Columbia University



Survey of Electronic Benefits Transfer (EBT) Card Services

to Participants in Temporary Assistance to Needy Families (TANF)

at Full Service Point of Service (FSPOS) Locations in New York City

Report Submitted to

New York State Attorney General's Office

Civil Rights Bureau

120 Broadway, 23rd Floor

New York, New York 10271

Prepared by

Dr. Héctor R. Cordero-Guzmán

Inter-American Consulting Corporation

67 Wall Street, Suite 2411

New York, New York 10005

Tel. 212-828-2489

Mobile 646-382-3717

E-mail: hcordero@

March 10, 2001

EXECUTIVE SUMMARY

The goal of this study was to determine whether participants in the Temporary Assistance to Needy Families (“TANF”) program are able to use their Electronic Benefit Transfer (“EBT”) cards at “Full Service Point of Sale” (“FSPOS”) locations throughout New York City to obtain their TANF benefits without paying surcharges and without making purchases. In order to conduct the study, audits were done of a sample of FSPOS establishments. The auditors were given actual EBT cards and visited the FSPOS establishments posing as TANF recipients. When visiting these locations, the auditors determined whether they were able to obtain $100 in cash without a fee and without making a purchase in each establishment. In cases in which they could not obtain $100 in cash without a fee and without purchase, they determined how much cash they could obtain, whether a fee was charged to obtain that cash, and whether a purchase was required to obtain cash. If no cash could be obtained at a particular location, the auditors determined whether the card could be used to make purchases or whether the card was not accepted at all.

$ The data for the project were collected from a stratified (by zip code) random sample of FSPOS establishments in New York City. The team of auditors set out to visit 627 of the 1172, or slightly more than 50%, FSPOS locations listed in a series of reports of establishments prepared for New York State entitled Cash Access Update[1] (“Cash Access Reports”). Since they were unable to locate 14 of the 627 locations, the actual number of audited establishments was 613.

$ Of the 613 locations visited, representing the population of 1172 FSPOS establishments, 48 establishments (representing 91 locations in the population) were willing to provide the auditors with $100 in cash without requiring a fee or a purchase. Based on the results from the sample of establishments visited, the study’s conclusion is that only 7.9% of the FSPOS locations in New York City will provide recipients with $100 in cash without requiring a fee or a purchase.

$ Broken down by borough, the percentage of the FSPOS establishments visited that were willing to provide the auditors with $100 in cash without requiring a fee or a purchase was: 22.2% in Staten Island, 8.9% in Queens, 7.8% in the Bronx, 6.9% in Manhattan and 6.7% in Brooklyn. If the sample is weighted to represent all of the FSPOS establishments in New York City, the percentages change slightly to: 20% in Staten Island, 9.1% in Queens, 8.4% in the Bronx, 6.9% in Manhattan, and 6.7% in Brooklyn.

$ Based on the results at locations audited, only 31.1% of FSPOS locations in New York City will provide TANF recipients with cash even if they agree to pay a fee, make a purchase, or accept an amount below $100. Conversely, 68.9% of FSPOS establishments in New York City will not provide cash under any circumstances to TANF recipients.

$ $100 in cash is available in .8% of FSPOS locations in New York City if recipients are willing to pay a fee. Fees are also charged at establishments unwilling to dispense $100, but instead willing to provide lesser amounts or only to allow purchases with the EBT card.

$ In the majority of establishments that charged fees (62.5%), the amount charged was $1.50. A fee of $2.00 was charged at 12.5% of establishments that charged fees.

$ In 11.2% of FSPOS establishments in New York City, less than $100 – usually much less – was available. The study found that 97.9% of the locations will provide $50 dollars or less, while 52.1% offered the auditors $30 dollars or less.

$ In 11.2% of FSPOS establishments in New York City, cash back is available only after purchase.

$ A large proportion of FSPOS establishments in New York City, 46.7%, accept the EBT card for purchases only.

$ In 12.2% of FSPOS establishments in New York City, recipients’ EBT cards are not accepted at all.

$ The study’s results strongly suggest that the current EBT system is not working to provide TANF recipients with consistent and reliable access to their cash benefits at FSPOS locations throughout New York City. Specifically, I conclude that very few establishments will provide $100 in cash without a fee or purchase, and a minority of establishments will dispense any cash at all, even with a fee or purchase. Thus, in a majority of FSPOS locations throughout New York City (68.9%), TANF recipients are unable to obtain any cash at all with their EBT cards.

I. Introduction

Prior to beginning my study, I was informed of the following background facts by attorneys from the Civil Rights Bureau of the New York State Attorney General’s Office.

New York State, through its Office of Temporary and Disability Assistance (“OTDA”), entered into a contractual arrangement with Citicorp Services, Inc. (“Citicorp”) to automate the delivery of food stamp and cash welfare benefits, including TANF benefits. The new system, which is known as EBT, was intended to provide cash to TANF recipients through two different mechanisms: Bank ATM’s and Point of Sale (“POS”) machines at various retail establishments. This system was designed by Citicorp, and its subcontractors, Lockheed Martin IMS and Deluxe Electronic Payment Systems, Inc., to replace the delivery of cash benefits by check cashers in neighborhoods throughout New York City. Recipients paid no fee to the check cashers to receive their benefits and all fees were paid by New York State.

Under the new EBT system, the State pays Citicorp, which in turn has guaranteed that the EBT system would provide at least the same access to “free” cash as the former system, and in fact has represented that the new system provides far superior cash access on a neighborhood by neighborhood basis. The new system was implemented in New York City starting in 1999 and was fully operational in 2000. Specifically, Citicorp’s system is supposed to ensure that a combination of ATM’s and POS locations will provide recipients with full access to their cash benefits in their home neighborhoods without having to pay surcharges to receive their money. Citicorp provides the State with periodic reports that list both ATM’s and POS locations in each borough (by zip code) that it represents will provide cash without a surcharge to TANF recipients. More specifically, in the case of POS locations, Citicorp’s cash access plan represents that certain establishments will provide up to $100 in cash without a surcharge and without recipients having to purchase items at those establishments. Those POS locations are known as FSPOS establishments. FSPOS retailers are the backbone of the system, particularly in the large proportion of neighborhoods with few ATM’s and large numbers of recipients.

I was further informed by attorneys from the New York State Attorney General’s Office that a percentage of TANF recipients have been paying surcharges each month to withdraw their benefits. Therefore, they requested that I design an audit study to examine whether the POS system was in actuality providing “free” cash to recipients. Specifically, I was asked to investigate whether TANF recipients in neighborhoods throughout the City are in fact able to withdraw their benefits with their EBT card without paying a surcharge or making a purchase at FSPOS locations.

II. Goals and Objectives of the Project

The goal of this study was to assess whether and how TANF recipients were able to use their EBT cards at FSPOS locations throughout New York City. As described above, FSPOS locations are establishments that have been represented to provide $100 or more in cash per customer without a fee and without requiring a purchase. The study tested the FSPOS locations listed on the August 2000 Cash Access Reports submitted by Citicorp’s subcontractor, Lockheed Martin IMS, to OTDA. The study examined whether recipients were in fact able use their cards at those locations, and, if so, what the allowed uses of the card were. Specifically, it examined whether locations provided cash, the amount of cash provided, and whether a fee was charged and/or a purchase required for recipients to obtain access to their cash.

III. Methodology

In order to determine whether and how EBT users can utilize their cards at FSPOS locations, I designed an audit study. For the study, I selected a stratified random sample of establishments throughout the City (explained below), and then hired and trained a team of auditors to conduct the study.

Instrument Design

The first stage of the project involved the design and preparation of a survey instrument that allowed each of the auditors to record her experiences in each of the locations. The survey instrument recorded whether auditors were able to use the cards, the cards’ allowed uses, and whether there were any fees charged for use of EBT cards. In addition, it recorded the name and location of the store, other relevant information about the establishment, and the experiences of the auditors during their visit.

Sampling

The data for the project were collected from a stratified (by zip code) random sample of FSPOS establishments in New York City. Stratified random sampling[2] is a method of sampling whereby every case in the population has the same probability of being selected into the sample but, in addition, cases are selected from known subsets of the population to reduce sampling error and ensure that the sample is representative of the particular trait that defines the population subsets. In this case, the representative trait is zip code, and therefore the stratified random sample was generated and the number of cases was selected to ensure that the final sample was distributed across zip codes in the same proportion as the FSPOS locations are distributed throughout New York City.

The first step in producing a statistically valid stratified sample was to obtain a list of the population (establishments by zip code in this case) from which the sample was going to be drawn. From the Cash Access Reports discussed above, I prepared a comprehensive list of establishments that were represented as providing $100 or more in cash per customer without requiring a fee or purchase (FSPOS establishments) throughout New York City broken down by zip code. I then selected a specific number of establishments per zip code into the stratified random sample. In order to do this, I divided the distribution of establishments by zip code, found the proportion of the establishments in each of the zip codes and determined how many I needed from each, assigned a number to each establishment in the population, and then randomly selected a specific number of establishments within each zip code. This ensures a random sample where the distribution of cases approximated the distribution of establishments in the population by zip code. Based on the proportion of establishments sampled in each zip code, I computed a sample weight[3] that was used to indicate the number of cases in the population represented by each sampled establishment. The weights allow me to prepare tables that include estimates of the various outcomes for the number of establishments in the population.[4]

My audit team planned to visit a little more than 50% (or 627) out of the 1172 FSPOS locations listed in the five boroughs of New York City. Aside from the usual time constraint and resource considerations, the number of establishments sampled was dictated by a preliminary analysis of the characteristics of the population of FSPOS locations by zip code. There are 60 zip codes in Queens, 40 in Brooklyn, 16 in Staten Island, 25 in the Bronx, and 45 in Manhattan, for a total of approximately 186 separate zip codes throughout New York City. I then determined that 159 zip codes had at least one FSPOS location. Also, in some instances (10 of 16 in Staten Island, three of 40 in Brooklyn, five of 25 in the Bronx, 29 out of 60 in Queens, 25 of 45 in Manhattan), there were only one or two establishments per zip code. This meant that using a strict sample of 50% would result in many zip codes not being audited at all and information not collected from those parts of the City.

An additional consideration was the fact that in a few zip codes, there were no TANF recipients. The data that I examined from New York State’s OTDA Internet site () included information on the number of public assistance cases in each zip code. In Manhattan, for example, of the 25 zip codes with two, one, or no establishments, only three did not have any recipients (10048, 10118, and 10280), and another five had fewer than 100 recipients. Sampling less than 50% of the establishments would have missed information from 17 zip codes in Manhattan with more than 100 recipients. Approximately 11.3% of the zip codes had one FSPOS establishment and all of these establishments were included in the sample (and assigned a sample weight of 1). Approximately 13.2% of the zip codes had two FSPOS establishments listed. For those zip codes, I sampled 50% (or one establishment) in each zip code and assigned those cases a weight of two. If a zip code had more than two establishments and the number of establishments was an even number (i.e., 2,4,6, etc.), I sampled exactly half of the establishments (i.e., 1,2,3, etc.). If the zip code had an odd number of establishments (i.e., 3,5,7, etc.), I sampled half of the establishments and then rounded up to the nearest whole number (i.e., 2 of 3, 3 of 5, 4 of 7, etc.).

In addition to the main study design, I added a matched pair component to the study to test for the reliability of the audit results. This involved sending two auditors to a randomly selected subgroup of locations (22% of the total FSPOS sample) and determining whether both auditors, only one, or neither of the two auditors were able to withdraw $100 in cash from the location without paying a fee or making a purchase.

Data Collection Methods

The field work for the study was conducted by five female auditors between the ages of 25 and 35. The auditors were asked to pose as TANF recipients who wanted to withdraw cash from their benefits using their EBT card. Each of the auditors received an EBT card with an open TANF and food stamp account and a PIN number but without any money in the accounts. If the cards were swiped by a POS machine, they were coded to read “insufficient funds.” The auditors saw a training video developed for actual TANF recipients that gave them general information on their TANF accounts and how the system works, how to use their EBT cards, and how to identify and conduct transactions in POS locations.

The auditors visited FSPOS establishments throughout New York City from early November to late December of 2000. They were trained to enter each location and ask the clerk or cashier whether they could use their EBT cards to withdraw $100 in cash from their TANF accounts. If an auditor was told that she could use her EBT card to withdraw $100, and also told that there was neither a fee charged nor a purchase required, she produced the card and it was swiped by store staff. The auditors were told to be as persistent as possible and to give clerks and store personnel an opportunity to conduct the transaction with them, including asking to see the store manager or visiting the customer service department at the location to clarify store policy on the use of EBT cards.

The findings of this study are based on the content of the interaction between the auditors and the staff of the various establishments. After each site visit was completed, the auditors recorded their interactions and the results in the survey instrument. The information collected included whether they were able to use the EBT card at each location, what the allowed uses of the card were, (cash withdrawal without purchase, purchase required for use of the card, or card accepted for purchases only), and whether there were any fees, minimum purchase amounts, or other transaction costs for the use of the EBT cards in the various FSPOS establishments. Every few days, I collected the completed survey instruments from each auditor and entered the information from the survey instrument into a statistical database for analysis.

IV. Results of Study of Cash Access at FSPOS Locations in New York City

Main Outcomes of the Study

Table 1 shows the number of establishments at which the auditors determined that a TANF recipient could obtain $100 without paying a fee or making a purchase. The auditors were told that they could get $100 in cash at 48 locations. Excluding the 14 locations at which the auditors did not locate the store, the number of establishments actually visited was 613. Using 613 establishments as the denominator, the 48 locations represent a success rate[5] of 7.8%.

Table 1A shows the number of establishments at which the auditors were told that they could get $100 without paying a fee or making a purchase for the sample weighted to represent the number of FSPOS establishments in New York City. My results indicate that recipients are able to get $100 in cash without paying a fee or making a purchase at 91 (or 7.9%) of the 1173 FSPOS establishments throughout New York City.

In addition to using the point estimate of 7.9% from the weighted sample, I used a second method for estimating the success rate in the population from the numbers in the sample. A second acceptable method for producing population estimates from samples using statistical theory is to estimate a confidence interval[6] of the population values based on sample numbers. Confidence intervals are usually estimated at the 95% level. That means that one can be 95% confident (only a 5% chance of error) that the real population value falls between the lower and upper parameter of the confidence interval. With the results from the sample, I estimated a 95% confidence interval around the proportion of establishments in the population at which the recipient can get $100 without paying a fee or making a purchase. The overall rate of 7.9% from the sample yields an estimate of the population value, with 95% confidence, between 5.7% and 10.0%. This means that I am 95% confident that the real population number (representing the 1173 FSPOS locations in New York City) is in that range.

Alternatively, I tested the hypothesis that our results (7.9%) could have come from a population of establishments with a success rate of 90% with a one sample t-test.[7] Our estimated t-test statistic was -75.7, indicating that, statistically, it is extremely unlikely that our sample could have come from a population where 90% of the establishments give $100 in cash without a fee and without a purchase. In fact, our statistical analysis rejects the null hypothesis that our sample comes from a population of establishments with a success rate of anything more than 10%. Statistical theory indicates that we are confident with a 95% level of certainty that the proportion of establishments in the population of FSPOS locations throughout New York City where a recipient can get $100 in cash without paying a fee or making a purchase (the success rate) is between 5.7% and 10.0%.

Main Outcomes by Borough

The main results by borough appear in Table 2 for the sample and Table 2A for the weighted sample. Table 2 presents the proportion of establishments at which a recipient can get $100 in cash with her EBT card without a fee and without having to make a purchase. I also show the number of establishments visited in each borough. The percentage of establishments at which a recipient can get $100 in cash without a fee in each borough was 22.2% in Staten Island, 8.9% in Queens, 7.8% in the Bronx, 6.9% in Manhattan, and 6.7% in Brooklyn.

Table 2A presents the results from the weighted sample representing the number and percent of establishments in the population at which a recipient can get $100 in cash without a fee and without a purchase with her EBT card. If the sample is weighted to represent all of the FSPOS establishments in New York City the success rates change slightly to: 20% in Staten Island, 9.1% in Queens, 8.4% in the Bronx, 6.9% in Manhattan, and 6.7% in Brooklyn. The number of establishments listed increases to reflect the number of establishments in the population, but the success rates in each borough are very similar to those of the unweighted sample.

I also estimated a 95% confidence interval for the percentage of establishments in each borough at which recipients can get $100 in cash without a fee and without a purchase. The 22.2% success rate for Staten Island has a 95% confidence interval between .9% and 43.5%. The confidence interval for Staten Island has a wide range because the percentage is estimated on the basis of a small number of visits. Other boroughs had more visits and this makes the estimate of the confidence interval more precise. The 8.9% success rate observed in Queens yields a 95% confidence interval between 4.0% and 13.8%. The 7.8% rate in the Bronx yields a population estimate with 95% confidence between 2.1% and 13.4%, while the 6.9% success rate in Manhattan yields a 95% confidence interval between 2.5% and 11.3%. Lastly, the 6.7% success rate in Brooklyn yields a 95% confidence interval between 3.5% and 9.9%. These results suggest that, using the most optimistic estimates, in none of the four largest boroughs in New York City the success rate in the population is higher than 14%.

Detailed Outcomes of the Study

Table 3 shows the detailed outcomes from the audit of the FSPOS locations in New York City. In addition to the results on how many locations were willing to dispense $100 in cash without fee or purchase, the auditors found a number of other outcomes during the course of the study. In five of the 613 locations sampled, or .8% of the establishments, the auditors were told that they could get $100 but they would have to pay a fee. In 69, or 11.3%, of the visits the auditors were told that they could get an amount less than $100. In 70, or 11.4%, of the establishments the auditors were told that they could get cash back only after purchases. In 285 establishments (or 46.5%) the auditors were told that the EBT card was accepted for purchases only at that location. The EBT card was not accepted for any use in 76 or 12.4% of the establishments visited, while in 42 establishments (or 6.9%) the auditors were told by store personnel that there was no POS machine, that the POS machine was broken, or that they did not know how to use the machine. In 18 locations (or 2.9%) the auditors found that the store was closed. The auditors did not find the establishment in 14 or 2.2% of the locations.

Table 3A shows the detailed outcomes from the audit for the weighted sample representing 1173 FSPOS locations throughout New York City. The number of establishments listed in the table increases to reflect the total number of FSPOS establishments throughout New York City. The detailed outcomes using the weighted sample will be presented throughout the discussion of the results.

Recipient Can Get $100 without a Fee or Purchase

The auditors were able to get $100 in cash without a fee in 48 of the 613 completed audits. This represents 91 establishments in the FSPOS population or a success rate of 7.9%. The following example from one of the surveys illustrates this result:

The sign was on the front door, indicating that cash access and food purchase was available through Quest. It was a small deli. The clerk was a [identifying information], late 20's to early 30's. I asked if I could take money out and she said yes, up to $100. I asked if there was a fee and she said no. At this time, I said wasn’t sure how much money I had left on my card but she offered to run it through so that I could see how much I had left, which she did. She gave me the receipt to review and then asked if I still wanted money. I said no, that my balance was too low but I thanked her for her help. She was very nice, very helpful. [Audit Number 61].[8]

Recipient Can Get $100 With a Fee

In five out of the 613 audits, representing 10 establishments in the population, (or .8%), the auditors were told that they could get $100 in cash but that they would have to pay a fee for the service. The following examples from the study help illustrate this result:

The first cashier said no, but then asked another guy. He said no. They were talking for a few minutes and then said yes, but the fee is $1.00 per $10. Therefore, $10 for $100. He was trying to discourage me by saying “The machine will charge me $10.” [Audit Number 480].

On the receipt it said $1.00 and I asked, he said it was the fee--that it was how much it costs. The POS machine was very large. I never saw one like that before. [Audit Number 563].

Recipient Can Get Smaller Amount

The weighted sample estimate indicates that in approximately 129 (or 11.2%) of the FSPOS establishments in New York City, recipients could get smaller amounts of cash-- usually much less than $100 with their EBT cards. In 97.3% of those establishments, the auditors were offered $50 or less. More specifically, 10% of the establishments offered the auditors $10, while 29% of the establishments said that they would give the auditors $20. Thirty dollars or less was available in 52.7% of the establishments. Some examples from the site visits demonstrate how the auditors were offered less than $100 in cash:

The store was located in an area that seemed to be mostly low-income [identifying information]. The two clerks were [identifying information], possible mother and daughter. The store was larger than a store front but there didn’t seem to be much in the way of stock in the store. I asked if I could use my card to withdraw money and the older of the two women looked at me a bit suspiciously and asked how much. I said $100 and they both almost looked like they wanted to laugh. The older women again answered and she said the most they could give was $30. I thanked her and said that I needed $100 and that I’d try to find another store. [Audit Number 63].

The clerk told me that he doesn’t give out that much money. He then suggested I go to [identifying information] where they give out $50. I asked how much he gives out and the reply was $10. [Audit Number 444].

Clerk initially said, “No problem,” she swiped the card and then manually input the card number. She then looked at me and said, “you want $1.00, right?” I told her “No, $100.” Then she said that the customer before me had “cleaned her out.” She then referred me to the [identifying information] supermarket which is one block away. They have an ATM machine with a $1.50 fee. [Audit Number 364].

He offered me $20 and told me that there was a $2.00 fee. I asked why, and he said that anywhere you go you will be charged a fee. If you go to an ATM, you will still be charged a fee [Audit Number 31].

Recipient Can Get Cash Back Only After Purchase

In 70 establishments, representing 128 (or 11.2%) of the FSPOS locations in the population, the auditors were told that the store would give cash back only after a purchase. Most of the establishments only required the purchase of one item. But, in only two of the 70 establishments (representing four of the 128 in the population) were the auditors told that they would be able to get $100 in cash after making a purchase. In the other 68 establishments (representing 124 locations in the population or 96.8%), auditors were told that they had to make a purchase in order to receive an amount less than $100 in cash. The following examples illustrate this result:

I went to the register and asked the clerk if I could use my card to access cash. She said yes and then asked how much. When I said $100, she looked a bit startled and even a bit afraid and said I had to ask the manager. When I told him the clerk had referred me to him, he asked what I wanted and I said I wanted to take $100 out on my card. He said the clerk was smart to send me to him because he could tell me that I could only access $40 and before I could do that I had to make a purchase. He was a bit disdainful. When asked if his ATM accepted Quest, he said, “I think you people need special banks for that sort of thing. I never heard of ATM machines taking those cards.” [Audit Number 349].

I asked the cashier, and she looked up at the manager. He said there is an ATM over there and to use that. Then he said “But you can take $20 out from us. The cashier then qualified that statement by saying that I had to buy something in order to get the $20. [Audit Number 450].

There was a sign that clearly stated that cash could be obtained from the register with no surcharge when using your EBT card. However, the clerk informed me that a purchase must be made and I could only get $20. [Audit Number 244].

Two women told me that I needed to make a purchase for cash back and that I would be charged a fee. [Audit Number 707].

EBT Card Accepted for Purchases Only

The auditors found that in 285 locations, a significant proportion of the 613 audited establishments, accepted the card only for purchases and did not give any cash back. This represents approximately 536 FSPOS establishments throughout New York City, or 46.7% of the total. The following examples from the study illustrate this result:

There was a Quest sticker on the door with a $ sign. The clerk was also the proprietor, I believe. I asked if I could use my card to get money and he started complaining that he had no money. He just paid for deliveries and “These people don’t fund me.” I have no idea what he meant by that. I asked if this was just a bad day to ask for cash. He basically said that they never have cash, but if I wanted to try back tomorrow I could. He was not encouraging. He did say that I could use the card for purchases with no fee. [Audit Number 253].

There was a sticker that had a $ sign on the front door but there was another sign that listed different EBT services the store could provide but the only one checked off was that of food purchase. The guy behind the counter didn’t seem to know much. I told him I needed cash. He seemed a bit nervous and said his boss wasn’t around and he didn’t know. I asked if he thought I could get money and he said he wasn’t sure, but his boss only told him about letting people use cards for purchases. [Audit Number 248].

There is a Quest sign with the $ insignia on front doors. It took a while to be able to communicate to the clerk exactly what I wanted. She kept saying, “Sure, we take card if you have cash on card.” Once she realized I wanted cash, “You give me card and I give you cash! Oh, no, no!” It was almost as if she thought it was the most insane idea. She almost started laughing. [Audit Number 739].

I asked the cashier and she told me that I can buy something but they don’t give out cash. She then asked the manager who said that I could use the ATM. There was a sign on the ATM saying that there was a $1.00 fee per transaction. [Audit Number 471].

The first clerk said that I cannot use my card to purchase something; the second clerk said I could. They both agreed that I could not take out cash. [Audit Number 457].

EBT Card Not Accepted at the Location

In 76 establishments, (representing 140 locations in the population or 12.2%), the auditors found that the EBT card was not accepted at all by the establishment. Their comments provide examples of how they determined that the EBT card was not accepted:

There was a hand-written sign above the clerk’s window behind the bullet-proof glass enclosed area that read, “We do not accept EBT.” When I asked whether they accepted the card at all, they said no. It should also be noted the sign stated that they do not accept Food Stamps either. [Audit Number 228].

When the clerk saw my EBT I.D., she said before I even asked, “We don’t take that here.” I asked about buying something and she said they were not set up for it. [Audit Number 148].

The clerk told me that they are “not set up to give cash” and that “they are a meat market and not a bank.” [Audit Number 660].

In nine out of the 613 audits, (or 1.5%), the location was not a food store, despite having been listed as such on the Cash Access Reports, and the EBT card was not accepted.

The store is no longer in existence. It is now the [identifying information] Senior Citizen Center Complex. [Audit Number 599].

Checked in the window and went in just to make sure since I was at a $.99 Store with Quest before. Store had no food. Just magazines and cigarettes. [Audit Number 332].

Broken\No Machine at the Location

The auditors found that, in 42 of the visits, representing 80 FSPOS establishments in the population, or 7.0%, the store personnel reported that there was no POS machine, that the POS machine was broken, or that they did not know how to use the machine. The following examples are illustrative:

I saw that there is an EBT option on the POS machine. The clerk said that their system has not been programmed yet, so I cannot use my card at all. [Audit Number 299].

Although there was a sign that stated cash benefits could be accessed at the register with no surcharge, this was not the case. The first clerk I asked simply said no. There was a second clerk. However, she said that they had a machine but they had no idea how it worked. They weren’t able to give services to anyone with an EBT card. He said there was an ATM machine if I needed to use it. [Audit Number 251].

Store Closed

In 18 visits, representing 34 locations in the population (or 3.0%), the auditors found that the store was closed. The following are typical examples:

I went to the hair salon next door and asked if the bodega at this address-[identifying information]- has been closed for a while. I asked how long it’s been closed and they said between 6 months to 10 months. [Audit Number 447].

The store was well lit but the two times I arrived there, but the door was locked. I arrived at 11:36 a.m. and 12:55 p.m. [Audit Number 153].

There was a big empty door. The building had [identifying information] colors and it obviously was a [identifying information] before. The windows were clouded and there were workers inside. There was a construction sign in the window that said “COMING SOON [identifying information].” [Audit Number 195].

It was a boarded up building-the 3 floors. You could tell that there was a store in the first level. The stores next to this were all closed. On the corner was a laundry and I asked the old man smoking there (it looked like he was doing laundry) and he said the landlord is from Long Island and doesn’t come around. I guess other people looking to buy have asked him before. Then he told me that it caught fire. I asked him what kind of store it was and he said grocery store. I asked if it was [identifying information] and he looked at me weird. I said the name and he didn’t know. [Audit Number 440].

Location Not Found

The auditors were not able to find 14 or 2.2% of the 627 locations in the sample. This represents 25 (or 2.1%) of the estimated 1173 FSPOS establishments throughout New York City. The following examples help to illustrate this:

No such address. [Audit Number 604].

Not a store. Address does not exist. There is a huge parking lot. [Audit Number 134].

Matched Pair Results

In addition to the main study design, I added a matched pair component to the study to test the reliability of the audit results. This involved sending two auditors to a randomly selected sub-group of locations (22% of the total FSPOS sample) and determining whether both auditors, only one, or neither of the two auditors were able to withdraw $100 in cash from the location without paying a fee or making a purchase. The results were as follows: of the 344 matched pair audits conducted, in 304 (or 88.4%), neither auditor determined that a recipient could withdraw $100 without a fee or purchase, in 20 audits (or 10 establishments) both auditors determined that a recipient could withdraw $100 cash from the location without a fee or purchase, and in another 20 audits (or 5.8%), one auditor determined that a recipient could withdraw $100 cash from the location without a fee or purchase and the other auditor determined that a recipient could not withdraw $100 without a fee or purchase at the same location. The fact that in 94.2% of the audits the outcomes were the same for both auditors suggests that the results observed during our visits to the establishments were consistent across auditors and can be considered reliable.

Additional Results: Fees Charged, Use of ATMs, and Inconsistent Cash Access Policies

Three additional sets of results were observed in the course of the study that merit some discussion. First, there were a number of establishments that charged fees for the use of the EBT card. Second, in many instances the auditors were referred to ATM machines in the location or at another location. Many of these ATM machines had fees usually of $1.00 or $1.50. Third, the auditors observed many inconsistencies in cash access policies in the various locations.

Fees Charged

As some of the preceding examples illustrate, a number of establishments charged fees for the use of EBT cards. We found that at least 23 of the establishments in the sample of FSPOS locations charged fees (representing 47 establishments in the population). In 62.5% of the establishments that charged fees, the rate was $1.50. In 12.5% of the establishments, the fee was $2.00. The average fee was $1.73. One establishment asked for a fee of $10 on a withdrawal of $100. The ubiquitousness of fees for some transactions is illustrated by the following example from one of the audits:

Clerk said, “Of course there is a fee. It is like a phone. Everytime you use it, you have to pay a fee depending on your withdrawal amount.” [Audit Number 59].

Auditors Sent to ATM

Also, a significant number of stores directed the auditors to ATM machines with fees in their location (or at other locations). The auditors were referred to an ATM machine in approximately 17.5% of the audits. In many instances the auditors were referred directly to the ATM machine at the store or at another location for any EBT card transactions, or there was a hand-written sign near the cashier indicating that all cash withdrawals had to be done at the ATM. The following examples illustrate auditors being referred to an ATM:

There was a Quest sign with the dollar insignia. When I got to the register, there was a sign that said “All cash advances, please use ATM. We no longer give out money at the register.” I confirmed this information with the clerk. There is a $1.50 charge at the ATM. [Audit Number 716].

There was a sign on the front door that clearly stated that cash could be obtained from the register with no surcharge. However, I was informed by the clerk that if any cash transactions had to be made with the EBT card at this location, it must be done at the ATM machine. The register only dealt with food purchases. The clerk was a bit snooty. I went to the ATM machine which had a Quest sign but also stated that any transaction would warrant a $1.50 charge on top of any fee from financial institutions. [Audit Number 242].

The sign on the front door looked like a blown-up circular from the store itself except it stated that EBT was accepted for food purchases and for cash withdrawals. I went to the register and the woman very nicely told me that cash benefits were dealt with by the ATM machine. I asked if I could get cash back if I bought something and she said “No, only at the ATM.” The machine had a $1.50 charge sign on it and it did accept Quest. [Audit Number 170].

1- I asked [identifying information], he did not know.

2- He asked [identifying information], she said, “NO, NO, NO! We don’t give out cash!”

3- I pointed to the Quest sign on the door. She said, “Oh, I know, but we don’t have cash right now.”

4- The [identifying information] then said, “Oh, you know what? You can use the ATM.” [Audit Number 269].

Inconsistent Cash Access Policies

In many instances when auditors asked to withdraw cash from their TANF accounts using their EBT card, they were sent to see store managers, were directed to the customer service department, or were told that store policy on when and how the card could be used and how much cash could be taken out, if any, was not consistent and depended on whether “the boss” or a supervisor was present and personally authorized the cash withdrawal and the amount. The following cases help to illustrate this:

The clerk said I could use the card for purchase but seemed a little nervous when I asked to access cash. There is a sign on the front door that says cash access at register, no surcharge. The clerk said her boss needs to approve all cash transactions but she wouldn’t be back until later. I asked whether I might get cash later and she said it depended on what the boss said. Sometimes she says yes and sometimes she says no. There is never any charge but the amount accessed depended on what the boss said. [Audit Number 710].

Sign on front door said they accept EBT for cash withdrawals but does not state whether there is a charge. There was also a Quest sign with the $ insignia. Asked if I could use the card to withdraw money he said not today. I asked whether I could come back and he said sure but he wasn’t sure if he’d have money. I asked if he usually allowed people to access cash and he said sure, up to $100. I asked how much he charged and he said it depended on how much I wanted to get. I said $100 and he said he’d charge me $1.50. [Audit Number 761].

Quest sign with the $ insignia on front door. The clerk was very friendly. Asked if I could use my card at this location he said yes but when I asked to take out money, he said the boss wasn’t around and he couldn’t give out cash unless his boss was there. I asked if I could come back later and he said sure but he didn’t think they’d have money later. Like I said, he was friendly but I couldn’t help feeling I was getting the run around. [Audit Number 752].

V. Conclusion

During approximately seven weeks of fieldwork, the team of five auditors visited 613 FSPOS locations in all of the zip codes that had at least one FSPOS location throughout New York City’s five boroughs. In addition, 172 locations were visited by two auditors, for a total of 799 audits. The auditors were told to be as persistent as possible and to afford clerks and store personnel ample time and opportunity to conduct transactions with them. The findings from the study are based on the results of the interactions between the auditors and the staff of the various establishments.

The results strongly suggest that the current EBT system is not working to provide TANF recipients with access to their cash benefits at FSPOS locations throughout New York City. Specifically, I conclude that 7.9% of FSPOS establishments will provide $100 in cash without requiring the recipient to pay a fee or make a purchase. In addition, I find that 31.1% of FSPOS establishments will dispense cash in any amount even after, in some cases, requiring the recipient to pay a fee or make a purchase. Thus, under the current EBT system, TANF recipients are not able to obtain any cash at all with their EBT cards in the overwhelming majority (68.9%) of FSPOS locations throughout New York City.

-----------------------

[1] The full title of the reports is Cash Access Update, Electronic Benefits Transfer Program, State of New York (August 1, 2000). There is one report prepared for each borough on a periodic basis.

[2] For a detailed description of stratified random sampling, see Chapter 6 in Joseph Healey, Statistics: A Tool for Social Research. New York: Wadsworth Publishers, 1998 (5th Edition).

[3] A weighted sample is one where the values have been adjusted to reflect the number of population units represented by each case. The sample weight is the reciprocal of the proportion of cases sampled. For example, if we sampled one of two (½) establishments, the weight would be the reciprocal of that fraction (2/1), and we would multiply the number of cases (1) by the weight (2/1) to estimate the number of establishments in the population (1 x 2/1 = 2).

[4] The number of establishments in the population was 1172. The weighted sample yields a total of 1173 establishments due to rounding.

[5] Success rate is the percent of establishments at which the auditors determined that a recipient can get $100 in cash without paying a fee or making a purchase.

[6] A confidence interval is a range of values from a sample statistic (mean or proportion)

that is likely to contain a population value at a given level of statistical probability determined by the confidence level. A confidence level (usually 95%) can be interpreted as the likelihood that the true population value is included in our estimate. See chapter 7 in Joseph Healey, Statistics: A Tool for Social Research. New York: Wadsworth Publishers, 1998 (5th Edition).

[7] A t-test is a test for the statistical significance of the results of a comparison between two group means or proportions. The null hypothesis is that the two numbers are not statistically different. The research hypothesis is that they are. Statisticians reject the null hypothesis of no difference if a t-test is larger than "ð 1.96. See Chapter 8 in Josephce if a t-test is larger than ∀ 1.96. See Chapter 8 in Joseph Healey, Statistics: A Tool for Social Research. New York: Wadsworth Publishers, 1998 (5th Edition).

[8] The number refers to the audit number from my database with the results.

................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download