Cashing in on Distress: The Expansion of Fringe Financial Institutions During the Great Recession

Abstract

The Great Recession was a consequence of widening inequality and the growth of a tiered financial services system, in which the rich and the poor have access to vastly different tools for wealth accumulation. The spatial organization of these dynamics created neighborhoods vulnerable to predation on behalf of subprime lenders and other fringe service providers. This project seeks to understand the reproduction of institutional marginalization in consumer finance. Results show that racially isolated neighborhoods in New York City, where subprime lending and foreclosures were common, were uniquely vulnerable during the Great Recession and were communities where check cashing outlets (CCOs) sprouted, highlighting a mechanism for the reproduction of inequality over time. CCOs cost more per transaction than a checking account—potentially totaling tens of thousands of dollars over a career. The link between widening financial services inequality and the recession’s consequences provides a strong impetus for safety net and community investment policies.

Keywords

financial services segregation Great Recession

Introduction

This article investigates the growth of check cashing across New York City during the Great Recession. The rise and fall of the housing market resulted in dramatic economic fallout. In many ways, the collapse of the housing market was a long-term consequence of two concurrent and reciprocal trends: widening inequality and the growth of a tiered financial services system, in which the rich and the poor have access to vastly different tools for wealth accumulation (Baradaran 2015; Caskey 1994; Kubrin et al. 2011; Smith, Smith, and Wackes 2008; Squires and O’Connor 1998). While the affluent are able to build home equity and retirement accounts via access to “mainstream” financial services, the poor are disproportionately reliant on “alternative” or “fringe” services such as check cashing outlets (CCOs), payday lending, and subprime mortgages (Faber 2013; Gramlich 2007; Williams, Nesiba, and McConnell 2005). There is an important spatial component to these two trends. Segregation by income and race makes disadvantaged neighborhoods more vulnerable to the effects of recessions (Massey and Denton 1993) and creates easily identifiable geographic markets for alternative financial service providers such as subprime lenders (Hwang, Hankinson, and Brown 2015; Rugh and Massey 2010). This all occurs in the context of limited regulation and the absence of public financial institutions, which could respond to demand in economically and racially isolated neighborhoods without incorporating the predatory dimensions of many fringe service providers.

CCOs are typically small storefront operations where patrons can cash checks for a fee.¹ Over the past few decades, they have become extremely common in New York City: In 2011, there were more CCOs (387) than Starbucks or McDonalds (González-Rivera 2013). CCOs have tended to locate in poor communities of color lacking mainstream financial services (Caskey 1994). Many scholars have pointed to a similar dynamic resulting in the geographic clustering of foreclosure-prone subprime mortgages: The lack of prime credit in certain communities created a market void, in which subprime lending grew (Gramlich 2007; Squires 2004). In addition to the historic factors linking these manifestations of financial services disparities, the Great Recession could have exacerbated neighborhood differences through demographically clustered job loss and erosion of trust in mainstream banks (Kenworthy and Owens 2011). Spatially differentiated treatment of certain neighborhoods by the mortgage industry also contributed to the geographic clustering of foreclosures (Chan et al. 2013). The concentration of unemployment (Hout, Levanon, and Cumberworth 2011), subprime lending (Faber 2013), foreclosures (Hall, Crowder, and Spring 2015; Dan Immergluck 2009), and wealth loss (Faber and Ellen 2016) in historically vulnerable communities provides further motivation for the study of the changing economic climate and prevalence of check cashing through the Great Recession.

Financial services are necessary for participation in social and economic life (Carruthers and Kim 2011; Dreier, Swanstrom, and Mollenkopf 2001). Fringe services provide stigmatized, low-quality products (Negro, Visentin, and Swaminathan 2014) to already-vulnerable populations (Caskey 1994) at a considerably higher cost than mainstream providers such as bank branches. In aggregate, these fees can cost communities millions of dollars, stunting economic growth and resilience in the face of negative shocks (Dreier, Swanstrom, and Mollenkopf 2001; Fernholz 2010; Kubrin et al. 2011; Rhine, Greene, and Toussaint-Comeau 2006; Smith, Smith, and Wackes 2008). These pecuniary penalties for those living in poor neighborhoods have been referred to as “poverty” or “ghetto” taxes (Gallmeyer and Roberts 2009; Wacquant 2008). Federal and local governments have acknowledged the connection between economic vulnerability and private financial institutions and the importance of expanding access to and adoption of mainstream financial services in lieu of fringe services (Federal Deposit Insurance Corporation [FDIC] 2012; New York City Department of Consumer Affairs [NYCDCA] 2010b).

The Great Recession has produced more academic scrutiny of alternative financial services (Gross, Hogarth, and Manohar 2012). However, despite the important role of finance in systems of stratification more broadly and the latest economic downturn in particular, fringe financial service providers have been inadequately studied (Carruthers and Kim 2011; Negro, Visentin, and Swaminathan 2014). We also know relatively little about how neighborhoods changed during the greatest economic downturn in almost a century (Owens and Sampson 2013). While unemployment and foreclosure have important implications for the study of inequality between individuals and households, their spatial organization could have lasting impacts given the growing evidence in support of neighborhood effects theories (Sharkey and Faber 2014). The recession offers a unique opportunity to investigate how rapid, locally organized changes in economic vulnerability led to demand for particular institutions. The goal of this project is to explore how the confluence of these factors created a spatial niche in which a specific kind of institution, CCOs, emerged. Furthermore, this article investigates how the geographic relationship between the recession’s impact and CCO growth shaped racial disparities in local access to financial services.

I leverage variation over time within census tracts to analyze changes in the financial services landscape across New York City. Surprisingly, given that previous work has typically argued that the CCO industry is vulnerable during economic downturns, I find remarkable growth in the number of CCOs between 2006 and 2011 (Figure 1). Using residential foreclosures as a proxy for the many impacts of the Great Recession (e.g., unemployment and wealth loss), I also find that economic vulnerability and applications for new check cashing licenses were concentrated in the same neighborhoods, which suggests CCOs may have been profiting from suffering communities. Furthermore, these neighborhoods tended to have larger minority presence. Mainstream bank presence also grew during this period, though dramatic racial disparities in local access to branches persisted. The geographic link between a recession caused by subprime credit and the subsequent growth of other predatory financial servicers serves as strong evidence for the argument that these actors exacerbate urban inequality.

Figure 1.

Change in CCO presence from 2006 to 2011.

Literature Review

The finance industry plays an important role in the stratification of space. Private financial institutions determine, often in partnership with local governments, where to invest in housing, economic development, and infrastructure. These monetary commitments to place are often drastically uneven, creating both winners (i.e., neighborhoods and municipalities with robust job growth and housing value appreciation) and losers (i.e., areas marred by crime and poverty) (Dreier, Swanstrom, and Mollenkopf 2001; Logan and Molotch 1987; Squires 1994, 2004). The decision of which neighborhoods a financial institution chooses to invest in has implications for access to basic financial services crucial to wealth generation (e.g., checking accounts). Research on the unbanked and underbanked has shown that convenience is one of the main reasons individuals choose alternative service providers (Caskey 1994). Without basic services, more sophisticated forms of wealth accumulation, such as investments in financial markets, homeownership, or retirement accounts, are out of reach.

“Alternative” or “fringe” service providers are the converse of mainstream banks. Just as the distribution of banks influences access to tools for wealth generation, CCOs, payday lenders, and pawnbrokers serve to leech assets from the neighborhoods in which they are located through comparably large fees and interest (Sawyer and Temkin 2004; Smith, Smith, and Wackes 2008). While mainstream banks are common in wealthy, suburban, and White areas, fringe services are more likely located in poorer, urban, communities of color (Gallmeyer and Roberts 2009; Smith, Smith, and Wackes 2008).

The expansion and segmentation of mainstream and fringe services in the financial sector since the 1980s has created a two-tiered system, in which the poor and the wealthy have access to dramatically different services (Baradaran 2015; Caskey 1994; Gallmeyer and Roberts 2009; Kubrin et al. 2011; Pew Charitable Trusts 2013; Squires and O’Connor 1998). The avoidance of neighborhoods by mainstream banks is implicated in both subprime lending and storefront fringe financial services such as CCOs (Aalbers 2009; Cover, Spring, and Kleit 2011; Graves 2003; Squires and O’Connor 1998). By creating vulnerable individuals and communities, the hierarchy of financial services serves as both a consequence and perpetuating factor of widening inequality (Aalbers 2012).

CCOs

Compared with mainstream banking, the use of CCOs is more expensive on a per-transaction basis (Bertrand, Mullainathan, and Shafir 2004; Caskey 1994; Cover, Spring, and Kleit 2011; Fox and Woodall 2006). CCO fees, which vary by state, are a percentage of the check’s value. CCOs also often offer additional services, such as paying utility bills, wiring money, and purchasing money orders—all of which require fees (Caskey 2002). Because CCOs provide many of the same services as typical bank branches (as opposed to payday lenders and other credit-based alternative financial services), their geography and pricing vis-à-vis mainstream banking are of particular interest to researchers and policy makers (Bradley et al. 2009).

CCO supporters justify the costs by pointing to the risks CCOs take such as staying open for long hours and operating in neighborhoods with higher crime rates (Cover, Spring, and Kleit 2011; Kubrin et al. 2011; Squires and O’Connor 1998). However, policy makers, inequality scholars, and advocates for the poor have been concerned with CCOs because these fees can add up to a substantial amount of money over time. Fellowes and Mabanta (2008) estimated that in some states, a full-time worker could lose $40,000 over a career through the reliance on check cashing. CCOs also typically locate in financially vulnerable areas, populated by people for whom this amount of money could be transformative.

Historically, CCOs were located primarily in large cities (Caskey 1994). Over the past few decades, however, their presence has grown dramatically throughout the country, aided in part by low regulatory barriers (Caskey 2005; Fox and Woodall 2006; Kubrin et al. 2011; Prager 2009). CCOs still typically remain more prevalent in urban areas (Squires and O’Connor 1998), though they have spread from central cities to some declining suburbs (Caskey 2002). Because most states do not provide data on CCO locations (Caskey 2005), only a small handful of studies have been able to systematically assess their geographic distribution. Those existing studies have consistently documented that CCOs are located in socioeconomically disadvantaged communities. They are more common in neighborhoods with non-White majorities and large immigrant populations. CCOs also tend to colocate with the working poor, though not the very poor, who may not have many checks to cash (Caskey 1994; Cover, Spring, and Kleit 2011; Gallmeyer and Roberts 2009; Graves 2003; Squires and O’Connor 1998).

Similarly, survey studies of individuals who use CCOs find their patrons have lower levels of educational attainment (Caskey 2002; Gross, Hogarth, and Manohar 2012; Lachance 2014; Prager 2009) and are more likely to use other fringe services (Fox and Woodall 2006). Racial minorities disproportionately use CCOs (Caskey 2002; FDIC 2012; Gross, Hogarth, and Manohar 2012; Prager 2009), even after controlling for having a bank account (Rhine, Greene, and Toussaint-Comeau 2006). Because the liquidity constraints of poverty make it more difficult to keep a minimum balance for a bank account, those relying on check cashers tend to be low income and/or unemployed (FDIC 2012; Gross, Hogarth, and Manohar 2012; Squires and O’Connor 1998).

Although a higher percentage of people without bank accounts use CCOs (Caskey 2002; FDIC 2012; Gross, Hogarth, and Manohar 2012), the effect of local bank branch presence on CCO patronage is unclear (Fellowes and Mabanta 2008). While the growth of CCOs in neighborhoods may have been due to a financial services void (Caskey 1994; Smith, Smith, and Wackes 2008; Squires 2004), check cashers today are often located in neighborhoods that also have banks (Fellowes and Mabanta 2008; Gross, Hogarth, and Manohar 2012; Sawyer and Temkin 2004). It is possible that the historical avoidance of disadvantaged communities by banks created local norms of reliance on fringe services.

Surveys have indicated that CCO users often lack comfort with mainstream institutions (Gross, Hogarth, and Manohar 2012; Rhine, Greene, and Toussaint-Comeau 2006), finding the rules and penalties associated with a bank account to be onerous (Bertrand, Mullainathan, and Shafir 2004; Caskey 2002). Ethnographic research at CCOs has found that many patrons were deliberately choosing CCOs for these reasons. There was also an immediacy of need for cash among patrons, which banks often cannot provide to the poor (Servon 2013, 2014). These issues may have become more salient during the Great Recession, as job losses have put more people at risk of increasingly expensive bank fees (Andriotis 2014; Dash 2009, 2011) for overdrafting an account or failing to keep a minimum balance. This work supports the argument that poor individuals have to make “hard choices” about their banking (Buckland 2012; Ratcliffe et al. 2015), which are often influenced by things out of their control such as decisions by banks of where to open branches, fee structures, and economic trends that affect cash flow and assets. This also suggests that CCOs are not inherently bad. If we eliminated all CCOs tomorrow, the communities in which they are located would likely be worse off. Nor are banks the good actors. Together, increasing bank fees and CCO growth are creating a system that is more expensive over time, especially for the poor.

Check Cashing in New York City

CCO presence has grown in New York over the past few decades. According to data provided by the New York State Department of Financial Services (NYSDFS), there were relatively few CCOs in the city until the 1970s. In the 1970s and 1980s, there were just over 100 applications for new CCOs. The number of new CCO licenses jumped to approximately 250 in the 1990s and almost 450 in the first decade of the twenty-first century. New York State has some of the strongest fringe financial service regulations in the country. Despite New York’s relatively low fees,² CCOs are still more expensive on a per-transaction basis than traditional checking accounts, which typically do not charge fees for depositing or cashing checks. Stated another way, the extent to which individuals are compelled to pay CCOs rather than use a free, identical service at a bank simply because they cannot afford a minimum balance constitutes a tax on poverty. Finally, although payday lending is currently illegal in the state (Abrams 2014; NYCDCA 2010a), some CCOs have lobbied for the right to provide the service, which many say is predatory (Beekman 2013).

Survey data can help characterize demand for financial services in New York City. In 2009 and 2011, the FDIC conducted the National Survey of Unbanked and Underbanked Households, which asked respondents what kinds of financial services they used and their reasons for using them (FDIC 2009, 2011). While not a complete overlap with the time period this article explores, these data do offer insight into how prevalent CCO use is and what motivates CCO use.

From 2009 to 2011, the population-weighted percentage of respondents with bank accounts increased slightly from 85.56% to 86.74% in New York City. Concurrently, the percentage of people who have “ever used check cashing” also increased from 13.24% to 15.02%. While CCO usage was most prominent and grew among those without bank accounts (32.27% in 2009 and 49.09% in 2011), patronage remained steady among those who did have bank accounts (10.24% in 2009 and 10.29% in 2011). In both years, the main reason cited for not having a bank account among unbanked New Yorkers was not having enough money. New Yorkers identified expediency (21.69% in 2009 and 26.12% in 2011), convenience (31.11% in 2009 and 24.92% in 2011), and lack of a bank account (28.04% in 2009 and 22.94% in 2011) as the main reasons for using CCOs.

CCO Growth During Recession

Economic downturns typically hurt CCO growth because unemployment results in fewer, smaller paychecks. However, a recession could create opportunity for CCOs in the form of greater numbers of families in dire financial straits. Many use check cashers because they do not have enough money to satisfy a bank’s minimum balance requirement or to draw down on to immediately cash a check at a bank. These are often individuals who are living paycheck-to-paycheck and need cash quickly (Caskey 1994, 2005; Squires 2004). The considerable increase in unemployment during the Great Recession likely increased the number of people in such a situation. Importantly, job loss was concentrated among those with less education, African-Americans, and immigrants (Hout, Levanon, and Cumberworth 2011)—the same people who disproportionately patronized CCOs (Caskey 2002). These same dynamics could have led to overlapping spatial and demographic clustering of foreclosures.

The decision to use a CCO to cash a check is partially a function of discomfort with and distrust in mainstream institutions (Bertrand, Mullainathan, and Shafir 2004; Gross, Hogarth, and Manohar 2012; Rhine, Greene, and Toussaint-Comeau 2006). In the 2009 FDIC survey, these factors were the main reason for not having a bank account for 8% of respondents (FDIC 2009). Some of this is due to requirements and fees viewed by some as unadvertised by banks (Caskey 1994; Rhine, Greene, and Toussaint-Comeau 2006). If subprime lending and foreclosures eroded trust in large, commercial banks, through ballooning mortgage payments, prepayment penalties, and other predatory characteristics, they could have also pushed more people to use fringe service providers. Not surprisingly, given the public perception of malfeasance, there was a substantial decrease in trust in banks and financial institutions during the recession (Kenworthy and Owens 2011). The practices that made communities vulnerable to the expansion of fringe financial services also positioned those same communities poorly to withstand the Great Recession (e.g., through an overreliance on subprime lending).

In addition to increasing economic insecurity, there are alternative reasons why CCOs may have expanded during the Great Recession. These institutions may be attracted to areas with growing populations. Because this time period was characterized by housing market tumult, population change may be correlated with the fallout of the foreclosure crisis. In addition, many banks increased fees to recoup losses incurred due to the collapse of the housing market (Andriotis 2014; Dash 2009, 2011), which may have pushed mainstream financial services out of reach for many individuals as unemployment was rising and incomes were falling.

The recession’s impact was highly spatially and racially organized, motivating an exploration of how rapid changes in economic vulnerability concentrated in space might lead to demand for fringe financial institutions. In addition to exploring the neighborhood conditions in which CCOs grew, I want to know whether disparities between communities increased over time. Given the long history of racial exclusion by the finance industry, it is important to understand the extent to which the increased segmentation of the industry has had a racialized impact. I anticipate that financial distress caused by the Great Recession and concentrated in particular communities served as a growth opportunity for CCOs, which can alleviate short-term cash flow emergencies—though at higher costs than banks. I further expect the expansion of these institutions to follow a highly racialized pattern and overlap with the prevalence of subprime lending during the housing boom.

Data

Although previous research has shown correlations between neighborhood socioeconomic disadvantage and the presence of alternative financial service providers (including CCOs), I leveraged data from several sources to connect the ecological and institutional aspects of the geography of check cashing to rapid changes in financial security (proxied by foreclosures). I focused on the years 2006 through 2011 because my foreclosure data began in 2006 and my CCO data ended in 2011. I used census tracts as the primary units of analysis, which offered imperfect measurement of neighborhoods but allowed me to bring together multiple data sets.

I acquired data on CCO locations via a Freedom of Information Law request to the NYSDFS. The data set contains the name, address, application date, and current status of every license during the period of study. Because I am interested in locations where individuals are able to cash checks, I only analyze “Full Service” offices (officials from NYSDFS indicated that “Unlicensed Headquarters” do not cash checks). I geocoded each address and identified the 2010 census tract in which it was located using ArcGIS 10. Once matched to tracts, I calculated the number of licensed CCOs present in each census tract at baseline (2005). I then identified the number of applications for new CCOs in each tract for each year between 2006 and 2011. Finally, I counted the number of lost CCO licenses for each tract by year. I verified this coding scheme with officials from NYSDFS.

Because CCOs offer many of the same services as banks (Bradley et al. 2009), I calculated the number of bank branches in each census tract by year. The FDIC provided the address of every branch by year in their Statistics on Depository Institutions (FDIC 2014), which I geocoded and aggregated at the census tract level. For another measure of the financial services environment, I calculated the subprime mortgage lending rate for each census tract across 2005 and 2006 using data provided by the Home Mortgage Disclosure Act (HMDA) (Federal Financial Institutions Examination Council [FFIEC] 2006, 2007) by dividing the number of approved purchase loans with interest rates three or more points above the federal treasury rate in both years by the total number of approved purchase loans in both years (Faber 2013).

Previous research has demonstrated a relationship between financial service mix and the demographic characteristics of communities. I used data from the National Historical Geographic Information System (NHGIS) (Minnesota Population Center 2011) to evaluate the salience of such features for CCO growth. Because tract-level data were not available for the starting year of the study (2006), I used linear interpolation to estimate local conditions for that year. Measures of population, racial makeup (i.e., percent non-Hispanic Black, non-Hispanic Asian, Hispanic/Latino, and non-Hispanic White), homeownership rate, and the number of housing units were from the 2000 and 2010 Decennial Census, whereas poverty rate, percent foreign born, and percent of those above 25 years old with a college degree were from the 2000 Census and 2008–2012 pooled American Community Survey. I used the 2000 to 2010 census tract crosswalk provided by the Longitudinal Tract Database (LTDB) (Logan, Xu, and Stults 2012) to match 2000 tracts to their 2010 equivalent for this interpolation.³

I added two controls of variation in commercial activity and viability across neighborhoods, which may have shaped CCO siting decisions. I used U.S. Postal Service (USPS) Office of Inspector General ZIP Code Crosswalk Files (U.S. Department of Housing and Urban Development [HUD] 2012) to estimate yearly, tract-level business activity based on data provided by the Census’s ZIP Code Business Patterns (U.S. Census Bureau 2012). The second measure was the time-invariant distance to the nearest subway stop (DOITT 2013). Using ArcGIS’s Network Analyst package, I calculated walking distance along city streets⁴ to the nearest subway entrance for each tract’s centroid.

A central goal of this project was to explore whether areas hit hardest by the Great Recession experienced changes different from places left relatively unscathed. The recession had a wide range of impact, including unemployment, lost assets, and declining home values. Unfortunately, it is difficult to get time-varying, neighborhood-level data on these measures. I used foreclosure data, collected by RealtyTrac, as a proxy for the local impact of the recession (for evidence connecting foreclosures to these and other measures of economic vulnerability, see, for example, Chan et al. 2013; Dwyer and Lassus 2015; Daniel Immergluck 2010; Rugh and Massey 2010). This data set includes the longitude, latitude, and date of every foreclosure from 2006 to 2011 in New York City. I spatially joined every residential unit of any size receiving a foreclosure action to the census tract in which it fell and aggregated foreclosure activity by tract for each year. If the same unit received multiple foreclosure actions as it progressed through the foreclosure process (e.g., a lis pendens followed by a sale), the unit was only counted once.

The unique housing market of New York City presents challenges with using these data: It is an expensive market, and most foreclosures were rental properties. However, the same factors led to foreclosure in New York as in the rest of the country: rising unemployment, depleted wages, and negative equity (Weselcouch 2014). I used foreclosures to stand in as an indicator of financial vulnerability. This is an imperfect measure, but it allowed me to estimate the changing impact of the recession at a level of temporal and geographic granularity unavailable with other data. My final sample consisted of 2,109 tracts across six years (12,654 tract-years). I excluded 37 tracts because NHGIS did not have information for them and an additional 22 tracts which did not have a path to a subway station that did not require crossing a body of water.

Methods

Because most of the aggregate change in CCO presence was driven by new CCOs opening, rather than existing ones closing, I estimate discrete-time event history models in which the unit of observation is the census tract-year and the “hazard” is whether there was an approved application for a new CCO within a particular tract-year. Because only a small minority of cases (38 of 12,660 tract-years) gained two or more CCOs in the same year, I use a binary variable to indicate the hazard, coded one if there was at least one application or zero if there were no new applications. I use logistic regression in which the tract-year is the unit of observation (Allison 2010, 2014; Singer and Willett 2003) to estimate this hazard (i.e., an outcome of one or zero) based on the covariates described above.

Because the foreclosure data began in 2006 and the check cashing data ended in 2011, the data set I created is both left and right censored. Fortunately, all tracts are right censored at the same point, which leaves the possibility that it is noninformative (Allison 2014). To reduce potential bias from left censoring, I controlled for the number of CCOs in each tract in the time immediately preceding the study period (i.e., 2005) (Singer and Willett 2003).

I include dummy variables for each borough (excluding Manhattan) and for each of the years from 2007 to 2011 (excluding 2006) to capture secular trends and any relationship between CCO growth and changes over time in fees charged by either CCOs or mainstream banks.⁵ Because this hazard is repeatable, I calculate gap time, which measures the number of years since the previous CCO application (Allison 2014). For example, if a census tract gains a CCO in 2007 and another in 2010, the gap time will be measured as zero in 2006, zero in 2007, one in 2008, two in 2009, three in 2010, and back to one in 2011. I also cluster standard errors at the tract level. These models estimate the likelihood of a census tract gaining a new CCO in any given year based on a set of baseline and time-varying characteristics.

Results

NYSDFS granted licenses for 231 new CCOs in New York City from 2006 to 2011: 75 CCOs opened in Queens, 63 in Brooklyn, 41 in the Bronx, 36 in Manhattan, and 16 in Staten Island. The city also lost 94 CCOs during this time period, so the city gained 137 CCOs in net—an increase of more than 50%. Figure 1 shows CCO growth was strongest in 2008 and 2009, which were also the only two years in which the city lost bank branches. Only in 2011 were there more lost licenses than those gained.⁶ The expansion of check cashing during the Great Recession was particularly interesting because this industry has historically retracted during economic downturns. The individual-level FDIC data described above indicated an increase in CCO usage among New Yorkers (though only from 2009 to 2011) and also pointed to economic vulnerability as a strong predictor of CCO patronage.

During the period, 185 census tracts experienced a net gain in CCOs, whereas 60 experienced a net loss, and the vast majority (1,864) experienced no change. Compared with tracts that experienced a net loss in CCO presence, those that gained CCOs had significantly larger Latino and foreign-born populations, smaller White populations, lower educational attainment, and fewer business establishments (Table 1). Subprime lending was also more prevalent during the housing boom in these areas, as were foreclosures during the recession. On average, tracts that gained CCOs had larger Latino populations, smaller White populations, lower homeownership rates, higher poverty rates, and lower educational attainment when compared with areas that did not experience any change.

Table 1.

Baseline Characteristics (2006) by Whether a Tract Gained, Lost, or Experienced No Change in Number of CCOs from 2006 to 2011.

	Net Loss of CCOs		No Change in CCOs		Net Gain of CCOs
	M	SD	M	SD	M	SD
Total foreclosures (2006–2011)	25.7	3.72	31.2	0.79	39.4*,^††	3.34
Tract characteristics in 2006
Percent Black	22	0.28	26	0.31	28	0.29
Percent Asian	13	0.16	12	0.14	13	0.15
Percent Latino	22	0.17	22	0.20	28**,^†††	0.21
Percent White	40	0.31	36	0.32	27***,^†††	0.27
Percent foreign born	34	0.16	36	0.16	38**	0.15
Population	4096	2419	3615	1967	3968	1990
Homeownership rate (%)	30	0.22	36	0.24	28^†††	0.21
Total housing units	1910	1429	1540	1069	1636	940
Poverty rate (%)	19	0.14	19	0.13	21^†	0.12
Percent with college degree (25+)	35	0.25	28	0.19	25***,^†	0.18
Subprime rate (%) (2005–2006)	8	0.08	11	0.08	12***	0.08
Business establishments	1760	1822	1054	979	1141**	1349
Distance to nearest subway station (miles)	0.60	0.74	0.79	0.94	0.73	1.00
Population change (2006–2011)	563	78	423	13	661^†††	50
	n	%	n	%	n	%
Number of tracts by borough
Bronx	8	2.4	288	86.8	36	10.8
Brooklyn	18	2.4	675	90.2	55	7.3
New York	15	5.4	237	85.6	25	9.0
Queens	18	2.8	569	88.1	59	9.1
Staten Island	1	0.9	97	89.8	10	9.4
Total	60	2.9	1,866	88.5	185	8.7

Note. CCO = check cashing outlet.

Significance of differences between areas that experienced a net loss or net gain in CCOs: *p < .05. **p < .01. ***p < .001. Significance of differences between areas that experienced no change and a net gain in CCOs: ^†p < .05. ^††p < .01. ^†††p < .001.

CCO Growth Was Concentrated in Communities of Color

Because CCOs were more likely to open in communities of color, disparities between these communities and White areas increased. Figure 2 categorizes neighborhoods by their racial majority⁷ (or if there is no majority) and shows the change in the average number of CCOs between 2006 and 2011. CCO presence increased across communities of all categories but most dramatically in Black, Latino, and mixed neighborhoods. By the end of period, non-White neighborhoods had approximately twice the number of CCOs, on average, as White neighborhoods.

Figure 2.

Change in average number of CCOs by neighborhood majority race.

CCO Presence Grew in Neighborhoods Hit Hardest by the Great Recession

In addition to exploring differences in changes in financial service environments across neighborhood racial categories, it is important to understand how neighborhoods changed during the recession. Again, the Great Recession had a wide range of impacts, and I used foreclosure activity as a neighborhood-level proxy for economic distress. Figure 3 shows trends in average foreclosure counts across neighborhoods that experienced a net gain (average of 6.6 foreclosures per year), loss (4.3), or no change (5.2) in the number of CCOs. At the peak of New York City’s foreclosure crisis (in 2009), areas in which CCOs expanded experienced 45% more foreclosures on average (9.9) than those that did not (6.7).

Figure 3.

Average number of foreclosures by whether a tract gained, lost, or experienced no change in number of CCOs from 2006 to 2011.

These figures suggest that the spatial organization of financial vulnerability to foreclosure was tied to the expansion of CCOs and New York City’s racial geography. Communities of color were often targets of subprime lending (Faber 2013; Hwang, Hankinson, and Brown 2015; Rugh, Albright, and Massey 2015) and experienced the most job loss (Hout, Levanon, and Cumberworth 2011), which led to the subsequent clustering of foreclosures. While these areas were accumulating foreclosures, they were also gaining CCOs.

Bank Branch Growth

Just as the number of CCOs increased throughout New York, so, too, was the number of mainstream financial service providers. In 2006, there were 1,255 bank branches in the city. By 2011, there were 1,732. The concurrent growth of CCOs and mainstream banks is reflected in the FDIC survey data, which show that the percentage of New Yorkers with bank accounts and those that used check cashing grew (FDIC 2011). In contrast to CCO growth, the increase in mainstream bank branches was concentrated in years with relatively little foreclosure activity (i.e., 2010 and 2011).⁸ Figure 4 shows the total number of branches in New York by year. Again, it is notable that CCO growth was strongest in the two years in which the city lost branches (i.e., 2008 and 2009).

Figure 4.

Number of bank branches in New York City from 2006 to 2011.

As with CCOs, there were substantial differences in bank presence across neighborhood racial makeup. Predominantly White tracts began the period with the largest average number of banks (1.04), and that number increased (to 1.35). Communities with Asian majorities also saw an increase in banks (from 0.91 to 1.04), as did majority Latino neighborhoods (from 0.28 to 0.47). Majority Black areas had the fewest number of banks in 2006 (0.16) and experienced the smallest increase (to 0.25). Even though the relative gains in bank presence were larger in Black (50% increase) and Latino (70%) tracts than in White (30%), the gaps between these neighborhoods did not substantively change. For every bank in a majority White census tract in 2006, Black and Latino tracts had 0.16 and 0.27 banks, respectively. In 2011, those ratios were 0.18 for Black neighborhoods and 0.35 for Latino neighborhoods. In Asian tracts, the ratio declined from 0.88 to 0.77. Compared with predominantly White communities, banks remained quite rare in communities of color. Combined with the racialized trends in CCO expansion, these gaps illustrate the vast differences in financial service environments between White neighborhoods and communities of color.

Event History Model Results

To investigate the relationships between New York’s racial geography, economic vulnerability during the Great Recession, and CCO growth, I estimated discrete-time event history models, in which the unit of observation was the tract-year and the outcome was the addition of a new CCO. These results can be interpreted as any other logistic regressions, in which the likelihood that a new CCO opens within a census tract in any year is estimated based on covariates of interest. Table 2 shows coefficients from these estimates, which can take any negative or positive values. Values less than zero indicate that an increase in a particular variable is associated with a decrease in the likelihood of a new CCO. These are not meant to be causal estimates, but rather an attempt to explore conditional correlations between CCO expansion and covariates of interest. By taking a yearly approach to the analysis of CCO expansion, these models offer additional insight into the specific relationship between CCO growth and foreclosures, which proxy the recession’s local impact in lieu of yearly data on unemployment and other measures of neighborhood economic health.

Table 2.

Discrete-Time Event History Models of Check Cashing Expansion in New York City from 2006 to 2011.

	Model 1		Model 2		Model 3		Model 4
Demographics in 2006
Percent Black^a	1.3259***	(0.398)	0.9484*	(0.442)	0.9975*	(0.493)	0.7785	(0.509)
Percent Asian	2.0060**	(0.752)	2.1042**	(0.740)	1.8206*	(0.891)	1.6707	(0.885)
Percent Latino	1.0528	(0.727)	0.9026	(0.712)	1.8711*	(0.799)	1.7682*	(0.817)
Percent foreign born	−1.0159	(0.768)	−0.9702	(0.781)	−1.2545	(0.979)	−1.0807	(0.988)
Log population	0.1951	(0.265)	0.0831	(0.250)	0.6058	(0.338)	0.5657	(0.323)
Homeownership rate	−1.7456**	(0.556)	−2.1120***	(0.618)	−1.9930**	(0.648)	−2.1120**	(0.659)
Total housing units	−0.0002	(0.000)	−0.0001	(0.000)	−0.0007**	(0.000)	−0.0007**	(0.000)
Poverty rate	−2.5618*	(1.176)	−2.4849*	(1.173)	−3.9681**	(1.534)	−4.1068**	(1.572)
Percent with college degree (25+)	−0.2743	(0.729)	−0.2624	(0.744)	−0.5692	(1.015)	−0.6410	(1.016)
Population growth (2006–2011)	0.0005***	(0.000)	0.0005***	(0.000)	0.0000	(0.000)	0.0000	(0.000)
Finance and the Great Recession
Business establishments	0.0003**	(0.000)	0.0002*	(0.000)	0.0002	(0.000)	0.0002	(0.000)
Distance to nearest subway (miles)	0.1140	(0.103)	0.1054	(0.099)	0.0871	(0.103)	0.0596	(0.104)
Foreclosures			0.0311*	(0.016)	0.0295*	(0.013)	0.0065	(0.011)
CCOs in 2005					2.9829***	(0.228)	2.9767***	(0.227)
Number of banks					−0.0020	(0.040)	−0.0039	(0.039)
Lost CCOs in prior year					0.3818	(0.395)	0.3899	(0.400)
Subprime lending rate (2005–2006)					0.3807	(1.228)	−0.8986	(1.334)
Foreclosures × Subprime Rate							0.2014*	(0.081)
Borough^b
Bronx	0.0664	(0.261)	0.0495	(0.264)	−0.3577	(0.332)	−0.2971	(0.334)
Brooklyn	0.1439	(0.245)	0.0872	(0.249)	−0.0807	(0.250)	−0.1013	(0.249)
Queens	0.5025	(0.264)	0.3773	(0.278)	−0.0194	(0.245)	−0.0862	(0.248)
Staten Island	0.7732*	(0.353)	0.4630	(0.392)	0.5298	(0.384)	0.5282	(0.377)
Time^c
2007	0.2989	(0.325)	0.1986	(0.335)	−0.3893	(0.378)	−0.4816	(0.379)
2008	1.4120***	(0.273)	1.3465***	(0.277)	0.9785**	(0.319)	0.9193**	(0.318)
2009	1.1194***	(0.273)	0.9631***	(0.282)	0.6584*	(0.317)	0.5912	(0.315)
2010	1.0552***	(0.282)	1.0581***	(0.282)	0.8553**	(0.297)	0.8383**	(0.297)
2011	0.0763	(0.339)	0.1434	(0.340)	0.0000	(0.0)	0.0000	(0.0)
Gap time	0.3801***	(0.087)	0.3789***	(0.087)	−0.4100*	(0.208)	−0.4107*	(0.206)
Constant	−6.3663**	(2.059)	−5.3050**	(1.968)	−8.7559***	(2.472)	−8.1352***	(2.340)
n	12,654		12,654		10,545		10,545

Note. Clustered standard errors in parentheses. CCO = check cashing outlet.

Percent White excluded due to multicollinearity.

Manhattan excluded as the reference group.

Year 2006 excluded as the reference group. An additional year of data was dropped when the lagged variable of lost CCOs in the previous year was added to the model.

p < .05. **p < .01. ***p < .001.

I began by modeling the relationship between CCO growth, tract sociodemographic characteristics, geography, and time (Model 1). Throughout the period of study, higher representation of Blacks and Asians within a tract was associated with significantly higher likelihood of a new CCO license within that tract. The relationships between CCO expansion and percent Asian was particularly strong, perhaps indicating that language barriers, a lack of familiarity with or trust in banks, or the need to send remittances abroad (a service available at some CCOs) made these communities attractive to CCOs. Percent foreign born was included in the model, though it was not significantly predictive of new CCOs, nor was it significant in supplemental models excluding the measures of racial makeup. Although the coefficient for percent Latino was positive, it was not significant in this model in part because it was highly correlated with population growth. When population growth was excluded from Model 1, percent Latino was strongly correlated with CCO growth. So there is evidence that new CCOs were more likely to open in neighborhoods with growing populations and predominantly Latino populations.

CCOs tended to expand in areas with lower homeownership rates, which may have operated as a measure of socioeconomic status. While other work has found CCOs to be more common in areas with lower educational attainment and higher poverty, no other social or economic factors were significantly associated with CCO growth in my model. To examine nonlinearities in the relationship between poverty and the likelihood of CCO expansion, I estimated a model with a quadratic term for poverty. In this model, the linear term was positive and the squared term was negative—neither was significant.⁹ The correlation between neighborhood poverty and other characteristics was covering up the relationship between CCO growth and poverty. When I excluded other characteristics, poverty was significant and positively associated with the likelihood of a new CCO opening. After controlling for other tract characteristics, commercial activity and population growth were strong predictors of new CCOs, though proximity to the nearest subway station was not. Net of demographics, tracts in Staten Island were significantly more likely to gain CCOs than those in Manhattan (the reference category). The gap time variable is also significant, suggesting the geographic clustering of CCOs over time. The coefficients for the year dummy variables, which decline in magnitude from 2008 through 2011, do not support the theory that CCOs were expanding to take advantage of increases in the maximum allowable fee, which rose during this time period.

The primary concern of this article was the connection between financial distress during the Great Recession and the expansion of alternative financial institutions. The economic collapse had a wide range of consequences, and I used residential foreclosure activity as a proxy for the local impact of the Great Recession. In Model 2, we see that this measure was positively associated with new CCOs opening net of demographic characteristics, geography, and time. Census tracts with more foreclosures within any particular year gained more CCOs than tracts left relatively unscathed by the foreclosure crisis, suggesting a connection between financial vulnerability and demand for CCOs, or, perhaps, a targeting of the hardest hit neighborhoods. Specifically, seven additional foreclosures (i.e., approximately one standard deviation of the yearly measure of foreclosure volume) within a tract-year were associated with approximately a 24% increase (i.e., e^0.0311×7) in the odds that a new CCO opened. This magnitude was not only statistically significant but substantively important. Although the strength of this relationship does not indicate that seven new customers would be enough to sustain a new CCO, even such a small number of foreclosures may be a signal of a much larger increase in economic insecurity within a neighborhood—and CCOs may have been attuned to these localized changes.

When I added additional variables measuring the financial environment in Model 3, the existing coefficients told a similar story—though now percent Latino was significantly correlated with CCO growth and the coefficient for population growth lost significance. The number of bank branches within a tract was not significantly related to a tract gaining a CCO, perhaps because bank fees rose during this period (Andriotis 2014; Dash 2009, 2011), making them less attractive for those suffering from the recession’s fallout. Although the number of lost CCOs in the prior year was not significantly correlated with CCO expansion, CCO presence in 2005 was a strong predictor of growth, which may have been indicative of already existing demand within neighborhoods or a dynamic wherein once a neighborhood was targeted by high-cost financial services, that space continued to be so. In addition, the fact that they grew in areas where they already existed was a sign of increasing inequality between places. The prevalence of subprime lending within a tract during 2005 and 2006 was not significantly correlated with the likelihood of a new CCO opening. Similar to poverty, the bivariate relationship between subprime lending and CCO growth was significant, while other neighborhood characteristics correlated with subprime lending (e.g., race) were also correlated with CCO expansion.

Model 4 investigates whether the relationship between CCO growth and foreclosures varied across the distribution of subprime lending. The interaction term greater than 1 shows that the relationship between foreclosures and CCO growth was strongest in tracts where subprime lending was more prevalent. We might expect that neighborhoods reliant on one fringe service would be more willing to use another. The statistically significant interaction suggests a vulnerability that is particularly activated during times of crisis, whereas the lack of significant main effect suggests heterogeneity in what the foreclosure variable means. Tracts with low rates of subprime lending but high rates of foreclosure were Whiter and had higher incomes than places with high subprime rates and foreclosure volume. Foreclosures in these areas may have been on second homes, strategic defaults, or among people who still have a financial cushion protecting them against the necessity of CCO use.

Alternative Estimates

To assess variation in the relationship between foreclosures and CCO growth over time, I interacted foreclosures with dummy variables for each year. None of the interaction terms were significant. When foreclosures were lagged by one year, they were significantly, positively associated with CCO growth, though when both lagged and contemporaneous foreclosures were included in the model, only the contemporaneous measure was significant. Interactions between neighborhood racial makeup and foreclosures were not significant.

I made several attempts to address the unique nature of New York. I estimated the main models using a sample excluding Manhattan, because the outer boroughs are more representative of the rest of the United States, and the findings were substantively identical. Some scholars have suggested that census tracts may not be appropriate analytical proxies for neighborhoods, as individuals regularly move across bureaucratically drawn boundaries (Faber and Sharkey 2015). The relatively small size of tracts in New York could accentuate this problem. To address the fact that residents of a tract are likely interacting with neighboring tracts, I estimated a series of spatial cross-regressive models, which include spatially lagged measures of variables (Anselin 2001, 2003; Crowder and South 2008). Adding spatially weighted measures of the outcome variable (i.e., new CCOs opening nearby), foreclosures, bank branches, CCOs in 2005, and/or lost CCOs did not substantively change the relationships between foreclosures, racial/ethnic makeup, and the likelihood of a new CCO opening within that tract. New CCOs in surrounding tracts were positively correlated with CCOs opening within a tract, which is indicative of the concentration of demand in space.

The spatial pattern of lost CCO licenses is also a part of changing financial services environments. I estimated CCO closings using a similar series of models as employed to estimate CCO growth with a sample restricted to those census tracts that had at least one CCO in 2005 (i.e., those that were at risk of losing a CCO). The results show that CCOs were more likely to close in census tracts with high homeownership rates and more banks, suggesting that bank branches may have replacing CCOs. The more CCOs within a tract in 2005 was also positively correlated with CCO loss, which we might expect as there would be more CCOs vulnerable to closing. None of the measures of racial or ethnic makeup were significantly predictive of CCO closures, nor was foreclosure activity.

Discussion

This article investigated whether New York City neighborhoods most impacted by the Great Recession also saw an increase in the presence of CCOs. The growth of the check cashing industry in recent decades is part of a wider trend of growing inequality and the marginalization of the poor through the emergence of a tiered financial services system (Caskey 1994, 2005; Kubrin et al. 2011; Squires 2004; Squires and O’Connor 1998). The spatial concentration of economic vulnerability creates a market for fringe institutions (Rugh and Massey 2010) as mainstream banking becomes increasingly expensive for poor customers (Dash 2009, 2011; Servon 2014). In the absence of mainstream institutions willing to meet the needs of low-income New Yorkers (Ratcliffe et al. 2015), other institutions spring up to fill the void.¹⁰

In New York, there was an increase of more than 50% in CCO presence between 2006 and 2011—a trend also reflected in survey data showing a rise in the percentage of New Yorkers relying on CCOs. CCOs expanded most in census tracts that were amassing foreclosures, which serve as a proxy for the local impact of the Great Recession. Specifically, a one standard deviation increase in foreclosures was associated with a 24% increase in the odds of a new CCO opening. Although foreclosures may be an imperfect measure of the distress caused by economic collapse, they approximate changing financial vulnerability at a level of geographic and temporal granularity unavailable with other data. The relationship between foreclosures and new CCOs remained significant after controlling for sociodemographic characteristics and each tract’s financial climate. Furthermore, this finding is particularly surprising because CCOs typically fare poorly during economic downturns.

Part of the story is economic instability. Patrons of check cashers tend to not have enough money to keep a minimum balance for a bank account or to draw upon to cash checks at banks (Caskey 1994). The dramatic increase in unemployment and loss of equity during the Great Recession, which both disproportionately impacted people of color (Faber and Ellen 2016; Hout, Levanon, and Cumberworth 2011), likely increased the population for whom this was the case and contributed to the clustering of foreclosures as households lost their ability to pay housing costs. The spatial overlap between economic vulnerability and erosion of trust in mainstream financial institutions implicated in the recession (Kenworthy and Owens 2011) may have offered a spatial niche for fringe services such as CCOs. An alternative to this “demand side” interpretation of the correlation between financial distress and CCO growth could be that these institutions were exploiting the Great Recession’s impact for profit. Although the maximum fee CCOs were allowed to charge increased during this time period, the change was little more than keeping up with the concurrent rise in Consumer Price Index (CPI). Therefore, it is unlikely that such a small change in potential profitability due directly to larger allowable fees would explain the substantial increase in CCO presence. In addition, the coefficients for the year dummy variables do not offer support for this alternative explanation.

In addition to CCO presence growing in tracts with more foreclosures, they were more likely to expand in neighborhoods with larger minority populations. Despite the concurrent growth of mainstream bank branches, racial disparities in financial services mix persisted (and reflect earlier findings; Sawyer and Temkin 2004). Although the relationships examined in this article are measured on the neighborhood level, these findings have important implications for our understanding for how race is lived on a day-to-day basis. The neighborhoods in which people of color live—and Blacks and Latinos, in particular—are vastly different from those inhabited by White New Yorkers. CCOs were also most likely to open within neighborhoods where they already existed prior to the study period, further indicating the widening of institutional inequality across spaces.

In addition to financial stress, one of the most commonly reported reasons among New Yorkers for using CCOs instead of banks is convenience (Caskey 1994; FDIC 2011). This further reinforces the importance of our understanding of the spatial distribution of financial institutions. Indeed, other work has connected the local concentration of alternative financial services makeup to the likelihood than an individual uses such services (Friedline and Kepple 2016). Although some studies of financial geography suggest mainstream and fringe financial service providers operate in different types of communities (Cover, Spring, and Kleit 2011; Squires 2004) and I find evidence of racial disparities, the presence of banks in a tract was not associated (positively or negatively) with the likelihood of the addition of a CCO in any given year net of other variables. The number of bank branches within a census tract was positively associated with CCO closings, however, suggesting that banks may have been replacing CCOs in some neighborhoods.

Reliance on data from New York City is a limitation of this study.¹¹ Unfortunately, because most states do not report location data for fringe institutions (Caskey 2005), studies of the geography of alternative finance are often limited to one city or state (Sawyer and Temkin 2004). This is particularly problematic because the lack of data makes it difficult to evaluate differences in state regulatory regimes. I restricted my analysis to the state’s largest city because the vast majority of CCO expansions in the state occurred there and it is home to the state’s largest communities of color. Despite recent growth of fringe financial institutions beyond cities, they remain a largely urban institution (Squires and O’Connor 1998). New York, in particular, has historic importance as one of the cities in which CCOs originated (Caskey 1994). I estimated models excluding tracts in Manhattan, and the findings were substantively identical. Although the site of this study may not be representative of the rest of the United States, at the very least, these findings raise an interesting set of questions about how neighborhoods changed during the economic downturn.

The subprime lending crisis showed that fringe financial products can have significant, negative consequences for the communities in which they are clustered and the broader economy. While check cashers likely do not pose the same economic threat as the subprime lending boom, their growth and concentration in neighborhoods hit hardest by the Great Recession should be a policy priority as they contribute to a trend of declining quality of life among the poor (Caskey 1994). In fact, both federal and local governments have acknowledged the importance of expanding access to and adoption of mainstream financial services in lieu of fringe services. The FDIC identified “safe, secure, and affordable banking services” as a national priority and noted that confidence in the banking system depends on how it serves “the nation’s diverse population” (FDIC 2012). The NYCDCA’s Office of Financial Empowerment launched an outreach campaign in 2010 to encourage unbanked New Yorkers to use bank accounts and move away from their reliance on fringe services (NYCDCA 2010b). Even if CCOs in New York are among the least costly in the nation, they are still more expensive than a traditional checking account, which allows patrons to cash checks for free, and some are lobbying the state legislature for the right to provide currently illegal and often-predatory payday loans (Beekman 2013).

Although there certainly is demand for these services, the fact that a subpopulation is consuming a particular good or service does not definitively indicate that they would not prefer something else (Squires 1994). Evidence from behavioral science indicates that predatory financial service providers use marketing to induce suboptimal behavior among vulnerable populations (McCoy 2004). Furthermore, although people from all socioeconomic strata make decisions that may not appear to be in their best financial interest, such mistakes may be more consequential for the poor, given that they have fewer resources with which to mitigate the effects (Bertrand, Mullainathan, and Shafir 2004).

The findings presented here explore one way in which spatial inequality deepened during the Great Recession. CCO presence grew rapidly in communities of color and in places amassing foreclosures. Although my analysis used data aggregated on the census tract level, micro-level data could deepen our understanding of the link between rapidly changing financial situations and fringe service providers. Are individuals experiencing foreclosure and adopting CCO use due to lost wealth or income? What individual and family characteristics moderate or mitigate this behavior? How does immigration status shape trust in financial institutions? The answers to these and related questions would not only illuminate specific mechanisms but also help further inform policy.

How to encourage individuals to move away from fringe service providers depends on what motivates people to use them in the first place. For example, if the link between foreclosure activity and CCO growth is due to subprime loans and eroded trust in mainstream banks, a theory supported by national public opinion data (Kenworthy and Owens 2011), then perhaps a campaign to restore trust would be effective. Alternatively, if foreclosures are signaling community-level vulnerability stemming from a lack of financial savvy, financial education programs could reduce demand for check cashing.

However, if CCOs are growing disproportionately in areas hardest hit by the Great Recession because people who would otherwise use mainstream banks are forced to use fringe services due to dire financial situations, then a more robust policy response is likely needed. A stronger safety net would mitigate the effects of job loss and help reduce the number of people who find themselves lacking the funds to keep a minimum balance for a bank account. Job creation and economic stability are elusive goals but important factors related to the changing geography of opportunity. Improved resiliency to economic shocks would leave fewer individuals relying on fringe financial services. While some CCOs are trying to expand into more mainstream operations, for example, by offering savings accounts, advocates for the poor argue this would just make customers more reliant on their high-fee services (Hu 2012).

In 2014, the inspector general of the USPS proposed allowing post offices to offer nonbank financial services, such as check cashing, money orders, and international transfers, at lower rates than currently offered by many private, nonbank financial institutions. Post offices have historically provided financial services in the United States and currently serve as the sites of social welfare disbursement in some European countries. Importantly, 38% of post offices are located in ZIP codes that do not have banks. This proposal could potentially provide low-cost services to a vulnerable population in a highly regulated environment throughout otherwise underserved and expensively served neighborhoods (Baradaran 2015; USPS Office of Inspector General 2014).

The fact that fringe financial services expanded in neighborhoods hardest hit by an economic downturn caused by the overuse of a fringe financial instrument (subprime mortgages) has serious implications for our understanding of place stratification. This link suggests that the various ways in which the financial services sector contributes to inequality can and do compound one another, widening the opportunity gaps between neighborhoods. These findings support the argument that the development and geographic organization of a tiered financial sector constitute a “ghetto” or “poverty” tax. While not the focus of this article, the geographic clustering of these phenomena is likely to have negative effects on the asset holdings of residents of these communities and contribute to an already historically wide wealth gap (Taylor et al. 2011), providing another mechanism through which neighborhood quality impacts life chances (Sharkey and Faber 2014).

Footnotes

Declaration of Conflicting Interests

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

Funding

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

Notes

Author Biography

Jacob W. Faber is an assistant professor at New York University’s Robert F. Wagner Graduate School of Public Service and a postdoctoral research associate in the Department of Sociology at Princeton University. His research focuses on spatial inequality. He leverages observational and experimental methods to study the mechanisms responsible for sorting individuals across space and how the distribution of people by race and class interacts with political, social, and ecological systems to create and sustain disparities.

References

Aalbers

Manuel

. 2009. “Geographies of the Financial Crisis.” Area 41 (1): 34–42.

Aalbers

Manuel

. 2012. “Subprime Cities and the Twin Cities.” In Subprime Cities: The Political Economy of Mortgage Markets, edited by Aalbers

Manuel

, 3–22. Malden: Blackwell.

Abrams

Rachel

. 2014. “New York State Makes New Efforts to Combat Payday Lenders.” The New York Times, Dealbook. Accessed July 7, 2014. http://dealbook.nytimes.com/2014/04/29/embargo-new-york-makes-new-efforts-to-combat-payday-lenders/.

Allison

Paul D.

2010. “Survival Analysis.” In The Reviewers Guide to Quantitative Methods in the Social Sciences, edited by Hancock

Gregory R.

Mueller

Ralph O.

, 413–24. New York: Routledge.

Allison

Paul D.

2014. Event History and Survival Analyses. 2nd ed. Thousand Oaks: SAGE.

Andriotis

Annamaria

. 2014. “Overdraft Fees at Banks Hit a High, Despite Curbs.” The Wall Street Journal, April 1. http://www.wsj.com/articles/SB10001424052702304157204579475573602576630

Anselin

Luc

. 2001. “Spatial Econometrics.” In A Companion to Theoretical Econometrics, edited by Baltagi

Badi H.

, 310–30. Malden: Blackwell.

Anselin

Luc

. 2003. “Spatial Externalities, Spatial Multipliers, and Spatial Econometrics.” International Regional Science Review 26:153–66.

Baradaran

Mehrsa

. 2015. How the Other Half Banks: Exclusion, Exploitation, and the Threat to Democracy. Cambridge: Harvard Univ. Press.

10.

Beekman

Daniel

. 2013. “Check Cashing Stores Push Albany Lawmakers to Allow 200% APR Loans.” Daily News, April 23. http://www.nydailynews.com/news/politics/check-cashing-stores-push-offer-200-apr-loans-article-1.1325661

11.

Bertrand

Marianne

Mullainathan

Sendhil

Shafir

Eldar

. 2004. “A Behavioral-Economics View of Poverty.” American Economic Review 94:419–23.

12.

Bradley

Christine

Burhouse

Susan

Gratton

Heather

Miller

Rae-Ann

. 2009. “Alternative Financial Services: A Primer.” FDIC Quarterly. Accessed September 20, 2016. https://www.fdic.gov/bank/analytical/quarterly/2009_vol3_1/AltFinServicesprimer.html.

13.

Buckland

Jerry

. 2012. Hard Choices: Financial Exclusion, Fringe Banks, and Poverty in Urban Canada. Toronto: Univ. of Toronto Press.

14.

Bureau of Labor Statistics. 2016. “Consumer Price Index—All Urban Consumers for New York-Northern New Jersey-Long Island, NY-NJ-CT-PA.” http://www.bls.gov/regions/new-york-new-jersey/news-release/consumerpriceindex_newyorkarea.htm

15.

Carruthers

Bruce G.

Kim

Jeong-Chul

. 2011. “The Sociology of Finance.” Annual Review of Sociology 37:239–59.

16.

Caskey

John P.

1994. Fringe Banking: Check-Cashing Outlets, Pawnshops, and the Poor. New York: Russell Sage Foundation.

17.

Caskey

John P.

2002. “Check-Cashing Outlets in a Changing Financial System.” Working Paper No. 02–4, Federal Reserve Bank of Philadelphia.

18.

Caskey

John P.

2005. “Fringe Banking and the Rise of Payday Lending.” In Credit Markets for the Poor, edited by Bolton

Patrick

Rosenthal

Howard

, vol. 17, 18–19. New York: Russell Sage Foundation.

19.

Chan

Sewin

Gedal

Michael

Been

Vicki

Haughwout

Andrew

. 2013. “The Role of Neighborhood Characteristics in Mortgage Default Risk: Evidence from New York City.” Journal of Housing Economics 22 (2): 100–18.

20.

Cover

Jane

Spring

Amy Fuhrman

Kleit

Rachel Garshick

. 2011. “Minorities on the Margins? The Spatial Organization of Fringe Banking Services.” Journal of Urban Affairs 33:317–44.

21.

Crowder

Kyle

South

Scott J.

2008. “Spatial Dynamics of White Flight: The Effects of Local and Extralocal Racial Conditions on Neighborhood Out-Migration.” American Sociological Review 73:792–812.

22.

Dash

Eric

. 2009. “Bank Fees Rise as Lenders Try to Offset Losses.” The New York Times, July 1, B1.

23.

Dash

Eric

. 2011. “Banks Quietly Ramping Up Costs to Consumers.” The New York Times, November 14, A1.

24.

Department of Information Technology & Telecommunications (DOITT). 2013. “Subway Entrances.” Accessed April 19, 2015. https://nycopendata.socrata.com/Transportation/Subway-Entrances/drex-xx56/about.

25.

Dreier

Peter

Swanstrom

Todd

Mollenkopf

John H.

2001. “Place Matters: Metropolitics for the Twenty-First Century.” University Press of Kansas, Lawrence.

26.

Dwyer

Rachel

Lassus

Lora A. Phillips

. 2015. “The Great Risk Shift and Precarity in the U.S. Housing Market.” The ANNALS of the American Academy of Political and Social Science 660 (1): 199–216.

27.

Faber

Jacob W.

2013. “Racial Dynamics of Subprime Mortgage Lending at the Peak Racial Dynamics of Subprime Mortgage Lending at the Peak.” Housing Policy Debate 23 (2): 328–49.

28.

Faber

Jacob W.

Ellen

Ingrid Gould

. 2016. “Race and the Housing Cycle: Differences in Home Equity Trends Among Long-Term Homeowners.” Housing Policy Debate 26 (3): 456–73.

29.

Faber

Jacob W.

Sharkey

Patrick

. 2015. “Neighborhood Effects.” In International Encyclopedia of Social & Behavioral Sciences, 2nd ed., edited by Wright

James D.

, 443–49. Waltham, MA: Elsevier.

30.

Federal Deposit Insurance Corporation. 2009. “2009 National Survey of Unbanked and Underbanked Households.” Accessed June 20, 2014. http://dataferrett.census.gov/.

31.

Federal Deposit Insurance Corporation. 2011. “2011 National Survey of Unbanked and Underbanked Households.” Accessed June 20, 2014. http://dataferrett.census.gov/.

32.

Federal Deposit Insurance Corporation. 2012. “2011 National Survey of Unbanked and Underbanked Households: Executive Summary.” https://www.fdic.gov/householdsurvey/2012_unbankedreport_execsumm.pdf

33.

Federal Deposit Insurance Corporation. 2014. “Statistics on Depository Institutions.” Accessed April 15, 2014. http://www2.fdic.gov/SDI/download_large_list_outside.asp.

34.

Federal Financial Institutions Examination Council. 2006. “HMDA Loan Application Register.” Accessed November 22, 2011. http://www.ffiec.gov/hmda/hmdaflat.htm.

35.

Federal Financial Institutions Examination Council. 2007. “HMDA Loan Application Register.” Accessed November 22, 2011. http://www.ffiec.gov/hmda/hmdaflat.htm.

36.

Fellowes

Matthew

Mabanta

Mia

. 2008. “Banking on Wealth: America’s New Retail Banking Infrastructure and Its Wealth-Building Potential.” Brookings Institution, Washington, DC.

37.

Fernholz

Tim

. 2010. “Making Bank: Are Simple Savings Plans the First Step to Combating Poverty?” The American Prospect, 21(5), June.

38.

Fox

Jean Ann

Woodall

Patrick

. 2006. “Cashed Out: Consumers Pay Steep Premium to ‘Bank’ at Check Cashing Outlets.” Consumer Federation of America, Washington, DC.

39.

Friedline

Terri

Kepple.

Nancy

2016. “Does Community Access to Alternative Financial Services Relate to Individuals’ Use of These Services? Beyond Individual Explanations.” Center on Assets, Education, and Inclusion, University Press of Kansas, Lawrence.

40.

Gallmeyer

Alice

Roberts

Wade T.

2009. “Payday Lenders and Economically Distressed Communities: A Spatial Analysis of Financial Predation.” The Social Science Journal 46 (3): 521–38.

41.

González-Rivera

Christian.

2013. “State of the Chains, 2013.” Center for an Urban Future, New York.

42.

Gramlich

Edward M.

2007. Subprime Mortgages: America’s Latest Boom and Bust. Washington, DC: The Urban Institute Press.

43.

Graves

Steven M.

2003. “Landscapes of Predation, Landscapes of Neglect: A Location Analysis of Payday Lenders and Banks.” Professional Geographer 55 (3): 303–17.

44.

Gross

Matthew B.

Hogarth

Jeanne M.

Manohar

Adithya

. 2012. “Who Uses Alternative Financial Services, and Why?” Consumer Interests Annual 58.

45.

Hall

Matthew

Crowder

Kyle

Spring

Amy

. 2015. “Neighborhood Foreclosures, Racial/Ethnic Transitions, and Residential Segregation.” American Sociological Review 80 (3): 526–49.

46.

Hout

Michael

Levanon

Asaf

Cumberworth

Erin

. 2011. “Job Loss and Unemployment During the Great Recession.” In The Great Recession, edited by Grusky

David B.

Western

Bruce

Wimer

Christopher

, 59–91. New York: Russell Sage Foundation.

47.

Winnie

. 2012. “Criticism Grows as Check-Cashing Stores Expand Services in Poorer Areas.” The New York Times, August 6, A18.

48.

Hwang

Jackelyn

Hankinson

Michael

Brown

Steven Kreg

. 2015. “Racial and Spatial Targeting: Segregation and Subprime Lending Within and Across Metropolitan Areas.” Social Forces 93 (3): 1081–1108.

49.

Immergluck

Dan

. 2009. Foreclosed: High-Risk Lending, Deregulation, and the Undermining of America’s Mortgage Market. Ithaca, NY: Cornell Univ. Press.

50.

Immergluck

Daniel

. 2010. “Neighborhoods in the Wake of the Debacle: Intrametropolitan Patterns of Foreclosed Properties.” Urban Affairs Review 46 (1): 3–36.

51.

Kenworthy

Lane

Owens

Lindsay A.

2011. “The Surprisingly Weak Effect of Recessions on Public Opinion.” In The Great Recession, edited by Grusky

David B.

Bruce

Western

Christopher

Wimer

, 196–219. New York: Russell Sage Foundation.

52.

Kubrin

Charis E.

Squires

Gregory D.

Ousey

Graham C.

2011. “Does Fringe Banking Exacerbate Neighborhood Crime Rates? Social Disorganization and the Ecology of Payday Lending.” Criminology & Public Policy 10 (2): 435–66.

53.

Lachance

Marie-Eve

. 2014. “Financial Literacy and Neighborhood Effects.” Journal of Consumer Affairs 48(2): 251–273.

54.

Logan

John R.

Molotch

Harvey L.

1987. Urban Fortunes: The Political Economy of Place. Berkeley: Univ. of California Press.

55.

Logan

John R.

Zengwang

Stults

Brian

. 2012. “Interpolating U.S. Decennial Census Tract Data from as Early as 1970 to 2010: A Longitudinal Tract Database.” Professional Geographer 66(3): 412–20.

56.

Massey

Douglas S.

Denton

Nancy

. 1993. American Apartheid: Segregation and the Making of the Underclass. Cambridge: Harvard Univ. Press.

57.

McCoy

Patricia A.

2004. “Predatory Lending Practices: Definition and Behavioral Implications.” In Why the Poor Pay More: How to Stop Predatory Lending, edited by Squires

Gregory D.

, 81–102. Westport: Praeger.

58.

Minnesota Population Center. 2011. “National Historical Geographic Information System: Version 2.0.” Univ. of Minnesota, Minneapolis.

59.

Negro

Giacomo

Visentin

Fabiana

Swaminathan

Anand

. 2014. “Resource Partitioning and the Organizational Dynamics of ‘Fringe Banking.’” American Sociological Review. Published electronically June 9. 79(4): 680–704.

60.

New York City Department of Consumer Affairs. 2010a. “Debt/Credit Help—Understand Loans: Tips About Payday Loans.”

61.

New York City Department of Consumer Affairs. 2010b. “More than 825,000 Adults in New York City Do Not Have Bank or Credit Union Accounts According to New Citywide Study.”

62.

New York State Department of Financial Services. 2006. “Industry Letter: Annual Maximum Check Casher Fee.” Accessed September 19, 2016. http://www.dfs.ny.gov/legal/industry/il060210.htm.

63.

New York State Department of Financial Services. 2011. “Industry Letter: Annual Maximum Check Casher Fee.” Accessed April 19, 2015. http://www.dfs.ny.gov/legal/industry/il110208.htm.

64.

Owens

Ann

Sampson

Robert J.

2013. “Community Well-Being and the Great Recession.” Pathways Spring. Stanford Center on Poverty and Inequality, Stanford.

65.

Pew Charitable Trusts. 2013. “Payday Lending in America: Policy Solutions.” http://www.pewtrusts.org/~/media/legacy/uploadedfiles/pcs_assets/2013/pewpaydaypolicysolutionsoct2013pdf.pdf

66.

Prager

Robin.

2009. “Determinants of the Locations of Payday Lenders, Pawnshops and Check-Cashing Outlets.” Federal Reserve Board, Washington, DC.

67.

Ratcliffe

Caroline

McKernan

Signe-Mary

Kalish

Emma

Martin.

Steven

2015. “Where Are the Unbanked and Underbanked in New York City?” Urban Institute, Washington, DC.

68.

Rhine

Sherrie L. W.

Greene

William H.

Toussaint-Comeau

Maude

. 2006. “The Importance of Check-Cashing Businesses to the Unbanked: Racial/Ethnic Differences.” The Review of Economics and Statistics 88 (1): 146–57.

69.

Rugh

Jacob S.

Albright

Len

Massey

Douglas S.

2015. “Race, Space, and Cumulative Disadvantage: A Case Study of the Subprime Lending Collapse.” Social Problems 62 (2): 186–218.

70.

Rugh

Jacob S.

Massey

Douglas S.

2010. “Racial Segregation and the American Foreclosure Crisis.” American Sociological Review 75 (5): 629–51.

71.

Sawyer

Noah

Temkin.

Kenneth

2004. “Analysis of Alternative Financial Service Providers.” Urban Institute, Washington, DC.

72.

Servon

Lisa

. 2013. “The High Cost, for the Poor, of Using a Bank.” The New Yorker, October 9, 2013. Accessed April 18, 2015. http://www.newyorker.com/business/currency/the-high-cost-for-the-poor-of-using-a-bank.

73.

Servon

Lisa

. 2014. “Are Banks Too Expensive to Use?” The New York Times, October 29, 2014 A31.

74.

Sharkey

Patrick

Faber

Jacob W.

2014. “Where, When, Why, and for Whom Do Residential Contexts Matter? Moving Away from the Dichotomous Understanding of Neighborhood Effects.” Annual Review of Sociology 40 (1): 559–79.

75.

Singer

Judith D.

Willett

John B.

2003. Applied Longitudinal Data Analysis: Modeling Change and Event Occurrence. London: Oxford Univ. Press.

76.

Smith

Tony E.

Smith

Marvin M.

Wackes

John

. 2008. “Alternative Financial Service Providers and the Spatial Void Hypothesis.” Regional Science & Urban Economics 38 (3): 205–27.

77.

Squires

Gregory D.

1994. Capital and Communities in Black and White: The Intersection of Race, Class, and Uneven Development. Albany: State Univ. of New York Press.

78.

Squires

Gregory D.

2004. “The New Redlining.” In Why the Poor Pay More: How to Stop Predatory Lending, edited by Squires

Gregory D.

, 1–24. Westport: Praeger.

79.

Squires

Gregory D.

O’Connor

Sally

. 1998. “Fringe Banking in Milwaukee: The Rise of Check-Cashing Businesses and the Emergence of a Two-Tiered Banking System.” Urban Affairs Review 34:126–49.

80.

Taylor

Paul

Kochhar

Rakesh

Fry

Richard

Velasco

Gabriel

Motel

Seth

. 2011. “Twenty-to-One: Wealth Gaps Rise to Record Highs Between Whites, Blacks and Hispanics.” Pew Social & Demographic Trends. Accessed June 10, 2013. http://www.pewsocialtrends.org/files/2011/07/SDT-Wealth-Report_7-26-11_FINAL.pdf.

81.

U.S. Census Bureau. 2012. “ZIP Code Business Patterns.” Accessed April 19, 2015. http://www.census.gov/econ/cbp/.

82.

U.S. Department of Housing and Urban Development. 2012. “HUD USPS ZIP Code Crosswalk Files.” Accessed 31 April 19, 2015. http://www.huduser.org/portal/datasets/usps_crosswalk.html.

83.

U.S. Postal Service Office of Inspector General. 2014. “Providing Non-bank Financial Services for the Underserved.” White Paper. U.S. Postal Service Office of Inspector General, Arlington.

84.

Wacquant

Loïc

. 2008. Urban Outcasts: A Comparative Sociology of Advanced Marginality. Polity.

85.

Weselcouch

Max.

2014. “Foreclosure Crisis in New York City: Causes and Consequences.” Presentation for the American Planning Association New York Metro Chapter, February 16, 2014.

86.

Williams

Richard

Nesiba

Reynold

Diaz McConnell

Eileen

. 2005. “The Changing Face of Inequality in Home Mortgage Lending.” Social Problems 52 (2): 181–208.