districts and luring consumers by improving the environmental aesthetic and safety. By the same logic, the proliferation of BIDs within a bounded retail market such as NYC may lead to organizational saturation, which arguably limits the scope of BID “attractiveness” by muting differentiating strategies, thus diminishing place-based benefits of the BID model.

How Does BID Theory Explain the Model’s Impact?

Planning, policy, and legal scholars have produced a prodigious body of BID literature, much of it employing qualitative methodologies and offering rich case studies that document BID histories, analyze BID governance, and infer correlation between BIDs and commercial performance (Briffault 2010; Gross 2005; Hoyt and Gopal-Agge 2007; Mitchell 2008; Morcol et al. 2008; Stokes 2006). There is a tenacious belief that the BID model is a useful tool for transforming disinvested downtown districts into alluring destinations for consumption, investment, and social interaction. In a national study of BID stakeholders, Briffault (1999) finds that building owners favor the BID model over public sector service delivery. The efficacy of BIDs date back to the 1970s and 1980s, when it was increasingly popular to create special assessment districts to facilitate revival in urban downtowns, tourist destinations, and large-scale real estate investments, such as Dongal Street in Toronto, Times Square in New York City, Downtown New Orleans, and Center City in Philadelphia. More recently, however, the BID model is increasingly being adopted by community stakeholders concerned with revitalizing relatively small retail strips in outlying neighborhoods (Briffault 2010; Gross 2005; Stokes 2006, 2010). Gross (2005, 175) warns that despite diffusion of the BID model to neighborhood retail strips, the lessons large downtown BIDs “offer for local economic development professionals with weaker resource bases situated in poorer neighborhoods are limited.”

As previously mentioned, the BID model is routinely touted as a planning and policy tool for revitalizing forlorn downtown districts and tourist destinations, yet relatively few studies measure the direct effects of BIDs on the social and economic environment. This section highlights two important impact studies, though not concerned with retail performance. First, Ellen, Schwartz, and Voicu (2007) analyze the impact of NYC BIDs on property values by comparing the price of commercial and residential properties in BIDs to prices of comparable properties outside of BIDs but in similar NYC neighborhoods. The researchers use hedonic regression models to explain sales price as a function of property characteristics and neighborhood attributes, and a difference-in-difference approach to determine whether price changes are associated with the BID designation. Drawing on Gross’s (2005) analysis of the function, resources, and practices of NYC BIDs, Ellen, Schwartz, and Voicu (2007) disaggregate effects by BID size. They find that large-scale Corporate BIDs significantly increase commercial property values, whereas medium-size and small community BIDs, with mostly retail and residential properties, seem to have no significant impact on property values. This was the first study to examine potential economic benefits of BIDs for property owners. But there is no reason to assume that retail tenants derive similar benefits or burdens as property owners.

In the second study, Brooks (2008) makes important methodological and empirical contributions for understanding the impact of BIDs in the city of Los Angeles on crime reduction. Using multiple estimation techniques, including propensity score matching and fixed effects models, and comparing BIDs to their neighbors, Brooks finds that BIDs significantly reduce the incidence and severity of crimes in Los Angeles. Brooks is not the first to find that BIDs reduce crime; Hoyt (2005) finds that BIDs in Philadelphia also lower the rate of property crime. However, Brooks (2008) is the first to use multiple matching techniques and estimation models, creating a quasi-experimental design for making reasonable inferences about the direct impact of BIDs on crime reduction relative to comparable areas of Los Angeles without BIDs. Though “safe and clean” shopping environments likely attract consumers and bolster commercial activity, we cannot infer that longstanding retailers will be made better off by the foot traffic spurred by crime reduction or that commercial district activity yields discernable economic effects.

New York City BIDs

In most cities, BIDs are first established in the central business district or downtown and then spread from neighborhood to neighborhood through polycentric diffusion processes. In NYC, however, the model first emerged during the 1970s in the outer boroughs of Brooklyn and Queens rather than Manhattan. Three special assessment districts, which are structurally and functionally similar to BIDs but allow the tax levy to be used for district projects not directly related to commercial development, were adopted to bolster the fledgling pedestrian-friendly Fulton Mall in Brooklyn and to stave off disinvestment and abandonment along Jamaica Avenue and 165th Street in Queens. Following requisite New York State enabling legislation (circa 1980) and subsequent to NYC Charter modifications, these three special assessment districts evolved into the city’s first BIDs and adopted an explicit focus on physical and social dimensions of commercial environments. By the end of the 1990s, thirty-seven additional BIDs of varied sizes, functions, capacity, and operations had been established throughout the city.

In an earlier study of NYC BIDs, Gross (2005) develops a useful typology³ that differentiates BIDs by physical form, property attributes, revenue, and services. One end of the spectrum is characterized by large “Corporate” BIDs. In NYC, Corporate BIDs—such as Grand Central Partnership, 34th Street BID, Metro Tech BID, Time Square BID, and the Alliance for Downtown New York—have large physical footprints, are located in economically stable environments with high property values, a dense retail base, and regular pedestrian foot traffic, such as downtown Manhattan, or surrounding popular tourist attractions and transit hubs. In fact, NYC’s seven Corporate BIDs span 1,127 block faces of Class-A real estate and, on average, generate more than $10 million in annual revenue from the property tax assessment. The large operating budgets of corporate BIDs covers an array of district services, from capital improvement projects to district enhancement and development, intended to attract new commercial investment and patrons.

Small “Community” BIDs are on the other end of the BID spectrum. In 2008, Community BIDs in NYC outnumbered Corporate BIDs 6:1, but covered only slightly more blocks, approximately 1,223, and operated with an average budget of just $300,000. Unlike Corporate BIDs, Community BIDs are typically located in weaker economic markets with higher rates of retail vacancy, less foot traffic, and more establishments that cater to local demand. Because of resource constraints, Community BIDs typically focus on district maintenance, upkeep, and retention of existing businesses rather than capital improvements or even security (Gross 2005).

In contrast, medium-sized “Destination” BIDs represent the third BID type. They are generally located along major commercial thoroughfares across the city—such as 14th Street, 23rd Street, 125th Street in Manhattan, Jamaica Avenue in Queens, Fordham Road in the Bronx, among others—that are easily accessible by mass transit and attract many national retailers and millions of consumers from outside the immediate area. The average Destination BID budget is approximately five times smaller than Corporate BIDs and six times larger than Community BIDs. Destination BIDs typically generate sufficient revenue to provide “safe and clean” services in addition to district marketing and support services for existing businesses.

The NYC Department of Small Business Services (NYCSBS), the agency that oversees the BID program, boasts that BIDs, undifferentiated by size, “increase property values, improve sales, and decrease commercial vacancy rates” (2009, 5). Yet studies of BIDs in NYC, Philadelphia, and elsewhere find that BID functions and capacities are highly variable and tend to be correlated with BID size. Hence, commercial vibrancy observable in Corporate BIDs may not be obtainable by much smaller Community BIDs (Gross 2005; Ellen, Schwartz, and Voicu 2007; see the symposium issue of the Drexel Law Review [2010] for an in-depth examination of BIDs in Philadelphia).

The stock of BIDs in NYC has steadily increased since the 1970s. However, between 2002 and 2008, the number of BIDs grew by an unprecedented 45 percent, disproportionately driven by new Community BIDs established in the boroughs of Brooklyn, Queens, and the Bronx. NYC’s proliferation of BIDs coincides with Mayor Michael Bloomberg’s tenure that began in January 2002. In Bloomberg’s State of the City address, he explicitly called for improving “the effectiveness of New York City’s BIDs and strengthening their relationship with the City” (NYCSBS 2009, 3). To that end, in 2002 Bloomberg appointed Robert Walsh, a former BID director, as the commissioner of NYCSBS and, according to news reports, encouraged Walsh to “grow the [BID] program, energize it, and get out of the way as much as you can and let these organizations develop” (Lefkowitz 2010). Consequently, twenty-three new BIDs were established by the end of Bloomberg’s second term in office—the highest number created under any mayor.⁴ To control for political environmental factors in NYC, this study estimates change in retail performance within BIDs established during the first two terms of the Bloomberg administration relative to retail performance we might expect given comparable neighborhood and retail conditions in 2000, prior to the Bloomberg administration and pre-BID formation. The time frame of this study (2000–2008) conveniently excludes Corporate BIDs since the city’s large-scale BIDs were all adopted prior to 2000. Before describing the research design, the next section compares rates of retail retention and attraction within BIDs to the rest of NYC. Figure 1 shows all BIDs, by size, established in NYC between 1976 and 2008. Figure 2 only shows BIDs established during the study period.

Figure 1.

All New York City BIDs, established 1976–2008.

Figure 2.

New York City BIDs in the study, established 2002–2008.

Retail Attraction and Retention

With resurgence in downtown shopping coupled with polycentric urban retailing, the number of retailers in NYC increased 53.7 percent between 2000 and 2008. Retail attraction, which includes both new business development and business relocation from outside of NYC, is a widely used indicator of commercial district vibrancy and BID effectiveness. That said, local economic development scholars routinely critique disproportionate attention to firm attraction while underemphasizing other critical factors such as business retention, growth, and retail mix or composition (Bradshaw and Blakely 1999; Reese and Rosenfeld 2004). The Center for an Urban Future (2010), an economic development think tank in NYC, tracks annual change in the number of national and regional chain stores in NYC since 2008. The Center documents dispersion of chain stores to the outer boroughs, and the preponderance of particular establishments, like Starbucks, Dunkin Donuts, McDonalds, and Subway. Overall, NYC’s retail landscape is typified by a dense tapestry of independently owned (and relatively small) retailers (e.g., bistros, bars, bookstores, bodegas, and other retail and service establishments). In fact, the number of independent retailers increased 60.3 percent (2000–2008) compared to a decline of 3 percentage points for chain stores citywide. Growth in independent retailers is noteworthy because they are more likely to oppose BID formation, commonly critiqued as double taxation or an unfair economic burden resulting from the tax assessment that property owners typically pass directly to retail tenants (Fung 2011a, 2011b; McLaughlin 2008; Stokes 2006).

Summary statistics in Table 1 compare BIDs established between 2000 and 2008 to non-BID areas of NYC, based on four dimensions: total number of retail establishments, rate of retail growth or attraction, retail retention, and rate of retention for independent retailers. Overall, the number of retailers increased markedly in NYC. On average, the rate of growth was significantly higher in areas that never adopted BIDs compared to those with BIDs: 54.5 percent versus 41.5 percent, respectively. This may be a function of higher rates of commercial vacancy and turnover or more physical space to accommodate retail churn in non-BID areas. This explanation seems to reconcile with the rate of retail retention (column 4), which, on average, is higher for BIDs. In other words, for retailers operating within BIDs in 2000, approximately 43.6 percent remained in place through 2008 compared to 41 percent for non-BID retailers. The 2.6 percentage point difference is small but statistically significant. Interestingly, when chain stores are excluded, differences in the rate of retention between BID and non-BID retailers essentially disappears (column 5), suggesting that BIDs provide marginal benefit for retaining independent retailers. One possible explanation is that commercial areas without BIDs have other organizational resources (e.g., community development corporations [CDCs] and merchant associations) that assist small businesses and retail corridors, muting the effect of BIDs. The next section describes the research design and data used to systematically measure the effect of BIDs on retail performance relative to statistically similar non-BID areas of the city, rather than citywide. This study emphasizes effects for independent retail since, in 2008, they constituted approximately 91 percent of the citywide retail landscape and 84 percent of retail composition in BIDs.

Table 1.

Retail Attraction and Retention.

BID	Borough	Year BID established	(1)Retail Count 2000	(2)Retail Count 2008	(3)Rate of Attraction 2000–2008	(4)Rate of Retention 2000–2008	(5)Retention for Independent Retailers
Columbus Avenue BID	Manhattan	2000	188	209	0.1117	0.5243	0.5000
86th Street–Bay Ridge BID	Brooklyn	2001	120	140	0.1667	0.5429	0.4907
Kingsbridge Road BID	Bronx	2001	127	172	0.3543	0.5233	0.5088
East Mid-Manhattan BID	Manhattan	2002	965	1,214	0.2580	0.5592	0.5319
Flushing BID	Queens	2003	296	536	0.8108	0.4692	0.4516
Fordham Rd BID	Bronx	2004	173	248	0.4335	0.5742	0.5543
Sutphin Boulevard BID	Queens	2004	46	63	0.3696	0.4742	0.4530
161st Street BID	Bronx	2005	78	103	0.3205	0.2194	0.2119
DUMBO BID	Brooklyn	2005	58	155	1.6724	0.5049	0.4586
Forest Ave BID	Staten Island	2005	86	105	0.2209	0.4600	0.4420
Myrtle Ave Brooklyn BID	Brooklyn	2005	82	147	0.7927	0.3475	0.3140
Queens Plaza Court Square	Queens	2005	69	79	0.1449	0.3190	0.2974
Bay Ridge–5th Avenue BID	Brooklyn	2006	240	300	0.2500	0.4395	0.4064
Flatbush–Nostrand Junction BID	Brooklyn	2006	113	150	0.3274	0.5048	0.5000
Flatiron/23rd Street BID	Manhattan	2006	676	1,131	0.6731	0.3985	0.3939
Bayside Village BID	Queens	2007	147	152	0.0340	0.4477	0.4304
Court–Livingston–Schermerhorn BID	Brooklyn	2007	153	194	0.2680	0.3605	0.3521
Southern Blvd BID	Bronx	2007	128	166	0.2969	0.4158	0.4175
Sunnyside BID	Queens	2007	180	249	0.3833	0.5063	0.4714
Belmont BID	Bronx	2008	167	260	0.5569	0.4578	0.4444
Fulton Street BID	Brooklyn	2008	87	133	0.5287	0.4418	0.4304
Park Slope–5th Avenue BID	Brooklyn	2008	279	404	0.4480	0.4127	0.4032
Total for BIDs			4,458	6,310	0.4154	0.4360	0.4092
Total for non-BID areas			67,006	103,534	0.5451	0.4108	0.3953

Source: BID names and boundaries come from the NYC Department of Small Business Services (NYCSBS) BID profile and the Department of City Planning (NYCDCP) shapefiles, as of June 2010. BID = business improvement district.

Note: Bold type denotes Destination BIDs.

Methods and Data

Observed physical and social improvements within BIDs are commonly used to signify the model’s effectiveness. Since BID financing structure disproportionately burdens retail tenants, it is reasonable to expect that they derive economic benefits following BID formation. Particularly, retailers located within BID boundaries prior to BID formation as they bear the direct burden of rent inflation as a result of the compulsory tax increment. This study estimates the effect of BIDs on the economic performance of neighborhood retailers using the difference before and after BID formation for annual sales and employment, conventional economic indicators of firm performance.

It is plausible that factors unrelated to BIDs, such as buoyancy in the regional economy, seasonal shifts, and sectoral trends, affect firm performance (Bartik and Bingham 1997; Bartik 2007). Estimating the direct impact of BIDs requires making reliable causal inferences. Randomized controlled experimental designs are the most rigorous method for doing so while reducing selection bias. However, for this study, randomization is infeasible since BIDs are not randomly assigned. They are adopted voluntarily by district stakeholders, such as community-based organizations, property owners, and others who demarcate potential BID boundaries and then petition municipal government for ratification. Assuming the stakeholder group satisfies the requirements specified in the BID-enabling legislation, local state agents sanction BID formation. In NYC, once BIDs are formally established, they typically remain intact indefinitely.

This study uses semiparametric and nonparametric matching techniques as an analog to randomization to address the selection bias inherent in BID adoption and to make valid causal inferences that yield unbiased estimates of BID impact. According to Dehejia and Wahba (1999, 2002), matching techniques reduce selection bias by constructing counterfactuals based on control groups derived from baseline information observed prior to “treatment” or in this case BID adoption. Matching also controls for observed differences between the treatment group “BIDs” and the control group “non-BID” census tracts that theory suggests may predispose neighborhoods to BID adoption. Matching BIDs to non-BID census tracts based on observed pre-BID attributes known to affect BID adoption (e.g., retail density, assessed property value, population) produces a credible control group of non-BID census tracts that have a high probability of BID adoption but have never done so. A combination of retail firm, neighborhood, and building attributes observed in 2000 constitute the baseline information used to make inferences about retail sales and employment we might expect, come 2008, had the BID model not been adopted.

Neighborhood-Level Analysis

Consistent with previous studies of the impact of place-based policies such as BIDs (Brooks 2008), empowerment zones (Oakley and Tsao 2006), and enterprise zones (Elvery 2008; O’Keefe 2004), treatment and control groups are constructed using census tract–level propensity score estimators to calculate the predicted probability that a BID will be established in any NYC neighborhood or census tract. Propensity score matching helps reduce selection bias and constructs valid control groups for making reliable estimates of BID effects. The primary assumption when using propensity score matching is that treatment assignment only depends on observable pretreatment variables (Dehejia and Wahba 1999; Rosenbaum and Rubin 1983; Sianesi 2004). Accordingly, propensity scores are derived based on observed neighborhood-level attributes in 2000 (prior to BID formation), whereas estimation models only include BIDs adopted between 2002 and 2008. This two-year time lag helps ensure that baseline observations are not influenced by the process of BID formation. In other words, treatment and control neighborhoods are observed to be similar prior to embarking on BID formation. It also situates estimates of retail performance within the same political administration.⁵ Hence, reasonable inferences can be made about the effect of BIDs on retailers.

According to previous studies, BID adoption is typically determined by a set of neighborhood characteristics related to civic infrastructure, organizational capacity, community participation, rates of home ownership, real estate value, and commercial density (Caruso and Weber 2006; Ellen, Schwartz, and Voicu 2007). This study models the predicted probability that any NYC neighborhood establishes a BID as a function of neighborhood-level covariates that include the local commercial environment (e.g., retail density, business age, and business relocation); civic characteristics (e.g., education, income, race, foreign-born, and home ownership); and real property traits (e.g., size and assessed property value). Dummy variables for each borough are also included. Propensity scores are calculated using logistic regression. The logit model takes the following form:

\Pr (y = 1) = \frac{e^{β ’ x}}{1 + e^{β ’ x}}

where y = 1 if tracts are fully or partially within BIDs and is equal to zero otherwise, and x is a matrix of neighborhood-level covariates from 2000.

The matching process identifies non-BID census tracts most similar to each of the 104 census tracts that correspond with BIDs referred to here as “BID tracts.” Only retailers firmly located within BID boundaries are retained for BID tracts. Proximate retailers that share BID census tracts but are located outside of BID boundaries are excluded from the analysis. Each BID tract is matched with two “nearest neighbor” non-BID tracts with the closest propensity scores to create a match trio, or cluster of census tracts similar to BID tracts based on retail density, property value, and demographic attributes. Nearest neighbor tracts, presumably, are likely to establish a BID but have filed or completed a BID application at the time of this study. While it is impossible to find neighborhoods with identical propensity scores, the nearest neighbor non-BID matches in this study fall within the conventional range .05 for treatment and control groups (O’Keefe 2004).

It is important to note that BID proliferation in NYC during the Bloomberg administration, coupled with the pool of preexisting BIDs established 1976–1999, limits the number of non-BID tracts available for one-to-one matches. As a result, I allow a single non-BID tract to serve as the best match for more than one BID tract. According to Dehejia and Wahba (2002) the process of matching with replacement is preferable to using a larger array of less well matched tracts because matching with replacement improves the estimate of the average treatment effect without raising the variance. Approximately 36 percent of non-BID tracts are matched more than once. It is also important to note that retail density within the three medium-size Destination BIDs in this study presents a methodological challenge of creating reliable matched trios. For Destination BIDs, producing comparable non-BID tracts requires matching with replacement. For both Community and Destination BIDs, statistical balance of covariates is achieved. However, matches for Destination BIDs are generally weaker than for Community BIDs. The map in Figure 3 shows matched trios that include Community and Destination BIDs and matched tracts that serve as controls.

Figure 3.

Propensity score matched BID and non-BID tracts.

Since factors affecting BID adoption are controlled for in the matching process, I use fixed effects regression for each matched trio to estimate BID impact on retail performance, controlling for unobserved time-invariant factors like transit access, zoning and land-use ordinances, and regulatory issues potentially correlated with sales or employment growth (Oakley and Tsao 2006; O’Keefe 2004). Fixed effects models also control for average differences across matched trios in order to estimate net within-match effects of the independent variables on percentage change of retail sales and employment. Time-variant confounding factors such as the total number of retailers and the proportion of independently owned business are controlled for in the models. The regression models take the form:

Δ Y_{i t} = α_{j} + β_{1} \cdot B I D_{a g e} + γ_{i} \cdot Z_{i} + ε_{i}

where $i$ indexes each census tract and $j$ indexes a matched trio composed of a BID tract and two “nearest neighbor” non-BID tracts based on the propensity score matching process highlighted previously and $α_{i}$ (i = 1, . . . , n) is the unknown intercept estimate for each matched trio. The coefficient $β_{1}$ approximates the effect of BID designation on change in retail sales or employment, also considered an economic growth model.

The fundamental premise is that BIDs should have a positive direct impact on retail sales $β_{1}$ (with secondary effects observable through employment growth) relative to comparable non-BID areas. Since BIDs are adopted at different times and BID maturity may affect economic performance, the variable BIDage_it^1–7 is used in lieu of a binary $B I D$ indicator variable in order to determine how BID effectiveness may change based on the maturation of the BID. BIDage_i^1–7 indicates 1–7 years since BID adoption, or 0 for non-BID matched tracts. $γ_{i}$ is a vector of 104 fixed effects for each match tract trio; Z represents a vector of time-variant controls that might affect growth, such as the total number of retailers by sector in each neighborhood in 2008; and $ε_{i}$ represents the model error term. To examine the effect of independent retailers in BIDs, I add the interaction BIDage × Independent to the model. Separate fixed effects models are conducted for Community BIDs and Destination BIDs.

Data

Data are drawn from a panel data set that I developed using the NETS data set, a proprietary firm-level data set derived from Dunn and Bradstreet archival records.⁶ NETS includes address-level data for all neighborhood retailers in NYC (2000–2008). The NETS data set includes attributes such as annual establishment sales and number of employees,⁷ business name and address, six-digit North American Industry Classification System code, start year and close year, and establishment structure (e.g., independently owned or chain store). I use ArcGIS and NYC Geo-support software to geocode neighborhood retailers to their corresponding building addresses⁸ for 2000 and 2008—83,396 and 127,960, respectively.⁹

The NETS data are with three additional data sets: the Neighborhood Change Database from Geolytics provides a balanced panel of 2,216 NYC tract-level census data for 1990 and 2000; the Real Property Assessed Data obtained from the NYC Department of Finance provides attributes (e.g., square feet, retail space, and assessed property value) of physical property in 2000 and 2008; and a database of BID traits (e.g., year established, revenue, expenses, number of city blocks, geographic boundaries) that I assembled using the NYCSBS annual report on BIDs for fiscal year 2009, coupled with publicly available BID shapefiles from the Department of Planning and made available through NYC Open Data.¹⁰

Key Variables

There are two primary outcome variables: rates of change in retail sales and employment. Not only are the outcome variables conventional measures of firm performance but sales tax revenue and employment are common indicators of local economic development. Both retail sales and employment represent the rate of change before and after BID formation for all retailers operational throughout the period. The primary predictor variable for the fixed-effect models is BID age in 2008, determined by subtracting the year the BID was adopted from the last year of the study.

Myriad factors endogenous and exogenous to neighborhoods affect firm performance, from business management and marketing decisions to industry trends, competition, and consumer preferences. Therefore, it is typically difficult to attribute economic growth to beautification programs, streetscape projects, public safety, sanitation services, and other BID resources and services. However, according to retail theory, the attractiveness of shopping districts is a critical determinant of consumer capture (Brown 1993; Teller and Reutterer 2008). Given the political and scholarly rhetoric on the benefit of BIDs for beautifying the environment, it is reasonable to expect BIDs to positively influence consumer shopping preferences. The degree to which BIDs affect consumer spending is most directly measured through sales growth, all else equal. For most firms, sales and employment growth are positively correlated. Both outcomes are used here because one might observe positive change in sales after BID adoption, but positive change in employment is a strong indication of sustained sales growth. Although employment growth is a secondary objective of BIDs (Caruso and Weber 2006), it is an important measure of local economic development. Therefore, estimating the impact of BIDs on both sales and employment allows one to test the robustness of findings and more fully examine the potential local economic development effects of the BID model.

Summary of Findings

Propensity Score Matching and Postmatching Balancing Diagnostic

As previously noted, I use propensity score matching to create a counterfactual for estimating the impact of BIDs compared to neighborhoods that never adopted BIDs but are most likely to do so. Characteristics of actual BIDs prior to formation are used to determine the likelihood that any neighborhood will adopt the model. Logistic regression results presented in Table 2 corroborate theories about BID formation. In other words, BIDs are more likely to be adopted in neighborhoods with higher retail density, assessed property values, dense population, and college-educated residents. Interestingly, BIDs are twice as likely to be adopted in neighborhoods with a greater proportion of Hispanic residents, but three times less likely in neighborhoods with a greater proportion of White residents. Other factors that seem to significantly reduce the likelihood of BID adoption in NYC include the number of times a business relocated by 2000 and a higher proportion of foreign-born residents. Factors that presumably inhibit BID formation arguably reflect corridor instability and reticence toward civic engagement with governmental institutions, on the one hand, and possibly predominately white neighborhoods that have less need for corridor revitalization tools, on the other.

Table 2.

Logistic Regression Measuring the Likelihood of BID Adoption (N = 1,286).

Business Improvement District	Coefficient	Standard Error	P > z
Retail density (per acre)	2.442531	0.288700	0.0000^***
Age of business	–0.024563	0.109458	0.8220
Relocation (1990–1999)	–0.147412	0.049125	0.0030^**
Average income (ln)	–1.295185	0.718822	0.0720
Change in RMI (1990–2000)^a	0.570377	0.622470	0.3600
Population (ln)	0.776901	0.361129	0.0310^*
White (%)	–3.181326	0.738518	0.0000^***
Hispanic (%)	2.021481	0.930127	0.0300^*
Foreign born (%)	–2.758014	1.149312	0.0160^**
College educated (%)	9.659029	1.865833	0.0000^***
Number of rental units	–0.000234	0.000267	0.3810
Assessed property value	0.526703	0.104075	0.0000^***
Manhattan	–9.628490	1.597416	0.0000^***
Brooklyn	–0.764662	0.648820	0.2390
Queens	–1.531403	0.721942	0.0340^*
Bronx	–2.470166	0.743594	0.0010^***
_cons	–1.465040	8.109992	0.8570

Relative mean income (RMI) is the average income for a census tract relative to the average income for the borough. BID = business improvement district.

p < 0.05, ^**p < 0.01, ^***p < 0.001.

A simple postmatching balancing diagnostic was conducted to determine how well propensity scores create fair comparison groups of non-BID matched tracts.¹¹ Balance was achieved between BIDs and non-BID matched tracts for all covariates. Therefore, it is reasonable to assume that propensity scores adequately reduce bias and create comparable comparison groups for inferential analysis of BID impact (Dehejia and Wahba 2002; Friedline, Elliott, and Nam 2011). It is worth noting that covariate balance is generally stronger for Community BIDs than Destination BIDs because of the dearth of possible non-BID tracts in NYC with comparable retail density to Destination BIDs. That said, caution is warranted for inferences drawn from analysis of Destination BIDs. Summary statistics in Table 3 show comparable covariate means between BIDs overall and non-BID matched tracts for the pre-BID baseline year. Negligible differences are observed for important predictors such as population, retail density per acre, proportion of independent retailers, the age of businesses, and assessed land and building value. Higher observed differences in retail count per acre between BID and non-BID areas can be explained by the large number of retailers accounted for by the three Destination BIDs.

Table 3.

Summary Statistics for BID Tracts, Matched Tracts, and NYC in 2000.

	NYC BIDs	Non-BID Match Tracts	New York City
Number of tracts	104	98	1,286
Average sales ($)	823,828.50	730,301.10	650,447.90
Number of employees	8.460	7.565	6.201
Number of relocations (1990s)	3.697	2.803	1.761
Business age (years)	7.593	7.719	7.604
Independently owned retail (%)	0.843	0.877	0.913
Chain store (%)	0.157	0.123	0.087
Retail density (per acre)	2.186	1.907	0.601
Retail count (per acre)	80.40	66.15	30.65
Average family income ($)	83,710.84	75,798.72	63,363.90
Population	4,046	4,013	3,638
White (%)	0.386	0.398	0.397
Black (%)	0.206	0.154	0.269
Hispanic (%)	0.293	0.341	0.223
Asian (%)	0.110	0.099	0.104
Foreign born (%)	0.336	0.333	0.351
College educated residents ≥25 years old (%)	0.344	0.338	0.233
Total housing units	1656	1650	1345
Rental units (%)	0.771	0.765	0.598
Assessed value land + building ($)	2,953,116.00	3,209,332.00	1,871,610.00
Manhattan	0.1616	0.1111
Brooklyn	0.3838	0.4091
Queens	0.1919	0.2727
Bronx	0.2222	0.1667

Note: Statistics exclude all NYC census tracts associated with BIDs that were established between 1976 and 1999, prior to the study period. BID = business improvement district; NYC = New York City.

BID-Level Effects

Are BIDs Good for Business?

This primary research question is addressed with two fixed effects regression models for each performance measure. As previously noted, this method controls for long-standing and unobservable neighborhood factors possibly correlated with sales or employment, such as vagrancies of privately negotiated commercial leases, small business regulations and code enforcement, and transit access. The models presented in Table 4 control for time-variant factors, including change in the overall number of retailers and the proportion of independently owned establishments before and after BID adoption.

Table 4.

Baseline Effects for the Rate of Change in Retail Sales and Employment.

	(1)	(2)	(3)	(4)
	Rate of Change in Sales	Rate of Change in Employment	Rate of Change in Sales	Rate of Change in Employment
BIDs (years)	0.0175	0.0283	0.197^†	0.186
	(0.0199)	(0.0216)	(0.107)	(0.117)
Independent retail (%)	–0.795^**	–0.633^*	–0.555^†	–0.422
	(0.247)	(0.268)	(0.284)	(0.309)
Retail count (ln)	0.155^***	0.156^**	0.151^***	0.153^**
	(0.0438)	(0.0475)	(0.0438)	(0.0475)
BIDs × Independent			–0.220^†	–0.194
			(0.129)	(0.141)
Constant	0.0433	–0.0219	–0.149	–0.191
	(0.272)	(0.295)	(0.294)	(0.319)
Observations	624	624	624	624
R ²	0.3598	0.3743	0.3633	0.3766

Note: Standard errors are in parentheses.

***

p < 0.001, ^**p < 0.01, ^*p < 0.05, ^†p < 0.10.

The coefficients in the baseline models (columns 1 and 2) suggest that BIDs, irrespective of size, have no statistically significant effect on either measure of retail performance relative to similar areas of the city that never adopted BIDs. On its own, this finding is surprising given widespread beliefs and claims in the literature that BIDs help improve business profitability (Hoffman and Houstoun 2010), generate higher sales tax revenues, and serve as a tool for local economic development (Gross 2005; Mitchell 2008). The baseline models also show that as the proportion of independent retail increases across the city, relative to chain stores, both sales and employment decline significantly, by 79 and 63 percent, respectively. On average, sales volume and employment of independent retailers are disproportionately low relative to chain stores. So these significant declines in performance are plausible amid growth in the proportion of independent retailers.

Since NYC’s commercial landscape is disproportionately populated by independent retailers, models 3 and 4 include an interaction term to specifically examine the effects of independent retail in BIDs. The results show that as the proportion of independent retail businesses increases within BIDs, sales declines by 22 percent and employment declines by 19 percent relative to similar non-BID areas. Change in sales is only marginally significant. We would expect effects on sales to be larger than employment, since sales is a more immediate indicator of firm performance. However, observed declines in sales and employment seem to corroborate critics who claim that the BID model, particularly small BIDs, fail to provide sufficient added value (McLaughlin 2008). Alternatively, the magnitude of the decline on interaction coefficients is notably less than the decline for independent retail overall. That is to say, as the proportion of independent retail grows within BIDs, sales and employment decline significantly less than the decline observed among independent retail overall. This may suggest that BIDs in fact provide a buffer against economic weakening. This proposition is tested in subsequent models by disaggregating BIDs by size, and testing the effects of Community BIDs and Destination BIDs separately.

How Do Effects Differ by Retail Structure and BID Size?

Models 1 and 2 in Table 5 suggest that growth of independent retail in Community BIDs drives observed decline in sales and employment each year, approximately 64 and 57 percent respectively, relative to similar non-BID areas of NYC. In contrast to the models presented previously, the slope of decline in sales and employment within Community BIDs is comparable to the overall decline for independent retail. That is to say, by disaggregating effects by BID size, we see that the magnitude of the coefficients for independent retail in Community BIDs is comparable to independent retail generally, thereby refuting the “buffer” hypothesis, at least for Community BIDs. Rather than serving as a buffer against economic decline for independent retailers, it seems that Community BIDs are a hindrance to sales and employment growth, relative to comparable areas. This finding is particularly noteworthy amid the proliferation of Community BIDs in NYC and elsewhere, and NYC’s concentration of independently owned goods and service establishments. For instance, during the study period, 86 percent of BIDs adopted in the city can be characterized as Community BIDs, and Community BIDs represent 67 percent of NYC BIDs.

Table 5.

Effects by BID Type and Retail Structure.

	(1)	(2)	(3)	(4)
	Community BIDs Rate of Change in Sales	Community BIDs Rate of Change in Employment	Destination BIDs Rate of Change in Sales	Destination BIDs Rate of Change in Employment
BIDs (years)	0.630^***	0.585^**	–0.149^*	–0.0936^†
	(0.163)	(0.179)	(0.0638)	(0.0509)
Independent retail (%)	–0.236	–0.143	–1.176^***	–0.878^***
	(0.351)	(0.386)	(0.264)	(0.211)
Retail count (ln)	0.179^**	0.181^**	0.00517	0.0124
	(0.0541)	(0.0596)	(0.0368)	(0.0294)
BID × Independent	–0.643^***	–0.577^**	0.152^†	0.0942
	(0.186)	(0.205)	(0.0810)	(0.0646)
Constant	–0.521	–0.533	0.933^**	0.745^**
	(0.354)	(0.390)	(0.303)	(0.242)
Observations	450	450	174	174
R ²	0.3948	0.4014	0.4150	0.3873

Note: Standard errors are in parentheses.

***

p < 0.001, ^**p < 0.01, ^*p < 0.05, ^†p < 0.10.

Findings for Destination BIDs contrast findings for Community BIDs. That is to say, Destination BIDs seem to positively affect sales and employment of independent retailers, 15 and 9 percent growth, respectively (column 3 and 4). Nevertheless, sales growth is only modestly significant. This study includes just three Destination BIDs, but approximately 44 percent of all retailers in this study are located within Destination BIDs. According to descriptive statistics for Destination BIDs, both sales and employment decline between 2000 and 2008. However, findings from the fixed effects models show positive growth for independent retail in Destination BIDs. In other words, observed economic decline in Destination BIDs seems to result from growth in chain retail, not independent retail. All else equal, Destination BIDs may be better able to absorb and economically advance independent retailers than Community BIDs. That said, employment effects for Destination BIDs are not significant and sales effects are only marginally significant.

Disparate observed outcomes for Community and Destination BIDs are generally consistent with previous studies that find BID size moderates effects on property value (Ellen, Schwartz, and Voicu 2007) and organizational effectiveness (Gross 2005). To check the robustness of the findings, regression models without fixed effects are conducted. Time variant and invariant predictor variables, including commercial square feet, retail density by sector, proportion independent retail, and dummy variables for each borough, are controlled for in the models. Findings for both Community and Destination BIDs remain consistent with strong and significant sales and employment decline in the former and relatively modest economic growth in the latter.¹²

Discussing the Effectiveness of Bids for Business Performance

The BID model is a mechanism of urban governance designed to enhance and sustain targeted commercial environments. The self-financing structure has made the model widely popular among municipal leaders, practitioners, and scholars, but the direct effect of BIDs for retailers—critical stakeholders in the commercial environments—had not been rigorously examined. This study begins to address the lacunae in the literature by testing how small and medium-size BIDs, established in NYC between 2002 and 2008, impact retail performance measured as change in sales and employment. Results from the balanced-panel fixed-effect regression models demonstrate that, overall, BIDs seem to have no significant impact on either sales or employment relative to statistically comparable non-BID areas of the city. This preliminary finding is somewhat surprising given the wide range of purported physical, social, and economic benefits associated with BIDs. However, when the effects of BIDs are disaggregated by size and retail structure, a more nuanced understanding of effectiveness is revealed.

This discussion begins with findings for Community BIDs because they are the fastest-growing type of BID in NYC. Two previous studies of BIDs in NYC and Philadelphia, Gross (2005) and Stokes (2006), characterize Community BIDs as being relatively small in physical size, having limited resources and a narrow scope of services, and being geographically located in relatively weak economic environments, thus contending that more critical attention be given to how BIDs shape local development processes. What is more, limited resources and a narrow scope of services and physical footprint in relatively weak economic environments arguably make NYC’s Community BIDs more comparable to BIDs in commercial markets of small and midsize cities. One important distinction is that in NYC, Community BIDs likely have a higher proportion of independent retailers. During the study period, independent retailers in Community BIDs increased from 87 to 93 percent. Growth in independent retail is concomitant with significant decline in sales and employment.

There are at least three explanations for the significant decline in sales and employment for independent retailers in Community BIDs: First, the number of independent retail in Community BIDs grew overall, whereas the number of chain stores declined by approximately 5 percent. The chain stores that closed or relocated previously may have had positive spillover effects for proximate independent retailers. Hence, closure of chains may in fact hurt some small businesses, contrary to much of the anti-chain critiques. However, there is no reason to believe that chain closure within Community BIDs occurred at a greater magnitude than what we might expect from comparable non-BID areas. According to a report from the Center for an Urban Future, the economic recession, which officially began in December 2007 and lasted until the first quarter of 2009, caused chain stores across the city—such as Circuit City, Burritoville, 1-800-Mattress, and KB Toys—to shutter between 2008 and 2009, while others reduced their corporate footprint. However, the report also states that dozens of chain stores expanded the number of branches across the city. It is reasonable to assume that chain store restructuring occurred fairly similarly in comparable neighborhoods across the city. According to the descriptive statistics, there is actually reason to believe that areas with BIDs were able to retain chain stores more than non-BID areas. However, it is not clear if retention is similar for Community and Destination BIDs.

A second possible explanation for observed sales and employment decline is that the type of retailers attracted to Community BIDs may slowly displace existing establishments. The retailers in this study have operated at least nine years (2000–2008) and, thus, arguably represent the city’s most stable enterprises. Nevertheless, new restaurants, cafés, chain stores, and other retailers attracted to Community BIDs may supplant the existing establishments through direct competition or modernization. The third explanation relates to the geographic dispersion of Community BIDs. In just six years (2002–2008), the stock of Community BIDs in NYC increased 64 percent, with 84 percent of the growth coming from new BIDs in Brooklyn, Queens, and the Bronx. Some new BIDs are proximate to preexisting BIDs in commercially dense areas of downtown Brooklyn and Queens, but most, however, are physically disconnected from other BIDs and situated along burgeoning commercial corridors of economically weaker neighborhoods, thus lacking the requisite organizational capacity and resources to provide a sufficient bundle of goods and services to provide substantive support for existing businesses. Small-scale Community BIDs in outlying neighborhoods have limited leverage for expanding consumption, thereby bolstering retail performance. If the magnitude of the effect of Community BID was close to zero, the model may arguably help maintain existing retailers. Decline in sales and employment after BIDs are established suggests retail erosion or modernization. Meaning, new retailers may aptly reflect a shifting neighborhood identity, thus attracting more patrons. However, effects for existing retailers are contingent on the structure of their commercial lease.

Findings for Destination BIDs contrast findings for Community BIDs. Most notably, as the proportion of independent retail increases in Destination BIDs from 76 to 87 percent, sales and employment also improve, although the magnitude of the effects are small and only modestly significant for sales. This finding corroborates earlier retail studies that contend that retail density, or urbanization economies, improves economic performance for proximate retailers. Since Destination BIDs are typically located along major commercial thoroughfares served by multiple modes of transit and may attract tens of thousands of potential consumers daily, commercial tenants typically pay significantly higher rent than Community BID tenants. Nevertheless, independent retailers in Destination BIDs seem to benefit from the volume of foot traffic and possible spillovers from chain stores that generally anchor busy thoroughfares. Although it is not evident in this analysis, the retail mix in Destination BIDs, even among independent retailers, is markedly different from Community BIDs. For instance, the large ground-floor retail spaces in Destination BIDs accommodate retailers of durable and semidurable goods—such as home furnishing, furniture stores, appliance and electronics stores, and apparel outlets—that may be part of a national or regional chain or independently owned. Whereas Community BIDs accommodate more local service retailers, such as beauty and barber shops, convenience stores, dry cleaners, grocery stores, cafés, and small clothing stores.

What Does This Analysis Suggest for Policymakers, Practitioners, and Scholars?

Although it may be tempting to interpret findings for Community BID as a call to focus on attracting and retaining chain stores to improve overall sales and employment, I contend that such an interpretation is shortsighted and will ultimately thwart sustainable local economic development. First, BID managers have negligible leverage with local branches of chain stores beholden to corporate strategy. Corporate decision makers respond to national or regional strategy and capricious markets rather than local needs. Second, there is often a mismatch between vacant commercial space within BID boundaries, which typically remain fixed over time, and needs of chain stores. Focusing on attracting chain stores would likely require storefront assemblage that would displace and otherwise adversely affect small independent retailers through direct competition and additional rent inflation. Further, a preponderance of chain stores in a locality is said to erode neighborhood character. This is not to unequivocally exclude chains from Community BIDs but to highlight the potential for diminishing returns of undifferentiated commercial districts. Stokes (2006) and others have questioned whether the BID model is equitable policy and suggested that uniform application of BIDs across neighborhood typologies deserves more critical attention. This study begins to substantiate this claim by showing disparate effects by BID. More alarmingly, this study identifies adverse effects of Community BIDs with a preponderance of independent retail. Future research could disaggregate effects by retail sector to better understand how different types of retailers perform in BIDs. Additionally, a firm-level analysis of retail performance could identify potential spillover effects to proximate non-BID areas, thereby addressing extant concerns regarding free-riding and the relocation of existing retailers seeking to evade the BID tax assessment.

Footnotes

Acknowledgements

I am grateful to a former master’s student in urban planning, Ben Huff, for his assistance with map development. I also appreciate the thoughtful comments of three anonymous reviewers.

Declaration of Conflicting Interests

The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author received no financial support for the research, authorship, and/or publication of this article.

Notes

Author Biography

Stacey A. Sutton is an assistant professor of Urban Planning in the Graduate School of Architecture, Planning and Preservation at Columbia University. Her research interests include local economic development, neighborhood change, small business dynamics, and the impact of city policy and planning.

References

Bartik

Timothy J.

Bingham

Richard D.

1997. “Can Economic Development Programs Be Evaluated?” In Dilemmas of Urban Economic Development: Issues in Theory and Practice, edited by Bingham

Richard D.

Mier

Robert

, 246–77. Newbury Park, CA: Sage.

Bartik

Timothy J.

2007. “Solving the Problems of Economic Development Incentives.” In Reining in the Competition for Capital, edited by Markusen

Ann

, 103–40. Kalamazoo, MI: W.E. Upjohn Institute for Employment Research.

Bradshaw

Ted

Blakely

Edward

. 1999. “What Are ‘Third Wave’ State Economic Development Efforts? From Incentives to Industrial Policy.” Economic Development Quarterly 13 (3): 229–44.

Briffault

Richard

. 2010. “The Business Improvement District Comes of Age.” Drexel Law Review 3 (1): 19–33.

Briffault

Richard

. 1999. “A Government for Our Time? Business Improvement Districts and Urban Governance.” Columbia Law Review 99(2): 365–477.

Brindisi

Francesco

Wilhelm

Lindsay M.

2008 “Employment Dynamics and Relocation Patterns in New York City 1993-2006.” Paper prepared for the 55th Annual Meeting of the North American Regional Science Association, Brooklyn, NY, November 2008.

Brooks

Leah

. 2008. “Volunteering to Be Taxed: Business Improvement Districts and the Extra-governmental Provision of Public Safety.” Journal of Public Economics 92:388–406.

Brown

Stephen

. 1993. “Retail Location Theory: Evolution and Evaluation.” International Review of Retail, Distribution and Consumer Research 3 (2): 185–229.

Caruso

Gina

Weber

Rachel

. 2006. “Getting the Max for the Tax: An Examination of BID Performance Measures.” International Journal of Public Administration 29:187–219.

10.

Center for an Urban Future. 2010. “A Chain Reaction: This Year’s Borough-by-Borough Analysis of New York City’s Largest Retailers.” (December 10), vol 1 issue 1.

11.

Dawson

J. A.

Kirby

D. A.

1980. “Urban Retail Provision and Consumer Behavior: Some Examples from Western Society.” Geography and the Urban Environment 3:87–132.

12.

Dehejia

Rajeev H.

Wahba

Sadek

. 1999. “Causal Effects in Non-experimental Studies: Reevaluating the Evaluation of Training Programs.” Journal of the American Statistical Association 94 (448): 1053–62.

13.

Dehejia

Rajeev H.

Wahba

Sadek

. 2002. “Propensity Score Matching Methods for Non-experimental Causal Studies.” Review of Economics and Statistics 84 (1): 151–61.

14.

Ellen

Ingrid Gould

Schwartz

Amy Ellen

Voicu

Ioan

. 2007. “The Impact of Business Improvement Districts on Property Values: Evidence from New York City.” Working Paper Series, The Furman Center for Real Estate and Urban Policy, New York University.

15.

Elvery

Joel A.

2009. “The Impact of Enterprise Zones on Resident Employment an Evaluation of the Enterprise Zone Programs of California and Florida.” Economic Development Quarterly 23 (1), 44–59.

16.

Fox

Edward

Postrel

Steven

McLaughlin

Amanda

. 2007. “The Impact of Retail Location on Retailer Revenues: An Empirical Investigation.” Mimeo, Southern Methodist University, Dallas, Texas.

17.

Friedline

Terri L.

Elliott

William

Nam

Ilsung

. 2011. “Predicting Savings from Adolescence to Young Adulthood: A Propensity Score Approach.” Journal of the Society for Social Work and Research 2 (1): 1–22.

18.

Fung

Amanda.

2011a. “Chinatown BID Moves a Step Closer to Takeoff.” Crain’s New York Business, March 2, 2011 (online).

19.

Fung

Amanda.

2011b. “Smaller BIDs Get Bigger Boost from City Hall.” Crain’s New York Business, June 21, 2012 (online).

20.

Frung

Gerald E.

2010. “The Seductions of Form.” Drexel Law Review, 3 (1): 11–17.

21.

Gopal-Agge

Devika

Hoyt

Lorlene

. 2008. “The Retail-Revitalization Nexus: The Business Improvement Model in Canada and the United States.” In Business Improvement Districts: Research, Theories and Controversies, edited by Morçöl

Goktug

Hoyt

Lorlene

Meek

Jack W.

Zimmermann

Ulf

. Oxford, UK: Taylor & Francis.

22.

Gross

Jill

. 2005. “Business Improvement Districts in New York City’s Low-Income and High-Income Neighborhoods.” Economic Development Quarterly 19 (2): 174–89.

23.

Hoffman

Daniel

Houstoun

Lawrence O.

Jr.

2010. “Business Improvement Districts as a Tool for Improving Philadelphia’s Economy.” Drexel Law Review 3 (1): 89–107.

24.

Hoyt

Lorlene

. 2005. “Do Business Improvement District Organizations Make a Difference? Crime in and around Commercial Areas in Philadelphia.” Journal of Planning Education and Research 25:185–99.

25.

Hoyt

Lorlene

Gopal-Agge

Devika

. 2007. “The Business Improvement District Model: A Balanced Review of Contemporary Debates.” Geography Compass 1/4 (2007): 946–58.

26.

Huff

D. L.

1962. “Determination of Intra-Urban Retail Trade Areas.” Real Estate Research Program, Graduate Schools of Business Administration, University of California, Los Angeles.

27.

Lefkowicz

Melanie

. 2010. “BIDS These Days: The Ups and Downs of Business Improvement Districts.” City Limits, October 26.

28.

Levy

Paul R.

2001. “Paying for Public Life.” Economic Development Quarterly 15 (2): 124–31.

29.

McLaughlin

Mike

. 2008. “BID-ing War on Fulton Street.” Brooklyn Paper 31 (48).

30.

Mitchell

Jerry

. 2001. “Business Improvement Districts and the ‘New’ Revitalization of Downtown.” Economic Development Quarterly, 15 (2): 115–23.

31.

Mitchell

Jerry

. 2008. Business Improvement Districts and the Shape of American Cities. Albany, NY: SUNY Series in Urban Public Policy.

32.

Morcol

Gogtug

Hoyt

Lorlene Meek

Jack

Zimmermann

Ulf

, eds. 2008. Business Improvement Districts: Research, Theories, and Controversies. Boca Raton, FL: Auerbach.

33.

Nevin

John R.

Michael J

Houston

. 1980. “Image as a Component of Attraction to Intraurban Shopping Areas.” Journal of Retailing, 56 (1), 77–93.

34.

New York City Department of Small Business Services. 2009. “Starting a Business Improvement District: A Step-by-Step Guide.” Agency document.

35.

Oakley

Deirdre

Tsao

Hui-Shien

. 2006. “A New Way of Revitalizing Distressed Urban Communities? Assessing the Impact of the Federal Empowerment Zone Program.” Journal of Urban Affairs 28 (5): 443–71.

36.

O’Keefe

Suzanne

. 2004. “Job Creation in California’s Enterprise Zones: A Comparison Using a Propensity Score Matching Model.” Journal of Urban Economics 55:131–50.

37.

Reese

Laura

Rosenfeld

Raymond

. 2004. “Local Economic Development in the United States and Canada: Institutionalizing Policy Approaches.” American Review of Public Administration September 34 (3): 277–92.

38.

Rosenbaum

Paul R.

Rubin

Donald B.

1983. “The Central Role of the Propensity Score in Observational Studies for Causal Effects.” Biometrika 70:41–55.

39.

Sianesi

Barbara

. 2004. “An Evaluation of the Swedish System of Active Labor Market Programs in the 1990s.” Review of Economics and Statistics 86:133–55.

40.

Stokes

Robert

. 2006. “Business Improvement Districts and Inner City Revitalization: The Case of Philadelphia’s Frankford Special Services District.” International Journal of Public Administration 29:173–86.

41.

Stokes

Robert

. 2010. “The Challenges of Using BIDs in Lower-Income Areas: The Case of Germantown, Philadelphia.” Drexel Law Review 3 (1): 325–37.

42.

Teller

Christoph

Reutterer

Thomas

. 2008. “The Evolving Concept of Retail Attractiveness: What Makes Retail Agglomerations Attractive When Customers Shop at Them?” Journal of Retailing and Consumer Services 15:127–43.